Async Replication error between galera clusters

I recieved the following error when using asnyc replciation between galera clusters:

Error 'Percona-XtraDB-Cluster prohibits use of DML command on a table (wsrep.membership) that resides in non-transactional storage engine with pxc_strict_mode = ENFORCING or MASTER' on query. Default database: ''. Query: 'DELETE FROM wsrep.membership'

Should I just set replicate-ignore-db=wsrep on the slave?

1 Like

I’m not familiar with the database ‘wsrep’. Is that a custom database for your application? If so, then you have a table in that database which is not InnoDB. You should convert this table to InnoDB storage engine.

1 Like

https://jira.percona.com/browse/PXC-1189

I have no idea how or when /usr/share/mysql/wsrep_notify is getting called as wsrep_notify_cmd is unset, but that script creates this table.

1 Like

From Percona developers in [PXC-1189] LP #1676771: wsrep_notify script does not work with pxc_strict_mode - Percona JIRA

I would NOT recommend to change the example script to use InnoDB tables instead, as it would introduce data inconsistency in the nodes, due to disabled replication in the function used.
Instead, I would rather drop the idea of using DB tables there, and either use local files or update external monitoring DB instance maybe