<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="">
Hi Fred,
<div class=""><br class="">
</div>
<div class="">I have a total of 48 NSDs served up by 8 NSD servers. 12 of those NSDs are in our small /home filesystem, which is performing just fine. The other 36 are in our ~1 PB /scratch and /data filesystem, which is where the problem is. Our max filesystem
block size parameter is set to 16 MB, but the aforementioned filesystem uses a 1 MB block size.</div>
<div class=""><br class="">
</div>
<div class="">nsdMaxWorkerThreads is set to 1024 as shown below. Since each NSD server serves an average of 6 NSDs and 6 x 12 = 72 we’re OK if I’m understanding the calculation correctly. Even multiplying 48 x 12 = 576, so we’re good?!?</div>
<div class=""><br class="">
</div>
<div class="">Your help is much appreciated! Thanks again…</div>
<div class=""><br class="">
</div>
<div class="">Kevin<br class="">
<div><br class="">
<blockquote type="cite" class="">
<div class="">On Jul 3, 2018, at 4:53 PM, Frederick Stock <<a href="mailto:stockf@us.ibm.com" class="">stockf@us.ibm.com</a>> wrote:</div>
<br class="Apple-interchange-newline">
<div class=""><font size="3" face="Arial" class="">How many NSDs are served by the NSD servers and what is your maximum file system block size? Have you confirmed that you have sufficient NSD worker threads to handle the maximum number of IOs you are configured
to have active? That would be the number of NSDs served times 12 (you have 12 threads per queue).</font><br class="">
<br class="">
<font size="3" face="sans-serif" class="">Fred<br class="">
__________________________________________________<br class="">
Fred Stock | IBM Pittsburgh Lab | 720-430-8821<br class="">
<a href="mailto:stockf@us.ibm.com" class="">stockf@us.ibm.com</a></font><br class="">
<br class="">
<br class="">
<br class="">
<font size="1" color="#5f5f5f" face="sans-serif" class="">From: </font><font size="1" face="sans-serif" class="">"Buterbaugh, Kevin L" <<a href="mailto:Kevin.Buterbaugh@Vanderbilt.Edu" class="">Kevin.Buterbaugh@Vanderbilt.Edu</a>></font><br class="">
<font size="1" color="#5f5f5f" face="sans-serif" class="">To: </font><font size="1" face="sans-serif" class="">gpfsug main discussion list <<a href="mailto:gpfsug-discuss@spectrumscale.org" class="">gpfsug-discuss@spectrumscale.org</a>></font><br class="">
<font size="1" color="#5f5f5f" face="sans-serif" class="">Date: </font><font size="1" face="sans-serif" class="">07/03/2018 05:41 PM</font><br class="">
<font size="1" color="#5f5f5f" face="sans-serif" class="">Subject: </font><font size="1" face="sans-serif" class="">Re: [gpfsug-discuss] High I/O wait times</font><br class="">
<font size="1" color="#5f5f5f" face="sans-serif" class="">Sent by: </font><font size="1" face="sans-serif" class=""><a href="mailto:gpfsug-discuss-bounces@spectrumscale.org" class="">gpfsug-discuss-bounces@spectrumscale.org</a></font><br class="">
<hr noshade="" class="">
<br class="">
<br class="">
<br class="">
<font size="3" class="">Hi Fred, </font><br class="">
<br class="">
<font size="3" class="">Thanks for the response. I have been looking at the “mmfsadm dump nsd” data from the two NSD servers that serve up the two NSDs that most commonly experience high wait times (although, again, this varies from time to time). In addition,
I have been reading:</font><br class="">
<br class="">
