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-11290 Lustre <p>Lock cancellation can be very time consuming when there are many pages under a lock.</p> <p>One easy area for improvement is the osc_page_gang_lookup callback functions - These work on individual pages from an array, but they can be trivially modified to work on the entire array at once, reducing overhead.</p> <p>I've got a simple patch for this, which improves performance of lock cancellation by about 10%, and truncate <span class="error">&#91;when pages are in cache&#93;</span> (which also uses the call back) by about 5%.</p> LU-11290

Batch callbacks in osc_page_gang_lookup

Bug Minor Resolved Fixed Dongyang Li Patrick Farrell Tue, 28 Aug 2018 19:51:39 +0000 Wed, 2 Jun 2021 23:26:35 +0000 Wed, 2 Jun 2021 23:26:35 +0000 Lustre 2.15.0 0 8 <p>Patrick Farrell (paf@cray.com) uploaded a new patch: <a href="https://review.whamcloud.com/33089" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/33089</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-11290" title="Batch callbacks in osc_page_gang_lookup" class="issue-link" data-issue-key="LU-11290"><del>LU-11290</del></a>: Batch gang_lookup cbs<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: 5c53ea14823b8fd710be967390a3bf84fc55c725</p> <p>Alexander Zarochentsev (alexander.zarochentsev@hpe.com) uploaded a new patch: <a href="https://review.whamcloud.com/39327" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/39327</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-11290" title="Batch callbacks in osc_page_gang_lookup" class="issue-link" data-issue-key="LU-11290"><del>LU-11290</del></a> ldlm: page discard speedup<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: 7731839accebd6f1dae77ad7757a074dedbefb0f</p> <p>pages discard speedup from <a href="https://review.whamcloud.com/39327" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/39327</a> can be seen in a simple test with a 30G cached file:</p> <div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent"> <pre>[root@c-lmo037 ~]# ls -lh /mnt/cslmo17/ total 31G -rw-r--r-- 1 root root 31G Jun 4 17:11 20G drwxrwxrwx 2 root root 4.0K Jun 3 18:18 d0.metabench [root@c-lmo037 ~]# </pre> </div></div> <p>no fix: </p> <div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent"> <pre>[root@c-lmo037 ~]# cat /mnt/testfs/20G &gt; /dev/null; time echo 1 &gt;&gt; /mnt/cslmo17/20G real 0m9.622s user 0m0.001s sys 0m0.000s [root@c-lmo037 ~]# cat /mnt/testfs/20G &gt; /dev/null; time echo 1 &gt;&gt; /mnt/cslmo17/20G real 0m9.421s user 0m0.000s sys 0m0.001s [root@c-lmo037 ~]# </pre> </div></div> <p>and with the fix:</p> <div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent"> <pre>[root@c-lmo037 ~]# cat /mnt/testfs/20G &gt; /dev/null; time echo 1 &gt;&gt; /mnt/cslmo17/20G real 0m7.140s user 0m0.001s sys 0m0.000s [root@c-lmo037 ~]# cat /mnt/testfs/20G &gt; /dev/null; time echo 1 &gt;&gt; /mnt/cslmo17/20G real 0m7.240s user 0m0.000s sys 0m0.001s [root@c-lmo037 ~]# </pre> </div></div> <p>23% less time spent.</p> <p>Oleg Drokin (green@whamcloud.com) merged in patch <a href="https://review.whamcloud.com/39327/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/39327/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-11290" title="Batch callbacks in osc_page_gang_lookup" class="issue-link" data-issue-key="LU-11290"><del>LU-11290</del></a> ldlm: page discard speedup<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: <br/> Commit: 0f48cd0b9856fe1ea920b8abab3579ded0b9511e</p> <p>Does the original patch from <a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=paf" class="user-hover" rel="paf">paf</a> still need to land or can it be abandoned?</p> <p>I think patch: <a href="https://review.whamcloud.com/33089" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/33089</a> "<tt><a href="https://jira.whamcloud.com/browse/LU-11290" title="Batch callbacks in osc_page_gang_lookup" class="issue-link" data-issue-key="LU-11290"><del>LU-11290</del></a>: Batch gang_lookup cbs</tt>" is still useful to land, as it improves a different part of the code.</p> <p>&gt;Batch gang_lookup cbs" is still useful to land, as it improves a different part of the code.<br/> yes it gives another 5% to lock cancel speed. I will upload it soon.</p> <p>Oleg Drokin (green@whamcloud.com) merged in patch <a href="https://review.whamcloud.com/33089/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/33089/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-11290" title="Batch callbacks in osc_page_gang_lookup" class="issue-link" data-issue-key="LU-11290"><del>LU-11290</del></a> osc: Batch gang_lookup cbs<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: <br/> Commit: 0d6d0b7bc95a82dee02d35d0a8a41d24692cad45</p> <p>Landed for 2.15</p> Related LU-14711 LU-9920 Development Rank 1|i001d3: Rank (Obsolete) 9223372036854775807 Severity