Details
Description
According to cross-reference report, it happens rarely, and almost only on bld-dan-* and p8 builders (there was one occasion on win64-release2, but it was long time ago).
main.log_tables w1 [ fail ]
Test ended at 2015-01-06 10:29:02
CURRENT_TEST: main.log_tables
mysqltest: At line 832: query 'call `db_17876.archiveSlowLog`()' failed: 1136: Column count doesn't match value count at row 1
The result from queries just before the failure was:
< snip >
INSERT INTO
`db_17876.general_log_data`
VALUES(event_time, user_host, thread_id, server_id,
command_type, argument);
END;
END IF;
END;
UNTIL done END REPEAT;
CLOSE cur1;
TRUNCATE mysql.general_log;
END //
SET GLOBAL general_log = ON;
SET GLOBAL slow_query_log = ON;
select "put something into general_log";
put something into general_log
put something into general_log
select "... and something more ...";
... and something more ...
... and something more ...
call `db_17876.archiveSlowLog`();
Gliffy Diagrams
Attachments
Issue Links
- relates to
-
MDEV-7069 Fix buildbot failures in main server trees
-
- Stalled
-
Activity
- All
- Comments
- Work Log
- History
- Activity
- Transitions
http://bazaar.launchpad.net/~maria-captains/maria/10.0/revision/4555
Committer: Elena Stepanova Date: 2015-01-15 16:15:50 UTC Revision ID: elenst@montyprogram.com-20150115161550-em88687tm2fs02g0 MDEV-7431 main.log_tables fails sporadically in buildbot Structure of the table created by the test to archive mysql.slow_log data didn't match the structure of mysql.slow_log. The failure only appeared if the slow_log was not empty, which was very rare. Updated the structure of the table.
It's easily reproducible if long-query-time is set to a low value:
$ perl ./mtr main.log_tables --mysqld=--long-query-time=0.001 ... worker[1] Using MTR_BUILD_THREAD 300, with reserved ports 16000..16019 main.log_tables [ fail ] Test ended at 2015-01-15 19:50:04 CURRENT_TEST: main.log_tables mysqltest: At line 832: query 'call `db_17876.archiveSlowLog`()' failed: 1136: Column count doesn't match value count at row 1 The result from queries just before the failure was: