Ignore unreachable client error

[donasiv]

When some hosts from a group are unreachable, sup actually exits on the first SSH error. I propose the option flag "ignore-error" that just logs (instead return) when an host is not reachable.

[VojtechVitek]

You can remove the host from the Supfile .. or filter it out in the CLI right now:

sup --except failing-host.com

EDIT: I'd rather not introduce new options. "Error" is too vague term for ignoring the SSH connection failures only (which ones?).

[donasiv]

In my case, I don't want (and cannot) modify the Supfile and/or CLI calls on the fly.
At the moment, I cannot use sup to perform automatic admin operation on a pool of servers, since it fails each time one host goes down. Classic ssh errors: no route to host, connection refused, connection timeout.
That's why I suggest a flag to bypass and log connection issues and run commands on the up and running hosts.

[VojtechVitek]

@donasiv Could you use inventory command and get list of active servers only?

[donasiv]

I thought about it.
But in my case (again), the inventory list will rely on a monitoring tool running on an host administrated by sup itself. That's not secure at all.