Persistent pop3s connections

[Alain Spineux]

I remember lot of post on cyrus-imap mailing list about this kind of problem.
There was a problem in the SSL connection establishment.
Someone wrote a patch.
Search the archive.

Regards.

On Feb 6, 2008 5:52 PM, Simon Barrett <simon.barrett at jinny.ie> wrote:
>
> Hello,
>
> Five days ago I moved our kolab install to a new machine (CentOS 4.6) because
> of system instability on the previous hardware.  I performed a native install
> on the new machine of the same version of Kolab (2.0.4), then followed the
> recommendations in
>
> http://wiki.kolab.org/index.php/Backups_for_kolab2
>
> to migrate the users.
> Everything is running well since the change, but today some users reported
> timeouts, both on the local lan and from remote.
>
> Among other steps I checked netstat to see what the current usage was,
>
> # netstat -ant | more
>
> and noticed that one external IP had an unual number of pop3s connections
> open, status 'ESTABLISHED'.
>
> # netstat -ant | grep 12.34.56.78 | wc -l
> 169
>
> I eventually blocked this IP on the firewall:
>
> # iptables -L | head
> Chain INPUT (policy DROP)
> target     prot opt source               destination
> DROP       all  --  12345678.example.com  anywhere
>
> but this has not made any difference to the number of established connections.
>
> I checked for authentication information in /kolab/var/imapd/log/pop3d.log and
> it shows that one of our users was connecting from that IP up until about 33
> hours ago.  It is his home DSL account.  He only has a laptop, which is
> sitting here in the office, and he does not use any other PC to collect mail.
> The PC has a different IP on the office lan (obviously), he is not connected
> remotely to his home lan and his laptop is not causing the same trouble from
> this side of the firewall.
>
> I got the pid for some of these external connections from
> # lsof -i
>
> and killed them off.  New processes were spawned in their place.  I've been
> keeping an eye on the number of connections and it seems to vary between 163
> and 170.
>
> I would guess that I'm missing something basic, but I'd welcome any insights
> as to why these connections are being maintained, what could cause the issue
> in the first place, whether this could be the source of timeouts for other
> users, and whether I should be concerned.  I plan on restarting the daemon
> after hours tonight.
>
>
> Thanks and regards,
>
> Simon Barrett
>
> _______________________________________________
> Kolab-users mailing list
> Kolab-users at kolab.org
> https://kolab.org/mailman/listinfo/kolab-users
>



-- 
Alain Spineux
aspineux gmail com
May the sources be with you