Bug 8984 - mysql-test-run --force - Too many tests(10) failed! Terminating. Only 221 of 2466 completed
Summary: mysql-test-run --force - Too many tests(10) failed! Terminating. Only 221 o...
Status: RESOLVED DUPLICATE of bug 8921
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: 2
Hardware: All Linux
Priority: Normal normal
Target Milestone: ---
Assignee: AL13N
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks: 8921
  Show dependency treegraph
 
Reported: 2013-02-07 12:54 CET by claire robinson
Modified: 2013-02-21 07:22 CET (History)
2 users (show)

See Also:
Source RPM: mariadb
CVE:
Status comment:


Attachments
Console output from test (5.09 KB, text/plain)
2013-02-07 23:05 CET, Dave Hodgins
Details
/usr/share/mysql/mysql-test/var/log/mysqld.1.err (2.47 KB, text/plain)
2013-02-07 23:11 CET, Dave Hodgins
Details

Description claire robinson 2013-02-07 12:54:31 CET
See bug 8921 comment 17

Many tests fail from the mariadb testsuite
claire robinson 2013-02-07 12:54:52 CET

Hardware: i586 => All
Version: Cauldron => 2

Comment 1 AL13N 2013-02-07 19:30:09 CET
please give detailed output of the first failed test:

[ ]$ ( cd /usr/share/mysql/mysql-test/ && ./mysql-test-run.pl )

then redo the one that fails:


[ ]$ ( cd /usr/share/mysql/mysql-test/ && ./mysql-test-run.pl <failing_test> )

this should create a log file, please attach the logfile
claire robinson 2013-02-07 19:40:14 CET

CC: (none) => davidwhodgins

Comment 2 Dave Hodgins 2013-02-07 23:05:07 CET
Created attachment 3496 [details]
Console output from test
Comment 3 Dave Hodgins 2013-02-07 23:11:09 CET
Created attachment 3497 [details]
/usr/share/mysql/mysql-test/var/log/mysqld.1.err
AL13N 2013-02-07 23:18:23 CET

CC: (none) => alien

Comment 4 AL13N 2013-02-08 01:30:12 CET
that looks especially nasty

trying to reproduce at install:

[root@localhost ~]# urpmi --media Core\ Updates\ Testing mariadb-bench
A requested package cannot be installed:
mariadb-bench-5.5.25-2.7.mga2.x86_64 (due to unsatisfied libaio.so.1()(64bit))


hmm... libaio? is there some kind of aio things being tested for updates?



running the tests.... it seems the federated one is the first one to run, and so far i haven't had any failed tests yet, with all federated and rpl tests completed.

(i did a KVM dvd x86_64 install as minimal as possible; installed vim, less, ssh, bash-completion, removed sound stuff + orphans, updated to latest; poweroff; made a template)

rebooted, installed mariadb-bench + mariadb itself (no suggests), (patch seemed to be needed (i'll have to add this as a suggest in cauldron))

cd /usr/share/mysql/mysql-test
./mysql-test-run.pl --vardir=/tmp --force --big-test



tbh, i don't know how you got this, but if that many tests were failing, i doubt mariadb would even be workable?

do you have some specific configuration?

I can't really reproduce...
Comment 5 Dave Hodgins 2013-02-08 02:56:33 CET
As discussed on irc, this is
http://bugs.mysql.com/bug.php?id=61575

Workaround for now is to run "service mysqld stop" and "killall mysqld"
before running the tests on i586.
Comment 6 AL13N 2013-02-08 13:52:08 CET
this likely only happens if you have alot of mysqld 's running and specific virtualization (which has less aio).

in any case, 1 test is still failing and i'm trying to provide updated binaries for this, so i might as well also patch this aio issue...

i submitted mariadb-5.5.25-2.8.mga2 that should fix both issues.

how to test:

cd /usr/share/mysql/mysql-test && ./mysql-test-run.pl rpl.rpl_mdev382
AL13N 2013-02-08 13:54:56 CET

Assignee: alien => qa-bugs

Comment 7 claire robinson 2013-02-08 19:24:24 CET
Assigning back to AL13N as we're handling this now in the original bug 8921

Blocks: (none) => 8921
Assignee: qa-bugs => alien

Comment 8 AL13N 2013-02-21 07:22:20 CET
fixes are comitted for the failing tests mga2 (for next release) and submitted for cauldron

*** This bug has been marked as a duplicate of bug 8921 ***

Status: NEW => RESOLVED
Resolution: (none) => DUPLICATE


Note You need to log in before you can comment on or make changes to this bug.