From pavel.pokorny at datera.cz  Thu Apr  9 22:00:07 2015
From: pavel.pokorny at datera.cz (Pavel Pokorny)
Date: Thu, 9 Apr 2015 23:00:07 +0200
Subject: [gpfsug-discuss] Share GPFS from Windows?
Message-ID: <CAHy3nHJ9=D_zYq3xxFO8NnJutdy7JAWxkN7s_PZds277Nx4rxA@mail.gmail.com>

Hello to all,
Is there any technical reason why it is not supported to export GPFS from
Windows nodes using CIFS? As stated in GPFs FAQ (
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html
):

*Exporting GPFS file systems as Server Message Block (SMB) shares (also
known as CIFS shares) from GPFS Windows nodes is not supported.*

Or it is limitation more related to licensing and business point of view.
Thank you for answers,
Pavel

--
Ing. Pavel Pokorn?
DATERA s.r.o. | Ovocn? trh 580/2 | Praha | Czech Republic
www.datera.cz | Mobil: +420 602 357 194 | E-mail: pavel.pokorny at datera.cz
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150409/6f6c75d7/attachment.htm>

From BOLIK at de.ibm.com  Fri Apr 10 12:24:14 2015
From: BOLIK at de.ibm.com (Christian Bolik)
Date: Fri, 10 Apr 2015 13:24:14 +0200
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
	multitude of GPFS clusters
Message-ID: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>


Just wanted to let you know that recently GPFS support has been added to
TPC, which is IBM's Tivoli Storage Productivity Center (soon to be renamed
to IBM Spectrum Control). As of now, TPC allows GPFS administrators to get
answers to the following questions, across any number of GPFS clusters
which have been added to TPC:

- Which of my clusters are running out of free space?
- Which of my clusters or nodes have a health problem?
- Which file systems and pools are running out of capacity?
- Which file systems are mounted on which nodes?
- How much space is occupied by snapshots? Are there any very old,
potentially obsolete ones?
- Which quotas are close to being exceeded or have already been exceeded?
- Which filesets are close to running out of free inodes?
- Which NSDs are at risk of becoming unavailable, or are unavailable?
- Are the volumes backing my NSDs performing OK?
- Are all nodes fulfilling critical roles in the cluster up and running?
- How can I be notified when nodes go offline or file systems fill up
beyond a threshold?

There's a short 6-minute video available on YouTube which shows how TPC
helps answering these questions:
https://www.youtube.com/watch?v=8Esk5U_cYw8&feature=youtu.be

For more information about TPC, please check out the product wiki on
developerWorks: http://ibm.co/1adWNFK

Thanks,
Christian Bolik
IBM Storage Software Development




From zgiles at gmail.com  Fri Apr 10 16:27:36 2015
From: zgiles at gmail.com (Zachary Giles)
Date: Fri, 10 Apr 2015 11:27:36 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>
References: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>
Message-ID: <CAMYZk=fKkAgeR-MW79a=BO46WpR9+Gm+=_kCLkQo4+v5KxgKqg@mail.gmail.com>

Christian:
Interesting and thanks for the latest news.

May I ask: Is there an intent moving forward that TPC and / or other Tivoli
products will be a required part of GPFS?
The concern I have is that GPFS is pretty straightforward at the moment and
has very logical requirements to operate (min servers, quorum, etc),
whereas there are many IBM products that require two or three more servers
just to manage the servers managing the service.. too much. It would be
nice to make sure, going forward, that the core of GPFS can still function
without additional web servers, Java, a suite of middleware, and a handful
of DB2 instance .. :)

-Zach



On Fri, Apr 10, 2015 at 7:24 AM, Christian Bolik <BOLIK at de.ibm.com> wrote:

>
> Just wanted to let you know that recently GPFS support has been added to
> TPC, which is IBM's Tivoli Storage Productivity Center (soon to be renamed
> to IBM Spectrum Control). As of now, TPC allows GPFS administrators to get
> answers to the following questions, across any number of GPFS clusters
> which have been added to TPC:
>
> - Which of my clusters are running out of free space?
> - Which of my clusters or nodes have a health problem?
> - Which file systems and pools are running out of capacity?
> - Which file systems are mounted on which nodes?
> - How much space is occupied by snapshots? Are there any very old,
> potentially obsolete ones?
> - Which quotas are close to being exceeded or have already been exceeded?
> - Which filesets are close to running out of free inodes?
> - Which NSDs are at risk of becoming unavailable, or are unavailable?
> - Are the volumes backing my NSDs performing OK?
> - Are all nodes fulfilling critical roles in the cluster up and running?
> - How can I be notified when nodes go offline or file systems fill up
> beyond a threshold?
>
> There's a short 6-minute video available on YouTube which shows how TPC
> helps answering these questions:
> https://www.youtube.com/watch?v=8Esk5U_cYw8&feature=youtu.be
>
> For more information about TPC, please check out the product wiki on
> developerWorks: http://ibm.co/1adWNFK
>
> Thanks,
> Christian Bolik
> IBM Storage Software Development
>
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>



-- 
Zach Giles
zgiles at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150410/c4ee09f6/attachment.htm>

From dhildeb at us.ibm.com  Fri Apr 10 18:45:14 2015
From: dhildeb at us.ibm.com (Dean Hildebrand)
Date: Fri, 10 Apr 2015 10:45:14 -0700
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <CAMYZk=fKkAgeR-MW79a=BO46WpR9+Gm+=_kCLkQo4+v5KxgKqg@mail.gmail.com>
References: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>
	<CAMYZk=fKkAgeR-MW79a=BO46WpR9+Gm+=_kCLkQo4+v5KxgKqg@mail.gmail.com>
Message-ID: <OFDFD967F3.9188C435-ON88257E23.0060AE82-88257E23.006186C8@us.ibm.com>


Hi Zach,

The summary is that GPFS is being integrated much more across the
portfolio...

With GPFS itself, there is a video below demonstrating the ESS/GSS GUI and
monitoring feature that is in the product today.  Moving forward, as you
can probably see, there is a push in IBM to move GPFS to software-defined,
which includes features such as the GUI as well...
https://www.youtube.com/watch?v=Mv9Sn-VYoGU

Dean




From:	Zachary Giles <zgiles at gmail.com>
To:	gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:	04/10/2015 08:27 AM
Subject:	Re: [gpfsug-discuss] Monitoring capacity and health status for
            a multitude of GPFS clusters
Sent by:	gpfsug-discuss-bounces at gpfsug.org



Christian:
Interesting and thanks for the latest news.

May I ask: Is there an intent moving forward that TPC and / or other Tivoli
products will be a required part of GPFS?
The concern I have is that GPFS is pretty straightforward at the moment and
has very logical requirements to operate (min servers, quorum, etc),
whereas there are many IBM products that require two or three more servers
just to manage the servers managing the service.. too much. It would be
nice to make sure, going forward, that the core of GPFS can still function
without additional web servers, Java, a suite of middleware, and a handful
of DB2 instance .. :)

-Zach



On Fri, Apr 10, 2015 at 7:24 AM, Christian Bolik <BOLIK at de.ibm.com> wrote:

  Just wanted to let you know that recently GPFS support has been added to
  TPC, which is IBM's Tivoli Storage Productivity Center (soon to be
  renamed
  to IBM Spectrum Control). As of now, TPC allows GPFS administrators to
  get
  answers to the following questions, across any number of GPFS clusters
  which have been added to TPC:

  - Which of my clusters are running out of free space?
  - Which of my clusters or nodes have a health problem?
  - Which file systems and pools are running out of capacity?
  - Which file systems are mounted on which nodes?
  - How much space is occupied by snapshots? Are there any very old,
  potentially obsolete ones?
  - Which quotas are close to being exceeded or have already been exceeded?
  - Which filesets are close to running out of free inodes?
  - Which NSDs are at risk of becoming unavailable, or are unavailable?
  - Are the volumes backing my NSDs performing OK?
  - Are all nodes fulfilling critical roles in the cluster up and running?
  - How can I be notified when nodes go offline or file systems fill up
  beyond a threshold?

  There's a short 6-minute video available on YouTube which shows how TPC
  helps answering these questions:
  https://www.youtube.com/watch?v=8Esk5U_cYw8&feature=youtu.be

  For more information about TPC, please check out the product wiki on
  developerWorks: http://ibm.co/1adWNFK

  Thanks,
  Christian Bolik
  IBM Storage Software Development


  _______________________________________________
  gpfsug-discuss mailing list
  gpfsug-discuss at gpfsug.org
  http://gpfsug.org/mailman/listinfo/gpfsug-discuss



--
Zach Giles
zgiles at gmail.com_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150410/9afe8bc4/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150410/9afe8bc4/attachment.gif>

From BOLIK at de.ibm.com  Mon Apr 13 14:11:01 2015
From: BOLIK at de.ibm.com (Christian Bolik)
Date: Mon, 13 Apr 2015 15:11:01 +0200
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
Message-ID: <OFE5AC06CE.E75B85AD-ONC1257E26.00399901-C1257E26.00486BB2@de.ibm.com>


Hi Zach,

I'm not aware of any intent to make TPC or any other Tivoli/IBM product a
prereq for GPFS, and I don't think any such plans exist.  Rather, as Dean
also pointed out, we're investing work to improve integration of
GPFS/Spectrum Scale into other products being members of the newly
announced IBM Spectrum Storage family, with the goal of improving
manageability of the individual components (rather than worsening it...).

Cheers,
Christian

> Christian:
> Interesting and thanks for the latest news.
>
> May I ask: Is there an intent moving forward that TPC and / or other
Tivoli
> products will be a required part of GPFS?
> The concern I have is that GPFS is pretty straightforward at the moment
and
> has very logical requirements to operate (min servers, quorum, etc),
> whereas there are many IBM products that require two or three more
servers
> just to manage the servers managing the service.. too much. It would be
> nice to make sure, going forward, that the core of GPFS can still
function
> without additional web servers, Java, a suite of middleware, and a
handful
> of DB2 instance .. :)
>
> -Zach

Christian Bolik

Software Defined Storage Development

IBM Deutschland Research & Development GmbH, Hechtsheimer Str. 2, 55131
Mainz, Germany
Vorsitzende des Aufsichtsrats: Martina Koederitz
Gesch?ftsf?hrung: Dirk Wittkopp
Sitz der Gesellschaft: B?blingen / Registergericht: Amtsgericht Stuttgart,
HRB 243294




From zgiles at gmail.com  Mon Apr 13 17:05:11 2015
From: zgiles at gmail.com (Zachary Giles)
Date: Mon, 13 Apr 2015 12:05:11 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <OFE5AC06CE.E75B85AD-ONC1257E26.00399901-C1257E26.00486BB2@de.ibm.com>
References: <OFE5AC06CE.E75B85AD-ONC1257E26.00399901-C1257E26.00486BB2@de.ibm.com>
Message-ID: <CAMYZk=eaQpwRcW6RM7WK9x4v8wO=o=GmZPrjCCq=dKa9K7H4Bg@mail.gmail.com>

Thanks for your replies. I can definitely appreciate the the goal of
improving management of components, and I agree that if GPFS will be using
within other products (which it is and will continue to be), then it would
be great for those products to be able to manage GPFS via an interface.

My fear with the idea of the above mentioned as a prereq is that the
"improvement" of management might looks like an improvement when you only
have 1 or 2 tiers of storage and 1 or 2 filesets called "database" and
another fileset called "user files", but if you have hundreds or thousands
of filesets and several tiers of storage, with both GSS and non-GSS systems
in the same cluster, then the GUI may actually be more cumbersome than the
original method. So, I just want to voice an opinion that we should
continue to be able to configure / maintain / monitor GPFS in a programatic
/ scriptable non-point-and-click way, if possible.



On Mon, Apr 13, 2015 at 9:11 AM, Christian Bolik <BOLIK at de.ibm.com> wrote:

>
> Hi Zach,
>
> I'm not aware of any intent to make TPC or any other Tivoli/IBM product a
> prereq for GPFS, and I don't think any such plans exist.  Rather, as Dean
> also pointed out, we're investing work to improve integration of
> GPFS/Spectrum Scale into other products being members of the newly
> announced IBM Spectrum Storage family, with the goal of improving
> manageability of the individual components (rather than worsening it...).
>
> Cheers,
> Christian
>
> > Christian:
> > Interesting and thanks for the latest news.
> >
> > May I ask: Is there an intent moving forward that TPC and / or other
> Tivoli
> > products will be a required part of GPFS?
> > The concern I have is that GPFS is pretty straightforward at the moment
> and
> > has very logical requirements to operate (min servers, quorum, etc),
> > whereas there are many IBM products that require two or three more
> servers
> > just to manage the servers managing the service.. too much. It would be
> > nice to make sure, going forward, that the core of GPFS can still
> function
> > without additional web servers, Java, a suite of middleware, and a
> handful
> > of DB2 instance .. :)
> >
> > -Zach
>
> Christian Bolik
>
> Software Defined Storage Development
>
> IBM Deutschland Research & Development GmbH, Hechtsheimer Str. 2, 55131
> Mainz, Germany
> Vorsitzende des Aufsichtsrats: Martina Koederitz
> Gesch?ftsf?hrung: Dirk Wittkopp
> Sitz der Gesellschaft: B?blingen / Registergericht: Amtsgericht Stuttgart,
> HRB 243294
>
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>



-- 
Zach Giles
zgiles at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150413/ec65870f/attachment.htm>

From jamiedavis at us.ibm.com  Tue Apr 14 13:23:59 2015
From: jamiedavis at us.ibm.com (James Davis)
Date: Tue, 14 Apr 2015 08:23:59 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <mailman.1.1429009201.21161.gpfsug-discuss@gpfsug.org>
References: <mailman.1.1429009201.21161.gpfsug-discuss@gpfsug.org>
Message-ID: <OF41183321.B20BDD25-ON85257E27.00438322-85257E27.00441D50@us.ibm.com>


Zach,

Not sure if I'm formatting this message properly (need to switch off of
digest mode), but I know of no plans to replace the GPFS command line
interface with a GUI. With the GPFS GUI, TPC monitoring, etc., we want to
enable a wider variety of users to effectively use and manage GPFS, but the
command line will of course remain available for power users,
script-writers, etc. I certainly don't intend to throw away any of my nice
mm-themed test programs :-)

Jamie Davis
GPFS Test

Date: Mon, 13 Apr 2015 12:05:11 -0400
From: Zachary Giles <zgiles at gmail.com>
To: gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Subject: Re: [gpfsug-discuss] Monitoring capacity and health status
		 for a multitude of GPFS clusters
Message-ID:

<CAMYZk=eaQpwRcW6RM7WK9x4v8wO=o=GmZPrjCCq=dKa9K7H4Bg at mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

Thanks for your replies. I can definitely appreciate the the goal of
improving management of components, and I agree that if GPFS will be using
within other products (which it is and will continue to be), then it would
be great for those products to be able to manage GPFS via an interface.

My fear with the idea of the above mentioned as a prereq is that the
"improvement" of management might looks like an improvement when you only
have 1 or 2 tiers of storage and 1 or 2 filesets called "database" and
another fileset called "user files", but if you have hundreds or thousands
of filesets and several tiers of storage, with both GSS and non-GSS systems
in the same cluster, then the GUI may actually be more cumbersome than the
original method. So, I just want to voice an opinion that we should
continue to be able to configure / maintain / monitor GPFS in a programatic
/ scriptable non-point-and-click way, if possible.


Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150414/9352bfbe/attachment.htm>

From ecgarris at iu.edu  Thu Apr 16 21:33:59 2015
From: ecgarris at iu.edu (Garrison, E Chris)
Date: Thu, 16 Apr 2015 20:33:59 +0000
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
Message-ID: <D15594F5.1ADED%ecgarris@iu.edu>

Hello,

My site is working up to upgrading our paired GridScaler system from GPFS 3.5 to 4.1. There is a mandate to provide encryption at rest, and that's an advertised feature of 4.1. We already have over 100 TB of data, synchronously replicated between two geographically separated sites, and we have concerns about how the upgrade, as well as the application of encryption to all that data, will go.

I'd like to hear from admins who've been through this upgrade. What gotchas should we look out for? Can it easily be done in place, or would we need some extra equipment to "slosh" our data to and from so that it is written to an encrypted GPFS?

Thank you for your time, and for any sage advice on this process.

Chris
--
Chris Garrison
Indiana University
Research Systems Storage
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150416/aeed3961/attachment.htm>

From zgiles at gmail.com  Thu Apr 16 21:39:51 2015
From: zgiles at gmail.com (Zachary Giles)
Date: Thu, 16 Apr 2015 16:39:51 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <OF41183321.B20BDD25-ON85257E27.00438322-85257E27.00441D50@us.ibm.com>
References: <mailman.1.1429009201.21161.gpfsug-discuss@gpfsug.org>
	<OF41183321.B20BDD25-ON85257E27.00438322-85257E27.00441D50@us.ibm.com>
Message-ID: <CAMYZk=dni=XHyMAcro6Y74tW+F-sVBRU4BCUm3vGOdxFDjOuaA@mail.gmail.com>

Thanks Jamie. I appreciate the input. Glad to hear it.

On Tue, Apr 14, 2015 at 8:23 AM, James Davis <jamiedavis at us.ibm.com> wrote:

> Zach,
>
> Not sure if I'm formatting this message properly (need to switch off of
> digest mode), but I know of no plans to replace the GPFS command line
> interface with a GUI. With the GPFS GUI, TPC monitoring, etc., we want to
> enable a wider variety of users to effectively use and manage GPFS, but the
> command line will of course remain available for power users,
> script-writers, etc. I certainly don't intend to throw away any of my nice
> mm-themed test programs :-)
>
> Jamie Davis
> GPFS Test
>
> Date: Mon, 13 Apr 2015 12:05:11 -0400
> From: Zachary Giles <zgiles at gmail.com>
> To: gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
> Subject: Re: [gpfsug-discuss] Monitoring capacity and health status
> for a multitude of GPFS clusters
> Message-ID:
> <CAMYZk=eaQpwRcW6RM7WK9x4v8wO=o=GmZPrjCCq=dKa9K7H4Bg at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Thanks for your replies. I can definitely appreciate the the goal of
> improving management of components, and I agree that if GPFS will be using
> within other products (which it is and will continue to be), then it would
> be great for those products to be able to manage GPFS via an interface.
>
> My fear with the idea of the above mentioned as a prereq is that the
> "improvement" of management might looks like an improvement when you only
> have 1 or 2 tiers of storage and 1 or 2 filesets called "database" and
> another fileset called "user files", but if you have hundreds or thousands
> of filesets and several tiers of storage, with both GSS and non-GSS systems
> in the same cluster, then the GUI may actually be more cumbersome than the
> original method. So, I just want to voice an opinion that we should
> continue to be able to configure / maintain / monitor GPFS in a programatic
> / scriptable non-point-and-click way, if possible.
>
>
> Jamie Davis
> GPFS Functional Verification Test (FVT)
> jamiedavis at us.ibm.com
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>
>


-- 
Zach Giles
zgiles at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150416/da9891e5/attachment.htm>

From fleers at gmail.com  Thu Apr 16 23:21:34 2015
From: fleers at gmail.com (Frank Leers)
Date: Thu, 16 Apr 2015 15:21:34 -0700
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <D15594F5.1ADED%ecgarris@iu.edu>
References: <D15594F5.1ADED%ecgarris@iu.edu>
Message-ID: <CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>

Hi Chris,

Since you mention GRIDScaler, I assume that you are a DDN customer and that
you have a support contract with them.  If not, then feel free to take the
advice that follows with a measure of salt although it mostly still applies
;-)
With 4.1, there are now 'Editions' of GPFS, which break out various feature
sets into a tiered arrangement, with each tier being licensed (and possibly
priced) differently.
Have a look here (Q 1.3) for the 4.1 licensing notes -
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html


With GPFS 3.5, you are most likely running the equivalent of the 'Standard
Edition' today.  The crypto feature set comes with the 'Advanced Edition',
which is licensed differently.

