[LU-13194] Interop: sanityn test 104 fails with '/mnt/lustre/f104.sanityn mtime (1580286855:1580286853) diff' Created: 03/Feb/20 Updated: 08/Feb/20 Resolved: 07/Feb/20 |
|
| Status: | Resolved |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | Lustre 2.10.8, Lustre 2.12.3 |
| Fix Version/s: | Lustre 2.14.0, Lustre 2.12.4 |
| Type: | Bug | Priority: | Minor |
| Reporter: | James Nunez (Inactive) | Assignee: | James Nunez (Inactive) |
| Resolution: | Fixed | Votes: | 0 |
| Labels: | interop | ||
| Severity: | 3 |
| Rank (Obsolete): | 9223372036854775807 |
| Description |
|
sanityn test_104 fails for clients with Lustre versions that are newer than the client version starting with Lustre 2.12.3 servers and 2.12.58.171 clients starting on 24 OCT 2019; see https://testing.whamcloud.com/test_sets/591b7932-f793-11e9-a197-52540065bddc and https://testing.whamcloud.com/test_sets/d58f7e54-f767-11e9-add9-52540065bddc. Looking at a recent failure, https://testing.whamcloud.com/test_sets/8d9bdf58-4333-11ea-86c7-52540065bddc, for Lustre servers 2.11.0 and clients 2.12.3.109, we see the mtimes produced CMD: trevis-51vm12 debugfs -c -R 'stat ROOT/f104.sanityn' /dev/mapper/mds1_flakey trevis-51vm12: debugfs 1.45.2.wc1 (27-May-2019) trevis-51vm12: /dev/mapper/mds1_flakey: catastrophic mode - not reading inode or group bitmaps STAT a|m|ctime 1580286853 1580286853 1580286853 MDT a|m|ctime 1580286853 1580286853 1580286853 1+0 records in 1+0 records out 1024 bytes (1.0 kB) copied, 0.00156288 s, 655 kB/s CMD: trevis-51vm12 debugfs -c -R 'stat ROOT/f104.sanityn' /dev/mapper/mds1_flakey trevis-51vm12: debugfs 1.45.2.wc1 (27-May-2019) trevis-51vm12: /dev/mapper/mds1_flakey: catastrophic mode - not reading inode or group bitmaps STAT a|m|ctime 1580286853 1580286855 1580286855 MDT a|m|ctime 1580286853 1580286855 1580286855 sanityn test_104: @@@@@@ FAIL: /mnt/lustre/f104.sanityn mtime (1580286855:1580286853) diff The patch for sanityn test 104 does not have a Lustre version check, but needs one. Master, future 2.14.0, interop testing runs with 2.12.3 and 2.13.0. 2.13.0 already has this patch and 2.12.4 has this patch. So, when we move interop testing for master to use 2.12.4, we don’t have to worry about master interop testing hitting this issue. The only other release we need to worry about is the LTS branch b2_12. The 2.12 branch runs interop testing with 2.11.0 and 2.10.8. Thus, if we land a patch to b2_12 that checks the server version to make sure it is 2.12.3.110 or later, then all b2_12 interop testing should skip this test when appropriate. |
| Comments |
| Comment by Gerrit Updater [ 04/Feb/20 ] |
|
James Nunez (jnunez@whamcloud.com) uploaded a new patch: https://review.whamcloud.com/37420 |
| Comment by Andreas Dilger [ 04/Feb/20 ] |
|
What is a bit confusing is that it looks like the test should be passing: STAT a|m|ctime 1580286853 1580286855 1580286855 MDT a|m|ctime 1580286853 1580286855 1580286855 sanityn test_104: @@@@@@ FAIL: /mnt/lustre/f104.sanityn mtime (1580286855:1580286853) diff The respective atimes, mtimes, and ctimes are the same between the client and MDT, even if they are not the same as each other. Either the STAT or MDT lines are being printed incorrectly, or the test is checking the wrong values? |
| Comment by Gerrit Updater [ 06/Feb/20 ] |
|
James Nunez (jnunez@whamcloud.com) uploaded a new patch: https://review.whamcloud.com/37461 |
| Comment by Gerrit Updater [ 06/Feb/20 ] |
|
Oleg Drokin (green@whamcloud.com) merged in patch https://review.whamcloud.com/37461/ |
| Comment by Peter Jones [ 07/Feb/20 ] |
|
Landed for 2.14 |
| Comment by Gerrit Updater [ 08/Feb/20 ] |
|
Oleg Drokin (green@whamcloud.com) merged in patch https://review.whamcloud.com/37420/ |