From: Man-wai Chang to The Door (24000bps) on
> I see now why they didn't show before. They are all used when receiving
> incoming email. Apparently the last time I checked none was coming in.
> Here is a sockstat that shows 615 and 747 in use: (no idea why they
> seem to be different ports each time)
>

Google "tcp port 615 747".

Does the link http://www.emsisoft.com/en/kb/portlist/ help?


--
@~@ Might, Courage, Vision, SINCERITY.
/ v \ Simplicity is Beauty! May the Force and Farce be with you!
/( _ )\ (x86_64 Ubuntu 9.10) Linux 2.6.32.8
^ ^ 20:21:01 up 3:50 2 users load average: 1.28 1.28 1.20
不借貸! 不詐騙! 不援交! 不打交! 不打劫! 不自殺! 請考慮綜援 (CSSA):
http://www.swd.gov.hk/tc/index/site_pubsvc/page_socsecu/sub_addressesa
From: Doug Hardie on
In article <4b83c8ac$0$283$14726298(a)news.sunsite.dk>,
"Man-wai Chang to The Door (24000bps)" <toylet.toylet(a)gmail.com>
wrote:

> > I see now why they didn't show before. They are all used when receiving
> > incoming email. Apparently the last time I checked none was coming in.
> > Here is a sockstat that shows 615 and 747 in use: (no idea why they
> > seem to be different ports each time)
> >
>
> Google "tcp port 615 747".
>
> Does the link http://www.emsisoft.com/en/kb/portlist/ help?


Only a couple are listed: Kerberos Administration, Fujitsu Device
Control etc. I don't use Kerberos or any Fujitsu drives.
From: ska on
Doug Hardie wrote:
> > > Here is a sockstat that shows 615 and 747 in use: (no idea why they
> > > seem to be different ports each time)

Well, "different ports each time" may suggest that these are
dynamically allocated for some transport, e.g. DNS, ... .
Do you have any milter running?

Can you trace a sendmail process through exec()? Maybe the time of
creation of the sockets or the traffic help.

-ska
From: Doug Hardie on
In article
<0532637f-f366-4e32-b2a0-d5bd85578fa9(a)x22g2000yqx.googlegroups.com>,
ska <skg(a)mail.inf.fh-brs.de> wrote:

> Doug Hardie wrote:
> > > > Here is a sockstat that shows 615 and 747 in use: (no idea why they
> > > > seem to be different ports each time)
>
> Well, "different ports each time" may suggest that these are
> dynamically allocated for some transport, e.g. DNS, ... .
> Do you have any milter running?
>
> Can you trace a sendmail process through exec()? Maybe the time of
> creation of the sockets or the traffic help.
>
> -ska

Good call. In the trace I found that sendmail and all the milters and
even mail.local were creating those sockets. I noticed that all of them
were sending what looked like NIS packets. I found in one program that
the only system call between two obvious trace entries where it occurred
was a call to getpwnam. A very simple test program with only a call to
getpwnam yields exactly the same issue with a weird UDP socket left
open. So this is an OS issue and not sendmail's problem. Thanks for
the help.