On 7/15/05, Bruno Wolff III <[EMAIL PROTECTED]> wrote:
> On Thu, Jul 14, 2005 at 16:29:58 -0600,
> Dan Harris <[EMAIL PROTECTED]> wrote:
> >
> > Ok, I tried this one. My ssh keeps getting cut off by a router
> > somewhere between me and the server due to inactivity timeouts, so
> > all I know is that both the select and explain analyze are taking
> > over an hour to run. Here's the explain select for that one, since
> > that's the best I can get.
>
> Are you using NAT at home? That's probably where the issue is. If you
> have control of that box you can probably increase the timeout to a
> couple of hours.
Some versions of ssh have such a configuration option (in .ssh/config):
Host *
ServerAliveInterval 600
...it means that ssh will send a "ping" packet to a sshd every 10 minutes
of inactivity. This way NAT will see activity and won't kill the session.
I'm using OpenSSH_4.1p1 for this...
Oh, and it doesn't have anything to do with TCP keep alive, which is
rather for finding dead connections than keeping connections alive. ;)
Regards,
Dawid
---------------------------(end of broadcast)---------------------------
TIP 4: Have you searched our list archives?
http://archives.postgresql.org
