重新同步 galera 节点

Question

I have galera cluster with several nodes. Some of the nodes were down from long time (4 days) and the binlog files get deleted daily. Now I am unable to start any of these nodes.

Dec 25 00:20:06 mysqld[14220]: 2022-12-25  0:20:01 0 [ERROR] mysqld: File '/var/log/mysql/master-bin.059377' not found (Errcode: 2 "No such file or directory")
Dec 25 00:20:06 mysqld[14220]: 2022-12-25  0:20:01 0 [ERROR] Failed to open log (file '/var/log/mysql/master-bin.059377', errno 2)
Dec 25 00:20:06 mysqld[14220]: 2022-12-25  0:20:01 0 [ERROR] Could not open log file
Dec 25 00:20:06 mysqld[14220]: 2022-12-25  0:20:01 0 [ERROR] Can't init tc log
Dec 25 00:20:06 mysqld[14220]: 2022-12-25  0:20:01 0 [ERROR] Aborting'

How can I force these nodes to start over as if they were new nodes?

Answer 1

I solved the issue and it worked. In my scenario, I have 4 nodes, 3 of them were down for a long time and a single node was in a healthy state and running. When I tried to start any of the broken nodes, they refused as they couldn't find the needed master-bin.x log files. I forced full sync by running the following on the broken nodes:

sudo -u mysql mysqld --skip-slave-start

Once the node stated changed to 'joined' and 'synced', I softly killed the process and started it again using systems.