<a href="https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.ibm.com%2Fdeveloperworks%2Fcommunity%2Fwikis%2Fhome%3Flang%3Den%23!%2Fwiki%2FGeneral+Parallel+File+System+(GPFS)%2Fpage%2FNSD+Server+Design+and+Tuning&data=02%7C01%7CKevin.Buterbaugh%40vanderbilt.edu%7C7658e1b458b147ad8a3908d5e12f6982%7Cba5a7f39e3be4ab3b45067fa80faecad%7C0%7C0%7C636662516110903567&sdata=cWw5UipcO7HgupLQTFgOWVwXF%2B9b8S%2Fw935%2FeqG6xIY%3D&reserved=0" originalsrc="https://www.ibm.com/developerworks/community/wikis/home?lang=en#!/wiki/General Parallel File System (GPFS)/page/NSD Server Design and Tuning" shash="yiyIM+n0cleiCGjr6iVBPpjzqsYco7rrUmYCugVWK0s7ZGBITUDKvmfgBgefKMlH5pKmegJOT7BDHKb9lCBERlMiR+fZpeQAuNqxIebaP0SbroAELopYJ6XrlA9JRDhQZJW8g7DSqlgPsnhAAllbR4RcEH2ch3YyA2K+Dqd/Atc=" class=""><font size="3" color="blue" class=""><u class="">https://www.ibm.com/developerworks/community/wikis/home?lang=en#!/wiki/General%20Parallel%20File%20System%20(GPFS)/page/NSD%20Server%20Design%20and%20Tuning</u></font></a><br class="">
<br class="">
<font size="3" class="">And:</font><br class="">
<br class="">
<a href="https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.ibm.com%2Fdeveloperworks%2Fcommunity%2Fwikis%2Fhome%3Flang%3Den%23!%2Fwiki%2FGeneral%2520Parallel%2520File%2520System%2520(GPFS)%2Fpage%2FNSD%2520Server%2520Tuning&data=02%7C01%7CKevin.Buterbaugh%40vanderbilt.edu%7C7658e1b458b147ad8a3908d5e12f6982%7Cba5a7f39e3be4ab3b45067fa80faecad%7C0%7C0%7C636662516110903567&sdata=CAuOPOhC1MXdZW2e2HaVOY0PmySwP6FzlsvNNlteWZw%3D&reserved=0" originalsrc="https://www.ibm.com/developerworks/community/wikis/home?lang=en#!/wiki/General%20Parallel%20File%20System%20(GPFS)/page/NSD%20Server%20Tuning" shash="myAxRzpQPE449qqW8iD74ZDdvftszJV/DBIaQM+iCKOO6b+xRHqz1qnhFUZ6OOQLfBTs7NDHpqxGZG0xugrPfpi3tr9KlcnKD8e9pVoMfpwcGmSbMlSfNO+mMxFtC5H9P/UEr3CnRRS3XluCXg+GAULo3cKEtyTEQOKazrNPfLE=" class=""><font size="3" color="blue" class=""><u class="">https://www.ibm.com/developerworks/community/wikis/home?lang=en#!/wiki/General%20Parallel%20File%20System%20(GPFS)/page/NSD%20Server%20Tuning</u></font></a><br class="">
<br class="">
<font size="3" class="">Which seem to be the most relevant documents on the Wiki.</font><br class="">
<br class="">
<font size="3" class="">I would like to do a more detailed analysis of the “mmfsadm dump nsd” output, but my preliminary looks at it seems to indicate that I see I/O’s queueing in the 50 - 100 range for the small queues and the 60 - 200 range on the large queues.</font><br class="">
<br class="">
<font size="3" class="">In addition, I am regularly seeing all 12 threads on the LARGE queues active, while it is much more rare that I see all - or even close to all - the threads on the SMALL queues active.</font><br class="">
<br class="">
<font size="3" class="">As far as the parameters Scott and Yuri mention, on our cluster they are set thusly:</font><br class="">
<br class="">
<font size="3" class="">[common]</font><br class="">
<font size="3" class="">nsdMaxWorkerThreads 640</font><br class="">
<font size="3" class="">[<all the GPFS servers listed here>]</font><br class="">
<font size="3" class="">nsdMaxWorkerThreads 1024</font><br class="">
<font size="3" class="">[common]</font><br class="">
<font size="3" class="">nsdThreadsPerQueue 4</font><br class="">
<font size="3" class="">[<all the GPFS servers listed here>]</font><br class="">
<font size="3" class="">nsdThreadsPerQueue 12</font><br class="">
<font size="3" class="">[common]</font><br class="">
<font size="3" class="">nsdSmallThreadRatio 3</font><br class="">
<font size="3" class="">[<all the GPFS servers listed here>]</font><br class="">
<font size="3" class="">nsdSmallThreadRatio 1</font><br class="">
<br class="">
<font size="3" class="">So to me it sounds like I need more resources on the LARGE queue side of things … i.e. it sure doesn’t sound like I want to change my small thread ratio. If I increase the amount of threads it sounds like that might help, but that also
takes more pagepool, and I’ve got limited RAM in these (old) NSD servers. I do have nsdbufspace set to 70, but I’ve only got 16-24 GB RAM each in these NSD servers. And a while back I did try increase the page pool on them (very slightly) and ended up causing
problems because then they ran out of physical RAM.</font><br class="">
<br class="">
