Bug #83711 The description of the crash safe slave with GTID in the manual is incorrect
Submitted: 7 Nov 2016 3:33 Modified: 11 Jul 2017 10:14
Reporter: huajun chen Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Server: Documentation Severity:S3 (Non-critical)
Version:5.6 5.7 OS:Any
Assigned to: CPU Architecture:Any
Tags: crash safe, GTID

[7 Nov 2016 3:33] huajun chen
Description:
According the manual,When using GTIDs and MASTER_AUTO_POSITION,should "set relay_log_recovery=0" for crash safe slave. But it seems to be a mistake, and is incorrect with the content of the table above. (The correct one should be "set relay_log_recovery=1")
 
http://dev.mysql.com/doc/refman/5.6/en/replication-solutions-unexpected-slave-halt.html
----------------------------------------- 
When using GTIDs and MASTER_AUTO_POSITION, set relay_log_recovery=0. With this configuration the setting of relay_log_info_repository and other variables does not impact on recovery. 
----------------------------------------- 

How to repeat:
always
[7 Nov 2016 4:38] MySQL Verification Team
Hello huajun chen,

Thank you for the report and feedback!

Thanks,
Umesh
[11 Jul 2017 10:14] Margaret Fisher
Posted by developer:
 
Thank you for the report. I have made the correction in the 5.6, 5.7, and 8.0 manuals.
[17 Nov 2018 14:35] Prasad N
I think there is still an inconsistency in this documentation: https://dev.mysql.com/doc/refman/5.7/en/replication-solutions-unexpected-slave-halt.html
between the table and the explanation.

While in the table it is given that relay_log_recovery can be 'Any' which I assume as either 0 or 1, for the case when using GTID and MASTER_AUTO_POSITION.

The explanation below states that relay_log_recovery has to be 1 in such a case.
------------------------------------------------------------------
As the table shows, when using a multithreaded slave the following configurations are most resilient to unexpected halts:

When using GTIDs and MASTER_AUTO_POSITION, set relay_log_recovery=1. With this configuration the setting of relay_log_info_repository and other variables does not impact on recovery.
----------------------------------------------------------------------

Please clarify this in the documentation. 

Thanks!