From WHManager, try &Update System Software&.... and update backend scripts as well.
Then when its complete try the service restart from WHManager--- otherwise try a hard reboot, and if it doesn't come up normally you'll have to have someone at the NOC do surgery on it unless you have another method of remote access.
Somebody at the NOC needs to update / reinstall ssh from the source or force the RPM in there......
unless you have telnet open in some secure fashion?
Otherwise, definitely have someone with physical access grab the latest rpm release for your OS...
rpm -Uvh rpm* --force
Problems might occur with those flags but for fastest access to SSH this might be the quickest way.
Not to un-necessarily bump this, but here's something I discovered:
I went to the data center the other night and couldn't fix it. So I enabled root login on the server, and was able to connect via root. I STILL can't connect via my regular username. It logs in, sends the MOTD and then disconnects.
While you are in with root, connect with regular shell users, and tail -f /var/log/messages.... this should give you a hint of WTfreaking hell is going on. (tail the messages log with one of your root shells that is)
Or at least some fun data to post back here so we can gallavant around with it and pretend we know what's going on.