Open
Description
Describe the bug
Queries will return wrong results during Ingester rolling.
To Reproduce
Steps to reproduce the behavior:
- Start Mimir 2.14 in K8s.
- Perform Operations(Read/Write/Others)
Firstly, restart a Ingester pod.
During the pod rolling (from the status 'Terminating' to 'Running 0/1'), some queries will return wrong results.
After the pod status becomes 'Running 0/1', the right results come back.
Right result:
Wrong result during Ingester rolling:
Expected behavior
The result should keep the same during the Ingester rolling.
Environment
- Infrastructure: k8s
- Deployment tool: helm
- some key config:
- ingester.ring.replication-factor = 2
- ingester.ring.unregister-on-shutdown = true
Others
I tried to change the config replication-factor and unregister-on-shutdown and test. Only 2 and true will reproduce this issue.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment