[MXS-296] Weird characters in the slave status on the slave servers Created: 2015-07-30  Updated: 2015-09-04  Resolved: 2015-09-04

Status: Closed
Project: MariaDB MaxScale
Component/s: binlogrouter
Affects Version/s: 1.1.1
Fix Version/s: 1.2.0

Type: Bug Priority: Major
Reporter: Tibs Assignee: Massimiliano Pinto (Inactive)
Resolution: Fixed Votes: 0
Labels: None
Environment:

maxscale-1.1.1-2.x86_64
Centos 6.6
mysql 5.6.24-log


 Description   

Hi,

I restarted the master database server and on the slave servers which are connecting to the MaxScale Binlog servers I could see a lot of weird characters in the slave status but the replication is working.
Master_Log_File: mysql-bin.004821A���
Read_Master_Log_Pos: 135635651
Relay_Log_File: mysql-relay-bin.000090
Relay_Log_Pos: 30742929
Relay_Master_Log_File: mysql-bin.004821A���

I checked the master status on master, that one is fine:
show master status;
------------------------------------------------------------------------

File Position Binlog_Do_DB Binlog_Ignore_DB Executed_Gtid_Set

------------------------------------------------------------------------

mysql-bin.004821 139901230      

------------------------------------------------------+-----------------

Master status on binlog server is fine:
show master status;
-----------------------------------------------------------------------

File Position Binlog_Do_DB Binlog_Ignore_DB Execute_Gtid_Set

-----------------------------------------------------------------------

mysql-bin.004821 143078963      

-----------------------------------------------------------------------

Slave status on binlog server is fine:
Master_Log_File: mysql-bin.004821
Read_Master_Log_Pos: 144237107
Relay_Log_File: mysql-bin.004821
Relay_Log_Pos: 144237107
Relay_Master_Log_File: mysql-bin.004821

The file on the binlog server is fine:
rw-rw---. 1 root root 141948913 Jul 30 09:43 mysql-bin.004821

So everything is fine except the binlog file names on the slaves which are connecting to the binlog server.

If I stop the slave on the slave server and I restart it the file names will be fine again.

It did not cause any problems but it is weird and I could not find any report about this problem.

Thanks

 Comments   
Comment by Dipti Joshi (Inactive) [ 2015-07-30 ]

Massimiliano Pinto Has this been already fixed in 1.2 ?

Comment by Massimiliano Pinto (Inactive) [ 2015-07-30 ]

Tibs There was a bug, fixed n 1.2 about rotate event in slaves when master & slaves are running without CRC32

Could you please check in your setup whether master & slaves (I guess all mysql 5.6) have binlog-checksum set to CRC32.

You might report anyway whether there is a particular condition that would cause the issue

Comment by Tibs [ 2015-07-30 ]

Hi,

Yes we are using 5.6 everywhere and:
mysql> show global variables like 'binlog_checksum';
----------------------+

Variable_name Value

----------------------+

binlog_checksum NONE

----------------------+

I will make a test with CRC32 to see can I reproduct the issue or not.

Thanks,
Tibi

Comment by Dipti Joshi (Inactive) [ 2015-08-04 ]

Tibs Have you been able to reproduce it with MaxScale 1.2 ?

Comment by Massimiliano Pinto (Inactive) [ 2015-08-24 ]

@tibs Hi, any news about the issue reported ? Current MaxScale, 1.2, fixes the issue

Generated at Thu Feb 08 03:58:14 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.