kubernetes pods getting evited with error "eviction manager: attempting to reclaim memory" even if memory consumption is far less

Question

Kubernetes pod running zookeeper is getting evicted quiet oftenly with error

***** eviction_manager.go:344] eviction manager: attempting to reclaim memory
***** eviction_manager.go:355] eviction manager: must evict pod(s) to reclaim memory

But memory used is far less. Can someone help me point the issue

Answer 1

You can resolve your issue by changing the Hard eviction threshold. You have to define thresholds for low RAM and low disk space, Kubernetes eviction policies act when those thresholds are reached. Kubernetes evicts Pods from a node to fix low RAM and low disk space problems.

Like below :

eviction-hard=”memory.available<600Mi”

Defines that when less than 600Mi RAM is available, Pods must be evicted HARD..immediately.

And eviction-pressure-transition-period=”30s”

Disk space eviction policies work the same as RAM eviction policies. Once you understand RAM eviction you will be able to easily apply your knowledge to disk space eviction. Please refer to Kubernetes Eviction Policies for Handling Low RAM and Disk Space Situations for more information.