MDEV-36340: Reset Connects_Tried with Master_Retry_Count=X
#3974
Conversation
| WHERE status_reset.Connects_Tried <> | ||
| IF(LENGTH(Connection_name), status_restart.Connects_Tried, 0); | ||
| IF(Connection_name = '', 0, status_change.Connects_Tried); |
There was a problem hiding this comment.
This format to match #3897.
Conflict Resolution: this wins.
There was a problem hiding this comment.
Hi @ParadoxV5 !
Thanks for a good patch and test - I just left one question.
| mi->retry_count= lex_mi->retry_count; | ||
| mi->connects_tried= 0; |
There was a problem hiding this comment.
Would it make sense to reset this for all CHANGE MASTER executions, rather than just when
MASTER_RETRY_COUNT is provided? That is, if we are changing the connection parameters (e.g. MASTER_HOST), it would make sense to me that we shouldn't report the old connects_tried for a new connection.
There was a problem hiding this comment.
I say naw, because other parameters, including MASTER_HOST and co., don’t relate to Connects_Tried, and so have no reason to clear it when it’s intended to remain past STOP REPLICA.
|
@bnestere says in JIRA that QA’s done. Now, (still) waiting for the main branch to switch to DST… 🫤 |
Rebase failed
Rebase failed
manual update time |
CHANGE MASTER can set `Master_Retry_Count` to be lower than `Connects_Tried`, which’ll look strange to those unaware of that CHANGE. Now, setting `Master_Retry_Count` also resets `Connects_Tried` to 0. Reviewed-by: Susil Behera <susil.behera@mariadb.com> Reviewed-by: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
Description
CHANGE MASTER can set
Master_Retry_Countto be lower thanConnects_Tried, which’ll look strange to those unaware of that CHANGE.Now, setting
Master_Retry_Countalso resetsConnects_Triedto 0.Release Notes
Amend MDEV-35304’s with:
Connects_Triedto 0, but RESET REPLICA does, and also does changingMaster_Retry_Countwith CHANGE MASTER.How can this PR be tested?
This PR adds a test case to
multi_source.connects_tried.This PR also
FROM status_before JOIN status_after USING(Connection_name); no function changewait_for_slave_to_start/stop.inctostart/stop_slave.incPR quality check
The bug is about a feature exclusive to the
mainbranch. Or is this a new sub-feature?mainbranch.