On Tuesday 19 August 2003 11.41, Vladimir wrote: > 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
> 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 Try using a lower peer_connect_timeout or masquerade of outgoing locally generated traffic. Most likely your dialup is assigned a dynamic IP address disturbing the probes initiated by Squid. > If there is less requests, my expirience is that > period to detect revived parent can be few times > longer. If there is no requests then Squid does not check. As soon as there is a request Squid checks if the parent is alive. Which Squid version is this? Regards Henrik
