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-1601 Lustre <p>Deadlock detection algorithm distinguish processes by pair pid+export. It is not correct when one process access files from different mount points.<br/> Proper entity should be pid+NID</p> <p>testcase patch attached:</p> <p> /local/work/lustre/lustre/tests/flocks_test 4 /mnt/lustre/file1 /mnt/lustre2/file2<br/> 6646: get lock2<br/> 6647: get lock1<br/> 6646: done<br/> 6647: done<br/> 6647: get lock2<br/> 6646: get lock1</p> <p>it waits forever, while works with files from one mount point</p> <p> /local/work/lustre/lustre/tests/flocks_test 4 /mnt/lustre/file1 /mnt/lustre/file2<br/> 6639: get lock2<br/> 6640: get lock1<br/> 6639: done<br/> 6640: done<br/> 6639: get lock1<br/> 6640: get lock2<br/> fcntl cmd 7 failed: Resource deadlock avoided<br/> 6640: done<br/> 6640: exit rc=0<br/> 6639: done<br/> 6639: exit rc=0</p> LU-1601

Flock deadlock isn't detected while accessing files from 2 mount points

Bug Minor Resolved Fixed Keith Mannthey Andriy Skulysh patch Thu, 5 Jul 2012 05:10:39 +0000 Fri, 4 Oct 2013 16:15:51 +0000 Thu, 19 Sep 2013 06:35:26 +0000 Lustre 2.5.0 0 3 <p>CODE: <a href="http://review.whamcloud.com/3276" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/3276</a></p> <p>Xyratex-bug-id: MRP-449</p> <p>Landed for 2.5.0</p> Related LU-4060 Development Rank 1|hzvuhb: Rank (Obsolete) 8917 Severity