From: David Luyer <>
Date: Thu, 20 Nov 1997 11:28:46 +0800

> Greetings,
> I am running Squid 1.1.17 on Linux 2.0.30.
> 99% of the time, the service runs perfectly. However, about two or three
> times a day, Squid loses contact with its parent for anything between 30
> seconds and 2 minutes. People browsing get cache hits perfectly, but any
> attempt to retrieve an uncached document results in an ERR-CANNOT-FETCH,
> and the user is told that there are "No peers to query, and the host is
> beyond the firewall".
> This occurs for no obvious reason. And it goes away again without any human
> intervention.
> A second machine accesses the same firewall. It runs Squid 1.1.16 on Linux
> 2.0.0. The problem never occurs on the second machine.

Either use the TCP connection refused patch I posted to the list a while back
or upgrade the 2.0.30 to 2.0.31 or 2.0.32. There's a really annoying bug in
the late 2.0.x's fixed in 2.0.31 that causes 'fake' connection failures
to bind to addr 0). Anyone who runs Linux should REALLY upgrade to 2.0.32
anyway to avoid the 'teardrop' problem.

Received on Wed Nov 19 1997 - 19:34:34 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:37:41 MST