Currently "Delete Clients" expect to have network
connectivity (ssh) to the nodes to be deleted such that
it can shut down some unnecessary services (eg. gmond).
This is bad if the nodes have already been taken out
of the network and/or powered down. We should
definitely check for some sort of connectivity before
performing the operation.
Seems like a bad bug and should be raised to priority 9.
Logged In: YES
user_id=879102
After looking into this a bit further, even though the
restarting of services failed, this does not seem to cause
major problems. I'll confirm this with the user and update
this issue.
Logged In: YES
user_id=1212811
Easy way to test if you can ping a node is run "ping -c 2 -l
2 -W 1 " before you try the SSH connections (-c 2 fires 2
packets, just incase you have network issues, -l 2 preloads
and fires both at the same time.. increasing the # of
packets and preload should make it more tolerant of network
issues)
errorlevel 0 is good, errorlevel 1 shows a no response on
the ping.
This should help for deletes of multiple clients, which
really lags because of the ssh timeouts.
Not really a huge issue, but gets annoying for large number
of machine deletes.
Logged In: YES
user_id=505737
Moving this to 5.0. We can revisit later.