<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body style="word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">
Dear Eric,
<div class=""><br class="">
</div>
<div class="">thanks a lot for this information.</div>
<div class="">And what about the gpfs_vfs metric group?</div>
<div class="">What is the difference beteween for example</div>
<div class=""><br class="">
</div>
<div class="">“gpfs_fis_read_calls" and “gpfs_vfs_read” ?</div>
<div class=""><br class="">
</div>
<div class="">Again I see the second one being tipically higher than the first one.</div>
<div class="">In addition gpfs_vfs_read is not related to a specific file system...</div>
<div class=""><br class="">
</div>
<div class="">
<div class="">[root@ela5 ~]# mmperfmon query gpfs_fis_read_calls -n1 -b 60</div>
<div class=""><br class="">
</div>
<div class="">Legend:</div>
<div class=""> 1:<span class="Apple-tab-span" style="white-space:pre"> </span><a href="http://ela5.cscs.ch" class="">ela5.cscs.ch</a>|GPFSFilesystemAPI|<a href="http://durand.cscs.ch" class="">durand.cscs.ch</a>|store|gpfs_fis_read_calls</div>
<div class=""> 2:<span class="Apple-tab-span" style="white-space:pre"> </span><a href="http://ela5.cscs.ch" class="">ela5.cscs.ch</a>|GPFSFilesystemAPI|<a href="http://por.login.cscs.ch" class="">por.login.cscs.ch</a>|apps|gpfs_fis_read_calls</div>
<div class=""> 3:<span class="Apple-tab-span" style="white-space:pre"> </span><a href="http://ela5.cscs.ch" class="">ela5.cscs.ch</a>|GPFSFilesystemAPI|<a href="http://por.login.cscs.ch" class="">por.login.cscs.ch</a>|project|gpfs_fis_read_calls</div>
<div class=""> 4:<span class="Apple-tab-span" style="white-space:pre"> </span><a href="http://ela5.cscs.ch" class="">ela5.cscs.ch</a>|GPFSFilesystemAPI|<a href="http://por.login.cscs.ch" class="">por.login.cscs.ch</a>|users|gpfs_fis_read_calls</div>
<div class=""><br class="">
</div>
<div class="">Row Timestamp gpfs_fis_read_calls gpfs_fis_read_calls gpfs_fis_read_calls gpfs_fis_read_calls</div>
<div class=""> 1 2018-07-03-10:03:00 0 0 7274 0</div>
<div class=""><br class="">
</div>
<div class="">[root@ela5 ~]# mmperfmon query gpfs_vfs_read -n1 -b 60</div>
<div class=""><br class="">
</div>
<div class="">Legend:</div>
<div class=""> 1:<span class="Apple-tab-span" style="white-space:pre"> </span><a href="http://ela5.cscs.ch" class="">ela5.cscs.ch</a>|GPFSVFS|gpfs_vfs_read</div>
<div class=""><br class="">
</div>
<div class="">Row Timestamp gpfs_vfs_read</div>
<div class=""> 1 2018-07-03-10:03:00 45123</div>
</div>
<div class=""><br class="">
</div>
<div class="">Cheers,</div>
<div class=""><br class="">
</div>
<div class="">Giuseppe</div>
<div class=""><br class="">
</div>
<div class="">
<div class="">
<div style="color: rgb(0, 0, 0); letter-spacing: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px; word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class="">
<p class="p1" style="font-family: Tahoma; font-size: 13px;"><span class="s1">***********************************************************************</span></p>
<p class="p1" style="font-family: Tahoma; font-size: 13px;"><span class="s1">Giuseppe Lo Re</span></p>
<p class="p1" style="font-family: Tahoma; font-size: 13px;"><span class="s1">CSCS - Swiss National Supercomputing Center</span></p>
<p class="p1" style="font-family: Tahoma; font-size: 13px;"><span class="s1">Via Trevano 131</span></p>
<p class="p1" style="font-family: Tahoma; font-size: 13px;"><span class="s1">CH-6900 Lugano (TI) Tel: + 41 (0)91 610 8225</span></p>
<p class="p1" style="font-family: Tahoma; font-size: 13px;"><span class="s1">Switzerland <a href="mailto:giuseppe.lore@cscs.ch" class="">Email: giuseppe.lore@cscs.ch</a> </span></p>
<p class="p1" style="font-family: Tahoma; font-size: 13px;">***********************************************************************</p>
</div>
</div>
<div><br class="">
<blockquote type="cite" class="">
<div class=""><br class="">
</div>
<div class="">
<div class="">Hello Giuseppe,<br class="">
<br class="">
Following was my attempt to answer a similar question some months ago.<br class="">
When reading about the different viewpoints of the Zimon sensors, please<br class="">
note that gpfs_fis_bytes_read is a metric provided by the GPFSFileSystemAPI<br class="">
sensor, while gpfs_fs_bytes_read is a metric provided by the GPFSFileSystem<br class="">
sensor. Therefore, gpfs_fis_bytes_read reflects application reads, while<br class="">
gpfs_fs_bytes_read reflects NSD reads.<br class="">
<br class="">
The GPFSFileSystemAPI and GPFSNodeAPI sensor metrics are from the point of<br class="">
view of "applications" in the sense that they provide stats about I/O<br class="">
requests made to files in GPFS file systems from user level applications<br class="">
using POSIX interfaces like open(), close(), read(), write(), etc.<br class="">
<br class="">
This is in contrast to similarly named sensors without the "API" suffix,<br class="">
like GPFSFilesystem and GPFSNode. Those sensors provide stats about I/O<br class="">
requests made by the GPFS code to NSDs (disks) making up GPFS file systems.<br class="">
<br class="">
The relationship between application I/O and disk I/O might or might not be<br class="">
obvious. Consider some examples. An application that starts sequentially<br class="">
