Hi. On January 16th we performed a Kubernetes upgrade. This involved rolling the kubernetes nodes in a given order such that the
nats nodes/pods would roll first then the nats streaming pods would roll with the two followers first and the leader last.

I see the following sequence:

1. Nats-streaming-0 becomes the new leader at 17:41:12
2. Nats-streaming-2 restores two channels at 17:43:26
3. Immediately after the restoration nats-streaming-2 gets runtime error: panic: runtime error: invalid memory address or nil pointer dereference
4. nats-streaming-2 restarts but very quickly shuts down due to "STREAM: Failed to start: log not found"
5. nats-streaming-2 restarts again but again shuts down for the same reason
6. and one more time
7. nats-streaming starts up again 1.75 hours later (no idea why such a delay). NO restoring is performed even though some channels restored in steps 4, 5 and 6.
8. Yesterday we rolled the nats-streaming-2 pod but again no channels were restored.

Is there a way to mitigate/correct such that nats-streaming-2 restores the channels. I worry that if there is a subsequent leader election, nats-streaming-2 may become the leader with potentially bad results.

Log file from nats-streaming-2:

nats-streaming-not-attempting-restore.csv