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a primer
...
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html

-frank


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> wrote:

>  Hello,
>
>  My site is working up to upgrading our paired GridScaler system from
> GPFS 3.5 to 4.1. There is a mandate to provide encryption at rest, and
> that's an advertised feature of 4.1. We already have over 100 TB of data,
> synchronously replicated between two geographically separated sites, and we
> have concerns about how the upgrade, as well as the application of
> encryption to all that data, will go.
>
>  I'd like to hear from admins who've been through this upgrade. What
> gotchas should we look out for? Can it easily be done in place, or would we
> need some extra equipment to "slosh" our data to and from so that it is
> written to an encrypted GPFS?
>
>  Thank you for your time, and for any sage advice on this process.
>
>  Chris
> --
> Chris Garrison
> Indiana University
> Research Systems Storage
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150416/38ac4616/attachment.htm>

From daniel.kidger at uk.ibm.com  Fri Apr 17 13:15:40 2015
From: daniel.kidger at uk.ibm.com (Daniel Kidger)
Date: Fri, 17 Apr 2015 13:15:40 +0100
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
References: <D15594F5.1ADED%ecgarris@iu.edu>
	<CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
Message-ID: <OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>

Hi Chris,

The Crypto feature of GPFS 4.1  requires the addition of just one extra 
RPM over the standard edition. Other RPMs remain the same.
It also requires licenses for  "Advanced Edition". These are of the order 
of 30% more expensive than the standard edition.

The model for GPFS encryption at rest is that the client node fetches the 
encrypted file from the fileserver. 
The file remains encrypted in transfer and is only decrypted by the client 
node using a key it holds.
A result is end-to-end encryption as well as encryption of data at rest,

Encryption can be on a per file level if desired. Hence a way to migrate 
from an existing non-encrypted. setup. 
note inodes should be 4kB to make sure there is enough room to store the 
encryption attributes.

A side effect of adding encryption is that you also need additional remote 
key management (RKM) servers to handle the key management. These need 
licensed software.
see 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/com.ibm.cluster.gpfs.v4r104.gpfs200.doc/bl1adv_encryptionsetupreqs.htm

I am sure DDN can help you with all of this.

Hope this helps,
Daniel



Dr.Daniel Kidger
No. 1 The Square,

Technical Specialist  SDI (formerly Platform Computing)
Temple Quay,
Bristol  BS1 6DG
Mobile:
+44-07818 522 266
United Kingdom
Landline:
+44-02392 564 121 (Internal ITN 3726 9250)
 

e-mail:
daniel.kidger at uk.ibm.com
 






From:   Frank Leers <fleers at gmail.com>
To:     gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:   16/04/2015 23:21
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to 
GPFS 4.1?
Sent by:        gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

Since you mention GRIDScaler, I assume that you are a DDN customer and 
that you have a support contract with them.  If not, then feel free to 
take the advice that follows with a measure of salt although it mostly 
still applies ;-)
With 4.1, there are now 'Editions' of GPFS, which break out various 
feature sets into a tiered arrangement, with each tier being licensed (and 
possibly priced) differently.
Have a look here (Q 1.3) for the 4.1 licensing notes - 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html
 

With GPFS 3.5, you are most likely running the equivalent of the 'Standard 
Edition' today.  The crypto feature set comes with the 'Advanced Edition', 
which is licensed differently.

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a primer 
...  
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html

-frank


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> 
wrote:
Hello,

My site is working up to upgrading our paired GridScaler system from GPFS 
3.5 to 4.1. There is a mandate to provide encryption at rest, and that's 
an advertised feature of 4.1. We already have over 100 TB of data, 
synchronously replicated between two geographically separated sites, and 
we have concerns about how the upgrade, as well as the application of 
encryption to all that data, will go.

I'd like to hear from admins who've been through this upgrade. What 
gotchas should we look out for? Can it easily be done in place, or would 
we need some extra equipment to "slosh" our data to and from so that it is 
written to an encrypted GPFS?

Thank you for your time, and for any sage advice on this process.

Chris
--
Chris Garrison
Indiana University
Research Systems Storage

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss


Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/a00ff70c/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 360 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/a00ff70c/attachment.gif>

From jamiedavis at us.ibm.com  Fri Apr 17 15:11:24 2015
From: jamiedavis at us.ibm.com (James Davis)
Date: Fri, 17 Apr 2015 10:11:24 -0400
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>
References: <D15594F5.1ADED%ecgarris@iu.edu>	<CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
	<OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>
Message-ID: <OFF10E0A53.76B236A9-ON85257E2A.004D54A7-85257E2A.004DF2E0@us.ibm.com>


Hi Chris,

Based on your question I think you are aware of this already, but just in
case...

There is not currently an encrypt-in-place solution for GPFS encryption. A
file's encryption state is determined at create-time. In order to encrypt
your existing 100TB of data, you will need to apply an encryption policy to
the current (or a new) GPFS file system(s) and then do something like:
  cp file file.enc #now file.new is encrypted
  mv file.enc file #replace the unencrypted file with the encrypted file

This can be done in parallel using mmapplypolicy if you want. In 4.1.1
(forthcoming) I plan to provide an improved version of
the /usr/lpp/mmfs/samples/ilm/mmfind (a find-esque interface to
mmapplypolicy) that shipped in the last release; this should be an
effective tool for the job.

Cheers,

Jamie

Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com



From:	Daniel Kidger <daniel.kidger at uk.ibm.com>
To:	gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:	17-04-15 08:16 AM
Subject:	Re: [gpfsug-discuss] Experiences upgrading in place to GPFS
            4.1?
Sent by:	gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

The Crypto feature of GPFS 4.1  requires the addition of just one extra RPM
over the standard edition. Other RPMs remain the same.
It also requires licenses for  "Advanced Edition". These are of the order
of 30% more expensive than the standard edition.

The model for GPFS encryption at rest is that the client node fetches the
encrypted file from the fileserver.
The file remains encrypted in transfer and is only decrypted by the client
node using a key it holds.
A result is end-to-end encryption as well as encryption of data at rest,

Encryption can be on a per file level if desired. Hence a way to migrate
from an existing non-encrypted. setup.
note inodes should be 4kB to make sure there is enough room to store the
encryption attributes.

A side effect of adding encryption is that you also need additional remote
key management (RKM) servers to handle the key management. These need
licensed software.
see
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/com.ibm.cluster.gpfs.v4r104.gpfs200.doc/bl1adv_encryptionsetupreqs.htm


I am sure DDN can help you with all of this.

Hope this helps,
Daniel
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
 Dr.Daniel                                   No. 1 The Square,                  
 Kidger                                                                         
                                                                                
 Technical                                   Temple Quay,                       
 Specialist                                  Bristol  BS1 6DG                   
 SDI                                                                            
 (formerly                                                                      
 Platform                                                                       
 Computing)                                                                     
                                                                                
                                                                                
 Mobile:    +44-07818 522 266                United Kingdom                     
                                                                                
 Landline:  +44-02392 564 121 (Internal ITN                                     
            3726 9250)                                                          
                                                                                
 e-mail:    daniel.kidger at uk.ibm.com                                            
                                                                                
                                                                                
                                                                                






From:        Frank Leers <fleers at gmail.com>
To:        gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:        16/04/2015 23:21
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to GPFS
4.1?
Sent by:        gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

Since you mention GRIDScaler, I assume that you are a DDN customer and that
you have a support contract with them.  If not, then feel free to take the
advice that follows with a measure of salt although it mostly still
applies ;-)
With 4.1, there are now 'Editions' of GPFS, which break out various feature
sets into a tiered arrangement, with each tier being licensed (and possibly
priced) differently.
Have a look here (Q 1.3) for the 4.1 licensing notes -
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html


With GPFS 3.5, you are most likely running the equivalent of the 'Standard
Edition' today.  The crypto feature set comes with the 'Advanced Edition',
which is licensed differently.

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a
primer ...
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html


-frank


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> wrote:

Hello,

My site is working up to upgrading our paired GridScaler system from GPFS
3.5 to 4.1. There is a mandate to provide encryption at rest, and that's an
advertised feature of 4.1. We already have over 100 TB of data,
synchronously replicated between two geographically separated sites, and we
have concerns about how the upgrade, as well as the application of
encryption to all that data, will go.

I'd like to hear from admins who've been through this upgrade. What gotchas
should we look out for? Can it easily be done in place, or would we need
some extra equipment to "slosh" our data to and from so that it is written
to an encrypted GPFS?

Thank you for your time, and for any sage advice on this process.

Chris
--
Chris Garrison
Indiana University
Research Systems Storage

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss


Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number
741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ecblank.gif
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment-0001.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 20386063.gif
Type: image/gif
Size: 360 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment-0002.gif>

From makaplan at us.ibm.com  Fri Apr 17 15:37:55 2015
From: makaplan at us.ibm.com (Marc A Kaplan)
Date: Fri, 17 Apr 2015 10:37:55 -0400
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <OFF10E0A53.76B236A9-ON85257E2A.004D54A7-85257E2A.004DF2E0@us.ibm.com>
References: <D15594F5.1ADED%ecgarris@iu.edu>	<CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
	<OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>
	<OFF10E0A53.76B236A9-ON85257E2A.004D54A7-85257E2A.004DF2E0@us.ibm.com>
Message-ID: <OF1D7F6052.1D03F850-ON85257E2A.004EEEB1-85257E2A.0050624D@us.ibm.com>

An "Encrypt-in-place" feature would not survive a serious, paranoid 
security audit.

If sensitive data ever was written to a disk, then the paranoid would say 
you can never be 100% sure that you have erased it.

That said, you decide how worried or paranoid you'd like to be and you can 
do an almost in place encryption as James Davis suggests by
simply copying the unencrypted file to a new file that will be subject to 
a GPFS encryption policy (SET ENCRYPTION) rule.

The safest way would be to copy-encrypt all the data to a new file system 
and then crush all the equipment that was used to store the clear-text 
files.

If crushing is too extreme, you might settle for a multipass sector by 
sector soft scrubbing by writing both carefully chosen and random data 
patterns.

Even then, unless you trust the manufacturer AND the manufacturer has 
provided you the means to "scrub" all the tracks/sectors of the disk you 
won't be sure...  Suppose that some sensitive data was written to a sector 
number that was later declared (partially) defective and remapped by the 
disk micro-code, and the original sector is put in a bad sector list that 
you can no longer address with standard disk driver software... 

Or it could be on some NVRAM or a disk that a service technician swapped 
out...  Ooops... it went out the door... and is now in the hands of the 
bad guys...





From:   James Davis/Poughkeepsie/IBM at IBMUS
To:     gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:   04/17/2015 10:12 AM
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to 
GPFS 4.1?
Sent by:        gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

Based on your question I think you are aware of this already, but just in 
case...

There is not currently an encrypt-in-place solution for GPFS encryption. A 
file's encryption state is determined at create-time. In order to encrypt 
your existing 100TB of data, you will need to apply an encryption policy 
to the current (or a new) GPFS file system(s) and then do something like:
  cp file file.enc #now file.new is encrypted
  mv file.enc file #replace the unencrypted file with the encrypted file

This can be done in parallel using mmapplypolicy if you want. In 4.1.1 
(forthcoming) I plan to provide an improved version of the 
/usr/lpp/mmfs/samples/ilm/mmfind (a find-esque interface to mmapplypolicy) 
that shipped in the last release; this should be an effective tool for the 
job.

Cheers,

Jamie

Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com

Daniel Kidger ---17-04-2015 08:16:54 AM---Hi Chris, The Crypto feature of 
GPFS 4.1  requires the addition of just one extra

From: Daniel Kidger <daniel.kidger at uk.ibm.com>
To: gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date: 17-04-15 08:16 AM
Subject: Re: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
Sent by: gpfsug-discuss-bounces at gpfsug.org



Hi Chris, 

The Crypto feature of GPFS 4.1  requires the addition of just one extra 
RPM over the standard edition. Other RPMs remain the same. 
It also requires licenses for  "Advanced Edition". These are of the order 
of 30% more expensive than the standard edition. 

The model for GPFS encryption at rest is that the client node fetches the 
encrypted file from the fileserver. 
The file remains encrypted in transfer and is only decrypted by the client 
node using a key it holds. 
A result is end-to-end encryption as well as encryption of data at rest, 

Encryption can be on a per file level if desired. Hence a way to migrate 
from an existing non-encrypted. setup. 
note inodes should be 4kB to make sure there is enough room to store the 
encryption attributes. 

A side effect of adding encryption is that you also need additional remote 
key management (RKM) servers to handle the key management. These need 
licensed software. 
see 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/com.ibm.cluster.gpfs.v4r104.gpfs200.doc/bl1adv_encryptionsetupreqs.htm 


I am sure DDN can help you with all of this. 

Hope this helps, 
Daniel 


Dr.Daniel Kidger 
No. 1 The Square, 

Technical Specialist  SDI (formerly Platform Computing) 
Temple Quay,
Bristol  BS1 6DG 
Mobile: 
+44-07818 522 266 
United Kingdom 
Landline: 
+44-02392 564 121 (Internal ITN 3726 9250) 
  

e-mail: 
daniel.kidger at uk.ibm.com 
 







From:        Frank Leers <fleers at gmail.com> 
To:        gpfsug main discussion list <gpfsug-discuss at gpfsug.org> 
Date:        16/04/2015 23:21 
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to 
GPFS 4.1? 
Sent by:        gpfsug-discuss-bounces at gpfsug.org 



Hi Chris, 

Since you mention GRIDScaler, I assume that you are a DDN customer and 
that you have a support contract with them.  If not, then feel free to 
take the advice that follows with a measure of salt although it mostly 
still applies ;-) 
With 4.1, there are now 'Editions' of GPFS, which break out various 
feature sets into a tiered arrangement, with each tier being licensed (and 
possibly priced) differently. 
Have a look here (Q 1.3) for the 4.1 licensing notes - 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html 
 

With GPFS 3.5, you are most likely running the equivalent of the 'Standard 
Edition' today.  The crypto feature set comes with the 'Advanced Edition', 
which is licensed differently. 

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a primer 
...  
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html 


-frank 


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> 
wrote: 
Hello, 

My site is working up to upgrading our paired GridScaler system from GPFS 
3.5 to 4.1. There is a mandate to provide encryption at rest, and that's 
an advertised feature of 4.1. We already have over 100 TB of data, 
synchronously replicated between two geographically separated sites, and 
we have concerns about how the upgrade, as well as the application of 
encryption to all that data, will go. 

I'd like to hear from admins who've been through this upgrade. What 
gotchas should we look out for? Can it easily be done in place, or would 
we need some extra equipment to "slosh" our data to and from so that it is 
written to an encrypted GPFS? 

Thank you for your time, and for any sage advice on this process. 

Chris 
-- 
Chris Garrison 
Indiana University 
Research Systems Storage 

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss


Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 21994 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0001.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0002.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 360 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0003.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0004.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0005.gif>

From secretary at gpfsug.org  Fri Apr 24 15:05:25 2015
From: secretary at gpfsug.org (Secretary GPFS UG)
Date: Fri, 24 Apr 2015 15:05:25 +0100
Subject: [gpfsug-discuss] GPFS User Group Meeting Agenda
Message-ID: <ebd76ace58f83d838895e0751950193b@webmail.gpfsug.org>

Dear Members,

The agenda for the next GPFS User Group Meeting is now available:
10:00 arrival for a 10:30 start
10:30                 Introductions - Jez Tucker, Group Chair & Claire 
Robson, Group Secretary
10:35                 Keynote - Doris Conti
10:50                 4.1.1 Roadmap / High-level futures - Scott Fadden
11:40                 Failure Events, Recovery & Problem determination - 
Scott Fadden
12:00                 Monitoring IBM Spectrum Scale using IBM Spectrum 
Control (VSC/TPC) - Christian Bolik
12:30                 Lunch
13:00                 User Experience from University of Birmingham & 
CLIMB - Simon Thompson
13:20                 User Experience from NERSC - Jason Hick
13:40                 AFM & Async DR Use Cases - Shankar Balasubramanian
14:25                 mmbackup + TSM Integration - Stefan Bender
15:10                 Break
15:25                 IBM Spectrum Scale (formerly GPFS) performance 
update, protocol performance and sizing - Sven Oehme
16:50                 Closing summary - Jez Tucker, Group Chair & Claire 
Robson, Group Secretary
17:00                 Ends

All attendees are invited to attend:
19:00 Buffet and drinks at York National Railway Museum

There are still places available for the event. If you have not 
registered and would like to attend, please email me, 
secretary at gpfsug.org with your name, job title, organisation, telephone 
and any dietary requirements.

We hope to see you in May!

Kind regards,

Claire

-- 
Claire Robson
GPFS User Group Secretary



From Dan.Foster at bristol.ac.uk  Wed Apr 29 08:51:03 2015
From: Dan.Foster at bristol.ac.uk (Dan Foster)
Date: Wed, 29 Apr 2015 08:51:03 +0100
Subject: [gpfsug-discuss] file system format version / software version
	compatibility matrix
Message-ID: <CALRUX+gQ2ZfN+7gWsadSUi2gUG+PaUkH0OrpEXAYDAUY-_LxRg@mail.gmail.com>

Hi All,

I'm trying to determine which versions of the GPFS file system format
are compatible with certain versions of the GPFS software.
Specifically in this instance I'm interested if file system format
v11.05 (from GPFS 3.3.0.2) is compatible with GPFS 3.5 .

The "File system format changes between versions of GPFS" [1] chapter
mentions format levels as old a v6, which infers that they are
compatible. But it would be useful to know for certain.

Thanks,
Dan.

[1] http://www-01.ibm.com/support/knowledgecenter/SSFKCN_3.5.0/com.ibm.cluster.gpfs.v3r5.gpfs100.doc/bl1adm_fsmigissues.htm?lang=en
-- 
Dan Foster | Senior Storage Systems Administrator | IT Services
e: dan.foster at bristol.ac.uk | t:  0117 3941170 [x41170]
m: Advanced Computing Research Centre, University of Bristol,
    8-10 Berkeley Square, Bristol BS8 1HH


From jamiedavis at us.ibm.com  Wed Apr 29 13:43:08 2015
From: jamiedavis at us.ibm.com (James Davis)
Date: Wed, 29 Apr 2015 08:43:08 -0400
Subject: [gpfsug-discuss] file system format version / software
	version	compatibility matrix
In-Reply-To: <CALRUX+gQ2ZfN+7gWsadSUi2gUG+PaUkH0OrpEXAYDAUY-_LxRg@mail.gmail.com>
References: <CALRUX+gQ2ZfN+7gWsadSUi2gUG+PaUkH0OrpEXAYDAUY-_LxRg@mail.gmail.com>
Message-ID: <OF27DFBC25.AE010664-ON85257E36.00453E1C-85257E36.0045DD6B@us.ibm.com>


Dan,

A "new" GPFS should be able to mount and interact with file systems with
"old" versions. Specifically I do not believe you will have trouble getting
GPFS 3.5 to talk to an FS created on GPFS 3.3.0.2.

The "Concepts, planning, and install guide" provides information about
migrating to GPFS 4.1 from GPFS 3.2 or earlier; this would implies file
systems created at GPFS 3.2 being supported on GPFS 4.1, which is even more
compatibility than you are asking about. The section to which I refer is
titled "Migrating to GPFS 4.1 from GPFS 3.2 or earlier releases of GPFS".

Note that without running mmchfs -V {full|compat} you will not have access
to some of the newer GPFS features. See the section titled "Completing the
migration to a new level of GPFS", also in the concepts guide.

Hope this helps,

Jamie

Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com



From:	Dan Foster <Dan.Foster at bristol.ac.uk>
To:	gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:	29-04-15 03:51 AM
Subject:	[gpfsug-discuss] file system format version / software version
            compatibility matrix
Sent by:	gpfsug-discuss-bounces at gpfsug.org



Hi All,

I'm trying to determine which versions of the GPFS file system format
are compatible with certain versions of the GPFS software.
Specifically in this instance I'm interested if file system format
v11.05 (from GPFS 3.3.0.2) is compatible with GPFS 3.5 .

The "File system format changes between versions of GPFS" [1] chapter
mentions format levels as old a v6, which infers that they are
compatible. But it would be useful to know for certain.

Thanks,
Dan.

