https://jira.whamcloud.com This file is an XML representation of an issue en-us 9.4.14 940014 05-12-2023 https://jira.whamcloud.com/browse/LU-9114 Lustre <p>It's somewhat widely seen in various logs that pacemaker complaints its thread was not scheduled for tens of seconds which is way too excessive.<br/> Indeed MDS is pretty cpu hungry, but we need to ensure we insert enough of schedule points so that other processes get a shot at CPU too.</p> <p>There are also some bandaids discussed like using numa settings to cordon off one cpu from use by Lustre, but those are just that - bandaids.</p> <p>We probably can play with various debug settings that warn about this and make the timeouts lower to try and catch more of the offenders. Likely have a bunch in flock code with its double loops</p> LU-9114

Make MDS (And other server threads?) hog CPU less

Improvement Minor Resolved Fixed Andreas Dilger Oleg Drokin Tue, 14 Feb 2017 00:28:25 +0000 Thu, 17 Dec 2020 18:29:58 +0000 Thu, 17 Dec 2020 18:29:58 +0000 Lustre 2.14.0 0 4 <p>In addition to checking if any one MDS thread running too long without scheduling, it may also be that the many MDS kernel threads are scheduled with a higher priority and prevent the userspace threads from being run. I think for pacemaker and such, it makes sense to <tt>mlock()</tt> the heartbeat daemons into memory (so they aren't swapped) and run them with realtime priority (or something like <tt>nice -15</tt>) so that they can always get CPU time even when all of the MDS threads are running.</p> <p>Dmitry</p> <p>Can you please investigate this area as a longer term task for 2018</p> <p>Peter</p> <p>Andreas Dilger (adilger@whamcloud.com) uploaded a new patch: <a href="https://review.whamcloud.com/39435" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/39435</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-9114" title="Make MDS (And other server threads?) hog CPU less" class="issue-link" data-issue-key="LU-9114"><del>LU-9114</del></a> ldlm: don't compute sumsq for pool stats<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: bbe08711a531af70404ef1ba5ffe17815cede034</p> <p>Oleg Drokin (green@whamcloud.com) merged in patch <a href="https://review.whamcloud.com/39435/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/39435/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-9114" title="Make MDS (And other server threads?) hog CPU less" class="issue-link" data-issue-key="LU-9114"><del>LU-9114</del></a> ldlm: don't compute sumsq for pool stats<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: <br/> Commit: 966f6bb550be52e4bf5dd7fd38a0d707fe2a5072</p> <p>Landed for 2.14</p> Related LU-4423 Development Rank 1|hzz3pz: Rank (Obsolete) 9223372036854775807