[Asterisk-Users] Problems starting Asterisk with TDM22B

[Dave Cotton]

On Mon, 2004-11-29 at 21:54 -0600, Mitchell S. Sharp wrote:

> You're having the same problem I had when I upgraded to Fedora Core 3
> (with the 2.6 kernel).  I have a TDM400P with 3 FXS modules.  I had to
> add some sleep statements between modules loads for some reason.  Try it
> out.  This is my edited init script:
> 
> [ clipped from /etc/rc.d/init.d/zaptel ]
> 
> MODULES="wcfxo wctdm"
> 
> RMODULES="wctdm wcfxo"
> 
> if [ "${DEBUG}" = "yes" ]; then
>         ARGS="debug=1"
> fi
> 
> # See how we were called.
> case "$1" in
>   start)
>         # Load drivers
>         rmmod wcusb >& /dev/null
>         rmmod wcfxsusb >& /dev/null
>         rmmod audio >& /dev/null
>         action "Loading zaptel framework: " modprobe zaptel
>         sleep 1
>         echo "Loading zaptel hardware modules: "
>         for x in $MODULES; do
>                 echo -n "     $x "
>                 modprobe ${x} >& /dev/null
>                 echo " - Sleep 1"
>                 sleep 1
>         done
>         echo "     Sleep 3"
>         sleep 3
>         echo
>         action "     Running ztcfg: " /sbin/ztcfg
>         RETVAL=$?
> 
>         [ $RETVAL -eq 0 ] && touch /var/lock/subsys/zaptel
> 
>         ;;
>   stop)
> 
> [ end clip ]

Looks to me like udev again. All start up scripts seem to be based on
the, at the time, correct assumption that all devices would be ready and
waiting nicely in /dev, udev has completely turned this upside down.

The Mandrake Cooker list has been discussing this recently. It would
seem that the answer is that the script should be altered to check if
the device is available before going on to the next stage. The downside
of this is that people will complain at the length of time their machine
takes to boot.

In the case of * is it really a problem? Surely a PBX is started once
and then left running 24/7.


-- 
Dave Cotton <dcotton at linuxautrement.com>