[MDEV-4491] FR: Indicate GTID-based start position in the error log on slave start Created: 2013-05-06  Updated: 2013-06-07  Resolved: 2013-06-07

Status: Closed
Project: MariaDB Server
Component/s: None
Fix Version/s: 10.0.4

Type: Task Priority: Minor
Reporter: Elena Stepanova Assignee: Kristian Nielsen
Resolution: Fixed Votes: 0
Labels: None

Issue Links:
Relates
relates to MDEV-26 Global transaction ID Closed

 Description   

When a slave starts in the old-style way, the record in the server error log contains information about start coordinates:

130506 18:25:24 [Note] Slave I/O thread: Start asynchronous replication to master 'root@127.0.0.1:3306' in log 'ubuntu12-04-bin.000002' at position 377
 
130506 18:25:24 [Note] Slave SQL thread initialized, starting replication in log 'ubuntu12-04-bin.000002' at position 377, relay log './ubuntu12-04-relay-bin.000001' position: 4
 
130506 18:25:24 [Note] Slave I/O thread: connected to master 'root@127.0.0.1:3306',replication started in log 'ubuntu12-04-bin.000002' at position 377

But with GTID, the position is always zero, both for SQL and IO thread:

130506 18:23:57 [Note] Slave I/O thread: Start asynchronous replication to master 'root@127.0.0.1:3306' in log 'FIRST' at position 0
 
130506 18:23:57 [Note] Slave SQL thread initialized, starting replication in log 'FIRST' at position 0, relay log './ubuntu12-04-relay-bin.000001' position: 4
 
130506 18:23:57 [Note] Slave I/O thread: connected to master 'root@127.0.0.1:3306',replication started in log 'FIRST' at position 0

I suppose it might be impossible to log the old-style position at this point since it might be unknown (right?), but then it would be good to log the GTID position which the slave attempts to acquire, otherwise it's totally impossible to understand where the slave was at the moment of (re)start.

bzr version-info

revision-id: knielsen@knielsen-hq.org-20130506123534-v8r8bhazhj5gr1he
 
revno: 3628
 
branch-nick: 10.0-base



 Comments   
Comment by Kristian Nielsen [ 2013-06-07 ]

The current code in 10.0-base now gives better information
about what it is doing, GTID-wise, when the slave is using GTID.

There are still a couple of printouts that are done by plugins (I
think semisync plugin). Those do not yet output any GTID info,
as the interface and structs used there do not include the new
GTID info. I think that is ok for now, there are other printouts
that do have the GTID info. (We can maybe extend the
plugin stuff later, or maybe leave it if not needed).

Generated at Thu Feb 08 06:56:51 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.