[MDEV-20413] Mariadb Galera Cluster Crash and won't restart Created: 2019-08-23  Updated: 2019-12-13  Resolved: 2019-12-13

Status: Closed
Project: MariaDB Server
Component/s: Galera, Storage Engine - InnoDB, wsrep
Affects Version/s: 10.3.17
Fix Version/s: N/A

Type: Bug Priority: Major
Reporter: Patrice CARMANT Assignee: Marko Mäkelä
Resolution: Incomplete Votes: 0
Labels: need_feedback
Environment:

Debian 8.11

Attachments: Text File node1_crash.txt     Text File node2_crash.txt     Text File node3_crash.txt     Text File node_innodb-force-recovery-6.txt     Text File node_wsrep-new-cluster.txt    
Issue Links:
Relates
relates to MDEV-19743 Crash while reorganizing an index page Closed
relates to MDEV-19783 Random crashes and corrupt data in IN... Closed
relates to MDEV-19916 Corruption after instant ADD/DROP and... Closed

 Description   

Our Galera Cluster crash (3 nodes successively) with the following error on the first node

All nodes are unable to restart.

We tried to bootstrap the cluster whith mysqld --wsrep-new-cluster without success

We also tried to bootstrap the cluster whith mysqld --wsrep-new-cluster --innodb-force-recovery=from 1 to 4 without success

on one node we tried --innodb-force-recovery=6 without success too

 Comments   
Comment by Patrice CARMANT [ 2019-08-25 ]

We tried to update one node to 10.4.7 but galera won't start anymore

Comment by Elena Stepanova [ 2019-08-26 ] 

ERROR] [FATAL] InnoDB: Rec offset 99, cur1 offset 10363, cur2 offset 16179

Comment by Marko Mäkelä [ 2019-11-15 ]

vialink_syst, did you upgrade to MariaDB 10.3.17 from some earlier 10.3 version?

If yes, the table could have been corrupted while running an earlier version. I suspect that you could have been the victim of MDEV-19916 (corruption of the table after using instant ADD COLUMN).

Generated at Thu Feb 08 08:59:17 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.