[LU-14312] Interop: sanity test 272b fails with 'failed to migrate to the new composite layout' Created: 08/Jan/21 Updated: 23/Jan/21 Resolved: 23/Jan/21 |
|
| Status: | Resolved |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | Lustre 2.14.0 |
| Fix Version/s: | Lustre 2.14.0 |
| Type: | Bug | Priority: | Minor |
| Reporter: | James Nunez (Inactive) | Assignee: | Mikhail Pershin |
| Resolution: | Fixed | Votes: | 0 |
| Labels: | interop | ||
| Severity: | 3 |
| Rank (Obsolete): | 9223372036854775807 |
| Description |
|
sanity test_272b fails for interop testing between a master, future 2.14.0, client and older (< 2.13.56 ) servers starting on 01 NOV 2020 for Lustre 2.13.56.63. Looking at a recent failure at https://testing.whamcloud.com/test_sets/6e4f8226-6269-4979-a03f-af7e98a714be, we see from the suite_log that there is an issue with getting a lock for a DoM file == sanity test 272b: DoM migration: DOM file to the OST-striped file (plain) ========================= 03:42:54 (1608781374) CMD: trevis-19vm4 lctl get_param -n osd*.*MDT0000.kbytesfree 1+0 records in 1+0 records out 2097152 bytes (2.1 MB, 2.0 MiB) copied, 0.0255055 s, 82.2 MB/s CMD: trevis-19vm4 lctl get_param -n osd*.*MDT0000.kbytesfree error: lfs migrate: /mnt/lustre/d272b.sanity/dom: data copy failed: No locks available sanity test_272b: @@@@@@ FAIL: failed to migrate to the new composite layout Trace dump: = /usr/lib64/lustre/tests/test-framework.sh:6273:error() = /usr/lib64/lustre/tests/sanity.sh:20792:test_272b() Looking at patches that landed during this time related to DoM and locks, we see two related patches: Logs for more failures are at |
| Comments |
| Comment by Peter Jones [ 09/Jan/21 ] |
|
Mike Thoughts on this one? Peter |
| Comment by Mikhail Pershin [ 11/Jan/21 ] |
|
This is compatibility issue with old servers, I will check that locally to get more details |
| Comment by Gerrit Updater [ 19/Jan/21 ] |
|
Mike Pershin (mpershin@whamcloud.com) uploaded a new patch: https://review.whamcloud.com/41268 |
| Comment by Mikhail Pershin [ 19/Jan/21 ] |
|
James, it that possible to check the patch somehow, if we have a configuration when problem occurs always or very often? I wasn't able to reproduce problem locally so need to confirm the patch solves that |
| Comment by James Nunez (Inactive) [ 19/Jan/21 ] |
|
Mike, It looks like this test fails 100% of the time for master clients and 2.12.5/6 and 2.13.0 severs. Let me add a test parameters line to the patch to mimic this set up. |
| Comment by Mikhail Pershin [ 20/Jan/21 ] |
|
James, from test results it looks like patch does the job, there are still failed interop tests in sanity which are not related to patch as first sign, could you check and confirm they are not something new but expected? |
| Comment by Gerrit Updater [ 23/Jan/21 ] |
|
Oleg Drokin (green@whamcloud.com) merged in patch https://review.whamcloud.com/41268/ |
| Comment by Peter Jones [ 23/Jan/21 ] |
|
Landed for 2.14 |