[MDEV-26549] GTID Replication. Slave doesn't start after switch master Created: 2021-09-06  Updated: 2021-09-06

Status: Open
Project: MariaDB Server
Component/s: Replication
Affects Version/s: 10.4.13, 10.4.15, 10.4.20
Fix Version/s: None

Type: Bug Priority: Major
Reporter: Streltsov Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: replication
Environment:

CentOS 7



 Description   

We have cascade model replication.
1 - > 2 -> 4 and 1 -> 3
We are using GTID for replication, using the same gtid_domain_id for all nodes, because only one node for record.
We want to change for 4 node master.
We stopped replication, changed master only master_host from 2 to 3 server, after start replication we have error Got fatal error 1236 from master when reading data from binary log: 'Could not find GTID state requested by slave in any binlog files. Probably the slave state is too old and required binlog files have been purged.'
But on 3 server in binlog files we found GTID position. We can start replication by specifying the file and position without GTID, and after start define master_use_gtid.
Why can't I change the master using only gtid?


Generated at Thu Feb 08 09:46:08 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.