[1]
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_3.5.0/com.ibm.cluster.gpfs.v3r5.gpfs100.doc/bl1adm_fsmigissues.htm?lang=en

--
Dan Foster | Senior Storage Systems Administrator | IT Services
e: dan.foster at bristol.ac.uk | t:  0117 3941170 [x41170]
m: Advanced Computing Research Centre, University of Bristol,
    8-10 Berkeley Square, Bristol BS8 1HH
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150429/23753578/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150429/23753578/attachment.gif>

From pavel.pokorny at datera.cz  Thu Apr  9 22:00:07 2015
From: pavel.pokorny at datera.cz (Pavel Pokorny)
Date: Thu, 9 Apr 2015 23:00:07 +0200
Subject: [gpfsug-discuss] Share GPFS from Windows?
Message-ID: <CAHy3nHJ9=D_zYq3xxFO8NnJutdy7JAWxkN7s_PZds277Nx4rxA@mail.gmail.com>

Hello to all,
Is there any technical reason why it is not supported to export GPFS from
Windows nodes using CIFS? As stated in GPFs FAQ (
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html
):

*Exporting GPFS file systems as Server Message Block (SMB) shares (also
known as CIFS shares) from GPFS Windows nodes is not supported.*

Or it is limitation more related to licensing and business point of view.
Thank you for answers,
Pavel

--
Ing. Pavel Pokorn?
DATERA s.r.o. | Ovocn? trh 580/2 | Praha | Czech Republic
www.datera.cz | Mobil: +420 602 357 194 | E-mail: pavel.pokorny at datera.cz
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150409/6f6c75d7/attachment-0001.htm>

From BOLIK at de.ibm.com  Fri Apr 10 12:24:14 2015
From: BOLIK at de.ibm.com (Christian Bolik)
Date: Fri, 10 Apr 2015 13:24:14 +0200
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
	multitude of GPFS clusters
Message-ID: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>


Just wanted to let you know that recently GPFS support has been added to
TPC, which is IBM's Tivoli Storage Productivity Center (soon to be renamed
to IBM Spectrum Control). As of now, TPC allows GPFS administrators to get
answers to the following questions, across any number of GPFS clusters
which have been added to TPC:

- Which of my clusters are running out of free space?
- Which of my clusters or nodes have a health problem?
- Which file systems and pools are running out of capacity?
- Which file systems are mounted on which nodes?
- How much space is occupied by snapshots? Are there any very old,
potentially obsolete ones?
- Which quotas are close to being exceeded or have already been exceeded?
- Which filesets are close to running out of free inodes?
- Which NSDs are at risk of becoming unavailable, or are unavailable?
- Are the volumes backing my NSDs performing OK?
- Are all nodes fulfilling critical roles in the cluster up and running?
- How can I be notified when nodes go offline or file systems fill up
beyond a threshold?

There's a short 6-minute video available on YouTube which shows how TPC
helps answering these questions:
https://www.youtube.com/watch?v=8Esk5U_cYw8&feature=youtu.be

For more information about TPC, please check out the product wiki on
developerWorks: http://ibm.co/1adWNFK

Thanks,
Christian Bolik
IBM Storage Software Development




From zgiles at gmail.com  Fri Apr 10 16:27:36 2015
From: zgiles at gmail.com (Zachary Giles)
Date: Fri, 10 Apr 2015 11:27:36 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>
References: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>
Message-ID: <CAMYZk=fKkAgeR-MW79a=BO46WpR9+Gm+=_kCLkQo4+v5KxgKqg@mail.gmail.com>

Christian:
Interesting and thanks for the latest news.

May I ask: Is there an intent moving forward that TPC and / or other Tivoli
products will be a required part of GPFS?
The concern I have is that GPFS is pretty straightforward at the moment and
has very logical requirements to operate (min servers, quorum, etc),
whereas there are many IBM products that require two or three more servers
just to manage the servers managing the service.. too much. It would be
nice to make sure, going forward, that the core of GPFS can still function
without additional web servers, Java, a suite of middleware, and a handful
of DB2 instance .. :)

-Zach



On Fri, Apr 10, 2015 at 7:24 AM, Christian Bolik <BOLIK at de.ibm.com> wrote:

>
> Just wanted to let you know that recently GPFS support has been added to
> TPC, which is IBM's Tivoli Storage Productivity Center (soon to be renamed
> to IBM Spectrum Control). As of now, TPC allows GPFS administrators to get
> answers to the following questions, across any number of GPFS clusters
> which have been added to TPC:
>
> - Which of my clusters are running out of free space?
> - Which of my clusters or nodes have a health problem?
> - Which file systems and pools are running out of capacity?
> - Which file systems are mounted on which nodes?
> - How much space is occupied by snapshots? Are there any very old,
> potentially obsolete ones?
> - Which quotas are close to being exceeded or have already been exceeded?
> - Which filesets are close to running out of free inodes?
> - Which NSDs are at risk of becoming unavailable, or are unavailable?
> - Are the volumes backing my NSDs performing OK?
> - Are all nodes fulfilling critical roles in the cluster up and running?
> - How can I be notified when nodes go offline or file systems fill up
> beyond a threshold?
>
> There's a short 6-minute video available on YouTube which shows how TPC
> helps answering these questions:
> https://www.youtube.com/watch?v=8Esk5U_cYw8&feature=youtu.be
>
> For more information about TPC, please check out the product wiki on
> developerWorks: http://ibm.co/1adWNFK
>
> Thanks,
> Christian Bolik
> IBM Storage Software Development
>
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>



-- 
Zach Giles
zgiles at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150410/c4ee09f6/attachment-0001.htm>

From dhildeb at us.ibm.com  Fri Apr 10 18:45:14 2015
From: dhildeb at us.ibm.com (Dean Hildebrand)
Date: Fri, 10 Apr 2015 10:45:14 -0700
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <CAMYZk=fKkAgeR-MW79a=BO46WpR9+Gm+=_kCLkQo4+v5KxgKqg@mail.gmail.com>
References: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>
	<CAMYZk=fKkAgeR-MW79a=BO46WpR9+Gm+=_kCLkQo4+v5KxgKqg@mail.gmail.com>
Message-ID: <OFDFD967F3.9188C435-ON88257E23.0060AE82-88257E23.006186C8@us.ibm.com>


Hi Zach,

The summary is that GPFS is being integrated much more across the
portfolio...

With GPFS itself, there is a video below demonstrating the ESS/GSS GUI and
monitoring feature that is in the product today.  Moving forward, as you
can probably see, there is a push in IBM to move GPFS to software-defined,
which includes features such as the GUI as well...
https://www.youtube.com/watch?v=Mv9Sn-VYoGU

Dean




From:	Zachary Giles <zgiles at gmail.com>
To:	gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:	04/10/2015 08:27 AM
Subject:	Re: [gpfsug-discuss] Monitoring capacity and health status for
            a multitude of GPFS clusters
Sent by:	gpfsug-discuss-bounces at gpfsug.org



Christian:
Interesting and thanks for the latest news.

May I ask: Is there an intent moving forward that TPC and / or other Tivoli
products will be a required part of GPFS?
The concern I have is that GPFS is pretty straightforward at the moment and
has very logical requirements to operate (min servers, quorum, etc),
whereas there are many IBM products that require two or three more servers
just to manage the servers managing the service.. too much. It would be
nice to make sure, going forward, that the core of GPFS can still function
without additional web servers, Java, a suite of middleware, and a handful
of DB2 instance .. :)

-Zach



On Fri, Apr 10, 2015 at 7:24 AM, Christian Bolik <BOLIK at de.ibm.com> wrote:

  Just wanted to let you know that recently GPFS support has been added to
  TPC, which is IBM's Tivoli Storage Productivity Center (soon to be
  renamed
  to IBM Spectrum Control). As of now, TPC allows GPFS administrators to
  get
  answers to the following questions, across any number of GPFS clusters
  which have been added to TPC:

  - Which of my clusters are running out of free space?
  - Which of my clusters or nodes have a health problem?
  - Which file systems and pools are running out of capacity?
  - Which file systems are mounted on which nodes?
  - How much space is occupied by snapshots? Are there any very old,
  potentially obsolete ones?
  - Which quotas are close to being exceeded or have already been exceeded?
  - Which filesets are close to running out of free inodes?
  - Which NSDs are at risk of becoming unavailable, or are unavailable?
  - Are the volumes backing my NSDs performing OK?
  - Are all nodes fulfilling critical roles in the cluster up and running?
  - How can I be notified when nodes go offline or file systems fill up
  beyond a threshold?

  There's a short 6-minute video available on YouTube which shows how TPC
  helps answering these questions:
  https://www.youtube.com/watch?v=8Esk5U_cYw8&feature=youtu.be

  For more information about TPC, please check out the product wiki on
  developerWorks: http://ibm.co/1adWNFK

  Thanks,
  Christian Bolik
  IBM Storage Software Development


  _______________________________________________
  gpfsug-discuss mailing list
  gpfsug-discuss at gpfsug.org
  http://gpfsug.org/mailman/listinfo/gpfsug-discuss



--
Zach Giles
zgiles at gmail.com_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150410/9afe8bc4/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150410/9afe8bc4/attachment-0001.gif>

From BOLIK at de.ibm.com  Mon Apr 13 14:11:01 2015
From: BOLIK at de.ibm.com (Christian Bolik)
Date: Mon, 13 Apr 2015 15:11:01 +0200
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
Message-ID: <OFE5AC06CE.E75B85AD-ONC1257E26.00399901-C1257E26.00486BB2@de.ibm.com>


Hi Zach,

I'm not aware of any intent to make TPC or any other Tivoli/IBM product a
prereq for GPFS, and I don't think any such plans exist.  Rather, as Dean
also pointed out, we're investing work to improve integration of
GPFS/Spectrum Scale into other products being members of the newly
announced IBM Spectrum Storage family, with the goal of improving
manageability of the individual components (rather than worsening it...).

Cheers,
Christian

> Christian:
> Interesting and thanks for the latest news.
>
> May I ask: Is there an intent moving forward that TPC and / or other
Tivoli
> products will be a required part of GPFS?
> The concern I have is that GPFS is pretty straightforward at the moment
and
> has very logical requirements to operate (min servers, quorum, etc),
> whereas there are many IBM products that require two or three more
servers
> just to manage the servers managing the service.. too much. It would be
> nice to make sure, going forward, that the core of GPFS can still
function
> without additional web servers, Java, a suite of middleware, and a
handful
> of DB2 instance .. :)
>
> -Zach

Christian Bolik

Software Defined Storage Development

IBM Deutschland Research & Development GmbH, Hechtsheimer Str. 2, 55131
Mainz, Germany
Vorsitzende des Aufsichtsrats: Martina Koederitz
Gesch?ftsf?hrung: Dirk Wittkopp
Sitz der Gesellschaft: B?blingen / Registergericht: Amtsgericht Stuttgart,
HRB 243294




From zgiles at gmail.com  Mon Apr 13 17:05:11 2015
From: zgiles at gmail.com (Zachary Giles)
Date: Mon, 13 Apr 2015 12:05:11 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <OFE5AC06CE.E75B85AD-ONC1257E26.00399901-C1257E26.00486BB2@de.ibm.com>
References: <OFE5AC06CE.E75B85AD-ONC1257E26.00399901-C1257E26.00486BB2@de.ibm.com>
Message-ID: <CAMYZk=eaQpwRcW6RM7WK9x4v8wO=o=GmZPrjCCq=dKa9K7H4Bg@mail.gmail.com>

Thanks for your replies. I can definitely appreciate the the goal of
improving management of components, and I agree that if GPFS will be using
within other products (which it is and will continue to be), then it would
be great for those products to be able to manage GPFS via an interface.

My fear with the idea of the above mentioned as a prereq is that the
"improvement" of management might looks like an improvement when you only
have 1 or 2 tiers of storage and 1 or 2 filesets called "database" and
another fileset called "user files", but if you have hundreds or thousands
of filesets and several tiers of storage, with both GSS and non-GSS systems
in the same cluster, then the GUI may actually be more cumbersome than the
original method. So, I just want to voice an opinion that we should
continue to be able to configure / maintain / monitor GPFS in a programatic
/ scriptable non-point-and-click way, if possible.



On Mon, Apr 13, 2015 at 9:11 AM, Christian Bolik <BOLIK at de.ibm.com> wrote:

>
> Hi Zach,
>
> I'm not aware of any intent to make TPC or any other Tivoli/IBM product a
> prereq for GPFS, and I don't think any such plans exist.  Rather, as Dean
> also pointed out, we're investing work to improve integration of
> GPFS/Spectrum Scale into other products being members of the newly
> announced IBM Spectrum Storage family, with the goal of improving
> manageability of the individual components (rather than worsening it...).
>
> Cheers,
> Christian
>
> > Christian:
> > Interesting and thanks for the latest news.
> >
> > May I ask: Is there an intent moving forward that TPC and / or other
> Tivoli
> > products will be a required part of GPFS?
> > The concern I have is that GPFS is pretty straightforward at the moment
> and
> > has very logical requirements to operate (min servers, quorum, etc),
> > whereas there are many IBM products that require two or three more
> servers
> > just to manage the servers managing the service.. too much. It would be
> > nice to make sure, going forward, that the core of GPFS can still
> function
> > without additional web servers, Java, a suite of middleware, and a
> handful
> > of DB2 instance .. :)
> >
> > -Zach
>
> Christian Bolik
>
> Software Defined Storage Development
>
> IBM Deutschland Research & Development GmbH, Hechtsheimer Str. 2, 55131
> Mainz, Germany
> Vorsitzende des Aufsichtsrats: Martina Koederitz
> Gesch?ftsf?hrung: Dirk Wittkopp
> Sitz der Gesellschaft: B?blingen / Registergericht: Amtsgericht Stuttgart,
> HRB 243294
>
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>



-- 
Zach Giles
zgiles at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150413/ec65870f/attachment-0001.htm>

From jamiedavis at us.ibm.com  Tue Apr 14 13:23:59 2015
From: jamiedavis at us.ibm.com (James Davis)
Date: Tue, 14 Apr 2015 08:23:59 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <mailman.1.1429009201.21161.gpfsug-discuss@gpfsug.org>
References: <mailman.1.1429009201.21161.gpfsug-discuss@gpfsug.org>
Message-ID: <OF41183321.B20BDD25-ON85257E27.00438322-85257E27.00441D50@us.ibm.com>


Zach,

Not sure if I'm formatting this message properly (need to switch off of
digest mode), but I know of no plans to replace the GPFS command line
interface with a GUI. With the GPFS GUI, TPC monitoring, etc., we want to
enable a wider variety of users to effectively use and manage GPFS, but the
command line will of course remain available for power users,
script-writers, etc. I certainly don't intend to throw away any of my nice
mm-themed test programs :-)

Jamie Davis
GPFS Test

Date: Mon, 13 Apr 2015 12:05:11 -0400
From: Zachary Giles <zgiles at gmail.com>
To: gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Subject: Re: [gpfsug-discuss] Monitoring capacity and health status
		 for a multitude of GPFS clusters
Message-ID:

<CAMYZk=eaQpwRcW6RM7WK9x4v8wO=o=GmZPrjCCq=dKa9K7H4Bg at mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

Thanks for your replies. I can definitely appreciate the the goal of
improving management of components, and I agree that if GPFS will be using
within other products (which it is and will continue to be), then it would
be great for those products to be able to manage GPFS via an interface.

My fear with the idea of the above mentioned as a prereq is that the
"improvement" of management might looks like an improvement when you only
have 1 or 2 tiers of storage and 1 or 2 filesets called "database" and
another fileset called "user files", but if you have hundreds or thousands
of filesets and several tiers of storage, with both GSS and non-GSS systems
in the same cluster, then the GUI may actually be more cumbersome than the
original method. So, I just want to voice an opinion that we should
continue to be able to configure / maintain / monitor GPFS in a programatic
/ scriptable non-point-and-click way, if possible.


Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150414/9352bfbe/attachment-0001.htm>

From ecgarris at iu.edu  Thu Apr 16 21:33:59 2015
From: ecgarris at iu.edu (Garrison, E Chris)
Date: Thu, 16 Apr 2015 20:33:59 +0000
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
Message-ID: <D15594F5.1ADED%ecgarris@iu.edu>

Hello,

My site is working up to upgrading our paired GridScaler system from GPFS 3.5 to 4.1. There is a mandate to provide encryption at rest, and that's an advertised feature of 4.1. We already have over 100 TB of data, synchronously replicated between two geographically separated sites, and we have concerns about how the upgrade, as well as the application of encryption to all that data, will go.

I'd like to hear from admins who've been through this upgrade. What gotchas should we look out for? Can it easily be done in place, or would we need some extra equipment to "slosh" our data to and from so that it is written to an encrypted GPFS?

Thank you for your time, and for any sage advice on this process.

Chris
--
Chris Garrison
Indiana University
Research Systems Storage
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150416/aeed3961/attachment-0001.htm>

From zgiles at gmail.com  Thu Apr 16 21:39:51 2015
From: zgiles at gmail.com (Zachary Giles)
Date: Thu, 16 Apr 2015 16:39:51 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <OF41183321.B20BDD25-ON85257E27.00438322-85257E27.00441D50@us.ibm.com>
References: <mailman.1.1429009201.21161.gpfsug-discuss@gpfsug.org>
	<OF41183321.B20BDD25-ON85257E27.00438322-85257E27.00441D50@us.ibm.com>
Message-ID: <CAMYZk=dni=XHyMAcro6Y74tW+F-sVBRU4BCUm3vGOdxFDjOuaA@mail.gmail.com>

Thanks Jamie. I appreciate the input. Glad to hear it.

On Tue, Apr 14, 2015 at 8:23 AM, James Davis <jamiedavis at us.ibm.com> wrote:

> Zach,
>
> Not sure if I'm formatting this message properly (need to switch off of
> digest mode), but I know of no plans to replace the GPFS command line
> interface with a GUI. With the GPFS GUI, TPC monitoring, etc., we want to
> enable a wider variety of users to effectively use and manage GPFS, but the
> command line will of course remain available for power users,
> script-writers, etc. I certainly don't intend to throw away any of my nice
> mm-themed test programs :-)
>
> Jamie Davis
> GPFS Test
>
> Date: Mon, 13 Apr 2015 12:05:11 -0400
> From: Zachary Giles <zgiles at gmail.com>
> To: gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
> Subject: Re: [gpfsug-discuss] Monitoring capacity and health status
> for a multitude of GPFS clusters
> Message-ID:
> <CAMYZk=eaQpwRcW6RM7WK9x4v8wO=o=GmZPrjCCq=dKa9K7H4Bg at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Thanks for your replies. I can definitely appreciate the the goal of
> improving management of components, and I agree that if GPFS will be using
> within other products (which it is and will continue to be), then it would
> be great for those products to be able to manage GPFS via an interface.
>
> My fear with the idea of the above mentioned as a prereq is that the
> "improvement" of management might looks like an improvement when you only
> have 1 or 2 tiers of storage and 1 or 2 filesets called "database" and
> another fileset called "user files", but if you have hundreds or thousands
> of filesets and several tiers of storage, with both GSS and non-GSS systems
> in the same cluster, then the GUI may actually be more cumbersome than the
> original method. So, I just want to voice an opinion that we should
> continue to be able to configure / maintain / monitor GPFS in a programatic
> / scriptable non-point-and-click way, if possible.
>
>
> Jamie Davis
> GPFS Functional Verification Test (FVT)
> jamiedavis at us.ibm.com
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>
>


-- 
Zach Giles
zgiles at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150416/da9891e5/attachment-0001.htm>

From fleers at gmail.com  Thu Apr 16 23:21:34 2015
From: fleers at gmail.com (Frank Leers)
Date: Thu, 16 Apr 2015 15:21:34 -0700
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <D15594F5.1ADED%ecgarris@iu.edu>
References: <D15594F5.1ADED%ecgarris@iu.edu>
Message-ID: <CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>

Hi Chris,

Since you mention GRIDScaler, I assume that you are a DDN customer and that
you have a support contract with them.  If not, then feel free to take the
advice that follows with a measure of salt although it mostly still applies
;-)
With 4.1, there are now 'Editions' of GPFS, which break out various feature
sets into a tiered arrangement, with each tier being licensed (and possibly
priced) differently.
Have a look here (Q 1.3) for the 4.1 licensing notes -
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html


With GPFS 3.5, you are most likely running the equivalent of the 'Standard
Edition' today.  The crypto feature set comes with the 'Advanced Edition',
which is licensed differently.

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a primer
...
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html

-frank


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> wrote:

>  Hello,
>
>  My site is working up to upgrading our paired GridScaler system from
> GPFS 3.5 to 4.1. There is a mandate to provide encryption at rest, and
> that's an advertised feature of 4.1. We already have over 100 TB of data,
> synchronously replicated between two geographically separated sites, and we
> have concerns about how the upgrade, as well as the application of
> encryption to all that data, will go.
>
>  I'd like to hear from admins who've been through this upgrade. What
> gotchas should we look out for? Can it easily be done in place, or would we
> need some extra equipment to "slosh" our data to and from so that it is
> written to an encrypted GPFS?
>
>  Thank you for your time, and for any sage advice on this process.
>
>  Chris
> --
> Chris Garrison
> Indiana University
> Research Systems Storage
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150416/38ac4616/attachment-0001.htm>

From daniel.kidger at uk.ibm.com  Fri Apr 17 13:15:40 2015
From: daniel.kidger at uk.ibm.com (Daniel Kidger)
Date: Fri, 17 Apr 2015 13:15:40 +0100
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
References: <D15594F5.1ADED%ecgarris@iu.edu>
	<CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
Message-ID: <OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>

Hi Chris,

The Crypto feature of GPFS 4.1  requires the addition of just one extra 
RPM over the standard edition. Other RPMs remain the same.
It also requires licenses for  "Advanced Edition". These are of the order 
of 30% more expensive than the standard edition.

