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/32ad6edf-a1b8-46ac-91ff-92eecbe0efe3
test_205a failed with the following error:
sanity test_205a: @@@@@@ FAIL: No jobstats for id.205a.dd.320 found on ost1::*.lustre-OST0000.job_stats
However, the dumped jobstats show that there is a result for id.205a.dd.320, but also a result from the previous "dd" write operation that has an ID that has a matching (but longer) string id.205a.dd.32075, so the "grep -c" in the test matches two jobids and fails the test:
- job_id: id.205a.dd.32075 snapshot_time: 1615208459 write_bytes: { samples: 1, unit: bytes, min: 1048576, max: 1048576, sum: 1048576, sumsq: 1099511627776 } write: { samples: 1, unit: usecs, min: 141, max: 141, sum: 141, sumsq: 19881 } punch: { samples: 1, unit: usecs, min: 31, max: 31, sum: 31, sumsq: 961 } sync: { samples: 1, unit: usecs, min: 42111, max: 42111, sum: 42111, sumsq: 1773336321 } - job_id: id.205a.dd.320 snapshot_time: 1615208461 read_bytes: { samples: 1, unit: bytes, min: 1048576, max: 1048576, sum: 1048576, sumsq: 1099511627776 } read: { samples: 1, unit: usecs, min: 53, max: 53, sum: 53, sumsq: 2809 }
I don't know the exact statistics of this happening, maybe around (4/32768 = 1/8192) since any 5-digit random number for the first "dd" has a 4-, 3-, 2-, and 1-digit substring that would also match. In any case, full-string matching should fix this problem.
VVVVVVV DO NOT REMOVE LINES BELOW, Added by Maloo for auto-association VVVVVVV
sanity test_205a - No jobstats for id.205a.dd.4 found on ost1::*.lustre-OST0000.job_stats