On Thu, 18 Nov 1999, Eric T. Jorgensen wrote: > > The outage of the shell server (and everything else that depends > on it) at about 9pm tonight was due to my own error. One of the > web servers was coming up with NFS errors from eskimo (the shell > server, and the location of the user directories among others). > Robert rebooted the machine seeing strange file system entries, > and after a minute or so the server started doing the same thing > again. I rebooted it again but still the errors were showing on > the screen. > > I killed the nfsd processes on eskimo and restarted them -- or so > I thought, since the other console windows only showed errors > during the couple seconds that elapsed between commands. However, > nfs is kernel-based on that machine, and it required a boot of > eskimo to restart it. Strong lesson learned there, of course... > > Luckily, after this, the original www console errors that > prompted the action in the first place also stopped. Sorry for > the downtime folks. > > ~ Eric Actually, Eskimo probably would have had to be rebooted even if you had not killed the nfs daemons.