The model for GPFS encryption at rest is that the client node fetches the 
encrypted file from the fileserver. 
The file remains encrypted in transfer and is only decrypted by the client 
node using a key it holds.
A result is end-to-end encryption as well as encryption of data at rest,

Encryption can be on a per file level if desired. Hence a way to migrate 
from an existing non-encrypted. setup. 
note inodes should be 4kB to make sure there is enough room to store the 
encryption attributes.

A side effect of adding encryption is that you also need additional remote 
key management (RKM) servers to handle the key management. These need 
licensed software.
see 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/com.ibm.cluster.gpfs.v4r104.gpfs200.doc/bl1adv_encryptionsetupreqs.htm

I am sure DDN can help you with all of this.

Hope this helps,
Daniel



Dr.Daniel Kidger
No. 1 The Square,

Technical Specialist  SDI (formerly Platform Computing)
Temple Quay,
Bristol  BS1 6DG
Mobile:
+44-07818 522 266
United Kingdom
Landline:
+44-02392 564 121 (Internal ITN 3726 9250)
 

e-mail:
daniel.kidger at uk.ibm.com
 






From:   Frank Leers <fleers at gmail.com>
To:     gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:   16/04/2015 23:21
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to 
GPFS 4.1?
Sent by:        gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

Since you mention GRIDScaler, I assume that you are a DDN customer and 
that you have a support contract with them.  If not, then feel free to 
take the advice that follows with a measure of salt although it mostly 
still applies ;-)
With 4.1, there are now 'Editions' of GPFS, which break out various 
feature sets into a tiered arrangement, with each tier being licensed (and 
possibly priced) differently.
Have a look here (Q 1.3) for the 4.1 licensing notes - 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html
 

With GPFS 3.5, you are most likely running the equivalent of the 'Standard 
Edition' today.  The crypto feature set comes with the 'Advanced Edition', 
which is licensed differently.

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a primer 
...  
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html

-frank


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> 
wrote:
Hello,

My site is working up to upgrading our paired GridScaler system from GPFS 
3.5 to 4.1. There is a mandate to provide encryption at rest, and that's 
an advertised feature of 4.1. We already have over 100 TB of data, 
synchronously replicated between two geographically separated sites, and 
we have concerns about how the upgrade, as well as the application of 
encryption to all that data, will go.

I'd like to hear from admins who've been through this upgrade. What 
gotchas should we look out for? Can it easily be done in place, or would 
we need some extra equipment to "slosh" our data to and from so that it is 
written to an encrypted GPFS?

Thank you for your time, and for any sage advice on this process.

Chris
--
Chris Garrison
Indiana University
Research Systems Storage

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss


Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/a00ff70c/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 360 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/a00ff70c/attachment-0001.gif>

From jamiedavis at us.ibm.com  Fri Apr 17 15:11:24 2015
From: jamiedavis at us.ibm.com (James Davis)
Date: Fri, 17 Apr 2015 10:11:24 -0400
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>
References: <D15594F5.1ADED%ecgarris@iu.edu>	<CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
	<OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>
Message-ID: <OFF10E0A53.76B236A9-ON85257E2A.004D54A7-85257E2A.004DF2E0@us.ibm.com>


Hi Chris,

Based on your question I think you are aware of this already, but just in
case...

There is not currently an encrypt-in-place solution for GPFS encryption. A
file's encryption state is determined at create-time. In order to encrypt
your existing 100TB of data, you will need to apply an encryption policy to
the current (or a new) GPFS file system(s) and then do something like:
  cp file file.enc #now file.new is encrypted
  mv file.enc file #replace the unencrypted file with the encrypted file

This can be done in parallel using mmapplypolicy if you want. In 4.1.1
(forthcoming) I plan to provide an improved version of
the /usr/lpp/mmfs/samples/ilm/mmfind (a find-esque interface to
mmapplypolicy) that shipped in the last release; this should be an
effective tool for the job.

Cheers,

Jamie

Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com



From:	Daniel Kidger <daniel.kidger at uk.ibm.com>
To:	gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:	17-04-15 08:16 AM
Subject:	Re: [gpfsug-discuss] Experiences upgrading in place to GPFS
            4.1?
Sent by:	gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

The Crypto feature of GPFS 4.1  requires the addition of just one extra RPM
over the standard edition. Other RPMs remain the same.
It also requires licenses for  "Advanced Edition". These are of the order
of 30% more expensive than the standard edition.

The model for GPFS encryption at rest is that the client node fetches the
encrypted file from the fileserver.
The file remains encrypted in transfer and is only decrypted by the client
node using a key it holds.
A result is end-to-end encryption as well as encryption of data at rest,

Encryption can be on a per file level if desired. Hence a way to migrate
from an existing non-encrypted. setup.
note inodes should be 4kB to make sure there is enough room to store the
encryption attributes.

A side effect of adding encryption is that you also need additional remote
key management (RKM) servers to handle the key management. These need
licensed software.
see
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/com.ibm.cluster.gpfs.v4r104.gpfs200.doc/bl1adv_encryptionsetupreqs.htm


I am sure DDN can help you with all of this.

Hope this helps,
Daniel
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
 Dr.Daniel                                   No. 1 The Square,                  
 Kidger                                                                         
                                                                                
 Technical                                   Temple Quay,                       
 Specialist                                  Bristol  BS1 6DG                   
 SDI                                                                            
 (formerly                                                                      
 Platform                                                                       
 Computing)                                                                     
                                                                                
                                                                                
 Mobile:    +44-07818 522 266                United Kingdom                     
                                                                                
 Landline:  +44-02392 564 121 (Internal ITN                                     
            3726 9250)                                                          
                                                                                
 e-mail:    daniel.kidger at uk.ibm.com                                            
                                                                                
                                                                                
                                                                                






From:        Frank Leers <fleers at gmail.com>
To:        gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:        16/04/2015 23:21
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to GPFS
4.1?
Sent by:        gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

Since you mention GRIDScaler, I assume that you are a DDN customer and that
you have a support contract with them.  If not, then feel free to take the
advice that follows with a measure of salt although it mostly still
applies ;-)
With 4.1, there are now 'Editions' of GPFS, which break out various feature
sets into a tiered arrangement, with each tier being licensed (and possibly
priced) differently.
Have a look here (Q 1.3) for the 4.1 licensing notes -
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html


With GPFS 3.5, you are most likely running the equivalent of the 'Standard
Edition' today.  The crypto feature set comes with the 'Advanced Edition',
which is licensed differently.

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a
primer ...
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html


-frank


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> wrote:

Hello,

My site is working up to upgrading our paired GridScaler system from GPFS
3.5 to 4.1. There is a mandate to provide encryption at rest, and that's an
advertised feature of 4.1. We already have over 100 TB of data,
synchronously replicated between two geographically separated sites, and we
have concerns about how the upgrade, as well as the application of
encryption to all that data, will go.

I'd like to hear from admins who've been through this upgrade. What gotchas
should we look out for? Can it easily be done in place, or would we need
some extra equipment to "slosh" our data to and from so that it is written
to an encrypted GPFS?

Thank you for your time, and for any sage advice on this process.

Chris
--
Chris Garrison
Indiana University
Research Systems Storage

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss


Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number
741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment-0003.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ecblank.gif
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment-0004.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 20386063.gif
Type: image/gif
Size: 360 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment-0005.gif>

From makaplan at us.ibm.com  Fri Apr 17 15:37:55 2015
From: makaplan at us.ibm.com (Marc A Kaplan)
Date: Fri, 17 Apr 2015 10:37:55 -0400
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <OFF10E0A53.76B236A9-ON85257E2A.004D54A7-85257E2A.004DF2E0@us.ibm.com>
References: <D15594F5.1ADED%ecgarris@iu.edu>	<CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
	<OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>
	<OFF10E0A53.76B236A9-ON85257E2A.004D54A7-85257E2A.004DF2E0@us.ibm.com>
Message-ID: <OF1D7F6052.1D03F850-ON85257E2A.004EEEB1-85257E2A.0050624D@us.ibm.com>

An "Encrypt-in-place" feature would not survive a serious, paranoid 
security audit.

If sensitive data ever was written to a disk, then the paranoid would say 
you can never be 100% sure that you have erased it.

That said, you decide how worried or paranoid you'd like to be and you can 
do an almost in place encryption as James Davis suggests by
simply copying the unencrypted file to a new file that will be subject to 
a GPFS encryption policy (SET ENCRYPTION) rule.

The safest way would be to copy-encrypt all the data to a new file system 
and then crush all the equipment that was used to store the clear-text 
files.

If crushing is too extreme, you might settle for a multipass sector by 
sector soft scrubbing by writing both carefully chosen and random data 
patterns.

Even then, unless you trust the manufacturer AND the manufacturer has 
provided you the means to "scrub" all the tracks/sectors of the disk you 
won't be sure...  Suppose that some sensitive data was written to a sector 
number that was later declared (partially) defective and remapped by the 
disk micro-code, and the original sector is put in a bad sector list that 
you can no longer address with standard disk driver software... 

Or it could be on some NVRAM or a disk that a service technician swapped 
out...  Ooops... it went out the door... and is now in the hands of the 
bad guys...





From:   James Davis/Poughkeepsie/IBM at IBMUS
To:     gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:   04/17/2015 10:12 AM
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to 
GPFS 4.1?
Sent by:        gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

Based on your question I think you are aware of this already, but just in 
case...

There is not currently an encrypt-in-place solution for GPFS encryption. A 
file's encryption state is determined at create-time. In order to encrypt 
your existing 100TB of data, you will need to apply an encryption policy 
to the current (or a new) GPFS file system(s) and then do something like:
  cp file file.enc #now file.new is encrypted
  mv file.enc file #replace the unencrypted file with the encrypted file

This can be done in parallel using mmapplypolicy if you want. In 4.1.1 
(forthcoming) I plan to provide an improved version of the 
/usr/lpp/mmfs/samples/ilm/mmfind (a find-esque interface to mmapplypolicy) 
that shipped in the last release; this should be an effective tool for the 
job.

Cheers,

Jamie

Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com

Daniel Kidger ---17-04-2015 08:16:54 AM---Hi Chris, The Crypto feature of 
GPFS 4.1  requires the addition of just one extra

From: Daniel Kidger <daniel.kidger at uk.ibm.com>
To: gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date: 17-04-15 08:16 AM
Subject: Re: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
Sent by: gpfsug-discuss-bounces at gpfsug.org



Hi Chris, 

The Crypto feature of GPFS 4.1  requires the addition of just one extra 
RPM over the standard edition. Other RPMs remain the same. 
It also requires licenses for  "Advanced Edition". These are of the order 
of 30% more expensive than the standard edition. 

The model for GPFS encryption at rest is that the client node fetches the 
encrypted file from the fileserver. 
The file remains encrypted in transfer and is only decrypted by the client 
node using a key it holds. 
A result is end-to-end encryption as well as encryption of data at rest, 

Encryption can be on a per file level if desired. Hence a way to migrate 
from an existing non-encrypted. setup. 
note inodes should be 4kB to make sure there is enough room to store the 
encryption attributes. 

A side effect of adding encryption is that you also need additional remote 
key management (RKM) servers to handle the key management. These need 
licensed software. 
see 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/com.ibm.cluster.gpfs.v4r104.gpfs200.doc/bl1adv_encryptionsetupreqs.htm 


I am sure DDN can help you with all of this. 

Hope this helps, 
Daniel 


Dr.Daniel Kidger 
No. 1 The Square, 

Technical Specialist  SDI (formerly Platform Computing) 
Temple Quay,
Bristol  BS1 6DG 
Mobile: 
+44-07818 522 266 
United Kingdom 
Landline: 
+44-02392 564 121 (Internal ITN 3726 9250) 
  

e-mail: 
daniel.kidger at uk.ibm.com 
 







From:        Frank Leers <fleers at gmail.com> 
To:        gpfsug main discussion list <gpfsug-discuss at gpfsug.org> 
Date:        16/04/2015 23:21 
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to 
GPFS 4.1? 
Sent by:        gpfsug-discuss-bounces at gpfsug.org 



Hi Chris, 

Since you mention GRIDScaler, I assume that you are a DDN customer and 
that you have a support contract with them.  If not, then feel free to 
take the advice that follows with a measure of salt although it mostly 
still applies ;-) 
With 4.1, there are now 'Editions' of GPFS, which break out various 
feature sets into a tiered arrangement, with each tier being licensed (and 
possibly priced) differently. 
Have a look here (Q 1.3) for the 4.1 licensing notes - 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html 
 

With GPFS 3.5, you are most likely running the equivalent of the 'Standard 
Edition' today.  The crypto feature set comes with the 'Advanced Edition', 
which is licensed differently. 

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a primer 
...  
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html 


-frank 


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> 
wrote: 
Hello, 

My site is working up to upgrading our paired GridScaler system from GPFS 
3.5 to 4.1. There is a mandate to provide encryption at rest, and that's 
an advertised feature of 4.1. We already have over 100 TB of data, 
synchronously replicated between two geographically separated sites, and 
we have concerns about how the upgrade, as well as the application of 
encryption to all that data, will go. 

I'd like to hear from admins who've been through this upgrade. What 
gotchas should we look out for? Can it easily be done in place, or would 
we need some extra equipment to "slosh" our data to and from so that it is 
written to an encrypted GPFS? 

Thank you for your time, and for any sage advice on this process. 

Chris 
-- 
Chris Garrison 
Indiana University 
Research Systems Storage 

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss


Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 21994 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0006.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0007.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0008.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 360 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0009.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0010.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0011.gif>

From secretary at gpfsug.org  Fri Apr 24 15:05:25 2015
From: secretary at gpfsug.org (Secretary GPFS UG)
Date: Fri, 24 Apr 2015 15:05:25 +0100
Subject: [gpfsug-discuss] GPFS User Group Meeting Agenda
Message-ID: <ebd76ace58f83d838895e0751950193b@webmail.gpfsug.org>

Dear Members,

The agenda for the next GPFS User Group Meeting is now available:
10:00 arrival for a 10:30 start
10:30                 Introductions - Jez Tucker, Group Chair & Claire 
Robson, Group Secretary
10:35                 Keynote - Doris Conti
10:50                 4.1.1 Roadmap / High-level futures - Scott Fadden
11:40                 Failure Events, Recovery & Problem determination - 
Scott Fadden
12:00                 Monitoring IBM Spectrum Scale using IBM Spectrum 
Control (VSC/TPC) - Christian Bolik
12:30                 Lunch
13:00                 User Experience from University of Birmingham & 
CLIMB - Simon Thompson
13:20                 User Experience from NERSC - Jason Hick
13:40                 AFM & Async DR Use Cases - Shankar Balasubramanian
14:25                 mmbackup + TSM Integration - Stefan Bender
15:10                 Break
15:25                 IBM Spectrum Scale (formerly GPFS) performance 
update, protocol performance and sizing - Sven Oehme
16:50                 Closing summary - Jez Tucker, Group Chair & Claire 
Robson, Group Secretary
17:00                 Ends

All attendees are invited to attend:
19:00 Buffet and drinks at York National Railway Museum

There are still places available for the event. If you have not 
registered and would like to attend, please email me, 
secretary at gpfsug.org with your name, job title, organisation, telephone 
and any dietary requirements.

We hope to see you in May!

Kind regards,

Claire

-- 
Claire Robson
GPFS User Group Secretary



From Dan.Foster at bristol.ac.uk  Wed Apr 29 08:51:03 2015
From: Dan.Foster at bristol.ac.uk (Dan Foster)
Date: Wed, 29 Apr 2015 08:51:03 +0100
Subject: [gpfsug-discuss] file system format version / software version
	compatibility matrix
Message-ID: <CALRUX+gQ2ZfN+7gWsadSUi2gUG+PaUkH0OrpEXAYDAUY-_LxRg@mail.gmail.com>

Hi All,

I'm trying to determine which versions of the GPFS file system format
are compatible with certain versions of the GPFS software.
Specifically in this instance I'm interested if file system format
v11.05 (from GPFS 3.3.0.2) is compatible with GPFS 3.5 .

The "File system format changes between versions of GPFS" [1] chapter
mentions format levels as old a v6, which infers that they are
compatible. But it would be useful to know for certain.

Thanks,
Dan.

[1] http://www-01.ibm.com/support/knowledgecenter/SSFKCN_3.5.0/com.ibm.cluster.gpfs.v3r5.gpfs100.doc/bl1adm_fsmigissues.htm?lang=en
-- 
Dan Foster | Senior Storage Systems Administrator | IT Services
e: dan.foster at bristol.ac.uk | t:  0117 3941170 [x41170]
m: Advanced Computing Research Centre, University of Bristol,
    8-10 Berkeley Square, Bristol BS8 1HH


From jamiedavis at us.ibm.com  Wed Apr 29 13:43:08 2015
From: jamiedavis at us.ibm.com (James Davis)
Date: Wed, 29 Apr 2015 08:43:08 -0400
Subject: [gpfsug-discuss] file system format version / software
	version	compatibility matrix
In-Reply-To: <CALRUX+gQ2ZfN+7gWsadSUi2gUG+PaUkH0OrpEXAYDAUY-_LxRg@mail.gmail.com>
References: <CALRUX+gQ2ZfN+7gWsadSUi2gUG+PaUkH0OrpEXAYDAUY-_LxRg@mail.gmail.com>
Message-ID: <OF27DFBC25.AE010664-ON85257E36.00453E1C-85257E36.0045DD6B@us.ibm.com>


Dan,

A "new" GPFS should be able to mount and interact with file systems with
"old" versions. Specifically I do not believe you will have trouble getting
GPFS 3.5 to talk to an FS created on GPFS 3.3.0.2.

The "Concepts, planning, and install guide" provides information about
migrating to GPFS 4.1 from GPFS 3.2 or earlier; this would implies file
systems created at GPFS 3.2 being supported on GPFS 4.1, which is even more
compatibility than you are asking about. The section to which I refer is
titled "Migrating to GPFS 4.1 from GPFS 3.2 or earlier releases of GPFS".

Note that without running mmchfs -V {full|compat} you will not have access
to some of the newer GPFS features. See the section titled "Completing the
migration to a new level of GPFS", also in the concepts guide.

Hope this helps,

Jamie

Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com



From:	Dan Foster <Dan.Foster at bristol.ac.uk>
To:	gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:	29-04-15 03:51 AM
Subject:	[gpfsug-discuss] file system format version / software version
            compatibility matrix
Sent by:	gpfsug-discuss-bounces at gpfsug.org



Hi All,

I'm trying to determine which versions of the GPFS file system format
are compatible with certain versions of the GPFS software.
Specifically in this instance I'm interested if file system format
v11.05 (from GPFS 3.3.0.2) is compatible with GPFS 3.5 .

The "File system format changes between versions of GPFS" [1] chapter
mentions format levels as old a v6, which infers that they are
compatible. But it would be useful to know for certain.

Thanks,
Dan.

