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-2051 Lustre <p>Per comments in <a href="http://review.whamcloud.com/4058" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/4058</a>, we need to have some discussion/investigation on how the lod_alloc_rr() code is behaving under various circumstances. It isn't clear that it is as flexible as the previous lov alloc_rr/osc_precreate behaviour, and this is an area that has needed a lot of tuning in the past to work well under a variety of conditions.</p> LU-2051

verify lod_alloc_rr() code is doing what we want

Technical task LU-1303 Critical Resolved Fixed Alex Zhuravlev Andreas Dilger Fri, 28 Sep 2012 18:30:23 +0000 Mon, 29 May 2017 04:10:04 +0000 Mon, 29 May 2017 04:10:04 +0000 Lustre 2.4.0 Lustre 2.4.0 0 2 <p>any specific concern ?</p> <p>Several previous bugs come to mind:</p> <ul class="alternate" type="square"> <li>ensure that precreate count is scaling with load</li> <li>precreate starts early enough so that MDS create doesn't stall</li> <li>skip inactive or degraded OSTs <ul class="alternate" type="square"> <li>ideally ZFS OSD can set this flag itself?</li> </ul> </li> <li>RR allocation is properly balancing objects on OSTs</li> <li>QOS space balance works properly with imbalanced OSTs</li> </ul> <p>I think there are tests foremost of these conditions, but it would be good to verify this. </p> <p>Andreas, do you think we still need to do this? there was a patch from Seagate (landed now) where they improved RR and they reported it's now doing much more predictable.</p> <p>The lod_alloc_rr() code itself was fixed to avoid races in updating the state, but it still isn't clear if things like OST object precreation is working optimally (e.g. precreate starts early enough). In particular, there was a benefit shown at LUG a long time ago to start precreate when only 1/4 of the MDS objects were used, rather than waiting until 1/2 were used.</p> <p>The ZFS DEGRADED flag setting is <a href="https://jira.whamcloud.com/browse/LU-4277" title="Integrate ZFS zpool resilver status with OFD OS_STATE_DEGRADED flag" class="issue-link" data-issue-key="LU-4277"><del>LU-4277</del></a>.<br/> The QOS balance and RR balance is <a href="https://jira.whamcloud.com/browse/LU-9" title="Optimize weighted QOS Round-Robin allocator" class="issue-link" data-issue-key="LU-9">LU-9</a>.</p> Development Rank 1|hzv4m7: Rank (Obsolete) 4281 Story Points 2.0