<font size="3" class="">Thoughts? Followup questions? Thanks!</font><br class="">
<br class="">
<font size="3" class="">Kevin</font><br class="">
<br class="">
<font size="3" class="">On Jul 3, 2018, at 3:11 PM, Frederick Stock <<a href="mailto:stockf@us.ibm.com" class="">stockf@us.ibm.com</a>> wrote:</font><br class="">
<br class="">
<font size="3" face="Arial" class="">Are you seeing similar values for all the nodes or just some of them? One possible issue is how the NSD queues are configured on the NSD servers. You can see this with the output of "mmfsadm dump nsd". There are queues
for LARGE IOs (greater than 64K) and queues for SMALL IOs (64K or less). Check the highest pending values to see if many IOs are queueing. There are a couple of options to fix this but rather than explain them I suggest you look for information about NSD
queueing on the developerWorks site. There has been information posted there that should prove helpful.</font><font size="3" class=""><br class="">
</font><font size="3" face="sans-serif" class=""><br class="">
Fred<br class="">
__________________________________________________<br class="">
Fred Stock | IBM Pittsburgh Lab | 720-430-8821<br class="">
<a href="mailto:stockf@us.ibm.com" class="">stockf@us.ibm.com</a></font><font size="3" class=""><br class="">
<br class="">
<br class="">
</font><font size="1" color="#5f5f5f" face="sans-serif" class=""><br class="">
From: </font><font size="1" face="sans-serif" class="">"Buterbaugh, Kevin L" <<a href="mailto:Kevin.Buterbaugh@Vanderbilt.Edu" class="">Kevin.Buterbaugh@Vanderbilt.Edu</a>></font><font size="1" color="#5f5f5f" face="sans-serif" class=""><br class="">
To: </font><font size="1" face="sans-serif" class="">gpfsug main discussion list <<a href="mailto:gpfsug-discuss@spectrumscale.org" class="">gpfsug-discuss@spectrumscale.org</a>></font><font size="1" color="#5f5f5f" face="sans-serif" class=""><br class="">
Date: </font><font size="1" face="sans-serif" class="">07/03/2018 03:49 PM</font><font size="1" color="#5f5f5f" face="sans-serif" class=""><br class="">
Subject: </font><font size="1" face="sans-serif" class="">[gpfsug-discuss] High I/O wait times</font><font size="1" color="#5f5f5f" face="sans-serif" class=""><br class="">
Sent by: </font><font size="1" face="sans-serif" class=""><a href="mailto:gpfsug-discuss-bounces@spectrumscale.org" class="">gpfsug-discuss-bounces@spectrumscale.org</a></font><font size="3" class=""><br class="">
</font>
<hr noshade="" class="">
<font size="3" class=""><br class="">
<br class="">
<br class="">
Hi all, <br class="">
<br class="">
We are experiencing some high I/O wait times (5 - 20 seconds!) on some of our NSDs as reported by “mmdiag —iohist" and are struggling to understand why. One of the confusing things is that, while certain NSDs tend to show the problem more than others, the
problem is not consistent … i.e. the problem tends to move around from NSD to NSD (and storage array to storage array) whenever we check … which is sometimes just a few minutes apart.<br class="">
<br class="">
In the past when I have seen “mmdiag —iohist” report high wait times like this it has *always* been hardware related. In our environment, the most common cause has been a battery backup unit on a storage array controller going bad and the storage array switching
to write straight to disk. But that’s *not* happening this time. <br class="">
<br class="">
Is there anything within GPFS / outside of a hardware issue that I should be looking for?? Thanks!<br class="">
<br class="">
—<br class="">
Kevin Buterbaugh - Senior System Administrator<br class="">
Vanderbilt University - Advanced Computing Center for Research and Education</font><font size="3" color="blue" class=""><u class=""><br class="">
</u></font><a href="mailto:Kevin.Buterbaugh@vanderbilt.edu" class=""><font size="3" color="blue" class=""><u class="">Kevin.Buterbaugh@vanderbilt.edu</u></font></a><font size="3" class="">- (615)875-9633<br class="">
