[Asterisk-Users] modprobe -r ztd-eth locks up machine...

[john]

>Did you ifdown the dynamic interfaces first ?
>
>Martin

Yes, this still results in a crash on the box with the tor2.

Is there any 'controlled' way to bring down a dynamic span?
... just for fun here is zttool (spans 3 & 4 are unused)...
Alarms          Span
OK              Tormenta 2 (PCI) Quad T1 Card 0 Span 1
OK              Tormenta 2 (PCI) Quad T1 Card 0 Span 2
RED             Tormenta 2 (PCI) Quad T1 Card 0 Span 3
RED             Tormenta 2 (PCI) Quad T1 Card 0 Span 4
OK              Dynamic 'eth' span at 'eth0/00:0A:5E:05

... and from zaptel.conf...
span=1,2,0,esf,b8zs
span=2,1,0,esf,b8zs
dynamic=eth,eth0/00:0A:5E:05:7E:89,24,0



> On Wed, 2003-12-17 at 10:36, john wrote:
> > Hi,
> >
> > I have just begun working with TDMoE running between 2 fiber nics the
> > dynamic span works great. In my main asterisk box's startup file I just
> > 'modprobe tor2', then start asterisk. The zaptel, ztdynamic & ztd-eth
> > modules all load by themselves when tor2 is loaded. If I stop asterisk
then
> > 'modprobe -r tor2' the  tor2 module is removed but the other three
remain.
> > If I then 'modprobe -r ztd-eth' it causes a complete lock up on the
machine.
> > The remote machine does not have any zap hardware in it yet and doesn't
have
> > these difficulties.
> >
> > I know I can just restart the machine but it is in a production
environment
> > (soon to increase from a few to ~30 simultaneous calls) and it is nice
to be
> > able to make changes and
> > cvs update installs without restarting.
> >
> > Has anyone experienced this or am I just missing a step or going in the
> > wrong order?
>
> Unloading of modules was of such a concern that it almost didn't make it
> into newer kernels. So you should probably not unload them. A production
> machine should have specified service windows available. Also decent
> hardware should be able to reboot fairly fast. The machine I have as our
> local asterisk machine can go from reset button to accepting new calls
> in under 50 seconds. Our remote machine is around 90 secs. Depending on
> y our call volume, and system setup, you should be able to handle this.
>
> --
> Steven Critchfield  <critch at basesys.com>
>

This e-mail was scanned and found clean by Monroe-Woodbury's Antivirus.