Details
-
Bug
-
Resolution: Fixed
-
Minor
-
None
-
None
-
3
-
9223372036854775807
Description
This issue was created by maloo for Andreas Dilger <adilger@whamcloud.com>
This issue relates to the following test suite run:
https://testing.whamcloud.com/test_sets/495acc8b-ba64-4b86-a856-a873da5ec286
https://testing.whamcloud.com/test_sets/452d1b80-f656-4a80-b893-d3c2382ccbad
test_metabench failed with the following error:
[08/24/2023 17:15:30] Leaving time_file_creation with proc_id = 7 totfile tot time create rate interval intv time intv rate ======= ========== ============ ======== ========== ============ 1000 32.3700 30.89 1000 32.3700 30.89 2000 63.7400 31.38 1000 31.3700 31.88 3000 100.4090 29.88 1000 36.6690 27.27 4000 136.2740 29.35 1000 35.8650 27.88 5000 170.7300 29.29 1000 34.4560 29.02 6000 204.9140 29.28 1000 34.1840 29.25 7000 244.4280 28.64 1000 39.5140 25.31 8000 280.8720 28.48 1000 36.4440 27.44 9000 318.2790 28.28 1000 37.4070 26.73 10000 355.4810 28.13 1000 37.2020 26.88 [08/24/2023 17:15:30] Entering par_create_multidir to create 1429 files in 1 dirs [08/24/2023 17:15:43] FATAL error on process 5 Proc 5: Could not create file number [788] named [VBqEA2OF1cLsZN]: No space left on device
Test session details:
clients: https://build.whamcloud.com/job/lustre-master/4455 - 4.18.0-477.15.1.el8_8.x86_64
servers: https://build.whamcloud.com/job/lustre-master/4455 - 4.18.0-477.15.1.el8_lustre.x86_64
It looks like this has been failing for a long time (at least since 2023-04-01 when I stopped looking) but no ticket has been filed for it.
This may be fallout from a previous test not cleaning up the files, so it would be useful to run "lfs df" and "lfs df -i" to this subtest if it fails to see if the filesystem is actually full, or if there is some other issue.
VVVVVVV DO NOT REMOVE LINES BELOW, Added by Maloo for auto-association VVVVVVV
parallel-scale-nfsv4 test_metabench - metabench failed! 1