If you run 'up2date --nox -u' from remote through an sshd connection, your connection dies when it updates the openssh-server rpm. At least going from fisher to wolverine it does. This would seem to me to be something many people would do, as I don't usually want to log onto the console of every machine I have to run up2date. One other thing I noticed was that sshd did not restart, I had to go upstairs and restart it on the server manually. Not certain if this is up2date or openssh related.
This defect is considered MUST-FIX for Florence Release-Candidate #2
Isnt this an issue with the openssh packages? Shouldnt they be handling the "update" themselves? I dont know anything that up2date would be doing, or could be doing to prevent this.
this is an openssh issue, but I don't know if it can be changed.
IIRC this was due to some changes in the way initscripts would restart services. I believe that this is fixed in the current tree (I have experimentally verified this, too). Please reopen this bug if you find that this continues to be a problem with initscripts-5.65 and later.