[LU-14998] conf-sanity test 35a fails with 'cleanup failed with 200'

Wed, 27 Oct 2021 16:43:27 +0000

conf-sanity test_35a started failing with the error 'cleanup failed with 200' on 06 MARCH 2021 for SLES15 SP2 only starting with Lustre 2.14.50.130.

Looking at a recent failure at https://testing.whamcloud.com/test_sets/e2e430d0-ba19-4c9a-86b9-9d9319c02b82, the last lines in the suite_log are

…
Started lustre-MDT0000
Wait for df (15649) ... 
done
debug=trace inode super tty malloc cache info ioctl neterror net warning buffs other dentry nettrace page dlmtrace error emerg ha rpctrace vfstrace reada mmap config console quota sec lfsck hsm snapshot layout
Debug log: 122 lines, 122 kept, 0 dropped, 0 bad.
umount lustre on /mnt/lustre.....
CMD: trevis-71vm6 grep -c /mnt/lustre' ' /proc/mounts
Stopping client trevis-71vm6 /mnt/lustre (opts:)
CMD: trevis-71vm6 lsof -t /mnt/lustre
CMD: trevis-71vm6 umount  /mnt/lustre 2>&1
/mnt/lustre is still busy, wait one second
CMD: trevis-71vm6 umount /mnt/lustre
umount: /mnt/lustre: target is busy.
 conf-sanity test_35a: @@@@@@ FAIL: cleanup failed with 200 
  Trace dump:
  = /usr/lib64/lustre/tests/test-framework.sh:6237:error()
  = /usr/lib64/lustre/tests/conf-sanity.sh:2711:test_35a()

The return code of 200 comes from the cleanup() routine and is an umount_client() failure

198 umount_client() {
 199         local mountpath=$1
 200         shift
 201         echo "umount lustre on $mountpath....."
 202         zconf_umount $HOSTNAME $mountpath $@ || return 97
 203 }
 204 
…
239 cleanup() {
 240         local force=""
 241         [ "x$1" != "x" ] && force='-f'
 242         umount_client $MOUNT $force|| return 200
 243         cleanup_nocli || return $?
 244 }

Logs for more of these failures are at
https://testing.whamcloud.com/test_sets/0a90aae9-c320-4c37-aae5-819ef5bc9ad1
https://testing.whamcloud.com/test_sets/526571d1-5455-4828-a9be-17792d5559ac
https://testing.whamcloud.com/test_sets/ab789fc0-f19a-4e8c-911b-50ef1d27f942

Whamcloud Community JIRA

[LU-14998] conf-sanity test 35a fails with 'cleanup failed with 200'