Samba 4.3.1 and messaging

[Volker Lendecke]

On Mon, Oct 26, 2015 at 02:36:00PM +0100, Thomas Bork wrote:
> Hi,
> 
> new version, new questions ;-)
> 
> Every restart of smbd the number of files in /etc/msg.sock/ and
> /run/lock/samba/msg.lock/ is growing. Is this normal?
> 
> testeis # /usr/bin/killall -SIGTERM smbd
> smbd: no process found
> testeis # rm -f /etc/msg.sock/?*
> testeis # rm -f /run/lock/samba/msg.lock/?*
> testeis # for i in `seq 1 10`;
> > do
> >     echo "round $i:"
> >     smbd -D
> >     sleep 3
> >     /usr/bin/killall -SIGTERM smbd
> >     sleep 3
> >     ls /etc/msg.sock/?* | wc -l
> >     ls /run/lock/samba/msg.lock/?* | wc -l
> > done

Is that your normal way to operate Samba? I did not see this scenario. I
have code that will convert this to a different messaging model, which
should then also cover restarting Samba every 3 seconds without spilling
files. In the current incarnation of Samba this use case WILL spill files,
I think it will take a few months before I can present the fixed tmond
streams based code. You should probably go back to 4.0 or so to make
Samba work better in this test.

Normally Samba cleans up every 15 minutes. If you restart Samba every 3
seconds, this will never kick in. Please set

messaging : messaging dgm cleanup interval = 1

in your [global] section to make Samba wipe old sockets every second to
cover your, admittedly a bit strange, use case.

With best regards,

Volker Lendecke

-- 
SerNet GmbH, Bahnhofsallee 1b, 37081 Göttingen
phone: +49-551-370000-0, fax: +49-551-370000-9
AG Göttingen, HRB 2816, GF: Dr. Johannes Loxen
http://www.sernet.de, mailto:kontakt at sernet.de

Besuchen Sie uns vom 10.-11.11.15 auf der ISSE!
Information Security Solutions Europe Conference
Hotel Palace Berlin, 20%-Rabattcode: "ISSE15SP"

Meet us at Information Security Conference ISSE!
November 10th - 11th 2015 in Hotel Palace Berlin
For 20% discount take voucher code:  "ISSE15SP"