[1]
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_3.5.0/com.ibm.cluster.gpfs.v3r5.gpfs100.doc/bl1adm_fsmigissues.htm?lang=en

--
Dan Foster | Senior Storage Systems Administrator | IT Services
e: dan.foster at bristol.ac.uk | t:  0117 3941170 [x41170]
m: Advanced Computing Research Centre, University of Bristol,
    8-10 Berkeley Square, Bristol BS8 1HH
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150429/23753578/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150429/23753578/attachment-0001.gif>

From pavel.pokorny at datera.cz  Thu Apr  9 22:00:07 2015
From: pavel.pokorny at datera.cz (Pavel Pokorny)
Date: Thu, 9 Apr 2015 23:00:07 +0200
Subject: [gpfsug-discuss] Share GPFS from Windows?
Message-ID: <CAHy3nHJ9=D_zYq3xxFO8NnJutdy7JAWxkN7s_PZds277Nx4rxA@mail.gmail.com>

Hello to all,
Is there any technical reason why it is not supported to export GPFS from
Windows nodes using CIFS? As stated in GPFs FAQ (
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html
):

*Exporting GPFS file systems as Server Message Block (SMB) shares (also
known as CIFS shares) from GPFS Windows nodes is not supported.*

Or it is limitation more related to licensing and business point of view.
Thank you for answers,
Pavel

--
Ing. Pavel Pokorn?
DATERA s.r.o. | Ovocn? trh 580/2 | Praha | Czech Republic
www.datera.cz | Mobil: +420 602 357 194 | E-mail: pavel.pokorny at datera.cz
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150409/6f6c75d7/attachment-0002.htm>

From BOLIK at de.ibm.com  Fri Apr 10 12:24:14 2015
From: BOLIK at de.ibm.com (Christian Bolik)
Date: Fri, 10 Apr 2015 13:24:14 +0200
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
	multitude of GPFS clusters
Message-ID: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>


Just wanted to let you know that recently GPFS support has been added to
TPC, which is IBM's Tivoli Storage Productivity Center (soon to be renamed
to IBM Spectrum Control). As of now, TPC allows GPFS administrators to get
answers to the following questions, across any number of GPFS clusters
which have been added to TPC:

- Which of my clusters are running out of free space?
- Which of my clusters or nodes have a health problem?
- Which file systems and pools are running out of capacity?
- Which file systems are mounted on which nodes?
- How much space is occupied by snapshots? Are there any very old,
potentially obsolete ones?
- Which quotas are close to being exceeded or have already been exceeded?
- Which filesets are close to running out of free inodes?
- Which NSDs are at risk of becoming unavailable, or are unavailable?
- Are the volumes backing my NSDs performing OK?
- Are all nodes fulfilling critical roles in the cluster up and running?
- How can I be notified when nodes go offline or file systems fill up
beyond a threshold?

There's a short 6-minute video available on YouTube which shows how TPC
helps answering these questions:
https://www.youtube.com/watch?v=8Esk5U_cYw8&feature=youtu.be

For more information about TPC, please check out the product wiki on
developerWorks: http://ibm.co/1adWNFK

Thanks,
Christian Bolik
IBM Storage Software Development




From zgiles at gmail.com  Fri Apr 10 16:27:36 2015
From: zgiles at gmail.com (Zachary Giles)
Date: Fri, 10 Apr 2015 11:27:36 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>
References: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>
Message-ID: <CAMYZk=fKkAgeR-MW79a=BO46WpR9+Gm+=_kCLkQo4+v5KxgKqg@mail.gmail.com>

Christian:
Interesting and thanks for the latest news.

May I ask: Is there an intent moving forward that TPC and / or other Tivoli
products will be a required part of GPFS?
The concern I have is that GPFS is pretty straightforward at the moment and
has very logical requirements to operate (min servers, quorum, etc),
whereas there are many IBM products that require two or three more servers
just to manage the servers managing the service.. too much. It would be
nice to make sure, going forward, that the core of GPFS can still function
without additional web servers, Java, a suite of middleware, and a handful
of DB2 instance .. :)

-Zach



On Fri, Apr 10, 2015 at 7:24 AM, Christian Bolik <BOLIK at de.ibm.com> wrote:

>
> Just wanted to let you know that recently GPFS support has been added to
> TPC, which is IBM's Tivoli Storage Productivity Center (soon to be renamed
> to IBM Spectrum Control). As of now, TPC allows GPFS administrators to get
> answers to the following questions, across any number of GPFS clusters
> which have been added to TPC:
>
> - Which of my clusters are running out of free space?
> - Which of my clusters or nodes have a health problem?
> - Which file systems and pools are running out of capacity?
> - Which file systems are mounted on which nodes?
> - How much space is occupied by snapshots? Are there any very old,
> potentially obsolete ones?
> - Which quotas are close to being exceeded or have already been exceeded?
> - Which filesets are close to running out of free inodes?
> - Which NSDs are at risk of becoming unavailable, or are unavailable?
> - Are the volumes backing my NSDs performing OK?
> - Are all nodes fulfilling critical roles in the cluster up and running?
> - How can I be notified when nodes go offline or file systems fill up
> beyond a threshold?
>
> There's a short 6-minute video available on YouTube which shows how TPC
> helps answering these questions:
> https://www.youtube.com/watch?v=8Esk5U_cYw8&feature=youtu.be
>
> For more information about TPC, please check out the product wiki on
> developerWorks: http://ibm.co/1adWNFK
>
> Thanks,
> Christian Bolik
> IBM Storage Software Development
>
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>



-- 
Zach Giles
zgiles at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150410/c4ee09f6/attachment-0002.htm>

From dhildeb at us.ibm.com  Fri Apr 10 18:45:14 2015
From: dhildeb at us.ibm.com (Dean Hildebrand)
Date: Fri, 10 Apr 2015 10:45:14 -0700
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <CAMYZk=fKkAgeR-MW79a=BO46WpR9+Gm+=_kCLkQo4+v5KxgKqg@mail.gmail.com>
References: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>
	<CAMYZk=fKkAgeR-MW79a=BO46WpR9+Gm+=_kCLkQo4+v5KxgKqg@mail.gmail.com>
Message-ID: <OFDFD967F3.9188C435-ON88257E23.0060AE82-88257E23.006186C8@us.ibm.com>


Hi Zach,

The summary is that GPFS is being integrated much more across the
portfolio...

With GPFS itself, there is a video below demonstrating the ESS/GSS GUI and
monitoring feature that is in the product today.  Moving forward, as you
can probably see, there is a push in IBM to move GPFS to software-defined,
which includes features such as the GUI as well...
https://www.youtube.com/watch?v=Mv9Sn-VYoGU

Dean




From:	Zachary Giles <zgiles at gmail.com>
To:	gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:	04/10/2015 08:27 AM
Subject:	Re: [gpfsug-discuss] Monitoring capacity and health status for
            a multitude of GPFS clusters
Sent by:	gpfsug-discuss-bounces at gpfsug.org



Christian:
Interesting and thanks for the latest news.

May I ask: Is there an intent moving forward that TPC and / or other Tivoli
products will be a required part of GPFS?
The concern I have is that GPFS is pretty straightforward at the moment and
has very logical requirements to operate (min servers, quorum, etc),
whereas there are many IBM products that require two or three more servers
just to manage the servers managing the service.. too much. It would be
nice to make sure, going forward, that the core of GPFS can still function
without additional web servers, Java, a suite of middleware, and a handful
of DB2 instance .. :)

-Zach



On Fri, Apr 10, 2015 at 7:24 AM, Christian Bolik <BOLIK at de.ibm.com> wrote:

  Just wanted to let you know that recently GPFS support has been added to
  TPC, which is IBM's Tivoli Storage Productivity Center (soon to be
  renamed
  to IBM Spectrum Control). As of now, TPC allows GPFS administrators to
  get
  answers to the following questions, across any number of GPFS clusters
  which have been added to TPC:

  - Which of my clusters are running out of free space?
  - Which of my clusters or nodes have a health problem?
  - Which file systems and pools are running out of capacity?
  - Which file systems are mounted on which nodes?
  - How much space is occupied by snapshots? Are there any very old,
  potentially obsolete ones?
  - Which quotas are close to being exceeded or have already been exceeded?
  - Which filesets are close to running out of free inodes?
  - Which NSDs are at risk of becoming unavailable, or are unavailable?
  - Are the volumes backing my NSDs performing OK?
  - Are all nodes fulfilling critical roles in the cluster up and running?
  - How can I be notified when nodes go offline or file systems fill up
  beyond a threshold?

  There's a short 6-minute video available on YouTube which shows how TPC
  helps answering these questions:
  https://www.youtube.com/watch?v=8Esk5U_cYw8&feature=youtu.be

  For more information about TPC, please check out the product wiki on
  developerWorks: http://ibm.co/1adWNFK

  Thanks,
  Christian Bolik
  IBM Storage Software Development


  _______________________________________________
  gpfsug-discuss mailing list
  gpfsug-discuss at gpfsug.org
  http://gpfsug.org/mailman/listinfo/gpfsug-discuss



--
Zach Giles
zgiles at gmail.com_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150410/9afe8bc4/attachment-0002.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150410/9afe8bc4/attachment-0002.gif>

From BOLIK at de.ibm.com  Mon Apr 13 14:11:01 2015
From: BOLIK at de.ibm.com (Christian Bolik)
Date: Mon, 13 Apr 2015 15:11:01 +0200
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
Message-ID: <OFE5AC06CE.E75B85AD-ONC1257E26.00399901-C1257E26.00486BB2@de.ibm.com>


Hi Zach,

I'm not aware of any intent to make TPC or any other Tivoli/IBM product a
prereq for GPFS, and I don't think any such plans exist.  Rather, as Dean
also pointed out, we're investing work to improve integration of
GPFS/Spectrum Scale into other products being members of the newly
announced IBM Spectrum Storage family, with the goal of improving
manageability of the individual components (rather than worsening it...).

Cheers,
Christian

> Christian:
> Interesting and thanks for the latest news.
>
> May I ask: Is there an intent moving forward that TPC and / or other
Tivoli
> products will be a required part of GPFS?
> The concern I have is that GPFS is pretty straightforward at the moment
and
> has very logical requirements to operate (min servers, quorum, etc),
> whereas there are many IBM products that require two or three more
servers
> just to manage the servers managing the service.. too much. It would be
> nice to make sure, going forward, that the core of GPFS can still
function
> without additional web servers, Java, a suite of middleware, and a
handful
> of DB2 instance .. :)
>
> -Zach

Christian Bolik

Software Defined Storage Development

IBM Deutschland Research & Development GmbH, Hechtsheimer Str. 2, 55131
Mainz, Germany
Vorsitzende des Aufsichtsrats: Martina Koederitz
Gesch?ftsf?hrung: Dirk Wittkopp
Sitz der Gesellschaft: B?blingen / Registergericht: Amtsgericht Stuttgart,
HRB 243294




From zgiles at gmail.com  Mon Apr 13 17:05:11 2015
From: zgiles at gmail.com (Zachary Giles)
Date: Mon, 13 Apr 2015 12:05:11 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <OFE5AC06CE.E75B85AD-ONC1257E26.00399901-C1257E26.00486BB2@de.ibm.com>
References: <OFE5AC06CE.E75B85AD-ONC1257E26.00399901-C1257E26.00486BB2@de.ibm.com>
Message-ID: <CAMYZk=eaQpwRcW6RM7WK9x4v8wO=o=GmZPrjCCq=dKa9K7H4Bg@mail.gmail.com>

Thanks for your replies. I can definitely appreciate the the goal of
improving management of components, and I agree that if GPFS will be using
within other products (which it is and will continue to be), then it would
be great for those products to be able to manage GPFS via an interface.

My fear with the idea of the above mentioned as a prereq is that the
"improvement" of management might looks like an improvement when you only
have 1 or 2 tiers of storage and 1 or 2 filesets called "database" and
another fileset called "user files", but if you have hundreds or thousands
of filesets and several tiers of storage, with both GSS and non-GSS systems
in the same cluster, then the GUI may actually be more cumbersome than the
original method. So, I just want to voice an opinion that we should
continue to be able to configure / maintain / monitor GPFS in a programatic
/ scriptable non-point-and-click way, if possible.



On Mon, Apr 13, 2015 at 9:11 AM, Christian Bolik <BOLIK at de.ibm.com> wrote:

>
> Hi Zach,
>
> I'm not aware of any intent to make TPC or any other Tivoli/IBM product a
> prereq for GPFS, and I don't think any such plans exist.  Rather, as Dean
> also pointed out, we're investing work to improve integration of
> GPFS/Spectrum Scale into other products being members of the newly
> announced IBM Spectrum Storage family, with the goal of improving
> manageability of the individual components (rather than worsening it...).
>
> Cheers,
> Christian
>
> > Christian:
> > Interesting and thanks for the latest news.
> >
> > May I ask: Is there an intent moving forward that TPC and / or other
> Tivoli
> > products will be a required part of GPFS?
> > The concern I have is that GPFS is pretty straightforward at the moment
> and
> > has very logical requirements to operate (min servers, quorum, etc),
> > whereas there are many IBM products that require two or three more
> servers
> > just to manage the servers managing the service.. too much. It would be
> > nice to make sure, going forward, that the core of GPFS can still
> function
> > without additional web servers, Java, a suite of middleware, and a
> handful
> > of DB2 instance .. :)
> >
> > -Zach
>
> Christian Bolik
>
> Software Defined Storage Development
>
> IBM Deutschland Research & Development GmbH, Hechtsheimer Str. 2, 55131
> Mainz, Germany
> Vorsitzende des Aufsichtsrats: Martina Koederitz
> Gesch?ftsf?hrung: Dirk Wittkopp
> Sitz der Gesellschaft: B?blingen / Registergericht: Amtsgericht Stuttgart,
> HRB 243294
>
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>



-- 
Zach Giles
zgiles at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150413/ec65870f/attachment-0002.htm>

From jamiedavis at us.ibm.com  Tue Apr 14 13:23:59 2015
From: jamiedavis at us.ibm.com (James Davis)
Date: Tue, 14 Apr 2015 08:23:59 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <mailman.1.1429009201.21161.gpfsug-discuss@gpfsug.org>
References: <mailman.1.1429009201.21161.gpfsug-discuss@gpfsug.org>
Message-ID: <OF41183321.B20BDD25-ON85257E27.00438322-85257E27.00441D50@us.ibm.com>


Zach,

Not sure if I'm formatting this message properly (need to switch off of
digest mode), but I know of no plans to replace the GPFS command line
interface with a GUI. With the GPFS GUI, TPC monitoring, etc., we want to
enable a wider variety of users to effectively use and manage GPFS, but the
command line will of course remain available for power users,
script-writers, etc. I certainly don't intend to throw away any of my nice
mm-themed test programs :-)

Jamie Davis
GPFS Test

Date: Mon, 13 Apr 2015 12:05:11 -0400
From: Zachary Giles <zgiles at gmail.com>
To: gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Subject: Re: [gpfsug-discuss] Monitoring capacity and health status
		 for a multitude of GPFS clusters
Message-ID:

<CAMYZk=eaQpwRcW6RM7WK9x4v8wO=o=GmZPrjCCq=dKa9K7H4Bg at mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

Thanks for your replies. I can definitely appreciate the the goal of
improving management of components, and I agree that if GPFS will be using
within other products (which it is and will continue to be), then it would
be great for those products to be able to manage GPFS via an interface.

My fear with the idea of the above mentioned as a prereq is that the
"improvement" of management might looks like an improvement when you only
have 1 or 2 tiers of storage and 1 or 2 filesets called "database" and
another fileset called "user files", but if you have hundreds or thousands
of filesets and several tiers of storage, with both GSS and non-GSS systems
in the same cluster, then the GUI may actually be more cumbersome than the
original method. So, I just want to voice an opinion that we should
continue to be able to configure / maintain / monitor GPFS in a programatic
/ scriptable non-point-and-click way, if possible.


Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150414/9352bfbe/attachment-0002.htm>

From ecgarris at iu.edu  Thu Apr 16 21:33:59 2015
From: ecgarris at iu.edu (Garrison, E Chris)
Date: Thu, 16 Apr 2015 20:33:59 +0000
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
Message-ID: <D15594F5.1ADED%ecgarris@iu.edu>

Hello,

My site is working up to upgrading our paired GridScaler system from GPFS 3.5 to 4.1. There is a mandate to provide encryption at rest, and that's an advertised feature of 4.1. We already have over 100 TB of data, synchronously replicated between two geographically separated sites, and we have concerns about how the upgrade, as well as the application of encryption to all that data, will go.

I'd like to hear from admins who've been through this upgrade. What gotchas should we look out for? Can it easily be done in place, or would we need some extra equipment to "slosh" our data to and from so that it is written to an encrypted GPFS?

Thank you for your time, and for any sage advice on this process.

Chris
--
Chris Garrison
Indiana University
Research Systems Storage
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150416/aeed3961/attachment-0002.htm>

From zgiles at gmail.com  Thu Apr 16 21:39:51 2015
From: zgiles at gmail.com (Zachary Giles)
Date: Thu, 16 Apr 2015 16:39:51 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <OF41183321.B20BDD25-ON85257E27.00438322-85257E27.00441D50@us.ibm.com>
References: <mailman.1.1429009201.21161.gpfsug-discuss@gpfsug.org>
	<OF41183321.B20BDD25-ON85257E27.00438322-85257E27.00441D50@us.ibm.com>
Message-ID: <CAMYZk=dni=XHyMAcro6Y74tW+F-sVBRU4BCUm3vGOdxFDjOuaA@mail.gmail.com>

Thanks Jamie. I appreciate the input. Glad to hear it.

On Tue, Apr 14, 2015 at 8:23 AM, James Davis <jamiedavis at us.ibm.com> wrote:

> Zach,
>
> Not sure if I'm formatting this message properly (need to switch off of
> digest mode), but I know of no plans to replace the GPFS command line
> interface with a GUI. With the GPFS GUI, TPC monitoring, etc., we want to
> enable a wider variety of users to effectively use and manage GPFS, but the
> command line will of course remain available for power users,
> script-writers, etc. I certainly don't intend to throw away any of my nice
> mm-themed test programs :-)
>
> Jamie Davis
> GPFS Test
>
> Date: Mon, 13 Apr 2015 12:05:11 -0400
> From: Zachary Giles <zgiles at gmail.com>
> To: gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
> Subject: Re: [gpfsug-discuss] Monitoring capacity and health status
> for a multitude of GPFS clusters
> Message-ID:
> <CAMYZk=eaQpwRcW6RM7WK9x4v8wO=o=GmZPrjCCq=dKa9K7H4Bg at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Thanks for your replies. I can definitely appreciate the the goal of
> improving management of components, and I agree that if GPFS will be using
> within other products (which it is and will continue to be), then it would
> be great for those products to be able to manage GPFS via an interface.
>
> My fear with the idea of the above mentioned as a prereq is that the
> "improvement" of management might looks like an improvement when you only
> have 1 or 2 tiers of storage and 1 or 2 filesets called "database" and
> another fileset called "user files", but if you have hundreds or thousands
> of filesets and several tiers of storage, with both GSS and non-GSS systems
> in the same cluster, then the GUI may actually be more cumbersome than the
> original method. So, I just want to voice an opinion that we should
> continue to be able to configure / maintain / monitor GPFS in a programatic
> / scriptable non-point-and-click way, if possible.
>
>
> Jamie Davis
> GPFS Functional Verification Test (FVT)
> jamiedavis at us.ibm.com
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>
>


-- 
Zach Giles
zgiles at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150416/da9891e5/attachment-0002.htm>

From fleers at gmail.com  Thu Apr 16 23:21:34 2015
From: fleers at gmail.com (Frank Leers)
Date: Thu, 16 Apr 2015 15:21:34 -0700
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <D15594F5.1ADED%ecgarris@iu.edu>
References: <D15594F5.1ADED%ecgarris@iu.edu>
Message-ID: <CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>

Hi Chris,

Since you mention GRIDScaler, I assume that you are a DDN customer and that
you have a support contract with them.  If not, then feel free to take the
advice that follows with a measure of salt although it mostly still applies
;-)
With 4.1, there are now 'Editions' of GPFS, which break out various feature
sets into a tiered arrangement, with each tier being licensed (and possibly
priced) differently.
Have a look here (Q 1.3) for the 4.1 licensing notes -
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html


With GPFS 3.5, you are most likely running the equivalent of the 'Standard
Edition' today.  The crypto feature set comes with the 'Advanced Edition',
which is licensed differently.

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a primer
...
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html

-frank


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> wrote:

>  Hello,
>
>  My site is working up to upgrading our paired GridScaler system from
> GPFS 3.5 to 4.1. There is a mandate to provide encryption at rest, and
> that's an advertised feature of 4.1. We already have over 100 TB of data,
> synchronously replicated between two geographically separated sites, and we
> have concerns about how the upgrade, as well as the application of
> encryption to all that data, will go.
>
>  I'd like to hear from admins who've been through this upgrade. What
> gotchas should we look out for? Can it easily be done in place, or would we
> need some extra equipment to "slosh" our data to and from so that it is
> written to an encrypted GPFS?
>
>  Thank you for your time, and for any sage advice on this process.
>
>  Chris
> --
> Chris Garrison
> Indiana University
> Research Systems Storage
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150416/38ac4616/attachment-0002.htm>

From daniel.kidger at uk.ibm.com  Fri Apr 17 13:15:40 2015
From: daniel.kidger at uk.ibm.com (Daniel Kidger)
Date: Fri, 17 Apr 2015 13:15:40 +0100
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
References: <D15594F5.1ADED%ecgarris@iu.edu>
	<CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
Message-ID: <OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>

Hi Chris,

The Crypto feature of GPFS 4.1  requires the addition of just one extra 
RPM over the standard edition. Other RPMs remain the same.
It also requires licenses for  "Advanced Edition". These are of the order 
of 30% more expensive than the standard edition.

