MySQL Bugs: #90935: Modify relay_log

Bug #90935	Modify relay_log_recovery Documentation
Submitted:	18 May 2018 18:33	Modified:	4 Sep 2019 14:43
Reporter:	Juan Arruti	Email Updates:
Status:	Closed	Impact on me:	None
Category:	MySQL Server: Documentation	Severity:	S4 (Feature request)
Version:	5.7, 8.0	OS:	Any
Assigned to:	Margaret Fisher	CPU Architecture:	Any

Description:
Since MySQL 5.7 onwards does not need to enable relay_log_recovery to guarantee recovery, please change documentation to reflect this change for versions 5.7 and 8.0 where it states the following:

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 find below the links to MySQL documentation to be updated:

https://dev.mysql.com/doc/refman/5.7/en/replication-solutions-unexpected-slave-halt.html
https://dev.mysql.com/doc/refman/8.0/en/replication-solutions-unexpected-slave-halt.html

Many thanks

How to repeat:
N/A

The statement in the description of this bug is false.  It is not true that "since MySQL 5.7 onwards does not need to enable relay_log_recovery to guarantee recovery".  So the documentation bug is about implying that relay_log_recovery is not needed.

I have reported this in Bug#92093 with an example where recovery fails without relay_log_recovery.

Hi,
I'm verifying that in combination with Bug#92093 this is wrong. So both bugs should be worked on together.

thanks
Bogdan

Hi Jean-Franç, Bogdan:

Thanks for your inputs. Since same page for MySQL 5.6 documentation does not show this issue, the intention of this bug was to report the contradiction between the chart, showing relay_log_recovery can be ANY to make the slave as crash safe, and the immediate paragraph below to the chart in the documentation, in order to review them. Since Jean-Franç already has a test case which shows relay_log_recovery is needed for a crash safe slave, the issue seems to be in the chart and not in the paragraph, showing the same as MySQL 5.6 documentation.

Many thanks.
JP

Posted by developer:
 
Thanks for raising this issue. I have updated the table in
https://dev.mysql.com/doc/refman/8.0/en/replication-solutions-unexpected-slave-halt.html
https://dev.mysql.com/doc/refman/5.7/en/replication-solutions-unexpected-slave-halt.html
to match the text. Please note also the fix for Bug #92882 in MySQL 8.0.18 and 5.7.28, which makes a multi-threaded slave skip relay log recovery if GTID auto-positioning is available.