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-7795 Lustre <p>In order to better adjust filesystem performance when using quotas we would like to have a tunable least_qunit option. </p> LU-7795

least_qunit should be tunnable

Improvement Minor Resolved Fixed Niu Yawei Mahmoud Hanafi Thu, 18 Feb 2016 20:25:33 +0000 Wed, 4 Dec 2019 09:07:29 +0000 Fri, 27 May 2016 12:47:00 +0000 Lustre 2.5.3 Lustre 2.9.0 0 6 <p>Niu</p> <p>Could you please comment on this suggestion?</p> <p>Thanks</p> <p>Peter</p> <p>Mahmoud, could you please explain a bit on how you want to tune the least_qunit? Do you want to increase it or decrease it, what performance impact do you see today, and how do you expect the tuning of least_qunit to change the performance?</p> <p>We have reports from users that when they exceed their soft quota there is significant performance impact. We had reported this in <a href="https://jira.whamcloud.com/browse/LU-4139" title="Significant perforamce issue when user over soft quota limit" class="issue-link" data-issue-key="LU-4139"><del>LU-4139</del></a> and was address to some extent, but we would want more controller over qunit allocation in such situations. </p> <p>Hi, Mahmoud , according to the numbers in <a href="https://jira.whamcloud.com/browse/LU-4139" title="Significant perforamce issue when user over soft quota limit" class="issue-link" data-issue-key="LU-4139"><del>LU-4139</del></a>, looks the performance is reasonable after the fixing:</p> <blockquote> <p>Direct I/O<br/> UnderSoftlimit: 383MB/sec<br/> OverSoftlimit: 359MB/sec<br/> Buffered I/O<br/> UnderSoftlimit:316MB.sec<br/> OverSoftlimit: 304MB/sec</p></blockquote> <p>Is your proposal to let user set a larger least_qunit so that the performance won't drop even after over soft limit? One thing worth noting is that larger least_qunit means worse limit/grace time accuracy, so we have to do some sort of compromise here. What do you think about?</p> <p>I think there are some IO patterns that are effected by the softquota limit then others. At least that is what our users are reporting. Since most of our quota are 100'sTB and we have large gap between soft and hard quota we are willing to give up limit/grace time accuracy for no lost of performance. </p> <p>I see, the requirement makes sense to me, but I don't think tuning the least_qunit is a proper way, because it affects the hard limit as well.</p> <p>A better way is to add a proc interface to quota master (qmt), and user may turn off the qunit adjustment for soft limit (no qunit adjustment even if soft limit is exceeded) via this interface. Of course, people should be warned that grace time accuracy will be lost when he turns it off.</p> <p>I can compose a patch if you don't against. Thanks.</p> <p>Are you purposing a per-user setting or global for a filesystem? Turn off qunit adjustment must a admin only function.</p> <p>It's global setting, and least qunit is also a global parameter.</p> <p>That sounds like what we would need.</p> <p>Niu Yawei (yawei.niu@intel.com) uploaded a new patch: <a href="http://review.whamcloud.com/18896" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/18896</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-7795" title="least_qunit should be tunnable" class="issue-link" data-issue-key="LU-7795"><del>LU-7795</del></a> quota: tuneable soft least qunit<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: 601fa84017a98450432d5cb3dbdea0cd8b82e056</p> <p>Can you provide a back port to 2.5.3-fe?<br/> It is not obvious to me how to resolve the conflicts on lprocfs_quota_qpi_vars declaration.</p> <p>Thanks!</p> <p>Here is the port to b2_5_fe: <a href="http://review.whamcloud.com/#/c/18976/" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/#/c/18976/</a></p> <p>With this patch when is client sync writes triggered?</p> <p>The client sync write will be triggered when approaching/exceeding hard limit, if there is no hard limit, sync write won't be triggered. </p> <p>Oleg Drokin (oleg.drokin@intel.com) merged in patch <a href="http://review.whamcloud.com/18896/" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/18896/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-7795" title="least_qunit should be tunnable" class="issue-link" data-issue-key="LU-7795"><del>LU-7795</del></a> quota: tuneable soft least qunit<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: <br/> Commit: 0cadce268a09a94824b82bc1a61e98963582c423</p> <p>Landed for 2.9</p> Related LU-4139 Development Rank 1|hzy1sn: Rank (Obsolete) 9223372036854775807