I am using helm chart for elastic 7.12.0 cluster (2 master , 2 data) deployment on K8s, with a stateful-set (for persistence i have a volume claim). getting a lot of unassigned shards, where the shards are corrupted.

*"unassigned_info": { "reason": "ALLOCATION_FAILED" "can_allocate": "no_valid_shard_copy" "type": "corrupt_index_exception"

"caused_by": { "type": "e_o_f_exception", "reason": "read past EOF: NIOFSIndexInput(path="/usr/share/elasticsearch/data/nodes/0/indices/q-23aUrvSI6Sy9UPttY3Yw/0/index/_lpy_1.fnm") buffer: java.nio.HeapByteBufferpos=0 lim=1024 cap=1024 chunkLen: 1024 end: 29477" }*

has anyone come across this type of scenario / any configuration that i should be aware of wen dealing with persistent volumes for th cluster data?

Thanks