Bug #47182 Timeouts in "embedded" test run happen specific to the SuSE 9 x86 RPM build
Submitted: 7 Sep 2009 20:01
Reporter: Joerg Bruehe Email Updates:
Status: Verified Impact on me:
Category:MySQL Server: Embedded Library ( libmysqld ) Severity:S3 (Non-critical)
Version:5.4.2-beta OS:Linux (SuSE 9, x86, )
Assigned to: CPU Architecture:Any
Triage: Triaged: D3 (Medium)

[7 Sep 2009 20:01] Joerg Bruehe
In the build of 5.4.2-beta, I have 10 timeouts in "embedded" test runs,
and they all happen on the same platform: SuSE-9-x86.

That cannot be explained with statistical variations, and the test run times are not so close to the 900 second timeout that this might really be a timing issue.

Affected tests:

All these tests failed again with timeout on retry.

As is usual with test timeouts, only very little information is given - like here:
main.innodb                              [ retry-fail ]  timeout after 900 seconds
        Test ended at YYYY-MM-DD HH:MM:SS

Test case timeout after 900 seconds

== /PATH/mysql-test/var/log/innodb.log ==

 - saving '/PATH/mysql-test/var/log/main.innodb/' to '/PATH/mysql-test/var/log/main.innodb/'

Here are the times of this test from the three SuSE RPM builds on x86 (SuSE 9, 10, and 11):
SuSE 9:
main.innodb       [ pass ]  28714
main.innodb       [ pass ]  29825
main.innodb       [ pass ]  27734
main.innodb       [ fail ]  timeout after 900 seconds
main.innodb       [ retry-fail ]  timeout after 900 seconds

SuSE 10:
main.innodb       [ pass ]  26357
main.innodb       [ pass ]  24283
main.innodb       [ pass ]  23418
main.innodb       [ pass ]  26427

SuSE 11:
main.innodb       [ pass ]  28284
main.innodb       [ pass ]  25413
main.innodb       [ pass ]  26439
main.innodb       [ pass ]  23605

In all these cases, the 4th run is "embedded", and its times (on SuSE 10 and 11) are close to that of client-server separated.
It would need a factor of (more than) 30 to bring these times to the 900 second test timeout.

These timeouts are mixed with many tests that pass, with decent times, so there is no symptom of a sudden machine overload or other crawl explaining this.

How to repeat:
It happened during the release build,
but I did that only once, so I can't guarantee it will re-occur.
[6 Jan 2010 15:04] Joerg Bruehe
That very same bug has occurred during the release builds of 5.1.40 and 5.5.1-m2.