Re: [squid-users] How to disable TCP dead detection?

Hi Henrik,

Well, i notice something different behaviour. Here is
cache.log extract:

-----------------------------------------------
2003/08/19 17:16:54| TCP connection to 127.0.0.1/8000
failed
2003/08/19 17:16:54| Detected DEAD Parent:
127.0.0.1/8000/0
2003/08/19 17:17:27| Failed to select source for
'http://www.b92.net/doc/aboutu'
2003/08/19 17:17:27| always_direct = 0
2003/08/19 17:17:27| never_direct = 1
2003/08/19 17:17:27| timedout = 0
2003/08/19 17:17:52| Failed to select source for
'http://www.b92.net/doc/aboutu'
2003/08/19 17:17:52| always_direct = 0
2003/08/19 17:17:52| never_direct = 1
2003/08/19 17:17:52| timedout = 0
2003/08/19 17:17:59| Failed to select source for
'http://www.b92.net/doc/aboutu'
2003/08/19 17:17:59| always_direct = 0
2003/08/19 17:17:59| never_direct = 1
2003/08/19 17:17:59| timedout = 0
2003/08/19 17:18:06| Failed to select source for
'http://www.danas.co.yu/'
2003/08/19 17:18:06| always_direct = 0
2003/08/19 17:18:06| never_direct = 1
2003/08/19 17:18:06| timedout = 0
2003/08/19 17:18:30| Failed to select source for
'http://www.cisco.com/'
2003/08/19 17:18:30| always_direct = 0
2003/08/19 17:18:30| never_direct = 1
2003/08/19 17:18:30| 08/19 17:18:30| Detected
REVIVED Parent: 127.0.0.1/8000

-------------------------------------------------

So, DEAD parent has been detected at 17:16:54, and
shortly after that parent was reachable. However,
there were 5-6 requests starting from 17:17:27 which
were not server, although parent was alive. We have
REVIVED message after one minute - at 17:18:30.

If there is less requests, my expirience is that
period to detect revived parent can be few times
longer.

Do you know any way do change this behaviour (to speed
up detecting of revived parent cache)?

Thanks
Vladimir

--- Henrik Nordstrom <hno@squid-cache.org> wrote:
> On Friday 15 August 2003 13.12, Vladimir wrote:
>
> > When there is delay in opening of ISDN line and
> ssh
> > channel, Squid can not reach parent proxy on
> address
> > localhost:8000 and it declares it as dead ( I
> think
> > this is famous TCP DEAD feature of Squid). After
> few
> > seconds, line is up, parent is reachable, but in
> the
> > browser I still get error message that all parent
> > proxies are down and request can not be forwarded.
> > But, parent proxy is alive and kicking. Only way
> to
> > fix it, is to wait something like 10 minutes or
> so,
>
> You should not need to wait at all. If you use
> never_direct then dead
> peers are supposed to be tried on each and every
> request as
> last-resort path.
>
> > My question is this: how to completely disable
> this
> > TCP dead feature, and force Squid to forward EACH
> > request to proxy no matter if it is dead or not,
> > because in my case this feature is not usefull at
> all,
> > and just makes troubles.
>
> It is supposed to do exacly just that in
> never_direct configurations.
>
> Regards
> Henrik
>
> --
> Donations welcome if you consider my Free Squid
> support helpful.
>
https://www.paypal.com/xclick/business=hno%40squid-cache.org
>
> If you need commercial Squid support or cost
> effective Squid or
> firewall appliances please refer to MARA Systems AB,
> Sweden
> http://www.marasystems.com/, info@marasystems.com

__________________________________
Do you Yahoo!?
Yahoo! SiteBuilder - Free, easy-to-use web site design software
http://sitebuilder.yahoo.com
Received on Tue Aug 19 2003 - 03:41:38 MDT