Search server being rebooted

[mark]

I'm doing some maintenance on our search database today. Slicehost has been poking at it for two days now to try to figure out why it has 75%+ packet loss to/from the admin server, but they have no clue what's wrong or why.

This is pretty terrible and makes me sad. We're starting to seriously look into physically colocating our servers in some nearby data center, because of the magnitude of random, unpredictable errors/issues Slicehost has been having. Unfortunately, that's a fairly pricey operation ($50,000+).

Anyway, if you get an error while doing journal search today, please wait a few minutes and try again. I'll keep the downtimes short.