Comment 5 for bug 811447

I was going to assume this is an issue of cold DB cache, but I actually doubt that now: did you manage to get it loaded at all for yourself, Andreas?

Least we can do is make the two slow, almost identical queries a single one (they differ in DISTINCT presense and what is being selected, but are otherwise compatible): http://pastebin.ubuntu.com/659339/ and http://pastebin.ubuntu.com/659341/

The way it works for me is that the two queries, instead of being 1200+ms each are roughly 150ms each as can be seen in OOPS-2043H61. Identical query when ran on staging took >100s (not ms, actual seconds)! Warm-cache run afterwards took 2.7s.