The following is an email sent from Dreamhost staff at 6:23PM EST. As I write this it is now 11:31PM EST. Our service has been down for five hours.
I am sorry but I had to disable your database (it was single-handedly
killing the database server - the load was around 20 and is now down
under 1 (4 is optimal) after disablement.
We can re-enable it if you can assure us you’ll take proper steps to curb
the usage (you need to be sure you are indexing things if you weren’t and
perhaps re-evaluate the efficiency of your code).
I emailed back immediately, and three additional times over the last few hours - I could certainly help investigate, index tables, and more, if the database was re-enabled. It hasn’t. Not only that, I can’t access the database to do a dump and backup.
Previously Philly Future was quarentined due to excessive CPU utilization. We solved those issues and we were moved back to a standard box.
This has occured without warning or indication of trouble. Looking at the db stats, I don’t see a trend that indicates a problem.
We run a version of CivicSpace, with minimal modules enabled, and with an improved caching system to keep CPU/db utilization down.
This situation is frightening. Can anyone help me?