Bug #69265 | -DBUILD_CONFIG=mysql_release -DWITH_DEBUG=ON fails 4 and skips 27 MTR tests | ||
---|---|---|---|
Submitted: | 17 May 2013 12:43 | Modified: | 19 Aug 2013 17:24 |
Reporter: | Laurynas Biveinis (OCA) | Email Updates: | |
Status: | Closed | Impact on me: | |
Category: | Tools: MTR / mysql-test-run | Severity: | S7 (Test Cases) |
Version: | 5.6.11 | OS: | Any |
Assigned to: | CPU Architecture: | Any | |
Tags: | mtr cmake |
[17 May 2013 12:43]
Laurynas Biveinis
[20 May 2013 8:14]
Laurynas Biveinis
I forgot to consider the option of running the testsuite on an installed server. The MTR code appears to work in that case and not for running the testsuite in the source tree. The proposed fix would reverse the situation and thus it's a wrong fix.
[23 May 2013 6:25]
Laurynas Biveinis
A better fix is (ignore the "Percona-Server" in the path). Installed server testsuite runs keep their behavior (didn't actually test), in-source testsuite runs are fixed. === modified file 'Percona-Server/mysql-test/mysql-test-run.pl' --- Percona-Server/mysql-test/mysql-test-run.pl 2013-05-12 09:13:00 +0000 +++ Percona-Server/mysql-test/mysql-test-run.pl 2013-05-23 06:22:07 +0000 @@ -2226,7 +2226,7 @@ mtr_error("Lines in $defs_file must have 3 or 4 items") unless $plug_var; # If running debug server, plugins will be in 'debug' subdirectory - $plug_file= "debug/$plug_file" if $running_debug; + $plug_file= "debug/$plug_file" if $running_debug && !$source_dist; my ($plugin)= find_plugin($plug_file, $plug_loc);
[6 Jun 2013 5:36]
Prabeen Kumar Pradhan
Thank you for the bug report. Verified as described
[18 Jun 2013 14:40]
Paul DuBois
Changes for test suite. No changelog entry needed.
[1 Aug 2013 15:56]
Laurynas Biveinis
It seems that the proposed fix was incorrect: laurynas@laurynas-ThinkPad-T410:~/percona/src/5.5$ bzr log -r 4378 ------------------------------------------------------------ revno: 4378 committer: sayantan dutta <sayantan.dutta@oracle.com> branch nick: mysql-5.5 timestamp: Fri 2013-06-07 19:29:56 +0530 message: Bug #16917425 -DBUILD_CONFIG=MYSQL_RELEASE -DWITH_DEBUG=ON FAILS 4 AND SKIPS 27 MTR TESTS laurynas@laurynas-ThinkPad-T410:~/percona/src/5.5$ bzr log -r 4379 ------------------------------------------------------------ revno: 4379 committer: Olav Sandstaa <olav.sandstaa@oracle.com> branch nick: 5.5-fix timestamp: Fri 2013-06-07 21:34:34 +0200 message: Bug #16917425 -DBUILD_CONFIG=MYSQL_RELEASE -DWITH_DEBUG=ON FAILS 4 AND SKIPS 27 MTR TESTS Revert fix since it caused mtr tests to not run in pushbuild.
[2 Aug 2013 10:18]
Prabeen Kumar Pradhan
This issue was fixed on 18-Jun-13. Kindly check and confirm if you still see this issue.
[2 Aug 2013 10:22]
Laurynas Biveinis
Prabeen - Could you please point me at the bzr revision? (or even the tree if it was not fixed on 5.5 perhaps). I see only the following revisions for Jun 18th. It's obviously not 4393, but maybe one of two mysterious other revisions? ------------------------------------------------------------ revno: 4393 committer: Vasil Dimov <vasil.dimov@oracle.com> branch nick: mysql-5.5 timestamp: Tue 2013-06-18 17:12:28 +0300 message: Fix Bug#16907783 5.5 STILL CRASHES IN DICT_UPDATE_STATISTICS WITH CONCURRENT DDL AND I_S QUERIES Skip partially created indexes (ones whose name starts with TEMP_INDEX_PREFIX) from stats gathering. Because InnoDB reports HA_INPLACE_ADD_INDEX_NO_WRITE to MySQL, the latter allows parallel execution of ha_innobase::add_index() and ha_innobase::info(). Reviewed by: Inaam (rb:2613) ------------------------------------------------------------ revno: 4392 [merge] committer: mysql-builder@oracle.com branch nick: clone timestamp: Tue 2013-06-18 15:49:13 +0530 message: ------------------------------------------------------------ revno: 4391 committer: mysql-builder@oracle.com branch nick: clone timestamp: Tue 2013-06-18 10:20:30 +0530 message: ------------------------------------------------------------
[2 Aug 2013 11:24]
Prabeen Kumar Pradhan
Below is the revision detail in which it was fixed: http://bazaar.launchpad.net/~mysql/mysql-server/5.6/revision/5217
[5 Aug 2013 3:39]
Laurynas Biveinis
The bug is _not_ fixed: the referenced fix merely replaces one failure mode with another, causing the tests to be skipped instead of failed, because the root cause of plugins being searched for in the wrong dirs is not addressed. Thus now instead of 4 failing and 27 skipped tests we have 31 skipped test instead: $ ./mysql-test-run --mem --debug audit_plugin ... main.audit_plugin [ skipped ] Audit_null plugin requires the environment variable \$AUDIT_NULL to be set (normally done by mtr) ... etc.
[6 Aug 2013 12:11]
Paul DuBois
Changes for test suite. No changelog entry needed.
[6 Aug 2013 13:05]
Prabeen Kumar Pradhan
We are currently working on this. Will update progress.
[8 Aug 2013 5:25]
Anitha Gopi
Laurynas, Please upload your suggested patch under contributions tag.
[8 Aug 2013 12:36]
Laurynas Biveinis
The bug tracker says "Contributions can be accepted to Open bugs only."
[8 Aug 2013 12:52]
Laurynas Biveinis
bug 69963.
[12 Aug 2013 9:27]
Laurynas Biveinis
Bug 69265 patch (*) I confirm the code being submitted is offered under the terms of the OCA, and that I am authorized to contribute it.
Contribution: bug69265.patch (text/x-patch), 572 bytes.
[12 Aug 2013 9:27]
Anitha Gopi
I have set status to Open. Please try now.
[12 Aug 2013 9:29]
Anitha Gopi
Got it now. Thanks.
[19 Aug 2013 17:24]
Paul DuBois
Changes for test suite. No changelog entry needed. Thanks for Laurynas Biveinis for the patch to fix this issue.