[gpfsug-discuss] CES, Ganesha, and Filesystem_id

Wed Jun 28 14:49:54 BST 2023

Hi Christof,

thanks! So the preferred way should be not to have Filesystem_Id. If I 
have understood correctly, this is anyhow set up by default in CES to 
666.666, so the suggested procedure during the CES setup should be to 
manually modify gpfs.ganesha.exports.conf and remove this parameter from 
all the exports, is that correct? Is there an easier way, or is there a 
plan to remove the 666.666 default value?

In our case, we do rely on two separated CES clusters (one in prod and 
one in stand by, so that we can perform upgrades with no downtime by 
migrating the IPs from one to the other), so it might be safer to 
explicilty set Filesystem_Id, to ensure consistency among the cluster - 
would that make sense?

Thanks again!

Regards

leo

Paul Scherrer Institut
Dr. Leonardo Sala
Group Leader Data Analysis and Research Infrastructure
Deputy Department Head a.i Science IT Infrastructure and Services department
Science IT Infrastructure and Services department (AWI)
WHGA/036
Forschungstrasse 111
5232 Villigen PSI
Switzerland

Phone: +41 56 310 3369
leonardo.sala at psi.ch
www.psi.ch

On 6/28/23 15:22, Christof Schmitt wrote:
> After another discussion it turns out that this parameter is not 
> required. While my previous comment is correct, that there is the need 
> to have unique handles across file systems, GPFS already provides that 
> information and Ganesha handles that correctly. So there is no need to 
> set the parameter in the Ganesha config.
> Regards,
>
> Christof
>
> On Wed, 2023-06-28 at 14:33 +0200, Leonardo Sala wrote:
>> Hi Christof, thanks a lot! In our case we are exporting multiple 
>> filesets from 2 filesystems, I guess we should fix unique Fileset_IDs 
>> for each fileset? What would happen in case we just remove the 
>> Fileset_id parameter, as suggested by the
>> ZjQcmQRYFpfptBannerStart
>> This Message Is From an External Sender
>> This message came from outside your organization.
>> Report Suspicious
>> <https://us-phishalarm-ewt.proofpoint.com/EWT/v1/PjiDSg!2k-hAt5_AJM6OsYbOMsPTZmPil8BIAHadeloM-6vntuXy4WCtpq_71hHj07-QEB3YfmuMTIe9Ezg-qdLzL_6DvzGmA12g21EJgQ$>
>> ZjQcmQRYFpfptBannerEnd
>>
>> Hi Christof,
>>
>> thanks a lot! In our case we are exporting multiple filesets from 2 
>> filesystems, I guess we should fix unique Fileset_IDs for each 
>> fileset? What would happen in case we just remove the Fileset_id 
>> parameter, as suggested by the ganesha docs?
>>
>> Regards
>>
>> leo
>>
>> Paul Scherrer Institut
>> Dr. Leonardo Sala
>> Group Leader Data Analysis and Research Infrastructure
>> Deputy Department Head a.i Science IT Infrastructure and Services department
>> Science IT Infrastructure and Services department (AWI)
>> WHGA/036
>> Forschungstrasse 111
>> 5232 Villigen PSI
>> Switzerland
>> Phone: +41 56 310 3369
>> leonardo.sala at psi.ch
>> www.psi.ch
>> On 6/28/23 14:22, Christof Schmitt wrote:
>>> The "FileSystem_Id" is a unique identifier for the file system. The 
>>> technical background is that Ganesha asks the file system for a file 
>>> handle, but that is only unique within the file system. If there are 
>>> NFS exports on different file systems, there needs to be a way to 
>>> make the file handles unique across multiple file systems. So if 
>>> there are NFS exports on different file systems, this parameter 
>>> should be set with a unique value for each file system. If there is 
>>> only one file system with NFS exports, then this should not be 
>>> necessary.
>>>
>>> Regards,
>>>
>>> Christof
>>>
>>> On Wed, 2023-06-28 at 08:53 +0200, Leonardo Sala wrote:
>>>> Hi Ed, thanks! In our case we do have unique export ids, but the 
>>>> same fsid, and this seems to create issues. Also, reading Ganesha 
>>>> docs, I can see [*]: FileSystem_ID EXPORT Option There is an EXPORT 
>>>> config option, FileSystem_ID. This really
>>>> ZjQcmQRYFpfptBannerStart
>>>> This Message Is From an External Sender
>>>> This message came from outside your organization.
>>>> Report Suspicious
>>>> <https://us-phishalarm-ewt.proofpoint.com/EWT/v1/PjiDSg!2k-ror9_yrma2gYRGMtvzfBw-f7-KrLp6ZZNwHChWJfOvNhVe0t6WfmouxHnuoc3gtMvXaxWV3AbFrgYAQT1JIEKoq1hoJ8cr1c$>
>>>> ZjQcmQRYFpfptBannerEnd
>>>>
>>>> Hi Ed,
>>>>
>>>> thanks! In our case we do have unique export ids, but the same 
>>>> fsid, and this seems to create issues. Also, reading Ganesha docs, 
>>>> I can see [*]:
>>>>
>>>>
>>>>       FileSystem_ID EXPORT Option
>>>>
>>>> There is an EXPORT config option, FileSystem_ID. This really should 
>>>> not be used, all it does it designate an fsid to be used with the 
>>>> attributes of all objects in the export. It will be folded to fit 
>>>> into NFSv3. Because it applies to the entire export, it prevents 
>>>> exporting multiple file systems since there will likely be issues 
>>>> with collision of inode numbers on the client.
>>>>
>>>> so before touching the defaults in GPFS CES configuration I would 
>>>> like some guidance or experiences from this mlist :)
>>>>
>>>> cheers
>>>>
>>>> leo
>>>>
>>>> [*] 
>>>> https://github.com/nfs-ganesha/nfs-ganesha/wiki/File-Systems#FileSystem_ID_EXPORT_Option
>>>>
>>>>
>>>> Paul Scherrer Institut
>>>> Dr. Leonardo Sala
>>>> Group Leader Data Analysis and Research Infrastructure
>>>> Deputy Department Head a.i Science IT Infrastructure and Services department
>>>> Science IT Infrastructure and Services department (AWI)
>>>> WHGA/036
>>>> Forschungstrasse 111
>>>> 5232 Villigen PSI
>>>> Switzerland
>>>> Phone: +41 56 310 3369
>>>> leonardo.sala at psi.ch
>>>> www.psi.ch
>>>> On 6/27/23 19:41, Wahl, Edward wrote:
>>>>>
>>>>> I vaguely recall seeing this and testing it.  My notes to myself 
>>>>> say: ‘As long as the export_id is unique, you are fine.’   See the 
>>>>> manuals, ganesha loves Camel Case so it’s more than likely 
>>>>> actually “Export_Id” or some such.
>>>>>
>>>>> Ed Wahl
>>>>>
>>>>> Ohio Supercomputer Center
>>>>>
>>>>> *From:*gpfsug-discuss <gpfsug-discuss-bounces at gpfsug.org> *On 
>>>>> Behalf Of *Leonardo Sala
>>>>> *Sent:* Tuesday, June 27, 2023 10:18 AM
>>>>> *To:* gpfsug-discuss at spectrumscale.org
>>>>> *Subject:* [gpfsug-discuss] CES, Ganesha, and Filesystem_id
>>>>>
>>>>> Hallo, we are checking our current CES configuration, and we 
>>>>> noticed that by default GPFS puts always Filesystem_Id=666. 666 
>>>>> [*], no matter which Export_Id value the export has. To my 
>>>>> understanding (which is poor!), this means that all clients
>>>>>
>>>>> Hallo,
>>>>>
>>>>> we are checking our current CES configuration, and we noticed that 
>>>>> by default GPFS puts always Filesystem_Id=666.666 [*], no matter 
>>>>> which Export_Id value the export has. To my understanding (which 
>>>>> is poor!), this means that all clients will see all our exports 
>>>>> (~20) with the same device number, creating various possible 
>>>>> issues (e.g. file state handles). Questions:
>>>>>
>>>>> * is there a reason for such default value? If we change it, are 
>>>>> there unpleasant effects we could see?
>>>>>
>>>>> * what would be a reasonable value? Looking around I saw that 
>>>>> Filesystem_Id = Export_Id.Export_Id is quite common, with the 
>>>>> possible issue of using the forbidden 152.152 [**]
>>>>>
>>>>> * what happens if we actually remove the Filesytem_Id parameter 
>>>>> from gpfs.ganesha.exports.conf?
>>>>>
>>>>> * is there a way to modify Filesystem_Id in 
>>>>> gpfs.ganesha.exports.conf without editing the file, eg using mmnfs 
>>>>> commands (seems not, but I might be mistaken)?
>>>>>
>>>>> Thanks a lot!
>>>>>
>>>>> cheers
>>>>>
>>>>> leo
>>>>>
>>>>> [*] 
>>>>> https://www.ibm.com/docs/en/storage-scale/5.0.4?topic=exports-making-bulk-changes-nfs
>>>>>
>>>>> [**] https://github.com/nfs-ganesha/nfs-ganesha/issues/615
>>>>>
>>>>> -- 
>>>>> Paul Scherrer Institut
>>>>> Dr. Leonardo Sala
>>>>> Group Leader Data Analysis and Research Infrastructure
>>>>> Deputy Department Head a.i Science IT Infrastructure and Services department
>>>>> Science IT Infrastructure and Services department (AWI)
>>>>> WHGA/036
>>>>> Forschungstrasse 111
>>>>> 5232 Villigen PSI
>>>>> Switzerland
>>>>>   
>>>>> Phone: +41 56 310 3369
>>>>> leonardo.sala at psi.ch
>>>>> www.psi.ch
>>>>> <http://www.psi.ch>
>>>>>
>>>>> _______________________________________________
>>>>> gpfsug-discuss mailing list
>>>>> gpfsug-discuss at gpfsug.org
>>>>> http://gpfsug.org/mailman/listinfo/gpfsug-discuss_gpfsug.org
>>>> _______________________________________________
>>>> gpfsug-discuss mailing list
>>>> gpfsug-discuss at gpfsug.org
>>>> http://gpfsug.org/mailman/listinfo/gpfsug-discuss_gpfsug.org
>>>> <http://gpfsug.org/mailman/listinfo/gpfsug-discuss_gpfsug.org>
>>>>   
>>>
>>> _______________________________________________
>>> gpfsug-discuss mailing list
>>> gpfsug-discuss at gpfsug.org
>>> http://gpfsug.org/mailman/listinfo/gpfsug-discuss_gpfsug.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20230628/4951c61a/attachment-0003.htm>