<div dir="ltr">Hi,<div><br></div><div>Yet again the wiki went down today, this time with a 500 which appeared to be a result of a the single backend of fast-cgi dying (a test we tried a few days ago in case of thread issues).</div>
<div><br></div><div>This time I tried to increase the backend count back to something a little higher with a quicker turn over on the child threads, as it would appear something in the ancient PHP/MW code is crashing.</div>
<div><br></div><div>I also created a my.cnf for mysql instead of using the stock settings (normally this is done, but was accidentally skipped by Sergey as he didn't know). Hopefully this helps get even more pressure off of the wiki and avoid hitting the queue limits. I make no promises though :)</div>
<div><br></div><div>Anyway, sorry for all the hassles! I will try keep an eye on it for a bit tonight and also watch MySQL, and probably restart it a few times to tweak some performance settings after it's been running for a while. Please mind any database errors if you happen to visit a page while I'm restarting stuff.</div>
<div><br></div><div>Cheers o/</div><div><br></div><div><br></div><div>Dan</div></div>