For some scitas clusters, ssh takes a very long time to figure out that it cannot connect (bellatrix, castor, deneb1, deneb2), while for others it immediately exits with "no route to host" (eltanin, fidis).
This can be remedied by passing "-o ConnectTimeout=3" to the call to ssh in the can_login function, limiting the timeout to 3s.
I cannot connect to lsmosrv1 and also have a timeout (deneb)/no route to host (fidis) problem. If it cannot connect without the proxy, I would make this default.
For some scitas clusters, ssh takes a very long time to figure out that it cannot connect (bellatrix, castor, deneb1, deneb2), while for others it immediately exits with "no route to host" (eltanin, fidis).
This can be remedied by passing "-o ConnectTimeout=3" to the call to ssh in the
can_login
function, limiting the timeout to 3s.