Closed narayanbhawar10 closed 1 year ago
Are you saying that if you increase a nifikop deployment to 2 replicas, only one actually does any work?
If so, this is by design. This works through leader election, ensuring there aren't concurrency issues processing individual CRDs (e.g. NifiCluster
, NifiDataflow
, etc.) in Kubernetes. Given a nifikop deployment with 2 replicas and if/when one nifikop crashes, the other will immediately take over responsibility for all work because it will become the leader.
This is how any workload in Kubernetes behaves which needs to avoid multiple pods processing a single shared resource 🙂
@mh013370 Thanks for making me understand and I understood. I really appreciate.
What steps will reproduce the bug?
Increase nifikop replica to reproduce issue.
What is the expected behavior?
Both replica should work in parallel
What do you see instead?
Only one replica working
Possible solution
No response
NiFiKop version
v1.0.0
Golang version
latest
Kubernetes version
latest
NiFi version
No response
Additional context
No response