Hey folks,
Due to a recent heatwave in Los Angeles, the cooling at the datacenter e621 (along with a bunch of other sites I host) was unable to keep up. This resulted in 2 disks failing in another server, and necessitated a move of a large number of virtual machines to another server in the same rack (which also happened to have better cooling). This didn't affect e621.
Simultaneous with this, a hardware failure developed in the main network switch which serves e621 (most likely also caused by the heatwave), which resulted in intermittent problems with some servers working, and some not, for roughly the last 24hrs. This faulty switch has now been replaced, and services are back to normal on e621 (and most of the other sites).
It fucking sucks that this happened, as the cooling at the datacenter in question was flaky to begin with; the datacenter is having another chiller added on the roof of the building which should dramatically improve the situation (which is expected to happen tonight). If the situation is not resolved within the next 2 days, we'll be moving everything to another facility and avoiding this bullshit which should've been dealt with earlier ever again.
The site should remain up and stable until we move all our equipment, and advance warning will be posted on the forums and in other places such as my twitter account here:
Ugh, these last 2 days have fucking sucked...
In other news, I have a new faster speed test result from the datacenter (see below).
http://www.speedtest.net/result/972078680.png
Updated