Those of you unfortunate enough to be on the Homie email cluster know that it’s been going down repeatedly for at least a month, and Dreamhost doesn’t seem able to fix it permanently. Moreover, frustrated users are kept in the dark as DH gets deluged with tickets and can’t respond to them – tickets related to the server outages are “separated out” and not answered, and users have to wait for a brief and belated update on the Status page.
They don’t even seem to have a handle on what the problems are: the status update they eventually posted today mentions “authentication” problems (true enough for some users), but many users cannot connect to the mailserver at all: see comments at http://www.dreamhoststatus.com/2009/05/06/homie-cluster-mail-authentication-issues/
This kind of downtime is unacceptable (and, in my 12+ years of hosting my domain on other servers, unprecedented).
If Dreamhost can’t fix Homie’s problems after trying for a month, it’s time to kill that cluster and migrate users to a new one that functions. Any downtime involved in the migration will be less than the outages we’re experiencing already in the middle of our workdays.