zrepl pull job fails after restarting service, also failures not shown in prometheus about zrepl

[deajan]

Well I've roamed the issues and found that mounting the replicas and doing a single ls -l will prevent zrepl from working, since atime will be modified, which is something I obviously did ;)
So far, I've rolledback the snapshots and replication works again.

My two cents:

1. It would be nice if recv placeholder could set readonly=on and atime=off to prevent these issues
2. There's still a problem with the prometheus support since nowhere the dashboard shows that replication fails. Could this be improved perhaps ?

[deajan]

Any thoughts on this ?