[MDEV-7083] sys_vars.innodb_sched_priority* tests fail in buildbot on work-amd64-valgrind Created: 2014-11-12  Updated: 2014-11-17  Resolved: 2014-11-14

Status: Closed
Project: MariaDB Server
Component/s: Storage Engine - XtraDB, Tests
Affects Version/s: 10.0
Fix Version/s: 10.0.15

Type: Bug Priority: Critical
Reporter: Elena Stepanova Assignee: Jan Lindström (Inactive)
Resolution: Fixed Votes: 0
Labels: buildbot, tests

Issue Links:
Blocks
blocks MDEV-7069 Fix buildbot failures in main server ... Stalled

 Description   

http://buildbot.askmonty.org/buildbot/builders/work-amd64-valgrind/builds/6310/steps/test/logs/stdio

sys_vars.innodb_sched_priority_cleaner_basic 'xtradb' w6 [ fail ]
 
        Test ended at 2014-11-11 05:01:29
 
 
 
CURRENT_TEST: sys_vars.innodb_sched_priority_cleaner_basic
 
--- /var/lib/buildbot/maria-slave/work-opensuse-amd64/build/mysql-test/suite/sys_vars/r/innodb_sched_priority_cleaner_basic.result	2014-11-11 00:35:08.000000000 +0200
 
+++ /var/lib/buildbot/maria-slave/work-opensuse-amd64/build/mysql-test/suite/sys_vars/r/innodb_sched_priority_cleaner_basic.reject	2014-11-11 05:01:29.000000000 +0200
 
@@ -5,6 +5,9 @@
 
 SELECT @@SESSION.innodb_sched_priority_cleaner;
 
 ERROR HY000: Variable 'innodb_sched_priority_cleaner' is a GLOBAL variable
 
 SET GLOBAL innodb_sched_priority_cleaner=19;
 
+Warnings:
 
+Warning	1210	Failed to set the LRU manager thread priority to 19,  the current priority is 9
 
+Warning	1210	Failed to set the page cleaner thread priority to 19,  the current priority is 9
 
 SELECT @@GLOBAL.innodb_sched_priority_cleaner;
 
 @@GLOBAL.innodb_sched_priority_cleaner
 
 19

Other tests fail in a similar way.
The failure is easily reproducible if you run MTR with nice 10, which happens on the work buildslave on whatever reason.

It could be classified as a configuration problem, but if you look at the failure, you'll see it's confusing regardless. The value of the variable is 19. But if you are trying to set it to 19, it throws a warning which says that the current value is 9. And it stays 19.

So, I'm not really sure what needs to be done this way, it might be 2-fold:

  • fix the code not to mix up values calculated based on the system nice with the visible values;
  • fix the build slave configuration.

It's all the same on Percona server, but I doubt they'll want to fix just for the sake of our builder (especially doubtful it would be done urgently).

 Comments   
Comment by Jan Lindström (Inactive) [ 2014-11-13 ]

revno: 4480
committer: Jan Lindström <jplindst@mariadb.org>
branch nick: 10.0-bugs
timestamp: Thu 2014-11-13 11:05:22 +0200
message:
MDEV-7083: sys_vars.innodb_sched_priority* tests fail in buildbot
on work-amd64-valgrind.

Fixed issue by finding out first the current used priority
for both treads and using that seeing did we really change
the priority or not.

Comment by Elena Stepanova [ 2014-11-14 ]

It looks like only sys_vars.innodb_sched_priority_cleaner_basic was fixed, the other 3 still fail in the same way.

http://buildbot.askmonty.org/buildbot/builders/work-amd64-valgrind/builds/6323/steps/test/logs/stdio

Comment by Jan Lindström (Inactive) [ 2014-11-14 ]

revno: 4493
committer: Jan Lindström <jplindst@mariadb.org>
branch nick: 10.0-bugs
timestamp: Fri 2014-11-14 09:13:13 +0200
message:
MDEV-7083: sys_vars.innodb_sched_priority* tests fail in
buildbot on work-amd64-valgrind

Fixed issue by finding out first the current used priority
for both treads and using that seeing did we really change
the priority or not.

Missed the other failing scheduling priority failures, this fixes rest of them.

Generated at Thu Feb 08 07:16:52 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.