The model for GPFS encryption at rest is that the client node fetches the 
encrypted file from the fileserver. 
The file remains encrypted in transfer and is only decrypted by the client 
node using a key it holds.
A result is end-to-end encryption as well as encryption of data at rest,

Encryption can be on a per file level if desired. Hence a way to migrate 
from an existing non-encrypted. setup. 
note inodes should be 4kB to make sure there is enough room to store the 
encryption attributes.

A side effect of adding encryption is that you also need additional remote 
key management (RKM) servers to handle the key management. These need 
licensed software.
see 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/com.ibm.cluster.gpfs.v4r104.gpfs200.doc/bl1adv_encryptionsetupreqs.htm

I am sure DDN can help you with all of this.

Hope this helps,
Daniel



Dr.Daniel Kidger
No. 1 The Square,

Technical Specialist  SDI (formerly Platform Computing)
Temple Quay,
Bristol  BS1 6DG
Mobile:
+44-07818 522 266
United Kingdom
Landline:
+44-02392 564 121 (Internal ITN 3726 9250)
 

e-mail:
daniel.kidger at uk.ibm.com
 






From:   Frank Leers <fleers at gmail.com>
To:     gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:   16/04/2015 23:21
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to 
GPFS 4.1?
Sent by:        gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

Since you mention GRIDScaler, I assume that you are a DDN customer and 
that you have a support contract with them.  If not, then feel free to 
take the advice that follows with a measure of salt although it mostly 
still applies ;-)
With 4.1, there are now 'Editions' of GPFS, which break out various 
feature sets into a tiered arrangement, with each tier being licensed (and 
possibly priced) differently.
Have a look here (Q 1.3) for the 4.1 licensing notes - 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html
 

With GPFS 3.5, you are most likely running the equivalent of the 'Standard 
Edition' today.  The crypto feature set comes with the 'Advanced Edition', 
which is licensed differently.

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a primer 
...  
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html

-frank


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> 
wrote:
Hello,

My site is working up to upgrading our paired GridScaler system from GPFS 
3.5 to 4.1. There is a mandate to provide encryption at rest, and that's 
an advertised feature of 4.1. We already have over 100 TB of data, 
synchronously replicated between two geographically separated sites, and 
we have concerns about how the upgrade, as well as the application of 
encryption to all that data, will go.

I'd like to hear from admins who've been through this upgrade. What 
gotchas should we look out for? Can it easily be done in place, or would 
we need some extra equipment to "slosh" our data to and from so that it is 
written to an encrypted GPFS?

Thank you for your time, and for any sage advice on this process.

Chris
--
Chris Garrison
Indiana University
Research Systems Storage

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss


Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/a00ff70c/attachment-0002.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 360 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/a00ff70c/attachment-0002.gif>

From jamiedavis at us.ibm.com  Fri Apr 17 15:11:24 2015
From: jamiedavis at us.ibm.com (James Davis)
Date: Fri, 17 Apr 2015 10:11:24 -0400
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>
References: <D15594F5.1ADED%ecgarris@iu.edu>	<CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
	<OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>
Message-ID: <OFF10E0A53.76B236A9-ON85257E2A.004D54A7-85257E2A.004DF2E0@us.ibm.com>


Hi Chris,

Based on your question I think you are aware of this already, but just in
case...

There is not currently an encrypt-in-place solution for GPFS encryption. A
file's encryption state is determined at create-time. In order to encrypt
your existing 100TB of data, you will need to apply an encryption policy to
the current (or a new) GPFS file system(s) and then do something like:
  cp file file.enc #now file.new is encrypted
  mv file.enc file #replace the unencrypted file with the encrypted file

This can be done in parallel using mmapplypolicy if you want. In 4.1.1
(forthcoming) I plan to provide an improved version of
the /usr/lpp/mmfs/samples/ilm/mmfind (a find-esque interface to
mmapplypolicy) that shipped in the last release; this should be an
effective tool for the job.

Cheers,

Jamie

Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com



From:	Daniel Kidger <daniel.kidger at uk.ibm.com>
To:	gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:	17-04-15 08:16 AM
Subject:	Re: [gpfsug-discuss] Experiences upgrading in place to GPFS
            4.1?
Sent by:	gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

The Crypto feature of GPFS 4.1  requires the addition of just one extra RPM
over the standard edition. Other RPMs remain the same.
It also requires licenses for  "Advanced Edition". These are of the order
of 30% more expensive than the standard edition.

The model for GPFS encryption at rest is that the client node fetches the
encrypted file from the fileserver.
The file remains encrypted in transfer and is only decrypted by the client
node using a key it holds.
A result is end-to-end encryption as well as encryption of data at rest,

Encryption can be on a per file level if desired. Hence a way to migrate
from an existing non-encrypted. setup.
note inodes should be 4kB to make sure there is enough room to store the
encryption attributes.

A side effect of adding encryption is that you also need additional remote
key management (RKM) servers to handle the key management. These need
licensed software.
see
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/com.ibm.cluster.gpfs.v4r104.gpfs200.doc/bl1adv_encryptionsetupreqs.htm


I am sure DDN can help you with all of this.

Hope this helps,
Daniel
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
 Dr.Daniel                                   No. 1 The Square,                  
 Kidger                                                                         
                                                                                
 Technical                                   Temple Quay,                       
 Specialist                                  Bristol  BS1 6DG                   
 SDI                                                                            
 (formerly                                                                      
 Platform                                                                       
 Computing)                                                                     
                                                                                
                                                                                
 Mobile:    +44-07818 522 266                United Kingdom                     
                                                                                
 Landline:  +44-02392 564 121 (Internal ITN                                     
            3726 9250)                                                          
                                                                                
 e-mail:    daniel.kidger at uk.ibm.com                                            
                                                                                
                                                                                
                                                                                






From:        Frank Leers <fleers at gmail.com>
To:        gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:        16/04/2015 23:21
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to GPFS
4.1?
Sent by:        gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

Since you mention GRIDScaler, I assume that you are a DDN customer and that
you have a support contract with them.  If not, then feel free to take the
advice that follows with a measure of salt although it mostly still
applies ;-)
With 4.1, there are now 'Editions' of GPFS, which break out various feature
sets into a tiered arrangement, with each tier being licensed (and possibly
priced) differently.
Have a look here (Q 1.3) for the 4.1 licensing notes -
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html


With GPFS 3.5, you are most likely running the equivalent of the 'Standard
Edition' today.  The crypto feature set comes with the 'Advanced Edition',
which is licensed differently.

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a
primer ...
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html


-frank


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> wrote:

Hello,

My site is working up to upgrading our paired GridScaler system from GPFS
3.5 to 4.1. There is a mandate to provide encryption at rest, and that's an
advertised feature of 4.1. We already have over 100 TB of data,
synchronously replicated between two geographically separated sites, and we
have concerns about how the upgrade, as well as the application of
encryption to all that data, will go.

I'd like to hear from admins who've been through this upgrade. What gotchas
should we look out for? Can it easily be done in place, or would we need
some extra equipment to "slosh" our data to and from so that it is written
to an encrypted GPFS?

Thank you for your time, and for any sage advice on this process.

Chris
--
Chris Garrison
Indiana University
Research Systems Storage

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss


Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number
741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment-0002.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment-0006.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ecblank.gif
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment-0007.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 20386063.gif
Type: image/gif
Size: 360 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment-0008.gif>

From makaplan at us.ibm.com  Fri Apr 17 15:37:55 2015
From: makaplan at us.ibm.com (Marc A Kaplan)
Date: Fri, 17 Apr 2015 10:37:55 -0400
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <OFF10E0A53.76B236A9-ON85257E2A.004D54A7-85257E2A.004DF2E0@us.ibm.com>
References: <D15594F5.1ADED%ecgarris@iu.edu>	<CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
	<OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>
	<OFF10E0A53.76B236A9-ON85257E2A.004D54A7-85257E2A.004DF2E0@us.ibm.com>
Message-ID: <OF1D7F6052.1D03F850-ON85257E2A.004EEEB1-85257E2A.0050624D@us.ibm.com>

An "Encrypt-in-place" feature would not survive a serious, paranoid 
security audit.

If sensitive data ever was written to a disk, then the paranoid would say 
you can never be 100% sure that you have erased it.

That said, you decide how worried or paranoid you'd like to be and you can 
do an almost in place encryption as James Davis suggests by
simply copying the unencrypted file to a new file that will be subject to 
a GPFS encryption policy (SET ENCRYPTION) rule.

The safest way would be to copy-encrypt all the data to a new file system 
and then crush all the equipment that was used to store the clear-text 
files.

If crushing is too extreme, you might settle for a multipass sector by 
sector soft scrubbing by writing both carefully chosen and random data 
patterns.

Even then, unless you trust the manufacturer AND the manufacturer has 
provided you the means to "scrub" all the tracks/sectors of the disk you 
won't be sure...  Suppose that some sensitive data was written to a sector 
number that was later declared (partially) defective and remapped by the 
disk micro-code, and the original sector is put in a bad sector list that 
you can no longer address with standard disk driver software... 

Or it could be on some NVRAM or a disk that a service technician swapped 
out...  Ooops... it went out the door... and is now in the hands of the 
bad guys...





From:   James Davis/Poughkeepsie/IBM at IBMUS
To:     gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:   04/17/2015 10:12 AM
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to 
GPFS 4.1?
Sent by:        gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

Based on your question I think you are aware of this already, but just in 
case...

There is not currently an encrypt-in-place solution for GPFS encryption. A 
file's encryption state is determined at create-time. In order to encrypt 
your existing 100TB of data, you will need to apply an encryption policy 
to the current (or a new) GPFS file system(s) and then do something like:
  cp file file.enc #now file.new is encrypted
  mv file.enc file #replace the unencrypted file with the encrypted file

This can be done in parallel using mmapplypolicy if you want. In 4.1.1 
(forthcoming) I plan to provide an improved version of the 
/usr/lpp/mmfs/samples/ilm/mmfind (a find-esque interface to mmapplypolicy) 
that shipped in the last release; this should be an effective tool for the 
job.

Cheers,

Jamie

Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com

Daniel Kidger ---17-04-2015 08:16:54 AM---Hi Chris, The Crypto feature of 
GPFS 4.1  requires the addition of just one extra

From: Daniel Kidger <daniel.kidger at uk.ibm.com>
To: gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date: 17-04-15 08:16 AM
Subject: Re: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
Sent by: gpfsug-discuss-bounces at gpfsug.org



Hi Chris, 

The Crypto feature of GPFS 4.1  requires the addition of just one extra 
RPM over the standard edition. Other RPMs remain the same. 
It also requires licenses for  "Advanced Edition". These are of the order 
of 30% more expensive than the standard edition. 

The model for GPFS encryption at rest is that the client node fetches the 
encrypted file from the fileserver. 
The file remains encrypted in transfer and is only decrypted by the client 
node using a key it holds. 
A result is end-to-end encryption as well as encryption of data at rest, 

Encryption can be on a per file level if desired. Hence a way to migrate 
from an existing non-encrypted. setup. 
note inodes should be 4kB to make sure there is enough room to store the 
encryption attributes. 

A side effect of adding encryption is that you also need additional remote 
key management (RKM) servers to handle the key management. These need 
licensed software. 
see 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/com.ibm.cluster.gpfs.v4r104.gpfs200.doc/bl1adv_encryptionsetupreqs.htm 


I am sure DDN can help you with all of this. 

Hope this helps, 
Daniel 


Dr.Daniel Kidger 
No. 1 The Square, 

Technical Specialist  SDI (formerly Platform Computing) 
Temple Quay,
Bristol  BS1 6DG 
Mobile: 
+44-07818 522 266 
United Kingdom 
Landline: 
+44-02392 564 121 (Internal ITN 3726 9250) 
  

e-mail: 
daniel.kidger at uk.ibm.com 
 







From:        Frank Leers <fleers at gmail.com> 
To:        gpfsug main discussion list <gpfsug-discuss at gpfsug.org> 
Date:        16/04/2015 23:21 
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to 
GPFS 4.1? 
Sent by:        gpfsug-discuss-bounces at gpfsug.org 



Hi Chris, 

Since you mention GRIDScaler, I assume that you are a DDN customer and 
that you have a support contract with them.  If not, then feel free to 
take the advice that follows with a measure of salt although it mostly 
still applies ;-) 
With 4.1, there are now 'Editions' of GPFS, which break out various 
feature sets into a tiered arrangement, with each tier being licensed (and 
possibly priced) differently. 
Have a look here (Q 1.3) for the 4.1 licensing notes - 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html 
 

With GPFS 3.5, you are most likely running the equivalent of the 'Standard 
Edition' today.  The crypto feature set comes with the 'Advanced Edition', 
which is licensed differently. 

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a primer 
...  
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html 


-frank 


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> 
wrote: 
Hello, 

My site is working up to upgrading our paired GridScaler system from GPFS 
3.5 to 4.1. There is a mandate to provide encryption at rest, and that's 
an advertised feature of 4.1. We already have over 100 TB of data, 
synchronously replicated between two geographically separated sites, and 
we have concerns about how the upgrade, as well as the application of 
encryption to all that data, will go. 

I'd like to hear from admins who've been through this upgrade. What 
gotchas should we look out for? Can it easily be done in place, or would 
we need some extra equipment to "slosh" our data to and from so that it is 
written to an encrypted GPFS? 

Thank you for your time, and for any sage advice on this process. 

Chris 
-- 
Chris Garrison 
Indiana University 
Research Systems Storage 

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss


Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0002.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 21994 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0012.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0013.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0014.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 360 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0015.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0016.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0017.gif>

From secretary at gpfsug.org  Fri Apr 24 15:05:25 2015
From: secretary at gpfsug.org (Secretary GPFS UG)
Date: Fri, 24 Apr 2015 15:05:25 +0100
Subject: [gpfsug-discuss] GPFS User Group Meeting Agenda
Message-ID: <ebd76ace58f83d838895e0751950193b@webmail.gpfsug.org>

Dear Members,

The agenda for the next GPFS User Group Meeting is now available:
10:00 arrival for a 10:30 start
10:30                 Introductions - Jez Tucker, Group Chair & Claire 
Robson, Group Secretary
10:35                 Keynote - Doris Conti
10:50                 4.1.1 Roadmap / High-level futures - Scott Fadden
11:40                 Failure Events, Recovery & Problem determination - 
Scott Fadden
12:00                 Monitoring IBM Spectrum Scale using IBM Spectrum 
Control (VSC/TPC) - Christian Bolik
12:30                 Lunch
13:00                 User Experience from University of Birmingham & 
CLIMB - Simon Thompson
13:20                 User Experience from NERSC - Jason Hick
13:40                 AFM & Async DR Use Cases - Shankar Balasubramanian
14:25                 mmbackup + TSM Integration - Stefan Bender
15:10                 Break
15:25                 IBM Spectrum Scale (formerly GPFS) performance 
update, protocol performance and sizing - Sven Oehme
16:50                 Closing summary - Jez Tucker, Group Chair & Claire 
Robson, Group Secretary
17:00                 Ends

All attendees are invited to attend:
19:00 Buffet and drinks at York National Railway Museum

There are still places available for the event. If you have not 
registered and would like to attend, please email me, 
secretary at gpfsug.org with your name, job title, organisation, telephone 
and any dietary requirements.

We hope to see you in May!

Kind regards,

Claire

-- 
Claire Robson
GPFS User Group Secretary



From Dan.Foster at bristol.ac.uk  Wed Apr 29 08:51:03 2015
From: Dan.Foster at bristol.ac.uk (Dan Foster)
Date: Wed, 29 Apr 2015 08:51:03 +0100
Subject: [gpfsug-discuss] file system format version / software version
	compatibility matrix
Message-ID: <CALRUX+gQ2ZfN+7gWsadSUi2gUG+PaUkH0OrpEXAYDAUY-_LxRg@mail.gmail.com>

Hi All,

I'm trying to determine which versions of the GPFS file system format
are compatible with certain versions of the GPFS software.
Specifically in this instance I'm interested if file system format
v11.05 (from GPFS 3.3.0.2) is compatible with GPFS 3.5 .

The "File system format changes between versions of GPFS" [1] chapter
mentions format levels as old a v6, which infers that they are
compatible. But it would be useful to know for certain.

Thanks,
Dan.

[1] http://www-01.ibm.com/support/knowledgecenter/SSFKCN_3.5.0/com.ibm.cluster.gpfs.v3r5.gpfs100.doc/bl1adm_fsmigissues.htm?lang=en
-- 
Dan Foster | Senior Storage Systems Administrator | IT Services
e: dan.foster at bristol.ac.uk | t:  0117 3941170 [x41170]
m: Advanced Computing Research Centre, University of Bristol,
    8-10 Berkeley Square, Bristol BS8 1HH


From jamiedavis at us.ibm.com  Wed Apr 29 13:43:08 2015
From: jamiedavis at us.ibm.com (James Davis)
Date: Wed, 29 Apr 2015 08:43:08 -0400
Subject: [gpfsug-discuss] file system format version / software
	version	compatibility matrix
In-Reply-To: <CALRUX+gQ2ZfN+7gWsadSUi2gUG+PaUkH0OrpEXAYDAUY-_LxRg@mail.gmail.com>
References: <CALRUX+gQ2ZfN+7gWsadSUi2gUG+PaUkH0OrpEXAYDAUY-_LxRg@mail.gmail.com>
Message-ID: <OF27DFBC25.AE010664-ON85257E36.00453E1C-85257E36.0045DD6B@us.ibm.com>


Dan,

A "new" GPFS should be able to mount and interact with file systems with
"old" versions. Specifically I do not believe you will have trouble getting
GPFS 3.5 to talk to an FS created on GPFS 3.3.0.2.

The "Concepts, planning, and install guide" provides information about
migrating to GPFS 4.1 from GPFS 3.2 or earlier; this would implies file
systems created at GPFS 3.2 being supported on GPFS 4.1, which is even more
compatibility than you are asking about. The section to which I refer is
titled "Migrating to GPFS 4.1 from GPFS 3.2 or earlier releases of GPFS".

Note that without running mmchfs -V {full|compat} you will not have access
to some of the newer GPFS features. See the section titled "Completing the
migration to a new level of GPFS", also in the concepts guide.

Hope this helps,

Jamie

Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com



From:	Dan Foster <Dan.Foster at bristol.ac.uk>
To:	gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:	29-04-15 03:51 AM
Subject:	[gpfsug-discuss] file system format version / software version
            compatibility matrix
Sent by:	gpfsug-discuss-bounces at gpfsug.org



Hi All,

I'm trying to determine which versions of the GPFS file system format
are compatible with certain versions of the GPFS software.
Specifically in this instance I'm interested if file system format
v11.05 (from GPFS 3.3.0.2) is compatible with GPFS 3.5 .

The "File system format changes between versions of GPFS" [1] chapter
mentions format levels as old a v6, which infers that they are
compatible. But it would be useful to know for certain.

Thanks,
Dan.

[1]
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_3.5.0/com.ibm.cluster.gpfs.v3r5.gpfs100.doc/bl1adm_fsmigissues.htm?lang=en

--
Dan Foster | Senior Storage Systems Administrator | IT Services
e: dan.foster at bristol.ac.uk | t:  0117 3941170 [x41170]
m: Advanced Computing Research Centre, University of Bristol,
    8-10 Berkeley Square, Bristol BS8 1HH
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150429/23753578/attachment-0002.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150429/23753578/attachment-0002.gif>

From pavel.pokorny at datera.cz  Thu Apr  9 22:00:07 2015
From: pavel.pokorny at datera.cz (Pavel Pokorny)
Date: Thu, 9 Apr 2015 23:00:07 +0200
Subject: [gpfsug-discuss] Share GPFS from Windows?
Message-ID: <CAHy3nHJ9=D_zYq3xxFO8NnJutdy7JAWxkN7s_PZds277Nx4rxA@mail.gmail.com>

Hello to all,
Is there any technical reason why it is not supported to export GPFS from
Windows nodes using CIFS? As stated in GPFs FAQ (
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html
):

*Exporting GPFS file systems as Server Message Block (SMB) shares (also
known as CIFS shares) from GPFS Windows nodes is not supported.*

Or it is limitation more related to licensing and business point of view.
Thank you for answers,
Pavel

--
Ing. Pavel Pokorn?
DATERA s.r.o. | Ovocn? trh 580/2 | Praha | Czech Republic
www.datera.cz | Mobil: +420 602 357 194 | E-mail: pavel.pokorny at datera.cz
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150409/6f6c75d7/attachment-0003.htm>

From BOLIK at de.ibm.com  Fri Apr 10 12:24:14 2015
From: BOLIK at de.ibm.com (Christian Bolik)
Date: Fri, 10 Apr 2015 13:24:14 +0200
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
	multitude of GPFS clusters
Message-ID: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>


Just wanted to let you know that recently GPFS support has been added to
TPC, which is IBM's Tivoli Storage Productivity Center (soon to be renamed
to IBM Spectrum Control). As of now, TPC allows GPFS administrators to get
answers to the following questions, across any number of GPFS clusters
which have been added to TPC:

- Which of my clusters are running out of free space?
- Which of my clusters or nodes have a health problem?
- Which file systems and pools are running out of capacity?
- Which file systems are mounted on which nodes?
- How much space is occupied by snapshots? Are there any very old,
potentially obsolete ones?
- Which quotas are close to being exceeded or have already been exceeded?
- Which filesets are close to running out of free inodes?
- Which NSDs are at risk of becoming unavailable, or are unavailable?
- Are the volumes backing my NSDs performing OK?
- Are all nodes fulfilling critical roles in the cluster up and running?
- How can I be notified when nodes go offline or file systems fill up
beyond a threshold?

