Something similar has been reported on RHEL7 too today: restraint-rhts-0.1.33-1.el7bkr.x86_64 restraint-0.1.33-1.el7bkr.x86_64 Submitted data is: path=%2Fdesktop%2Fci%2Fprepare%2Fnm_debug&result=PASS&score= Note that 'score' is empty string, presumably because of wrapper script, that quotes last parameter, which is optional: rstrnt-report-result --outputfile="$3" "$1" "$2" "$4" On server side it hits: beaker-server beaker-server[72030]: OperationalError: (OperationalError) (1366, "Incorrect decimal value: '' for column 'score' at row 1") 'INSERT INTO recipe_task_result (recipe_task_id, path, result, score, log, start_time) VALUES (%s, %s, %s, %s, %s, %s)' (73233377L, '/desktop/ci/prepare/nm_debug', 'Pass', '', None, datetime.datetime(2018, 6, 5, 13, 6, 56, 446427)) This allowed R:5237866 to work again: -rstrnt-report-result --outputfile="$3" "$1" "$2" "$4" +rstrnt-report-result --outputfile="$3" "$1" "$2" $4
comment 0 appears to be related to beakerlib BZ: Bug 1571784 - rstrnt-report-result is incompatible with beakerlib but set as command for reporting result task from comment 1 is not using beakerlib
The root cause would be the cutover to Mariadb in our production Beaker instance yesterday. Mariadb runs with "strict mode" applied by default. We thought we had found all instances where we were violating the strict mode but this is obviously one we did not catch. We definitely need to fix Beaker to accept non-int values for score, for compatibility with the previous behaviour (I assume non-strict MySQL used to map these values to 0 -- need to verify this). The issues with restraint+beakerlib putting its arguments in the wrong order is indeed a separate issue, as noted by Jan: bug 1571784.
*** Bug 1587970 has been marked as a duplicate of this bug. ***
As noted on beaker-user-list, if you are affected by this bug because of restraint sending the wrong parameter as score, you can work around it by using a fixed version of restraint: <repos> <repo name="myharness" url="http://brew-task-repos.usersys.redhat.com/repos/scratch/bpeck/restraint/0.1.34/1.git.12.c32c83a.el7/x86_64/"/> </repos> Thanks to bpeck for building a patched version.
*** Bug 1588737 has been marked as a duplicate of this bug. ***
This was deployed to production this morning and should get all tests who have seen this issue going again.
This was a Beaker bug, fixed in 25.5, not a bug in qa-tools-workstation (although a workaround may have been included in it).