Apparently, once instance of this data corruption problem is caused
by an unnamed bandwidth management system.  It runs as a bridge,
and does not show up in traceroute etc. output.  We were able to
estimate its location (at 5 ms round-trip time from one endpoint)
by analyzing packet arrival times.

Until now, this TCP data corruption problem has been observed only
when one of the connection endpoints runs a recent LINUX version.
Sightings have been reported by sites in Germany and in France.

Only recent LINUX versions request the use of timestamp options
that cause the tell-tale patterns of "01 01 08 0a" in TCP packets,
and that end up being regurgitated as ^A^A^H^J data.

I have updated the analysis at ftp://ftp.porcupine.org/pub/debugging/

        Wietse

Wietse Venema:
> This note is about a subtle data corruption problem with TCP data
> streams that may bite people as more and more (LINUX) systems are
> sending network traffic with TCP-level options turned on.
>
> Last week, several Postfix users reported mail delivery failures
> because sequences of control characters (for example, ^A^A^H) were
> being inserted into their SMTP connections, resulting in SMTP
> protocol errors and non-delivery of email.
>
> These data corruption problems are not host specific: they are
> observed with both Linux and BSD/OS systems, and with mail sent to
> and/or received from systems running Postfix, Sendmail and qmail.
>
> Over the weekend of March 18, 2000, a few people left tcpdump
> running on their machines, in order to record some of these corrupted
> SMTP sessions.  This note is based on an analysis of that data.

Reply via email to