[MXS-2787] Incorrect implementation of Galera replication Created: 2019-12-03  Updated: 2020-07-03  Resolved: 2020-07-03

Status: Closed
Project: MariaDB MaxScale
Component/s: binlogrouter
Affects Version/s: None
Fix Version/s: N/A

Type: Bug Priority: Major
Reporter: Michaël de groot Assignee: Unassigned
Resolution: Won't Do Votes: 1
Labels: None

Issue Links:
Problem/Incident
is caused by MXS-1980 Support Galera cluster nodes as maste... Closed
Relates
relates to MDEV-20720 Galera: Replicate MariaDB GTID to oth... Closed

 Description   

Hi,

The implementation of MDEV-1980 is incorrect. The GTID is not synced within the Galera cluster, therefore the gtid_slave_pos cannot be used for reconnecting the replication stream.

This implementation reads to inconsistent slave and not working point in time recovery.

The way to do this is to read the binlog that was retrieved from the old node, find the 'Xid' that belongs to the most recent transaction, then connect to a next node and find the same 'Xid' on the next node's binlog.

The procedure is risky because at SST all binlogs are removed on the joiner node.

 Comments   
Comment by markus makela [ 2019-12-17 ]

This looks more like a side-effect of non-functional GTID replication in Galera.

Generated at Thu Feb 08 04:16:41 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.