reading a file might, at least initially, cause more disk I/O than expected<br class="">
because GPFS has decided to prefetch data. An application write() might<br class="">
not immediately cause the writing of disk blocks, due to the operation of<br class="">
the pagepool. Ultimately, application write()s might cause twice as much<br class="">
data written to disk due to the replication factor of the file system.<br class="">
Application I/O concerns itself with user data; disk I/O might have to<br class="">
occur to handle the user data and associated file system metadata (like<br class="">
inodes and indirect blocks).<br class="">
<br class="">
The difference between GPFSFileSystemAPI and GPFSNodeAPI: GPFSFileSystemAPI<br class="">
reports stats for application I/O per filesystem per node; GPFSNodeAPI<br class="">
reports application I/O stats per node. Similarly, GPFSFilesystem reports<br class="">
stats for disk I/O per filesystem per node; GPFSNode reports disk I/O stats<br class="">
per node.<br class="">
<br class="">
<br class="">
Eric M. Agar<br class="">
<a href="mailto:agar@us.ibm.com" class="">agar@us.ibm.com</a><br class="">
<br class="">
IBM Spectrum Scale Level 2<br class="">
Software Defined Infrastructure, IBM Systems<br class="">
<br class="">
<br class="">
<br class="">
<br class="">
<br class="">
From:<span class="Apple-tab-span" style="white-space:pre"> </span>Kristy Kallback-Rose <kkr@lbl.gov><br class="">
To:<span class="Apple-tab-span" style="white-space:pre"> </span>gpfsug main discussion list <gpfsug-discuss@spectrumscale.org><br class="">
Date:<span class="Apple-tab-span" style="white-space:pre"> </span>07/02/2018 10:06 AM<br class="">
Subject:<span class="Apple-tab-span" style="white-space:pre"> </span>Re: [gpfsug-discuss] Zimon metrics details<br class="">
Sent by:<span class="Apple-tab-span" style="white-space:pre"> </span>gpfsug-discuss-bounces@spectrumscale.org<br class="">
<br class="">
<br class="">
<br class="">
+1<br class="">
<br class="">
Would love to see more detailed descriptions on Zimon metrics.<br class="">
<br class="">
Sent from my iPhone<br class="">
<br class="">
On Jul 2, 2018, at 6:50 AM, Lo Re Giuseppe <lore@cscs.ch> wrote:<br class="">
<br class="">
Hi everybody,<br class="">
<br class="">
I am extracting the Zimon performance data and uploading them to our<br class="">
elasticsearch cluster.<br class="">
Now that I have the mechanism in place it?s time to understand what I<br class="">
am actually uploading ;)<br class="">
<br class="">
Maybe this has been already asked.. where can I find a (as much as<br class="">
possible) detailed explaination of the different Zimon metrics?<br class="">
The SS probelm determination guide doens?t spend more than half a<br class="">
line for each.<br class="">
<br class="">
In particular I would like to understand the difference between these<br class="">
ones:<br class="">
<br class="">
- gpfs_fs_bytes_read<br class="">
- gpfs_fis_bytes_read<br class="">
<br class="">
The second gives tipically higher values than the first one.<br class="">
<br class="">
Thanks for any hit.<br class="">
<br class="">
Regards,<br class="">
<br class="">
Giuseppe<br class="">
<br class="">
<br class="">
<br class="">
***********************************************************************<br class="">
<br class="">
<br class="">
Giuseppe Lo Re<br class="">
<br class="">
<br class="">
CSCS - Swiss National Supercomputing Center<br class="">
<br class="">
<br class="">
Via Trevano 131<br class="">
<br class="">
<br class="">
CH-6900 Lugano (TI) Tel: + 41 (0)91 610 8225<br class="">
<br class="">
<br class="">
Switzerland Email:<br class="">
giuseppe.lore@cscs.ch<br class="">
<br class="">
<br class="">
***********************************************************************<br class="">
<br class="">
<br class="">
<br class="">
_______________________________________________<br class="">
gpfsug-discuss mailing list<br class="">
gpfsug-discuss at spectrumscale.org<br class="">
http://gpfsug.org/mailman/listinfo/gpfsug-discuss<br class="">
_______________________________________________<br class="">
gpfsug-discuss mailing list<br class="">
gpfsug-discuss at spectrumscale.org<br class="">
http://gpfsug.org/mailman/listinfo/gpfsug-discuss<br class="">
<br class="">
<br class="">
<br class="">
-------------- next part --------------<br class="">
An HTML attachment was scrubbed...<br class="">
URL: <http://gpfsug.org/pipermail/gpfsug-discuss/attachments/20180702/3aa08500/attachment.html><br class="">
-------------- next part --------------<br class="">
A non-text attachment was scrubbed...<br class="">
Name: graycol.gif<br class="">
Type: image/gif<br class="">
Size: 105 bytes<br class="">
Desc: not available<br class="">
URL: <http://gpfsug.org/pipermail/gpfsug-discuss/attachments/20180702/3aa08500/attachment.gif><br class="">
<br class="">
------------------------------<br class="">
<br class="">
_______________________________________________<br class="">
gpfsug-discuss mailing list<br class="">
gpfsug-discuss at spectrumscale.org<br class="">
http://gpfsug.org/mailman/listinfo/gpfsug-discuss<br class="">
<br class="">
<br class="">
End of gpfsug-discuss Digest, Vol 78, Issue 6<br class="">
*********************************************<br class="">
</div>
</div>
</blockquote>
</div>
<br class="">
</div>
</body>
</html>