There's a short 6-minute video available on YouTube which shows how TPC
helps answering these questions:
https://www.youtube.com/watch?v=8Esk5U_cYw8&feature=youtu.be

For more information about TPC, please check out the product wiki on
developerWorks: http://ibm.co/1adWNFK

Thanks,
Christian Bolik
IBM Storage Software Development




From zgiles at gmail.com  Fri Apr 10 16:27:36 2015
From: zgiles at gmail.com (Zachary Giles)
Date: Fri, 10 Apr 2015 11:27:36 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>
References: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>
Message-ID: <CAMYZk=fKkAgeR-MW79a=BO46WpR9+Gm+=_kCLkQo4+v5KxgKqg@mail.gmail.com>

Christian:
Interesting and thanks for the latest news.

May I ask: Is there an intent moving forward that TPC and / or other Tivoli
products will be a required part of GPFS?
The concern I have is that GPFS is pretty straightforward at the moment and
has very logical requirements to operate (min servers, quorum, etc),
whereas there are many IBM products that require two or three more servers
just to manage the servers managing the service.. too much. It would be
nice to make sure, going forward, that the core of GPFS can still function
without additional web servers, Java, a suite of middleware, and a handful
of DB2 instance .. :)

-Zach



On Fri, Apr 10, 2015 at 7:24 AM, Christian Bolik <BOLIK at de.ibm.com> wrote:

>
> Just wanted to let you know that recently GPFS support has been added to
> TPC, which is IBM's Tivoli Storage Productivity Center (soon to be renamed
> to IBM Spectrum Control). As of now, TPC allows GPFS administrators to get
> answers to the following questions, across any number of GPFS clusters
> which have been added to TPC:
>
> - Which of my clusters are running out of free space?
> - Which of my clusters or nodes have a health problem?
> - Which file systems and pools are running out of capacity?
> - Which file systems are mounted on which nodes?
> - How much space is occupied by snapshots? Are there any very old,
> potentially obsolete ones?
> - Which quotas are close to being exceeded or have already been exceeded?
> - Which filesets are close to running out of free inodes?
> - Which NSDs are at risk of becoming unavailable, or are unavailable?
> - Are the volumes backing my NSDs performing OK?
> - Are all nodes fulfilling critical roles in the cluster up and running?
> - How can I be notified when nodes go offline or file systems fill up
> beyond a threshold?
>
> There's a short 6-minute video available on YouTube which shows how TPC
> helps answering these questions:
> https://www.youtube.com/watch?v=8Esk5U_cYw8&feature=youtu.be
>
> For more information about TPC, please check out the product wiki on
> developerWorks: http://ibm.co/1adWNFK
>
> Thanks,
> Christian Bolik
> IBM Storage Software Development
>
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>



-- 
Zach Giles
zgiles at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150410/c4ee09f6/attachment-0003.htm>

From dhildeb at us.ibm.com  Fri Apr 10 18:45:14 2015
From: dhildeb at us.ibm.com (Dean Hildebrand)
Date: Fri, 10 Apr 2015 10:45:14 -0700
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <CAMYZk=fKkAgeR-MW79a=BO46WpR9+Gm+=_kCLkQo4+v5KxgKqg@mail.gmail.com>
References: <OFC5DEFC6B.D2342AA0-ONC1257E23.003E127B-C1257E23.003EA4C2@de.ibm.com>
	<CAMYZk=fKkAgeR-MW79a=BO46WpR9+Gm+=_kCLkQo4+v5KxgKqg@mail.gmail.com>
Message-ID: <OFDFD967F3.9188C435-ON88257E23.0060AE82-88257E23.006186C8@us.ibm.com>


Hi Zach,

The summary is that GPFS is being integrated much more across the
portfolio...

With GPFS itself, there is a video below demonstrating the ESS/GSS GUI and
monitoring feature that is in the product today.  Moving forward, as you
can probably see, there is a push in IBM to move GPFS to software-defined,
which includes features such as the GUI as well...
https://www.youtube.com/watch?v=Mv9Sn-VYoGU

Dean




From:	Zachary Giles <zgiles at gmail.com>
To:	gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:	04/10/2015 08:27 AM
Subject:	Re: [gpfsug-discuss] Monitoring capacity and health status for
            a multitude of GPFS clusters
Sent by:	gpfsug-discuss-bounces at gpfsug.org



Christian:
Interesting and thanks for the latest news.

May I ask: Is there an intent moving forward that TPC and / or other Tivoli
products will be a required part of GPFS?
The concern I have is that GPFS is pretty straightforward at the moment and
has very logical requirements to operate (min servers, quorum, etc),
whereas there are many IBM products that require two or three more servers
just to manage the servers managing the service.. too much. It would be
nice to make sure, going forward, that the core of GPFS can still function
without additional web servers, Java, a suite of middleware, and a handful
of DB2 instance .. :)

-Zach



On Fri, Apr 10, 2015 at 7:24 AM, Christian Bolik <BOLIK at de.ibm.com> wrote:

  Just wanted to let you know that recently GPFS support has been added to
  TPC, which is IBM's Tivoli Storage Productivity Center (soon to be
  renamed
  to IBM Spectrum Control). As of now, TPC allows GPFS administrators to
  get
  answers to the following questions, across any number of GPFS clusters
  which have been added to TPC:

  - Which of my clusters are running out of free space?
  - Which of my clusters or nodes have a health problem?
  - Which file systems and pools are running out of capacity?
  - Which file systems are mounted on which nodes?
  - How much space is occupied by snapshots? Are there any very old,
  potentially obsolete ones?
  - Which quotas are close to being exceeded or have already been exceeded?
  - Which filesets are close to running out of free inodes?
  - Which NSDs are at risk of becoming unavailable, or are unavailable?
  - Are the volumes backing my NSDs performing OK?
  - Are all nodes fulfilling critical roles in the cluster up and running?
  - How can I be notified when nodes go offline or file systems fill up
  beyond a threshold?

  There's a short 6-minute video available on YouTube which shows how TPC
  helps answering these questions:
  https://www.youtube.com/watch?v=8Esk5U_cYw8&feature=youtu.be

  For more information about TPC, please check out the product wiki on
  developerWorks: http://ibm.co/1adWNFK

  Thanks,
  Christian Bolik
  IBM Storage Software Development


  _______________________________________________
  gpfsug-discuss mailing list
  gpfsug-discuss at gpfsug.org
  http://gpfsug.org/mailman/listinfo/gpfsug-discuss



--
Zach Giles
zgiles at gmail.com_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150410/9afe8bc4/attachment-0003.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150410/9afe8bc4/attachment-0003.gif>

From BOLIK at de.ibm.com  Mon Apr 13 14:11:01 2015
From: BOLIK at de.ibm.com (Christian Bolik)
Date: Mon, 13 Apr 2015 15:11:01 +0200
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
Message-ID: <OFE5AC06CE.E75B85AD-ONC1257E26.00399901-C1257E26.00486BB2@de.ibm.com>


Hi Zach,

I'm not aware of any intent to make TPC or any other Tivoli/IBM product a
prereq for GPFS, and I don't think any such plans exist.  Rather, as Dean
also pointed out, we're investing work to improve integration of
GPFS/Spectrum Scale into other products being members of the newly
announced IBM Spectrum Storage family, with the goal of improving
manageability of the individual components (rather than worsening it...).

Cheers,
Christian

> Christian:
> Interesting and thanks for the latest news.
>
> May I ask: Is there an intent moving forward that TPC and / or other
Tivoli
> products will be a required part of GPFS?
> The concern I have is that GPFS is pretty straightforward at the moment
and
> has very logical requirements to operate (min servers, quorum, etc),
> whereas there are many IBM products that require two or three more
servers
> just to manage the servers managing the service.. too much. It would be
> nice to make sure, going forward, that the core of GPFS can still
function
> without additional web servers, Java, a suite of middleware, and a
handful
> of DB2 instance .. :)
>
> -Zach

Christian Bolik

Software Defined Storage Development

IBM Deutschland Research & Development GmbH, Hechtsheimer Str. 2, 55131
Mainz, Germany
Vorsitzende des Aufsichtsrats: Martina Koederitz
Gesch?ftsf?hrung: Dirk Wittkopp
Sitz der Gesellschaft: B?blingen / Registergericht: Amtsgericht Stuttgart,
HRB 243294




From zgiles at gmail.com  Mon Apr 13 17:05:11 2015
From: zgiles at gmail.com (Zachary Giles)
Date: Mon, 13 Apr 2015 12:05:11 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <OFE5AC06CE.E75B85AD-ONC1257E26.00399901-C1257E26.00486BB2@de.ibm.com>
References: <OFE5AC06CE.E75B85AD-ONC1257E26.00399901-C1257E26.00486BB2@de.ibm.com>
Message-ID: <CAMYZk=eaQpwRcW6RM7WK9x4v8wO=o=GmZPrjCCq=dKa9K7H4Bg@mail.gmail.com>

Thanks for your replies. I can definitely appreciate the the goal of
improving management of components, and I agree that if GPFS will be using
within other products (which it is and will continue to be), then it would
be great for those products to be able to manage GPFS via an interface.

My fear with the idea of the above mentioned as a prereq is that the
"improvement" of management might looks like an improvement when you only
have 1 or 2 tiers of storage and 1 or 2 filesets called "database" and
another fileset called "user files", but if you have hundreds or thousands
of filesets and several tiers of storage, with both GSS and non-GSS systems
in the same cluster, then the GUI may actually be more cumbersome than the
original method. So, I just want to voice an opinion that we should
continue to be able to configure / maintain / monitor GPFS in a programatic
/ scriptable non-point-and-click way, if possible.



On Mon, Apr 13, 2015 at 9:11 AM, Christian Bolik <BOLIK at de.ibm.com> wrote:

>
> Hi Zach,
>
> I'm not aware of any intent to make TPC or any other Tivoli/IBM product a
> prereq for GPFS, and I don't think any such plans exist.  Rather, as Dean
> also pointed out, we're investing work to improve integration of
> GPFS/Spectrum Scale into other products being members of the newly
> announced IBM Spectrum Storage family, with the goal of improving
> manageability of the individual components (rather than worsening it...).
>
> Cheers,
> Christian
>
> > Christian:
> > Interesting and thanks for the latest news.
> >
> > May I ask: Is there an intent moving forward that TPC and / or other
> Tivoli
> > products will be a required part of GPFS?
> > The concern I have is that GPFS is pretty straightforward at the moment
> and
> > has very logical requirements to operate (min servers, quorum, etc),
> > whereas there are many IBM products that require two or three more
> servers
> > just to manage the servers managing the service.. too much. It would be
> > nice to make sure, going forward, that the core of GPFS can still
> function
> > without additional web servers, Java, a suite of middleware, and a
> handful
> > of DB2 instance .. :)
> >
> > -Zach
>
> Christian Bolik
>
> Software Defined Storage Development
>
> IBM Deutschland Research & Development GmbH, Hechtsheimer Str. 2, 55131
> Mainz, Germany
> Vorsitzende des Aufsichtsrats: Martina Koederitz
> Gesch?ftsf?hrung: Dirk Wittkopp
> Sitz der Gesellschaft: B?blingen / Registergericht: Amtsgericht Stuttgart,
> HRB 243294
>
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>



-- 
Zach Giles
zgiles at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150413/ec65870f/attachment-0003.htm>

From jamiedavis at us.ibm.com  Tue Apr 14 13:23:59 2015
From: jamiedavis at us.ibm.com (James Davis)
Date: Tue, 14 Apr 2015 08:23:59 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <mailman.1.1429009201.21161.gpfsug-discuss@gpfsug.org>
References: <mailman.1.1429009201.21161.gpfsug-discuss@gpfsug.org>
Message-ID: <OF41183321.B20BDD25-ON85257E27.00438322-85257E27.00441D50@us.ibm.com>


Zach,

Not sure if I'm formatting this message properly (need to switch off of
digest mode), but I know of no plans to replace the GPFS command line
interface with a GUI. With the GPFS GUI, TPC monitoring, etc., we want to
enable a wider variety of users to effectively use and manage GPFS, but the
command line will of course remain available for power users,
script-writers, etc. I certainly don't intend to throw away any of my nice
mm-themed test programs :-)

Jamie Davis
GPFS Test

Date: Mon, 13 Apr 2015 12:05:11 -0400
From: Zachary Giles <zgiles at gmail.com>
To: gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Subject: Re: [gpfsug-discuss] Monitoring capacity and health status
		 for a multitude of GPFS clusters
Message-ID:

<CAMYZk=eaQpwRcW6RM7WK9x4v8wO=o=GmZPrjCCq=dKa9K7H4Bg at mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

Thanks for your replies. I can definitely appreciate the the goal of
improving management of components, and I agree that if GPFS will be using
within other products (which it is and will continue to be), then it would
be great for those products to be able to manage GPFS via an interface.

My fear with the idea of the above mentioned as a prereq is that the
"improvement" of management might looks like an improvement when you only
have 1 or 2 tiers of storage and 1 or 2 filesets called "database" and
another fileset called "user files", but if you have hundreds or thousands
of filesets and several tiers of storage, with both GSS and non-GSS systems
in the same cluster, then the GUI may actually be more cumbersome than the
original method. So, I just want to voice an opinion that we should
continue to be able to configure / maintain / monitor GPFS in a programatic
/ scriptable non-point-and-click way, if possible.


Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150414/9352bfbe/attachment-0003.htm>

From ecgarris at iu.edu  Thu Apr 16 21:33:59 2015
From: ecgarris at iu.edu (Garrison, E Chris)
Date: Thu, 16 Apr 2015 20:33:59 +0000
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
Message-ID: <D15594F5.1ADED%ecgarris@iu.edu>

Hello,

My site is working up to upgrading our paired GridScaler system from GPFS 3.5 to 4.1. There is a mandate to provide encryption at rest, and that's an advertised feature of 4.1. We already have over 100 TB of data, synchronously replicated between two geographically separated sites, and we have concerns about how the upgrade, as well as the application of encryption to all that data, will go.

I'd like to hear from admins who've been through this upgrade. What gotchas should we look out for? Can it easily be done in place, or would we need some extra equipment to "slosh" our data to and from so that it is written to an encrypted GPFS?

Thank you for your time, and for any sage advice on this process.

Chris
--
Chris Garrison
Indiana University
Research Systems Storage
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150416/aeed3961/attachment-0003.htm>

From zgiles at gmail.com  Thu Apr 16 21:39:51 2015
From: zgiles at gmail.com (Zachary Giles)
Date: Thu, 16 Apr 2015 16:39:51 -0400
Subject: [gpfsug-discuss] Monitoring capacity and health status for a
 multitude of GPFS clusters
In-Reply-To: <OF41183321.B20BDD25-ON85257E27.00438322-85257E27.00441D50@us.ibm.com>
References: <mailman.1.1429009201.21161.gpfsug-discuss@gpfsug.org>
	<OF41183321.B20BDD25-ON85257E27.00438322-85257E27.00441D50@us.ibm.com>
Message-ID: <CAMYZk=dni=XHyMAcro6Y74tW+F-sVBRU4BCUm3vGOdxFDjOuaA@mail.gmail.com>

Thanks Jamie. I appreciate the input. Glad to hear it.

On Tue, Apr 14, 2015 at 8:23 AM, James Davis <jamiedavis at us.ibm.com> wrote:

> Zach,
>
> Not sure if I'm formatting this message properly (need to switch off of
> digest mode), but I know of no plans to replace the GPFS command line
> interface with a GUI. With the GPFS GUI, TPC monitoring, etc., we want to
> enable a wider variety of users to effectively use and manage GPFS, but the
> command line will of course remain available for power users,
> script-writers, etc. I certainly don't intend to throw away any of my nice
> mm-themed test programs :-)
>
> Jamie Davis
> GPFS Test
>
> Date: Mon, 13 Apr 2015 12:05:11 -0400
> From: Zachary Giles <zgiles at gmail.com>
> To: gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
> Subject: Re: [gpfsug-discuss] Monitoring capacity and health status
> for a multitude of GPFS clusters
> Message-ID:
> <CAMYZk=eaQpwRcW6RM7WK9x4v8wO=o=GmZPrjCCq=dKa9K7H4Bg at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Thanks for your replies. I can definitely appreciate the the goal of
> improving management of components, and I agree that if GPFS will be using
> within other products (which it is and will continue to be), then it would
> be great for those products to be able to manage GPFS via an interface.
>
> My fear with the idea of the above mentioned as a prereq is that the
> "improvement" of management might looks like an improvement when you only
> have 1 or 2 tiers of storage and 1 or 2 filesets called "database" and
> another fileset called "user files", but if you have hundreds or thousands
> of filesets and several tiers of storage, with both GSS and non-GSS systems
> in the same cluster, then the GUI may actually be more cumbersome than the
> original method. So, I just want to voice an opinion that we should
> continue to be able to configure / maintain / monitor GPFS in a programatic
> / scriptable non-point-and-click way, if possible.
>
>
> Jamie Davis
> GPFS Functional Verification Test (FVT)
> jamiedavis at us.ibm.com
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>
>


-- 
Zach Giles
zgiles at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150416/da9891e5/attachment-0003.htm>

From fleers at gmail.com  Thu Apr 16 23:21:34 2015
From: fleers at gmail.com (Frank Leers)
Date: Thu, 16 Apr 2015 15:21:34 -0700
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <D15594F5.1ADED%ecgarris@iu.edu>
References: <D15594F5.1ADED%ecgarris@iu.edu>
Message-ID: <CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>

Hi Chris,

Since you mention GRIDScaler, I assume that you are a DDN customer and that
you have a support contract with them.  If not, then feel free to take the
advice that follows with a measure of salt although it mostly still applies
;-)
With 4.1, there are now 'Editions' of GPFS, which break out various feature
sets into a tiered arrangement, with each tier being licensed (and possibly
priced) differently.
Have a look here (Q 1.3) for the 4.1 licensing notes -
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html


With GPFS 3.5, you are most likely running the equivalent of the 'Standard
Edition' today.  The crypto feature set comes with the 'Advanced Edition',
which is licensed differently.

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a primer
...
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html

-frank


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> wrote:

>  Hello,
>
>  My site is working up to upgrading our paired GridScaler system from
> GPFS 3.5 to 4.1. There is a mandate to provide encryption at rest, and
> that's an advertised feature of 4.1. We already have over 100 TB of data,
> synchronously replicated between two geographically separated sites, and we
> have concerns about how the upgrade, as well as the application of
> encryption to all that data, will go.
>
>  I'd like to hear from admins who've been through this upgrade. What
> gotchas should we look out for? Can it easily be done in place, or would we
> need some extra equipment to "slosh" our data to and from so that it is
> written to an encrypted GPFS?
>
>  Thank you for your time, and for any sage advice on this process.
>
>  Chris
> --
> Chris Garrison
> Indiana University
> Research Systems Storage
>
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at gpfsug.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150416/38ac4616/attachment-0003.htm>

From daniel.kidger at uk.ibm.com  Fri Apr 17 13:15:40 2015
From: daniel.kidger at uk.ibm.com (Daniel Kidger)
Date: Fri, 17 Apr 2015 13:15:40 +0100
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
References: <D15594F5.1ADED%ecgarris@iu.edu>
	<CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
Message-ID: <OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>

Hi Chris,

The Crypto feature of GPFS 4.1  requires the addition of just one extra 
RPM over the standard edition. Other RPMs remain the same.
It also requires licenses for  "Advanced Edition". These are of the order 
of 30% more expensive than the standard edition.

