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-13086 Lustre <p>commit 3c7aca747 <a href="https://jira.whamcloud.com/browse/LU-12395" title="Failed dependencies while installing lustre-client-tests and lustre-iokit rpms on RHEL 8.0 client" class="issue-link" data-issue-key="LU-12395"><del>LU-12395</del></a> adds "--oversubscribe" MPI run option which causes the MPI tests built with mpich fail:</p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java"> [mpiexec@fre1307] match_arg (./utils/args/args.c:160): unrecognized argument oversubscribe [mpiexec@fre1307] HYDU_parse_array (./utils/args/args.c:175): argument matching returned error </pre> </div></div> LU-13086

3c7aca747 LU-12395 breaks compatibility mpi tests with mpich

Bug Minor Resolved Fixed Elena Gryaznova Elena Gryaznova Wed, 18 Dec 2019 14:15:02 +0000 Fri, 17 Sep 2021 14:40:04 +0000 Fri, 17 Sep 2021 14:40:04 +0000 Lustre 2.15.0 0 7 <p>Andreas,<br/> Jian,<br/> we need your advice how to fix this. I see only the way to remove this option or make it optional.</p> <p>Please advice.<br/> Thanks.</p> <p>Added Minh, since he was the author for that patch.</p> <p>Elena, I don't mind to make this optional. It seems we could also specify it as part of <tt>$MPIRUN_OPITIONS</tt> from the environment, but maybe this has some side-effect that I'm not aware of?</p> <p>Elena Gryaznova (c17455@cray.com) uploaded a new patch: <a href="https://review.whamcloud.com/38130" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/38130</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-13086" title="3c7aca747 LU-12395 breaks compatibility mpi tests with mpich" class="issue-link" data-issue-key="LU-13086"><del>LU-13086</del></a> tests: restore compatibility with mpich<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: fc31416862a5c6a066ee62429e896ffa1062c557</p> <p><a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=mdiep" class="user-hover" rel="mdiep">mdiep</a>, we could use your assistance with some questions in the review of <a href="https://review.whamcloud.com/#/c/38130/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/#/c/38130/</a>. Thanks!</p> <p><a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=mdiep" class="user-hover" rel="mdiep">mdiep</a> and <a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=jamesanunez" class="user-hover" rel="jamesanunez">jamesanunez</a> can you get together and reconcile your opinions about --oversubscribe and where you set it? James, it sounds like you can't get --oversubscribe in your env with the patch as-is. Is that right? If so, will you and Minh both be happy if it is moved to cfg/local.sh? Please update and clarify your comments in the Gerrit review. Thanks!</p> <p>Elena Gryaznova (c17455@cray.com) uploaded a new patch: <a href="https://review.whamcloud.com/38689" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/38689</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-13086" title="3c7aca747 LU-12395 breaks compatibility mpi tests with mpich" class="issue-link" data-issue-key="LU-13086"><del>LU-13086</del></a> tests: restore compatibility with mpich<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: 0ad3e4795d6490a630432950184b42f8a83811e3</p> <p><a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=mdiep" class="user-hover" rel="mdiep">mdiep</a> and <a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=jamesanunez" class="user-hover" rel="jamesanunez">jamesanunez</a>, now we have options between <a href="https://review.whamcloud.com/#/c/38130/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/#/c/38130/</a> and <a href="https://review.whamcloud.com/#/c/38689/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/#/c/38689/</a>, but reviews have gone stale. Can you please express what approach you want to proceed with and why?</p> <p><a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=mdiep" class="user-hover" rel="mdiep">mdiep</a>, <a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=colmstea" class="user-hover" rel="colmstea">colmstea</a>, and <a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=jamesanunez" class="user-hover" rel="jamesanunez">jamesanunez</a>, with the activity at <a href="https://review.whamcloud.com/#/c/38689/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/#/c/38689/</a> are we to assume that it is the preferred direction? And should <a href="https://review.whamcloud.com/#/c/38130/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/#/c/38130/</a> be abandoned?</p> <p><a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=mdiep" class="user-hover" rel="mdiep">mdiep</a>, <a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=colmstea" class="user-hover" rel="colmstea">colmstea</a>, and <a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=jamesanunez" class="user-hover" rel="jamesanunez">jamesanunez</a>, so <a href="https://review.whamcloud.com/#/c/38130/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/#/c/38130/</a> should be abandoned? And in <a href="https://review.whamcloud.com/#/c/38689/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/#/c/38689/</a> Elena has proposed to "totally get rid of --oversubscribe". Can you agree?</p> <p><a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=spitzcor" class="user-hover" rel="spitzcor">spitzcor</a>, I've abandoned 38130 and updated 38689 to address the minor defect therein. It needs a second review and testing to finish before it can land. It looks like a reasonable compromise to include <tt>--oversubscribe</tt> so that the testing works out-of-the-box for RHEL (which is far-and-away the most common distro used with Lustre), but still allow the external config file to specify different options based on the MPI version.</p> <p>Andreas Dilger (adilger@whamcloud.com) uploaded a new patch: <a href="https://review.whamcloud.com/41082" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/41082</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-13086" title="3c7aca747 LU-12395 breaks compatibility mpi tests with mpich" class="issue-link" data-issue-key="LU-13086"><del>LU-13086</del></a> config: add --oversubscribe to MPIRUN_OPTIONS<br/> Project: private/autotest<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: cca5447c7ee27ec2c7bd916f6a634ef8371b5cbd</p> <p>I don't have permissions to view <a href="https://review.whamcloud.com/#/c/41082/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/#/c/41082/</a>.</p> <p>Is <a href="https://review.whamcloud.com/#/c/41082/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/#/c/41082/</a> supposed to be a replacement for <a href="https://review.whamcloud.com/#/c/38689/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/#/c/38689/</a> ?</p> <p>"Charlie Olmstead &lt;charlie@whamcloud.com&gt;" merged in patch <a href="https://review.whamcloud.com/41082/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/41082/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-13086" title="3c7aca747 LU-12395 breaks compatibility mpi tests with mpich" class="issue-link" data-issue-key="LU-13086"><del>LU-13086</del></a> config: add --oversubscribe to MPIRUN_OPTIONS<br/> Project: private/autotest<br/> Branch: master<br/> Current Patch Set: <br/> Commit: c01e4db25d8a47a9c27cada04d8e3b4ea83292d4</p> <p>AT Patch has been deployed</p> <p>"Oleg Drokin &lt;green@whamcloud.com&gt;" merged in patch <a href="https://review.whamcloud.com/38689/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/38689/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-13086" title="3c7aca747 LU-12395 breaks compatibility mpi tests with mpich" class="issue-link" data-issue-key="LU-13086"><del>LU-13086</del></a> tests: restore compatibility with mpich<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: <br/> Commit: e418f47688facf07f2e9bd6535b71d484af4f8ac</p> <p>Landed for 2.15</p> Duplicate LU-13348 Related LU-12395 Development Rank 1|i00r5z: Rank (Obsolete) 9223372036854775807 Severity