[MDEV-27567] SegFault occurs after upgrading from MariaDB 10.1.35 with Galera Cluster to MariaDB 10.5.9 Created: 2022-01-21  Updated: 2023-08-07  Resolved: 2023-08-07

Status: Closed
Project: MariaDB Server
Component/s: Galera
Affects Version/s: 10.5.9
Fix Version/s: N/A

Type: Bug Priority: Major
Reporter: Venkata Satturi Assignee: Jan Lindström
Resolution: Incomplete Votes: 0
Labels: galera
Environment:

Platform : Drupal 9
OS : Redhat Enterprise Linux Server 7.6 on VMWare (Version)
Language : PHP 7.3.15
Web Server : Apache 2.4.6
Database : MariaDB 10.5.9 with Galera Cluster
Search Engine : Solr 6.6.5
BIG-IP Virtual Edition: Linux 3.10.0-514.26.2.el7.ve.x86_64: BIG-IP software release 13.1.3.2, build 0.0.4
BIG-IP Virtual Edition: Linux 3.10.0-514.26.2.el7.ve.x86_64: BIG-IP software release 13.1.3.2, build 0.0.4

we are on a three node cluster environment.


Attachments: Text File core-dump-backtrace-20220119.txt    

 Description   

After recent upgrade, all of sudden The MariaDB database is getting crashed on first VM, and one of the two databases were also getting crashed on second VM. To resolve impact, we had to stop the remaining database on second VM, edit the /data/database/mysql/grastate.dat file on first VM to say it's safe to bootstrap, bring up the database on First VM with a new cluster, and then start the databases on second VM.

Attaching the core dumps here.

Please let us know if the above description is not clear.

 Comments   
Comment by Jan Lindström [ 2023-06-06 ]

venkat557 Can you describe more detailed way how did you upgrade your servers. In this case we can't support rolling upgrade i.e. all nodes need to be shutdown and restarted one-by-one. Finally, only safe way to upgrade is 10.1->10.2->10.3->10.4->10.5 and following upgrade instructions.

Generated at Thu Feb 08 09:53:53 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.