PowerStore: NAS docker panic during NDU
Summary: NAS docker panics while updating remote system updateRemoteSystem.
Symptoms
The issue occurs when the following conditions are met:
- A File Remote system is configured.
- Both systems (source and remote) are upgraded almost around same time, rather than sequentially.
- One of the systems got upgraded first and the other one is still under upgrade and NAS is down for NDU or because of any node reboot.
Cause
SDNAS periodically updates the remote system version and when such request is received from remote appliance(already upgraded) and one of the NAS Nodes is down on local appliance for an extended period such as during an NDU process or Node reboot, the request waits for the node to be up, however the remotes sends the next request with same task id to the local appliance which completes immediately and free up the resources which are still in use by previous request and this leads to NAS docker panic.
Resolution
The workaround is to perform NDU one site at a time if File Remote System is configured. It is an assumption that if users have such a configuration, they will not try to upgrade both primary and backup sites at same time.