You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Leader election is critical for high availability for production environment. However, the extender does not need leader election, when it goes down, a new leader will take place since liveness probe works.
2. Steps to reproduce the issue
In case the single HAMi scheduler were not healthy or restarting, the total cluster would not schedule any pod since the scheduler is using extender with Bind and the Bind stage is only suitable for one implement. The scheduler itself can be HA, so this issue will focus on liveness probe on deployment of the HAMi scheduler.
3. Information to attach (optional if deemed irrelevant)
1. Issue or feature description
Leader election is critical for high availability for production environment. However, the extender does not need leader election, when it goes down, a new leader will take place since liveness probe works.
2. Steps to reproduce the issue
In case the single HAMi scheduler were not healthy or restarting, the total cluster would not schedule any pod since the scheduler is using extender with
Bind
and theBind
stage is only suitable for one implement. The scheduler itself can be HA, so this issue will focus on liveness probe on deployment of the HAMi scheduler.3. Information to attach (optional if deemed irrelevant)
Reference: By Tencent Cloud
The text was updated successfully, but these errors were encountered: