rule-checker does not consider the engine label constraint #6662
Labels
affects-7.1
This bug affects the 7.1.x(LTS) versions.
severity/critical
type/bug
The issue is confirmed as a bug.
Comments
JmPotato
added
affects-6.5
This was referenced Jun 24, 2023
JmPotato
removed
affects-6.5
ti-chi-bot bot
pushed a commit
that referenced
this issue
Jun 25, 2023
…#6660) close #6662 Signed-off-by: Ryan Leung <[email protected]>
rleungx
added
affects-6.5
rleungx
added a commit
to rleungx/pd
that referenced
this issue
Jul 10, 2023
…tikv#6660) close tikv#6662 Signed-off-by: Ryan Leung <[email protected]>
ti-chi-bot bot
pushed a commit
that referenced
this issue
Jul 10, 2023
…#6660) (#6778) close #6662 Signed-off-by: Ryan Leung <[email protected]>
ti-chi-bot
pushed a commit
to ti-chi-bot/pd
that referenced
this issue
Aug 14, 2023
close tikv#6662 Signed-off-by: ti-chi-bot <[email protected]>
|
We don't need to pick this fix to release-6.5 since it doesn't support survival preference in the placement rule in SQL. |
rleungx
removed
affects-6.5
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Bug Report
#6633 broke the constrain that we can only move peer between the stores that satisfy the rule, as a result,
rule-checkerwill move the learner peer away to a TiKV store while keep adding rule peer in the TiFlash stores. This will cause a TiFlash replica of a region never to be replicated well while bringing many redundant replicas to other TiKV nodes.In monitoring, its phenomenon may be as shown in the figure above, with a large number of
add-rule-peerandmove-to-better-locationat the same time. And in the log, you can see that the two act in the same region successively.The text was updated successfully, but these errors were encountered: