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-6148 Lustre <p>This ticket is a placeholder for further information forthcoming.</p> <p>Briefly, strided locking is a proposed new feature to make improved shared file IO performance possible in certain situations. This is being proposed at the Lustre developers meeting at LLNL (currently ongoing).</p> <p>Details will be forthcoming, for now, this ticket is mostly to allow me to push the prototype to Gerrit for ease of viewing.</p> LU-6148

Strided lock proposal - Feature proposal for 2.8

New Feature Minor Resolved Won't Do WC Triage Patrick Farrell Thu, 22 Jan 2015 17:14:04 +0000 Sat, 16 Sep 2017 07:58:47 +0000 Sat, 16 Sep 2017 07:58:33 +0000 0 6 <p>Patrick Farrell (paf@cray.com) uploaded a new patch: <a href="http://review.whamcloud.com/13498" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/13498</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-6148" title="Strided lock proposal - Feature proposal for 2.8" class="issue-link" data-issue-key="LU-6148"><del>LU-6148</del></a> ldlm: Strided locking prototype<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: 590c2b30d66021aac4f4a31fb05754036092c0dd</p> <p>Prototype/proof of concept here:<br/> <a href="http://review.whamcloud.com/13498" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/13498</a></p> <p>Presentation slides from developer's conference.</p> <p>what about group locks?</p> <p>Alex - The presentation PDF explains why not in a bit more detail, giving a few reasons.</p> <p>The main reason is what I'm calling write visibility:</p> <p>Client A and B both have the same group lock.<br/> Client A reads some pages, then client B writes to those pages.<br/> Client A reads the pages again, but because it has a lock on the file, it does not realize the pages have been modified &amp; will re-use its cached pages.</p> <p>Some of the more complex data formats used in HPC - Such as HFD5 - write out in file metadata as part of writing out a file, and it requires write visibility in the course of the IO.</p> <p>Jinshan suggested implementing an ioctl for user space to request LDLM locks, allowing user space to 'lock ahead' of the actual IO, and we agreed I'd implement that and test it for performance before considering strided locks further. If that can get most of the performance of file per process, then there isn't much point to strided locks, and we do not need to add the complexity. We agreed there that if that doesn't close the gap, strided locks are worth considering.</p> <p>A prototype of lock ahead is here:<br/> <a href="https://jira.whamcloud.com/browse/LU-6179" title="Lock ahead - Request extent locks from userspace" class="issue-link" data-issue-key="LU-6179"><del>LU-6179</del></a></p> <p>This ticket's going to be on hold until we have more results there.</p> <p>This approach was abandoned in favour of lockahead.</p> Related LU-6179 LU-6180 Development Rank 1|hzx4mf: Rank (Obsolete) 17170 Severity