[MDEV-13792] innodb.purge_thread_shutdown failed in buildbot with wrong result (sporadic) Created: 2017-09-13  Updated: 2023-12-11  Resolved: 2023-12-11

Status: Closed
Project: MariaDB Server
Component/s: Tests
Affects Version/s: 10.2, 10.3
Fix Version/s: 10.4.33, 10.5.24, 10.6.17, 10.11.7, 11.0.5, 11.1.4, 11.2.3

Type: Bug Priority: Minor
Reporter: Alice Sherepa Assignee: Kristian Nielsen
Resolution: Fixed Votes: 0
Labels: None


 Description   

http://buildbot.askmonty.org/buildbot/builders/kvm-fulltest2/builds/9567/steps/test_2/logs/stdio

innodb.purge_thread_shutdown 'innodb'    w4 [ fail ]
        Test ended at 2017-09-13 01:37:34
 
CURRENT_TEST: innodb.purge_thread_shutdown
--- /mnt/buildbot/build/mariadb-10.2.9/mysql-test/suite/innodb/r/purge_thread_shutdown.result	2017-09-13 00:04:28.000000000 +0300
+++ /mnt/buildbot/build/mariadb-10.2.9/mysql-test/suite/innodb/r/purge_thread_shutdown.reject	2017-09-13 01:37:33.000000000 +0300
@@ -5,11 +5,11 @@
 select user,state from information_schema.processlist order by 2;
 user	state
 root	
+system user	closing tables
 root	Filling schema table
 system user	InnoDB purge coordinator
 system user	InnoDB purge worker
 system user	InnoDB purge worker
-system user	InnoDB purge worker
 system user	InnoDB shutdown handler
 set global debug_dbug='+d,only_kill_system_threads';
 set global innodb_fast_shutdown=0;
 
mysqltest: Result length mismatch



 Comments   
Comment by Kristian Nielsen [ 2023-12-11 ]

From Zulip discussions:

The test wants to show the list of running processes, to show which system threads and connection threads are running.

The problem is that using the state to identify the thread doesn't work, as it changes as the thread is doing stuff (enter_state()). I would suggest to change the query to SELECT user FROM information_schema.processlist order by 1, this way the user-vs-system threads can be seen which I think is good enough.

Or alternatively, include/wait_condition.inc with some condition that counts each kind of thread and compares with expected count. That would be equivalent, but not sure if that complexity is needed

Comment by Kristian Nielsen [ 2023-12-11 ]

Pushed to 10.4.
(Marko removed the test from 10.5, as it is no longer meaningful there.)

Generated at Thu Feb 08 08:08:23 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.