[MDEV-26874] how to check what causing replication full in Galera? Created: 2021-10-21  Updated: 2022-10-04  Resolved: 2022-10-04

Status: Closed
Project: MariaDB Server
Component/s: Galera
Affects Version/s: 10.5.12, 10.5
Fix Version/s: N/A

Type: Bug Priority: Major
Reporter: William Wong Assignee: Jan Lindström (Inactive)
Resolution: Incomplete Votes: 1
Labels: None
Environment:

redhat



 Description   

Hi all,

In case production incident, found replication is full by checking status variable "wsrep_flow_control_active". But we don't know what to check further.

More than 50% chance the replication will be resumed within 15mins. However, we have several cases that replication is not resumed after 3 hours.

In case of replication not resumed long time, re reboot the DB node to get service resume.

Understand the problem may be caused by additonal workload transaction. But we want to check further. Anything we can check in next occurance?

Regards,
William Wong



 Comments   
Comment by Jan Lindström (Inactive) [ 2022-10-03 ]

Can you provide us full error log, node configuration and show processlist when this happens?

Comment by William Wong [ 2022-10-04 ]

Hi Jan,

The case was nearly 1 year ago. Log are not available already. We will capture information next occurrence.

Regards,
William Wong

Generated at Thu Feb 08 09:48:36 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.