[MDEV-15989] MariaDB server deadlock Created: 2018-04-23  Updated: 2018-07-02  Resolved: 2018-07-02

Status: Closed
Project: MariaDB Server
Component/s: OTHER
Affects Version/s: 5.5.54
Fix Version/s: N/A

Type: Bug Priority: Major
Reporter: Michael Graf Assignee: Unassigned
Resolution: Incomplete Votes: 0
Labels: need_feedback
Environment:

Windows Server 2008 R2 Enterprise


Attachments: Zip Archive Q100SPW-MZIS-H1.zip     File my.ini     Zip Archive mysqld.exe_180423_095842.zip    

 Description   

*Fetter Text*MariaDB hungs during shutdown.
After 10 hours it was still in status "shutting down".

The attachment contains the errorlog, a minidump and the used my.ini.



 Comments   
Comment by Elena Stepanova [ 2018-05-31 ]

Going through the error log, so

  • <lot of previous activity>
  • 171004 12:17:16:5.5.52 server started, buffer pool 512G, and on some reason assumed it needed recovery, even although the previous one, according to the log, shut down;
  • 171028 2:00:00 it still existed, then at unknown point it got lost;
  • 180221 12:10:14 5.5.52 server started, buffer pool 384G, ran recovery;
  • 180307 1:00:00 it still existed, then at unknown point it got lost;
  • 180405 10:06:55 5.5.52 server started, buffer pool 384G, ran recovery, at some point complained about a log semaphore wait;
  • 180410 8:50:51 started shutdown, no sign of finishing;
  • 180410 9:02:38 5.5.52 server started, buffer pool 384G, ran recovery;
  • 180410 9:12:18 it still existed, then at unknown point it got lost;
  • 180420 7:49:32 5.5.54 server started, buffer pool 384G, InnoDB complained about a fatal memory error (InnoDB: VirtualAlloc(2281701376 bytes) failed; Windows error 1455), server shut down;
  • 180420 7:51:46 5.5.54 server started again, buffer pool 64G, ran recovery;
  • 180420 8:40:44 normal shutdown (done);
  • 180420 8:46:42 5.5.54 server started, 64G buffer pool;
  • 180420 14:13:30 started shutdown, never finished;
  • 180420 14:17:56 5.5.54 server started, 64G buffer pool (no signs of crash recovery);
  • 180421 2:25:26 the server started complaining about Out of memory (Needed 2099232 bytes);
  • 180421 21:01:23 last complaint about OOM, InnoDB started complaining about a long semaphore wait;
  • 180423 0:00:04 normal shutdown started, numerous complaints till the end of the log

    180423  0:00:08 [ERROR] mysqld.exe: Fehler beim Schreiben eines Kommunikationspakets
    180423  0:00:08 [Warning] d:\leutek\mariadb\bin\mysqld.exe: Thread 14454 zwangsweise beendet. Benutzer: 'leutek'
    

So, which failure to stop are we looking into here, the last one?

Generated at Thu Feb 08 08:25:32 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.