Status (by brong at Wed Nov 19 08:12 UTC)
Due to a configuration change on one machine (which wasn’t actually supposed to be back in production yet) the supposedly redundant filestorage backend started timing out, causing everything to restart itself at the same time.
This is now fixed by changing the configuration, and will be fixed more completely by removing the timeout issue before I do any more changes.
The changes are due to 4 new much more powerful IMAP servers being phased in to replace our 4 oldest machines.
