Details
-
Bug
-
Status: Done
-
High
-
Resolution: Fixed
-
5.7.26-29, 8.0.16-7, Not 5.6.x
-
None
Description
Hi,
It looked like we already fixed this bug: https://bugs.mysql.com/bug.php?id=90351
But in MySQL 8.0.16 it still happening, we upgraded to 8.0.16 because basically many graphs were ruined because of this bug, but still happening, but not that frequent anymore maybe once in every few hours previously it was happening all the time:
I have started Iwo's script on the master server to monitor the statistics and yes we can see it is happening:
5468630320 < 5468650972 (Handler_commit) 5468628952 5468650972 5468630320 12136264379 < 12136310039 (Handler_external_lock) 12136261349 12136310039 12136264379 165444232 < 165444616 (Handler_prepare) 165444214 165444616 165444232 357644957 < 357646500 (Handler_read_first) 357644853 357646500 357644957 42022468474 < 42022518876 (Handler_read_key) 42022464955 42022518876 42022468474 85712119745 < 85712174300 (Handler_read_next) 85712105821 85712174300 85712119745 8663652932 < 8663667454 (Handler_read_prev) 8663651572 8663667454 8663652932
But because we took MySQL 8 code I am pretty sure this could happen on Percona Server 8 as well, and it ruins most of the PMM graphs.
We should review's oracle fix , why is this happening in MySQL 8 again.