Eskimo crashed shortly after 9am. Unfortunately, file systems didn't fsck clean the first time around and it took multiple passes. When it did come up, the web servers didn't recover properly. Unfortunately, an ongoing bug in Linux seems to be that if the NFS server goes away, sometimes when it goes back the machines still won't talk. It's an intermittant problem, it doesn't happen every time and when it does there aren't really any indications of why they aren't talking. Everything should be back to normal now.