zalando / zalando/postgres-operator
Retry unsuccessful failover on unschedulable nodes
Nobody has claimed this yet.
- Dominant language
- Go
- Stars
- 5.2k
- Forks
- 1.1k
- Avg merge
- 2d 16h
- Merged PRs (30d)
- 3
Description
- Which image of the operator are you using? ghcr.io/zalando/postgres-operator:v1.12.2
- Where do you run it - cloud or metal? Kubernetes or OpenShift? Kubernetes
- Are you running Postgres Operator in production? yes
- Type of issue? Feature request
When draining a node where the leader is running before any replica has become ready the failover will not succeed. That is good. However, if the replica then becomes ready the failover is never retried and you have to uncordon and redo the drain for it to succeed.
I believe the relevant part is here: https://github.com/zalando/postgres-operator/blob/2e398120d2d0b3bb2b8bb239c6d49011ebe37e88/pkg/controller/node.go#L68-L72
Would you be open to change this behavior? Is the harm in letting the failover retry if it the node is still not ready?
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start at pkg/controller/node.go around lines 68-72 and trace the node-drain failover path, including what happens when no replica is ready. Reproduce or inspect the behavior with an unschedulable node; done means a later-ready replica causes failover to be retried without uncordoning and redraining the node.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- go, kubernetes, postgresql
- Domain
- databases, devops, infrastructure
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 52/100