Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Fixed
-
24.02.2
-
None
-
MXS-SPRINT-229
Description
In the scenario we are using Maxscale as a binlog even listener in conjunction with the cdc client to get any updates happening in mariadb.
During startup of maxscale we use current gtid by getting the one from 'maxsctrl list server' output which it is getting from mariadb provided in the server list.
Incase of mariadb restart, the maxscale reconnects with mariadb to get latest events. But the problem we are seeing is that the current_gtid.txt, which has the older gtid value we are seeing following error and the updates are not received to maxscale and cdc.
Below are the errors : _
2025-02-13 21:37:09 error : Failed to read replicated event: 1236, 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._
I notice following logs in mariadb logs :
2025-02-13 13:42:01 177 [Note] Start binlog_dump to slave_server(1), pos(, 0), using_gtid(1), gtid('0-1-108,1-1-1')
Is there a way to let maxscale sync from the latest gtid with mariadb during restart without need to provide that in the current_gtid.txt ?
Is it possible for maxscale to get the latest gtid from mariadb and start its current gtid with that by a providing a config parameter ?
Attachments
Issue Links
- relates to
-
MXS-5539 Maxscale error due to sync starting from old GTID rather than newest even with 'gtid_start_pos=newest'
-
- Closed
-