Deleting node-0 in 3 node cluster after upgrade from 8.2.5 to 8.3.0 results in data loss.

100 entries put into persisted cache before upgrade are available right away after the upgrade. But upon deletion of node-0 after the upgrade only two thirds of the entries are available.

Nodes doesn't seems to have updated ___global.state file right away after the upgrade but seems to updated be after addition restart.

___global.state file right away after upgrade:

$ cat ___global.state 
@version=12.1.7.Final-redhat-00001
version-major=12
@timestamp=2021-12-08T17\:37\:42.391884Z
uuid=d0218fac-3e8d-4dde-b798-9f88744763fe

___global.state file after additional restart (taken from different node):

@version=13.0.2.Final-redhat-00002
version-major=13
@timestamp=2021-12-08T17\:44\:59.218123Z
uuid=38707135-df99-4cf2-b045-acba68556a72

Cache configuration:

{
  "distributed-cache": {
    "mode":"SYNC", 
    "encoding": {
      "media-type": "application/json"
    },
    "persistence": {
       "file-store":{}
    },
    "statistics":true
  }
}