TCP timeout

[Nikos Vassiliadis]

On Tuesday 04 July 2006 07:01, scuba at centroin.com.br wrote:
> Hi folks,
>
>  	I'm facing some troubles with mail server on a busy slow link.
>
>  	I'm getting a lot of "sendmail: SYSERR: collect: read timeout on
> connection from" entries in the logs.
>
>  	I found this page showing an workaround for Solaris. Is it
> applicable  for Free BSD 5.4? If positive, is there any sysctl variable to
> tun it?
>
> -----------------xxxxxx-----------------
>
> TCP/IP connections time out too soon, especially on slow links.
>
> The tcp/ip abort interval in Solaris 2.x is too short, the default value
> is 2 minutes. The result is that when an ACK isn't received in 2 minutes,
> the connection is closed. This is most often seen by sendmail, which will
> log
>
> sendmail: SYSERR: collect: read timeout on connection from ...
>
> You can fix this by running following command which increases the timeout
> to 8 minutes (unit is millisec), which is the Solaris 2.4+ (and patched
> 2.3) default.
>
> /usr/sbin/ndd -set /dev/tcp tcp_ip_abort_interval 480000
>
> -----------------xxxxxx-----------------
>

Are you sure that's the case? Isn't Solaris 2 ancient(circa 1992)?
that means that internet was very different...
It doesn't seem normal these days for a server not responding for
75 seconds, which is the default timeout value for not established
connections.

anyway
sysctl net.inet.tcp.keepinit to see current value
sysctl net.inet.tcp.keepinit=timeout_in_ms to change it
and /etc/sysctl.conf to make it permanent.

HTH, Nikos