On Thu, Feb 17, 2000 at 03:41:28PM +0900, kumon@flab.fujitsu.co.jp wrote:
> The problem exist AFTER "nic.funet.fi".
> nic.funet.fi -> OUR DOMAIN
> "nic.funet.fi" tries to send mail to our domain but massive resends
> occur in this transfer for over 10 hours as follows:
As I have my fingers at nic.funet.fi's MTA, I am interested
in this case..
> nic.funet.fi-> fgwmail.fujitsu.co.jp Mon, 14 Feb 2000 13:55:27 +0900
> nic.funet.fi-> fgwnews.fujitsu.co.jp Mon, 14 Feb 2000 13:56:03 +0900
> nic.funet.fi-> fgwmail6.fujitsu.co.jp Mon, 14 Feb 2000 13:56:18 +0900
This rapid skip-around doesn't make sense; NIC sends things
strictly serialized.
(Digging syslogs.. Oh..)
Feb 14 06:55:58 nic smtp[20906]: S3799AbQBNEyA: to=<kumon@flab.fujitsu.co.jp>,
delay=00:01:58, xdelay=00:00:33, mailer=smtp,
relay=fgwmail.fujitsu.co.jp ([164.71.1.133|25|193.166.0.145|57913]),
stat=deferred smtp; 500 (Error on SMTP response read: Connection reset
by peer, Cmd: .) cvt=NONE
A few tens, if not hundreds of those.
(hmm.. 0-5-12 per hour per troubled gw at syslogs..)
And it seems to re-occur:
Feb 16 07:13:10 nic smtp[29593]: S18933AbQBPFDj: to=<kumon@flab.fujitsu.co.jp>,
delay=00:09:31, xdelay=00:00:03, mailer=smtp,
relay=fgwmail.fujitsu.co.jp ([164.71.1.133|25|193.166.0.145|47112]),
stat=deferred smtp; 500 (Error on SMTP response read: Connection reset
by peer, Cmd: .) cvt=NONE
Especially fgwmail and fgwnews seem to exhibit this
behaviour, fgwmail5 and fgwmail6 don't show up at "bad boys"
statistics.
289 fgwnews.all
54 fgwnews.not-ok
730 fgwmail.all
116 fgwmail.not-ok
84 fgwmail5.all
0 fgwmail5.not-ok
73 fgwmail6.all
0 fgwmail6.not-ok
This covers from Feb 13 02:48:48 (UTC +0200) until snapshot
moment (Feb 17, 13:11, UTC +0200).
I would say fwgmail is having SERIOUS problems, and fgwnews
nearly as bad.
Out of memory (RAM+SWAP) ? Out disk-space ?
I don't know what mechanisms can cause sendmail to groak.
...
> nic.funet.fi-> fgwmail.fujitsu.co.jp Tue, 15 Feb 2000 03:32:44 +0900
> Resend occurs 241 times in the above case.
Hmm.. Only about 170 instances covering a lot longer timespan..
> As described above, I suspected "nic.funet.fi" behavior.
Running http://www.zmailer.org/mxverify-cgi.html with
your address doesn't reveal *currently* any problems.
And for that matter, it never would reveal problems, where
DATA+dot gets accepted, but not acknowledged.
> --
> Computer Systems Laboratory, Fujitsu Labs.
> kumon@flab.fujitsu.co.jp
/Matti Aarnio <matti.aarnio@sonera.fi> <mea@nic.funet.fi>
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Feb 23 2000 - 21:00:18 EST