[App_rpt] TDM400

[Steve Rodgers]

Eric,

The segfault thing has to be something else in your asterisk configuration. 
The problem we experienced with the TDM400 was not a segfault issue. In a 
nutshell the problem is this:

On-hook and off-hook voltage transients cause the TDM400 hardware to get into 
a strange state and then the whole system (including Linux) crashes or 
freezes. Not every card will do this though. Node 2060 (KE6UPI) on the 
allstar link uses a TDM400 and it seems to be working fine. 

If you asterisk configuration is segfaulting, then this has be something else 
in your asterisk configuration. (mixed versions of asterisk modules maybe?) 
follow the instructions posted on bugs.digium.com, or voip-info.org and 
prepare a backtrace file with GDB on the coredump. If you look at the 
backtrace file, you should be able to see exactly what was happening just 
before the segfault occurred. This will give you a clue as to what happened. 




Steve, WA6ZFT


On Thursday 02 March 2006 21:40, Eric Tichansky wrote:
> I finally acquired all the RF components for my new 2M/440 crossband
> repeater, using app_rpt of course!
>
> Anyways, just wanted to get a better idea and perhaps more details
> concerning the problems surrounding the use of the TDM400 series cards
> with the Analog Radio Interface Board (ARIB).
>
> It seems to work fine as long as the system is active.  Once idle for an
> extended period of time, an almost certain segfault results.  Using the
> safe_asterisk script to respawn asterisk after a crash works OK to
> recover for continued RF communications, but is useless when an incoming
> phone patch call triggers the segfault -- the call is dumped and just
> simply annoying.
>
> Just wondering what debugging has been done and conclusions as to what
> aspect of the system seems to be the culprit (TDM 400 hardware,
> Asterisk, app_rpt)?  Are there any other Zaptel compatible FXS cards
> available for a modest price?