Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

evict-leader-scheduler may fail to persist configuration #2572

Closed
disksing opened this issue Jun 23, 2020 · 4 comments · Fixed by #3787
Closed

evict-leader-scheduler may fail to persist configuration #2572

disksing opened this issue Jun 23, 2020 · 4 comments · Fixed by #3787
Labels
severity/moderate status/TODO The issue will be done in the future. type/bug The issue is confirmed as a bug.

Comments

@disksing
Copy link
Contributor

It is a hypothetically speculated bug: when remove a store id from evict-leader-scheduler, PD first update the in-memory configuration and then persist it. If it fails to persist to etcd, the operation looks successful, but later when PD leader changes, the removed store id will come up again.
@disksing disksing added the type/bug The issue is confirmed as a bug. label Jun 23, 2020
@rleungx rleungx added the status/TODO The issue will be done in the future. label Jun 24, 2020
@rleungx
Copy link
Member

rleungx commented Aug 25, 2020

Does #2817 solve this problem?

@disksing
Copy link
Contributor Author

No.

@nolouch
Copy link
Contributor

nolouch commented Sep 4, 2020

maybe we can use etcd STM?

@nolouch
Copy link
Contributor

nolouch commented Sep 9, 2020

Could you take a look? @ZenoTan

@rleungx rleungx mentioned this issue Jun 17, 2021
Merged
This was referenced Jun 30, 2021
Merged
Merged
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
severity/moderate status/TODO The issue will be done in the future. type/bug The issue is confirmed as a bug.
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

schedulers: rollback the scheduler when failing to persist rleungx/pd
3 participants
@nolouch @disksing @rleungx