[LU-12868] sanity-quota test 65 100% failure Created: 16/Oct/19 Updated: 18/Oct/19 |
|
| Status: | Open |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | Lustre 2.13.0 |
| Fix Version/s: | None |
| Type: | Bug | Priority: | Minor |
| Reporter: | Oleg Drokin | Assignee: | WC Triage |
| Resolution: | Unresolved | Votes: | 0 |
| Labels: | None | ||
| Issue Links: |
|
||||||||
| Severity: | 3 | ||||||||
| Rank (Obsolete): | 9223372036854775807 | ||||||||
| Description |
|
In my testing sanity-quota test 65 fails 100% of the time lately. == sanity-quota test 65: Check lfs quota result ====================================================== 22:37:17 (1571193437) Waiting for local destroys to complete Creating test directory fail_val=0 fail_loc=0 debug=+quota debug=+quota Write... running as uid/gid/euid/egid 60000/60000/60000/60000, groups: [dd] [if=/dev/zero] [bs=1M] [of=/mnt/lustre/d65.sanity-quota/f65.sanity-quota-0] [count=10] dd: error writing '/mnt/lustre/d65.sanity-quota/f65.sanity-quota-0': Disk quota exceeded 8+0 records in 7+0 records out 7340032 bytes (7.3 MB) copied, 0.0647576 s, 113 MB/s sanity-quota test_65: @@@@@@ FAIL: failed to write I suspect somewhere we forget to clear quota? I run some queries and this test actually never passed in my testing (And I apparently missed that even though there was a clear warning from the testing env). |
| Comments |
| Comment by Gerrit Updater [ 16/Oct/19 ] |
|
Oleg Drokin (green@whamcloud.com) uploaded a new patch: https://review.whamcloud.com/36462 |
| Comment by Oleg Drokin [ 16/Oct/19 ] |
|
hm, ok, so the test patch shows the quota is cleared which makes the whole failure somewhat mysterious. Hongchao, can you please take a look? http://testing.linuxhacker.ru:3333/lustre-reports/3721/testresults/sanity-quota-ldiskfs-DNE-centos7_x86_64-centos7_x86_64/ is probably the good one to look at |
| Comment by Hongchao Zhang [ 17/Oct/19 ] |
|
sorry, I can't open the URL, it shows "Unable to round-trip http request to upstream: dial tcp 73.108.203.87:3333: i/o timeout". The quota will be cleared at the end of each test by "cleanup_quota_test", which means there will be no quota limits on IDs. btw, I searched on Maloo and found there is no such error in it. |
| Comment by Oleg Drokin [ 18/Oct/19 ] |
|
ah, yes, I heard that my node is somehow not accessible from China. I wonder how to best get you debug logs from a test run? the issue only happens on my test systems but not on maloo. possibly because I have exclusion list, I always exclude these santy-quota tests: 2, 4a and 63 |