Hi there!
Thanks very much for the help. With Exim 3.20 it was indeed similar to the
sendmail timeout Gary wrote about.
It turned out that ident requests to the server attempting to send mail
were being firewalled, with the result that they timed out at 30s instead
of failing immediately with 'connection refused' or returning a response.
As we don't use ident information for anything, I disabled it with
rfc1413_query_timeout = 0s
A far better solution would be of course to set up the firewall to return an
ICMP port unreachable for accesses to the ident port.
Stops ident delays for any protocol (and a fair few use them).
On 11 Jan 01, at 18:37, gary wrote:
> I do have this problem b4 with my RedHat 6.2 and sendmail-8.9.3-20.
> is really took 30 sec for connection...
> But I dunno whether it is helpful for u???
>
> Anyway, what I did is just change parameter in sendmail.cf as below:
> from
> O Timeout.ident=30s
> to
> O Timeout.ident=0s
>
> It solved my problem, hope it will work for u too, :-)...
>
> rdgs,
> gary
>
> ----- Original Message -----
> From: <[EMAIL PROTECTED]>
> To: <[EMAIL PROTECTED]>
> Cc: "Holger Greindl" <[EMAIL PROTECTED]>
> Sent: Wednesday, January 10, 2001 7:39 PM
> Subject: smtp-connect delays about 30 seconds
>
>
> > Hi,
> >
> > sorry if this kind of problem has been posted before, but I didn't know
> > right how to find the right solution and it is urgent.
> >
> > I have RH6.2 running as mailserver with Exim 3.20 as MTA. The system is
> located
> > in front of a firewall which translates our internal IP-addresses into one
> official one.
> > The mailserves has also an official IP and uses the DNS server of our
> provider for
> > address-resolution.
> >
> > The problem: Everytime a user wants to send mail it takes about 30 seconds
> to get
> > a smtp-connection from the the (NT-)workstations to the mailserver.
> According to
> > the Exim log-file there is no problem but when I run tcpdump on the
> machine
> > I found a strange delay and I not even know what happens there :-(((
> >
> > Further down comes the output of tcpdump - apollo is the mailserver, fwint
> the firewall.
> > I'd appreciate any help. I am really blank how I shpuld get rid of this
> delay and
> > the people getting angry...
> >
> > Best regards,
> > Michael
> >
> > tcpdump: listening on eth0
> > 14:37:11.263039 FwInt.klumpp.de.61333 > apollo.klumpp.de.smtp:
> > S1417123735:1417123735(0) win 32120 <mss 1460,sackOK,timestamp
> > 86409152[|tcp]> (DF) (ttl 63, id 8916)
> > 14:37:11.263136 apollo.klumpp.de.smtp > FwInt.klumpp.de.61333:
> > S2295099218:2295099218(0) ack 1417123736 win 30660 <mss
> > 1460,nop,nop,sackOK,nop,wscale 0> (DF) (ttl 64, id 30710)
> > 14:37:11.264309 FwInt.klumpp.de.61333 > apollo.klumpp.de.smtp:
> > . ack 1 win 32120 (DF) (ttl 63, id 8917)
> > 14:37:16.258483 arp who-has apollo.klumpp.de tell FwInt.klumpp.de
> > 14:37:16.258526 arp reply apollo.klumpp.de is-at 0:0:1c:b5:b3:bc
> > 14:37:41.265312 apollo.klumpp.de.1296 > dns1.riodata.de.domain:
> > 18237+ (47) (ttl 64, id 30719)
> > 14:37:41.268503 apollo.klumpp.de.1297 > dns1.riodata.de.domain:
> > 18497+ (43) (ttl 64, id 30720)
> > 14:37:41.424769 dns1.riodata.de.domain > apollo.klumpp.de.1298:
> > 18239 NXDomain* q: 51.146.16.62. 0/1/0 (130) (ttl 24, id38492)
> > 14:37:41.425528 apollo.klumpp.de.smtp > FwInt.klumpp.de.61333:
> > P1:74(73) ack 1 win 32120 (DF) (ttl 64, id 30723)
> > 14:37:41.426845 FwInt.klumpp.de.61333 > apollo.klumpp.de.smtp:
> > . ack 74 win 32120 (DF) (ttl 63, id 8918)
> >
> >
> >
> > _______________________________________________
> > Redhat-list mailing list
> > [EMAIL PROTECTED]
> > https://listman.redhat.com/mailman/listinfo/redhat-list
>
>
>
> _______________________________________________
> Redhat-list mailing list
> [EMAIL PROTECTED]
> https://listman.redhat.com/mailman/listinfo/redhat-list
_______________________________________________
Redhat-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-list