<br class="">
</font><tt class=""><font size="2" class=""><br class="">
_______________________________________________<br class="">
gpfsug-discuss mailing list<br class="">
gpfsug-discuss at <a href="http://spectrumscale.org" class="">spectrumscale.org</a></font></tt><font size="3" color="blue" class=""><u class=""><br class="">
</u></font><a href="https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fgpfsug.org%2Fmailman%2Flistinfo%2Fgpfsug-discuss&data=02%7C01%7CKevin.Buterbaugh%40vanderbilt.edu%7Cd3d7ff675bb440286cb908d5e1212b66%7Cba5a7f39e3be4ab3b45067fa80faecad%7C0%7C0%7C636662454938066010&sdata=jL0pB5MEaWtJZjMbS8JzhsKGvwmYB6qV%2FVyosdUKcSU%3D&reserved=0" class=""><tt class=""><font size="2" color="blue" class=""><u class="">http://gpfsug.org/mailman/listinfo/gpfsug-discuss</u></font></tt></a><font size="3" class=""><br class="">
<br class="">
<br class="">
<br class="">
_______________________________________________<br class="">
gpfsug-discuss mailing list<br class="">
gpfsug-discuss at <a href="http://spectrumscale.org" class="">spectrumscale.org</a><br class="">
</font><a href="https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fgpfsug.org%2Fmailman%2Flistinfo%2Fgpfsug-discuss&data=02%7C01%7CKevin.Buterbaugh%40vanderbilt.edu%7Cd3d7ff675bb440286cb908d5e1212b66%7Cba5a7f39e3be4ab3b45067fa80faecad%7C0%7C0%7C636662454938076014&sdata=wIyB66HoqvL13I3LX0Ott%2Btr7HQQdInZ028QUp0QMhE%3D&reserved=0" class=""><font size="3" class="">https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fgpfsug.org%2Fmailman%2Flistinfo%2Fgpfsug-discuss&data=02%7C01%7CKevin.Buterbaugh%40vanderbilt.edu%7Cd3d7ff675bb440286cb908d5e1212b66%7Cba5a7f39e3be4ab3b45067fa80faecad%7C0%7C0%7C636662454938076014&sdata=wIyB66HoqvL13I3LX0Ott%2Btr7HQQdInZ028QUp0QMhE%3D&reserved=0</font></a><br class="">
<tt class=""><font size="2" class="">_______________________________________________<br class="">
gpfsug-discuss mailing list<br class="">
gpfsug-discuss at <a href="http://spectrumscale.org" class="">spectrumscale.org</a><br class="">
</font></tt><a href="https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fgpfsug.org%2Fmailman%2Flistinfo%2Fgpfsug-discuss&data=02%7C01%7CKevin.Buterbaugh%40vanderbilt.edu%7C7658e1b458b147ad8a3908d5e12f6982%7Cba5a7f39e3be4ab3b45067fa80faecad%7C0%7C0%7C636662516110923583&sdata=wbJuSS8E2n66iUoS3JiKjnly%2BESuuDpqlCLKGdM61rc%3D&reserved=0" originalsrc="http://gpfsug.org/mailman/listinfo/gpfsug-discuss" shash="x4NtVTjqpv/PxRsbKCVNmQN/PeVRQfkliyRwtzLFxGkyEWcgVCxaLClAtiiEvOoqDy2QeptAhrQEKYAGkT20ww42qy9kNAo0qrD7LXJ6MmYBBnRfcqFJzrs9Z86PC+yWe13gg/s0IO5D25skqKNBdH5WuJkg931QUGakVBZBJiA=" class=""><tt class=""><font size="2" class="">http://gpfsug.org/mailman/listinfo/gpfsug-discuss</font></tt></a><tt class=""><font size="2" class=""><br class="">
</font></tt><br class="">
<br class="">
<br class="">
_______________________________________________<br class="">
gpfsug-discuss mailing list<br class="">
gpfsug-discuss at <a href="http://spectrumscale.org" class="">spectrumscale.org</a><br class="">
<a href="https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fgpfsug.org%2Fmailman%2Flistinfo%2Fgpfsug-discuss&data=02%7C01%7CKevin.Buterbaugh%40vanderbilt.edu%7C7658e1b458b147ad8a3908d5e12f6982%7Cba5a7f39e3be4ab3b45067fa80faecad%7C0%7C0%7C636662516110933587&sdata=RKuWKLRGoBRMSDHkrMsKsuU6JkiFgruK4e7gGafxAGc%3D&reserved=0" class="">https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fgpfsug.org%2Fmailman%2Flistinfo%2Fgpfsug-discuss&data=02%7C01%7CKevin.Buterbaugh%40vanderbilt.edu%7C7658e1b458b147ad8a3908d5e12f6982%7Cba5a7f39e3be4ab3b45067fa80faecad%7C0%7C0%7C636662516110933587&sdata=RKuWKLRGoBRMSDHkrMsKsuU6JkiFgruK4e7gGafxAGc%3D&reserved=0</a><br class="">
</div>
</blockquote>
</div>
<br class="">
</div>
</body>
</html>