On 24 Jan 2001, Mark Longair wrote:
> It turned out that this was caused by using autofw to forward a range
> of ports (2300-2400 in this case.) It seems that these ports aren't
> reserved in any way, so eventually the server tries to use one as a
> local port on an outgoing connection.
>
> I'm
On 24 Jan 2001, Mark Longair wrote:
> Mark Longair <[EMAIL PROTECTED]> writes:
> [..]
> > I'm having a problem where twice a day or so, any new tcp connection
> > it gets stuck in SYN_SENT. Eventually this situation rights itself,
> > but obviously in the meantime many services (e.g. squid, X) a
Mark Longair <[EMAIL PROTECTED]> writes:
[..]
> I'm having a problem where twice a day or so, any new tcp connection
> it gets stuck in SYN_SENT. Eventually this situation rights itself,
> but obviously in the meantime many services (e.g. squid, X) are
> broken. The machine does IP masquerdading
On Fri, 12 Jan 2001, Mark Longair wrote:
> On Thursday 11 January, Richard B. Johnson wrote ("Re: [2.2.18] outgoing connections
>getting stuck in SYN_SENT"):
> [...]
> > You probably compiled your kernel with "CONFIG_INET_ECN" set.
> > If so,
On Thursday 11 January, Richard B. Johnson wrote ("Re: [2.2.18] outgoing connections
getting stuck in SYN_SENT"):
[...]
> You probably compiled your kernel with "CONFIG_INET_ECN" set.
> If so, you need to turn it OFF in /proc/sys/net/...something_ecn.
I don't
On Fri, 12 Jan 2001, Mark Longair wrote:
> I'm having a problem where twice a day or so, any new tcp connection
> it gets stuck in SYN_SENT. Eventually this situation rights itself,
> but obviously in the meantime many services (e.g. squid, X) are
> broken. The machine does IP masquerdading wit
I'm having a problem where twice a day or so, any new tcp connection
it gets stuck in SYN_SENT. Eventually this situation rights itself,
but obviously in the meantime many services (e.g. squid, X) are
broken. The machine does IP masquerdading with ipchains, and
masqueraded connections through it
7 matches
Mail list logo