The model for GPFS encryption at rest is that the client node fetches the 
encrypted file from the fileserver. 
The file remains encrypted in transfer and is only decrypted by the client 
node using a key it holds.
A result is end-to-end encryption as well as encryption of data at rest,

Encryption can be on a per file level if desired. Hence a way to migrate 
from an existing non-encrypted. setup. 
note inodes should be 4kB to make sure there is enough room to store the 
encryption attributes.

A side effect of adding encryption is that you also need additional remote 
key management (RKM) servers to handle the key management. These need 
licensed software.
see 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/com.ibm.cluster.gpfs.v4r104.gpfs200.doc/bl1adv_encryptionsetupreqs.htm

I am sure DDN can help you with all of this.

Hope this helps,
Daniel



Dr.Daniel Kidger
No. 1 The Square,

Technical Specialist  SDI (formerly Platform Computing)
Temple Quay,
Bristol  BS1 6DG
Mobile:
+44-07818 522 266
United Kingdom
Landline:
+44-02392 564 121 (Internal ITN 3726 9250)
 

e-mail:
daniel.kidger at uk.ibm.com
 






From:   Frank Leers <fleers at gmail.com>
To:     gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:   16/04/2015 23:21
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to 
GPFS 4.1?
Sent by:        gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

Since you mention GRIDScaler, I assume that you are a DDN customer and 
that you have a support contract with them.  If not, then feel free to 
take the advice that follows with a measure of salt although it mostly 
still applies ;-)
With 4.1, there are now 'Editions' of GPFS, which break out various 
feature sets into a tiered arrangement, with each tier being licensed (and 
possibly priced) differently.
Have a look here (Q 1.3) for the 4.1 licensing notes - 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html
 

With GPFS 3.5, you are most likely running the equivalent of the 'Standard 
Edition' today.  The crypto feature set comes with the 'Advanced Edition', 
which is licensed differently.

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a primer 
...  
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html

-frank


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> 
wrote:
Hello,

My site is working up to upgrading our paired GridScaler system from GPFS 
3.5 to 4.1. There is a mandate to provide encryption at rest, and that's 
an advertised feature of 4.1. We already have over 100 TB of data, 
synchronously replicated between two geographically separated sites, and 
we have concerns about how the upgrade, as well as the application of 
encryption to all that data, will go.

I'd like to hear from admins who've been through this upgrade. What 
gotchas should we look out for? Can it easily be done in place, or would 
we need some extra equipment to "slosh" our data to and from so that it is 
written to an encrypted GPFS?

Thank you for your time, and for any sage advice on this process.

Chris
--
Chris Garrison
Indiana University
Research Systems Storage

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss


Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/a00ff70c/attachment-0003.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 360 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/a00ff70c/attachment-0003.gif>

From jamiedavis at us.ibm.com  Fri Apr 17 15:11:24 2015
From: jamiedavis at us.ibm.com (James Davis)
Date: Fri, 17 Apr 2015 10:11:24 -0400
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>
References: <D15594F5.1ADED%ecgarris@iu.edu>	<CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
	<OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>
Message-ID: <OFF10E0A53.76B236A9-ON85257E2A.004D54A7-85257E2A.004DF2E0@us.ibm.com>


Hi Chris,

Based on your question I think you are aware of this already, but just in
case...

There is not currently an encrypt-in-place solution for GPFS encryption. A
file's encryption state is determined at create-time. In order to encrypt
your existing 100TB of data, you will need to apply an encryption policy to
the current (or a new) GPFS file system(s) and then do something like:
  cp file file.enc #now file.new is encrypted
  mv file.enc file #replace the unencrypted file with the encrypted file

This can be done in parallel using mmapplypolicy if you want. In 4.1.1
(forthcoming) I plan to provide an improved version of
the /usr/lpp/mmfs/samples/ilm/mmfind (a find-esque interface to
mmapplypolicy) that shipped in the last release; this should be an
effective tool for the job.

Cheers,

Jamie

Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com



From:	Daniel Kidger <daniel.kidger at uk.ibm.com>
To:	gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:	17-04-15 08:16 AM
Subject:	Re: [gpfsug-discuss] Experiences upgrading in place to GPFS
            4.1?
Sent by:	gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

The Crypto feature of GPFS 4.1  requires the addition of just one extra RPM
over the standard edition. Other RPMs remain the same.
It also requires licenses for  "Advanced Edition". These are of the order
of 30% more expensive than the standard edition.

The model for GPFS encryption at rest is that the client node fetches the
encrypted file from the fileserver.
The file remains encrypted in transfer and is only decrypted by the client
node using a key it holds.
A result is end-to-end encryption as well as encryption of data at rest,

Encryption can be on a per file level if desired. Hence a way to migrate
from an existing non-encrypted. setup.
note inodes should be 4kB to make sure there is enough room to store the
encryption attributes.

A side effect of adding encryption is that you also need additional remote
key management (RKM) servers to handle the key management. These need
licensed software.
see
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/com.ibm.cluster.gpfs.v4r104.gpfs200.doc/bl1adv_encryptionsetupreqs.htm


I am sure DDN can help you with all of this.

Hope this helps,
Daniel
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
 Dr.Daniel                                   No. 1 The Square,                  
 Kidger                                                                         
                                                                                
 Technical                                   Temple Quay,                       
 Specialist                                  Bristol  BS1 6DG                   
 SDI                                                                            
 (formerly                                                                      
 Platform                                                                       
 Computing)                                                                     
                                                                                
                                                                                
 Mobile:    +44-07818 522 266                United Kingdom                     
                                                                                
 Landline:  +44-02392 564 121 (Internal ITN                                     
            3726 9250)                                                          
                                                                                
 e-mail:    daniel.kidger at uk.ibm.com                                            
                                                                                
                                                                                
                                                                                






From:        Frank Leers <fleers at gmail.com>
To:        gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:        16/04/2015 23:21
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to GPFS
4.1?
Sent by:        gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

Since you mention GRIDScaler, I assume that you are a DDN customer and that
you have a support contract with them.  If not, then feel free to take the
advice that follows with a measure of salt although it mostly still
applies ;-)
With 4.1, there are now 'Editions' of GPFS, which break out various feature
sets into a tiered arrangement, with each tier being licensed (and possibly
priced) differently.
Have a look here (Q 1.3) for the 4.1 licensing notes -
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html


With GPFS 3.5, you are most likely running the equivalent of the 'Standard
Edition' today.  The crypto feature set comes with the 'Advanced Edition',
which is licensed differently.

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a
primer ...
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html


-frank


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> wrote:

Hello,

My site is working up to upgrading our paired GridScaler system from GPFS
3.5 to 4.1. There is a mandate to provide encryption at rest, and that's an
advertised feature of 4.1. We already have over 100 TB of data,
synchronously replicated between two geographically separated sites, and we
have concerns about how the upgrade, as well as the application of
encryption to all that data, will go.

I'd like to hear from admins who've been through this upgrade. What gotchas
should we look out for? Can it easily be done in place, or would we need
some extra equipment to "slosh" our data to and from so that it is written
to an encrypted GPFS?

Thank you for your time, and for any sage advice on this process.

Chris
--
Chris Garrison
Indiana University
Research Systems Storage

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss


Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number
741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment-0003.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment-0009.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ecblank.gif
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment-0010.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 20386063.gif
Type: image/gif
Size: 360 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/d0bab395/attachment-0011.gif>

From makaplan at us.ibm.com  Fri Apr 17 15:37:55 2015
From: makaplan at us.ibm.com (Marc A Kaplan)
Date: Fri, 17 Apr 2015 10:37:55 -0400
Subject: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
In-Reply-To: <OFF10E0A53.76B236A9-ON85257E2A.004D54A7-85257E2A.004DF2E0@us.ibm.com>
References: <D15594F5.1ADED%ecgarris@iu.edu>	<CA+bjhhSbn_4th=K0EAWB73v+i59WdYZioA0aFx5-RpGeqQ=q+Q@mail.gmail.com>
	<OFF72F4342.0B67EA18-ON80257E2A.0040DE46-80257E2A.00435AE7@uk.ibm.com>
	<OFF10E0A53.76B236A9-ON85257E2A.004D54A7-85257E2A.004DF2E0@us.ibm.com>
Message-ID: <OF1D7F6052.1D03F850-ON85257E2A.004EEEB1-85257E2A.0050624D@us.ibm.com>

An "Encrypt-in-place" feature would not survive a serious, paranoid 
security audit.

If sensitive data ever was written to a disk, then the paranoid would say 
you can never be 100% sure that you have erased it.

That said, you decide how worried or paranoid you'd like to be and you can 
do an almost in place encryption as James Davis suggests by
simply copying the unencrypted file to a new file that will be subject to 
a GPFS encryption policy (SET ENCRYPTION) rule.

The safest way would be to copy-encrypt all the data to a new file system 
and then crush all the equipment that was used to store the clear-text 
files.

If crushing is too extreme, you might settle for a multipass sector by 
sector soft scrubbing by writing both carefully chosen and random data 
patterns.

Even then, unless you trust the manufacturer AND the manufacturer has 
provided you the means to "scrub" all the tracks/sectors of the disk you 
won't be sure...  Suppose that some sensitive data was written to a sector 
number that was later declared (partially) defective and remapped by the 
disk micro-code, and the original sector is put in a bad sector list that 
you can no longer address with standard disk driver software... 

Or it could be on some NVRAM or a disk that a service technician swapped 
out...  Ooops... it went out the door... and is now in the hands of the 
bad guys...





From:   James Davis/Poughkeepsie/IBM at IBMUS
To:     gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:   04/17/2015 10:12 AM
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to 
GPFS 4.1?
Sent by:        gpfsug-discuss-bounces at gpfsug.org



Hi Chris,

Based on your question I think you are aware of this already, but just in 
case...

There is not currently an encrypt-in-place solution for GPFS encryption. A 
file's encryption state is determined at create-time. In order to encrypt 
your existing 100TB of data, you will need to apply an encryption policy 
to the current (or a new) GPFS file system(s) and then do something like:
  cp file file.enc #now file.new is encrypted
  mv file.enc file #replace the unencrypted file with the encrypted file

This can be done in parallel using mmapplypolicy if you want. In 4.1.1 
(forthcoming) I plan to provide an improved version of the 
/usr/lpp/mmfs/samples/ilm/mmfind (a find-esque interface to mmapplypolicy) 
that shipped in the last release; this should be an effective tool for the 
job.

Cheers,

Jamie

Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com

Daniel Kidger ---17-04-2015 08:16:54 AM---Hi Chris, The Crypto feature of 
GPFS 4.1  requires the addition of just one extra

From: Daniel Kidger <daniel.kidger at uk.ibm.com>
To: gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date: 17-04-15 08:16 AM
Subject: Re: [gpfsug-discuss] Experiences upgrading in place to GPFS 4.1?
Sent by: gpfsug-discuss-bounces at gpfsug.org



Hi Chris, 

The Crypto feature of GPFS 4.1  requires the addition of just one extra 
RPM over the standard edition. Other RPMs remain the same. 
It also requires licenses for  "Advanced Edition". These are of the order 
of 30% more expensive than the standard edition. 

The model for GPFS encryption at rest is that the client node fetches the 
encrypted file from the fileserver. 
The file remains encrypted in transfer and is only decrypted by the client 
node using a key it holds. 
A result is end-to-end encryption as well as encryption of data at rest, 

Encryption can be on a per file level if desired. Hence a way to migrate 
from an existing non-encrypted. setup. 
note inodes should be 4kB to make sure there is enough room to store the 
encryption attributes. 

A side effect of adding encryption is that you also need additional remote 
key management (RKM) servers to handle the key management. These need 
licensed software. 
see 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/com.ibm.cluster.gpfs.v4r104.gpfs200.doc/bl1adv_encryptionsetupreqs.htm 


I am sure DDN can help you with all of this. 

Hope this helps, 
Daniel 


Dr.Daniel Kidger 
No. 1 The Square, 

Technical Specialist  SDI (formerly Platform Computing) 
Temple Quay,
Bristol  BS1 6DG 
Mobile: 
+44-07818 522 266 
United Kingdom 
Landline: 
+44-02392 564 121 (Internal ITN 3726 9250) 
  

e-mail: 
daniel.kidger at uk.ibm.com 
 







From:        Frank Leers <fleers at gmail.com> 
To:        gpfsug main discussion list <gpfsug-discuss at gpfsug.org> 
Date:        16/04/2015 23:21 
Subject:        Re: [gpfsug-discuss] Experiences upgrading in place to 
GPFS 4.1? 
Sent by:        gpfsug-discuss-bounces at gpfsug.org 



Hi Chris, 

Since you mention GRIDScaler, I assume that you are a DDN customer and 
that you have a support contract with them.  If not, then feel free to 
take the advice that follows with a measure of salt although it mostly 
still applies ;-) 
With 4.1, there are now 'Editions' of GPFS, which break out various 
feature sets into a tiered arrangement, with each tier being licensed (and 
possibly priced) differently. 
Have a look here (Q 1.3) for the 4.1 licensing notes - 
http://www-01.ibm.com/support/knowledgecenter/SSFKCN/com.ibm.cluster.gpfs.doc/gpfs_faqs/gpfsclustersfaq.html 
 

With GPFS 3.5, you are most likely running the equivalent of the 'Standard 
Edition' today.  The crypto feature set comes with the 'Advanced Edition', 
which is licensed differently. 

Have a look at Chapter 15 of the Advanced Admin Guide for 4.1 as a primer 
...  
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_4.1.0.4/gpfs4104_content.html 


-frank 


On Thu, Apr 16, 2015 at 1:33 PM, Garrison, E Chris <ecgarris at iu.edu> 
wrote: 
Hello, 

My site is working up to upgrading our paired GridScaler system from GPFS 
3.5 to 4.1. There is a mandate to provide encryption at rest, and that's 
an advertised feature of 4.1. We already have over 100 TB of data, 
synchronously replicated between two geographically separated sites, and 
we have concerns about how the upgrade, as well as the application of 
encryption to all that data, will go. 

I'd like to hear from admins who've been through this upgrade. What 
gotchas should we look out for? Can it easily be done in place, or would 
we need some extra equipment to "slosh" our data to and from so that it is 
written to an encrypted GPFS? 

Thank you for your time, and for any sage advice on this process. 

Chris 
-- 
Chris Garrison 
Indiana University 
Research Systems Storage 

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss


Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0003.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 21994 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0018.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0019.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0020.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 360 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0021.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0022.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150417/04bba1fb/attachment-0023.gif>

From secretary at gpfsug.org  Fri Apr 24 15:05:25 2015
From: secretary at gpfsug.org (Secretary GPFS UG)
Date: Fri, 24 Apr 2015 15:05:25 +0100
Subject: [gpfsug-discuss] GPFS User Group Meeting Agenda
Message-ID: <ebd76ace58f83d838895e0751950193b@webmail.gpfsug.org>

Dear Members,

The agenda for the next GPFS User Group Meeting is now available:
10:00 arrival for a 10:30 start
10:30                 Introductions - Jez Tucker, Group Chair & Claire 
Robson, Group Secretary
10:35                 Keynote - Doris Conti
10:50                 4.1.1 Roadmap / High-level futures - Scott Fadden
11:40                 Failure Events, Recovery & Problem determination - 
Scott Fadden
12:00                 Monitoring IBM Spectrum Scale using IBM Spectrum 
Control (VSC/TPC) - Christian Bolik
12:30                 Lunch
13:00                 User Experience from University of Birmingham & 
CLIMB - Simon Thompson
13:20                 User Experience from NERSC - Jason Hick
13:40                 AFM & Async DR Use Cases - Shankar Balasubramanian
14:25                 mmbackup + TSM Integration - Stefan Bender
15:10                 Break
15:25                 IBM Spectrum Scale (formerly GPFS) performance 
update, protocol performance and sizing - Sven Oehme
16:50                 Closing summary - Jez Tucker, Group Chair & Claire 
Robson, Group Secretary
17:00                 Ends

All attendees are invited to attend:
19:00 Buffet and drinks at York National Railway Museum

There are still places available for the event. If you have not 
registered and would like to attend, please email me, 
secretary at gpfsug.org with your name, job title, organisation, telephone 
and any dietary requirements.

We hope to see you in May!

Kind regards,

Claire

-- 
Claire Robson
GPFS User Group Secretary



From Dan.Foster at bristol.ac.uk  Wed Apr 29 08:51:03 2015
From: Dan.Foster at bristol.ac.uk (Dan Foster)
Date: Wed, 29 Apr 2015 08:51:03 +0100
Subject: [gpfsug-discuss] file system format version / software version
	compatibility matrix
Message-ID: <CALRUX+gQ2ZfN+7gWsadSUi2gUG+PaUkH0OrpEXAYDAUY-_LxRg@mail.gmail.com>

Hi All,

I'm trying to determine which versions of the GPFS file system format
are compatible with certain versions of the GPFS software.
Specifically in this instance I'm interested if file system format
v11.05 (from GPFS 3.3.0.2) is compatible with GPFS 3.5 .

The "File system format changes between versions of GPFS" [1] chapter
mentions format levels as old a v6, which infers that they are
compatible. But it would be useful to know for certain.

Thanks,
Dan.

[1] http://www-01.ibm.com/support/knowledgecenter/SSFKCN_3.5.0/com.ibm.cluster.gpfs.v3r5.gpfs100.doc/bl1adm_fsmigissues.htm?lang=en
-- 
Dan Foster | Senior Storage Systems Administrator | IT Services
e: dan.foster at bristol.ac.uk | t:  0117 3941170 [x41170]
m: Advanced Computing Research Centre, University of Bristol,
    8-10 Berkeley Square, Bristol BS8 1HH


From jamiedavis at us.ibm.com  Wed Apr 29 13:43:08 2015
From: jamiedavis at us.ibm.com (James Davis)
Date: Wed, 29 Apr 2015 08:43:08 -0400
Subject: [gpfsug-discuss] file system format version / software
	version	compatibility matrix
In-Reply-To: <CALRUX+gQ2ZfN+7gWsadSUi2gUG+PaUkH0OrpEXAYDAUY-_LxRg@mail.gmail.com>
References: <CALRUX+gQ2ZfN+7gWsadSUi2gUG+PaUkH0OrpEXAYDAUY-_LxRg@mail.gmail.com>
Message-ID: <OF27DFBC25.AE010664-ON85257E36.00453E1C-85257E36.0045DD6B@us.ibm.com>


Dan,

A "new" GPFS should be able to mount and interact with file systems with
"old" versions. Specifically I do not believe you will have trouble getting
GPFS 3.5 to talk to an FS created on GPFS 3.3.0.2.

The "Concepts, planning, and install guide" provides information about
migrating to GPFS 4.1 from GPFS 3.2 or earlier; this would implies file
systems created at GPFS 3.2 being supported on GPFS 4.1, which is even more
compatibility than you are asking about. The section to which I refer is
titled "Migrating to GPFS 4.1 from GPFS 3.2 or earlier releases of GPFS".

Note that without running mmchfs -V {full|compat} you will not have access
to some of the newer GPFS features. See the section titled "Completing the
migration to a new level of GPFS", also in the concepts guide.

Hope this helps,

Jamie

Jamie Davis
GPFS Functional Verification Test (FVT)
jamiedavis at us.ibm.com



From:	Dan Foster <Dan.Foster at bristol.ac.uk>
To:	gpfsug main discussion list <gpfsug-discuss at gpfsug.org>
Date:	29-04-15 03:51 AM
Subject:	[gpfsug-discuss] file system format version / software version
            compatibility matrix
Sent by:	gpfsug-discuss-bounces at gpfsug.org



Hi All,

I'm trying to determine which versions of the GPFS file system format
are compatible with certain versions of the GPFS software.
Specifically in this instance I'm interested if file system format
v11.05 (from GPFS 3.3.0.2) is compatible with GPFS 3.5 .

The "File system format changes between versions of GPFS" [1] chapter
mentions format levels as old a v6, which infers that they are
compatible. But it would be useful to know for certain.

Thanks,
Dan.

[1]
http://www-01.ibm.com/support/knowledgecenter/SSFKCN_3.5.0/com.ibm.cluster.gpfs.v3r5.gpfs100.doc/bl1adm_fsmigissues.htm?lang=en

--
Dan Foster | Senior Storage Systems Administrator | IT Services
e: dan.foster at bristol.ac.uk | t:  0117 3941170 [x41170]
m: Advanced Computing Research Centre, University of Bristol,
    8-10 Berkeley Square, Bristol BS8 1HH
_______________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at gpfsug.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150429/23753578/attachment-0003.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20150429/23753578/attachment-0003.gif>