Horton

Re: Stats collector performance improvement

Hannu Krosing <hannu@skype.net>

View thread
From: Hannu Krosing <hannu@skype.net>
To: Greg Stark <gsstark@mit.edu>
Cc: Simon Riggs <simon@2ndquadrant.com>, Tom Lane <tgl@sss.pgh.pa.us>, Qingqing Zhou <zhouqq@cs.toronto.edu>, pgsql-hackers@postgresql.org
Date: 2006-01-09T15:48:21Z
Lists: pgsql-hackers, pgsql-performance 
Ühel kenal päeval, P, 2006-01-08 kell 11:49, kirjutas Greg Stark:
> Hannu Krosing <hannu@skype.net> writes:
> 
> > Interestingly I use pg_stat_activity view to watch for stuck backends,
> > "stuck" in the sense that they have not noticed when client want away
> > and are now waitin the TCP timeout to happen. I query for backends which
> > have been in "<IDLE>" state for longer than XX seconds. I guess that at
> > least some kind of indication for this should be available.
> 
> You mean like the tcp_keepalives_idle option?
> 
> http://www.postgresql.org/docs/8.1/interactive/runtime-config-connection.html#GUC-TCP-KEEPALIVES-IDLE
> 

Kind of, only I'd like to be able to set timeouts less than 120 minutes.

from:
http://developer.apple.com/documentation/mac/NetworkingOT/NetworkingWOT-390.html#HEADING390-0

kp_timeout
        Set the requested timeout value, in minutes. Specify a value of
        T_UNSPEC to use the default value. You may specify any positive
        value for this field of 120 minutes or greater. The timeout
        value is not an absolute requirement; if you specify a value
        less than 120 minutes, TCP will renegotiate a timeout of 120
        minutes.
        
-----------
Hannu