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-2869 Lustre <p>The patch implements an extended attribute cache for<br/> a Lustre client. It is organized as a write-through<br/> cache: reads are performed from cache, updates are sent<br/> synchronously to the MDS. An additional inode bit<br/> is added to protect the cache.</p> <p>The patch itself will be uploaded shortly.</p> LU-2869

extended attribute cache for Lustre

Improvement Minor Resolved Fixed James Nunez Andrew Perepechko patch xattr Tue, 26 Feb 2013 17:12:46 +0000 Sat, 28 Oct 2023 07:43:50 +0000 Mon, 29 Jul 2013 22:50:36 +0000 Lustre 2.5.0 0 9 <p><a href="http://review.whamcloud.com/#change,5537" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/#change,5537</a></p> <p>xattr cache performance for the simplest Lustre configuration (2 nodes), read case is faster for the cached version as compared with the uncached by 4.5 times</p> <p>Xyratex <a href="http://jira-nss.xy01.xyratex.com:8080/browse/MRP-57" class="external-link" target="_blank" rel="nofollow noopener">MRP-57</a></p> <p>Nathan the xyratex url is broken (server not found error), can you put a valid/public one ?</p> <p>Sorry JC, the Xyratex tickets are really only useful for Xyratex engineers. We have posted the patch in the first comment above.</p> <p>Will this patch be inspected? It's almost three months from the patch upload. I had to rebase to a recent version.</p> <p>The master branch should open for new features soon.</p> <p>Thanks, James!</p> <p>Landed for 2.5</p> <p>can you explain why mot_xattr_sem is needed, please?</p> <p>The semaphore was added to protect the case when the client does not take the XATTR lock and requests FLXATTRALL.<br/> FLXATTRALL is implemented as getting the list and then getting the values, so setting/removing an xattr in the middle<br/> can lead to failure. Currently, there is no FLXATTRLS user that does not take the XATTR lock, so there is no use<br/> case for the semaphore being contended.</p> <p>well, if this failure is -ENOENT, then it should be easy to check and skip? on the otherhand we carry this structure for every object in MDT cache. a bit expensive for such a case?</p> <p>Besides -ENOENT, due to races an FLXATTRLS user can get an inconsistent list of attribute names and values, which the object never had from another client viewpoint. But since there is no use case for that, the locking can be simply dropped.</p> <p>not sure what kind of inconsistency FLXATTRLS can observe? any plans to make a patch to drop it?</p> <blockquote><p>not sure what kind of inconsistency FLXATTRLS can observe? </p></blockquote> <p>Like, setxattr "a=x", setxattr "b=y". An unlocked FLXATTRLS can pack "b=y" update without packing "a=x" because of update races.</p> <blockquote><p>any plans to make a patch to drop it?</p></blockquote> <p>I'll update <a href="http://review.whamcloud.com/#/c/7208/" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/#/c/7208/</a> with this fix in a while.</p> <p>great. thanks!</p> Related LU-3795 LU-3670 LU-3669 LU-549 LU-3583 LU-10276 LU-17238 Development Rank 1|hzvjtz: Rank (Obsolete) 6935