[OSM-dev] API very slow this morning
steve at asklater.com
Thu Mar 22 11:30:44 GMT 2007
I'll bet money it's the bad /traces/ SQL code. Sit on db and mytop -s 1
and you'll see all the tables locked due to that code. This is fixed in
rails and also why we should move to something that doesn't lock tables
Nick Hill wrote:
> Thanks for the pointer. makes sense, will do.
> Styno wrote:
>> In my programming experience (with C software, not Apache or Ruby), this
>> vague behavior can be caused by lack of other resources e.g. file
>> descriptors or sockets which does not show up on top. Leaking of these
>> resources is mainly caused when they aren't properly cleaned-up after use.
>> Maybe you can try commands like 'netstat' and 'lsof' next time...
>> Nick Hill wrote:
>>> Restarting apache on www, which runs the API in mod_ruby seems to have
>>> sped things up.
>>> This wasn't a memory or CPU hogging issue. It's as though apache
>>> processes simply become unresponsive.
>>> Nick Whitelegg wrote:
>>>> Hello Steve and Nick,
>>>> API seems very slow, particularly for 0900 GMT, which should be a
>>>> quiet time - took best part of a minute to download a 0.03 x 0.05
>>>> degree area. Is there an issue?
>>> dev mailing list
>>> dev at openstreetmap.org
> dev mailing list
> dev at openstreetmap.org
SteveC steve at asklater.com http://www.asklater.com/steve/
More information about the dev