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
We currently start pods from the lowest index to highest index (0->2), which does not match the behavior of how StatefulSet controller applies changes (2->0).
Why is this needed?
If user needs to make a change to get cluster to healthy state, we want to start the nodes that have received latest changes first. Currently, they would get started as the last nodes thus potentially leaving us in unfortunate states (for example if the old nodes have issues).
┆Issue is synchronized with this Jira Story by Unito
┆Issue Number: CASS-90
The text was updated successfully, but these errors were encountered:
What is missing?
We currently start pods from the lowest index to highest index (0->2), which does not match the behavior of how StatefulSet controller applies changes (2->0).
Why is this needed?
If user needs to make a change to get cluster to healthy state, we want to start the nodes that have received latest changes first. Currently, they would get started as the last nodes thus potentially leaving us in unfortunate states (for example if the old nodes have issues).
┆Issue is synchronized with this Jira Story by Unito
┆Issue Number: CASS-90
The text was updated successfully, but these errors were encountered: