On Mar 27, 2009, at 2:36 AM, Simon Riggs wrote:
> Not really. I want to understand the actual problem with
> idle-in-transaction so we can consider all ways to solve it, rather
> than
> just focus on one method.
I have to distinct problems with idle in transaction. One is
reporting users / the tools they're using. I'll often find
transactions that have been open for minutes or hours. But, that's
not a big deal for me, because that's only impacting londiste slaves,
and I have no problem just killing those backends.
What does concern me is seeing idle in transaction from our web
servers that lasts anything more than a few fractions of a second.
Those cases worry me because I have to wonder if that's happening due
to bad code. Right now I can't think of any way to figure out if
that's the case other than a lot of complex logfile processing
(assuming that would even work). But if I knew what the previous
query was, I'd at least have half a chance to know what portion of
the code was responsible, and could then look at the code to see if
the idle state was expected or not.
--
Decibel!, aka Jim C. Nasby, Database Architect decibel@decibel.org
Give your computer some brain candy! www.distributed.net Team #1828
