Re: Postfix 20010228 on OpenBSD 2.9 could loose mail ?

[wietse@porcupine.org (Wietse Venema)]

In recent Postfix snapshots, the trigger client will connect()
write() then wait before close().

The problem that you are referring to just screws up the flush
service, but does not result in lost mail.

If am depressed by gratuitous changes in socket semantics from
"graceful close" (do not lose data that the other side hasn't
received yet) into "ungraceful close".

	Wietse

R?mi Guyomarch:
> On Wed, Mar 07, 2001 at 04:20:35PM -0500, Wietse Venema wrote:
> (full message here: http://marc.theaimsgroup.com/?l=postfix-users&m=98400011427323&w=2)
> 
> > Several parts of Postfix do: connect() write() close(), where the
> > close() may happen before the server has accept()ed the connection.
> > 
> > Due to an incompatible change in FreeBSD 4.2-STABLE, this causes
> > accept() after close() to fail. The already written data is lost.
> > [...]
> 
> I'm experiencing the same problem with postfix-20010228.pl02 on
> OpenBSD-2.9. I've filled a bug-report and provided a kernel patch :
> http://cvs.openbsd.org/cgi-bin/wwwgnats.pl/full?pr=1840
> 
> In short, if I clearly understand the issue, Postfix 20010228 on
> OpenBSD 2.9 *MAY* loose mail due to a kernel bug. Am I right ? I hope
> I'm not ... :-(
> 
> -- 
> R?mi
> 
> -
> To unsubscribe, send mail to majordomo@postfix.org with content
> (not subject): unsubscribe postfix-users
> 
>