cndi icon indicating copy to clipboard operation
cndi copied to clipboard

Published 20 hours ago • verified publisher icon polyseam

[Enhancement]: Ensure Leader Node can Scale Vertically, no downtime

Open johnstonmatt opened this issue 2 years ago • 1 comments

Version

cndi v1.13.4 kubeseal v0.21.0 terraform v1.5.5

Please provide a summary of the enhancement you are proposing:

Prevent the loss of uptime or data when a leader node is scaled vertically.

Please provide the motivation or use case for this enhancement:

Scaling a leader node is likely a common thing, and downtime shouldn't be the result

How can we best workaround this issue so far?

No response

How would you approach solving this problem within CNDI?

No response

Code of Conduct

  • [X] I agree to follow this project's Code of Conduct

johnstonmatt avatar Sep 13 '23 20:09 johnstonmatt

If this solution involves setting a leader node using array index 0 in infrastructure.cndi.nodes for the initial deployment, and not setting a node as role: leader that is totally fine.

This may be required because if a leader is upgraded in place, it may not be the leader when it comes back online.

johnstonmatt avatar Sep 29 '23 03:09 johnstonmatt