[Asterisk-Dev] TDM04b failures (xpost on purpose)

[Rich Adamson]

System: RHv9 cvs-head-11/23/04 with TDM04B and T100P (isdn pri).

We've all seen many posts over the last several months regarding
the stability of the TDM card. Today, the above system's TDM04B 
card (4 fxo ports) failed at approx 15:54:49. 

[root at asterisk asterisk]# cat /proc/zaptel/2
Span 2: WCTDM/0 "Wildcard TDM400P REV E/F Board 1"

          25 WCTDM/0/0 FXSKS (In use)
          26 WCTDM/0/1 FXSKS (In use)
          27 WCTDM/0/2 FXSKS (In use)
          28 WCTDM/0/3 FXSKS (In use)
[root at asterisk asterisk]#

The /var/log/asterisk/messages log included the following entries:

Dec 30 13:07:05 WARNING[10297]: Can't change device with no technology!
Dec 30 13:07:05 WARNING[10297]: Can't change device with no technology!
Dec 30 15:54:49 NOTICE[10297]: Removed default indication country 'us'
Dec 30 15:54:49 WARNING[10297]: Ignoring switchtype
Dec 30 15:54:49 WARNING[10297]: Ignoring pridialplan
Dec 30 15:54:49 WARNING[10297]: Ignoring signalling
Dec 30 15:54:49 WARNING[10297]: Ignoring switchtype
Dec 30 15:54:49 WARNING[10297]: Ignoring signalling
Dec 30 15:54:49 WARNING[10297]: Ignoring switchtype
Dec 30 15:54:49 WARNING[10297]: Ignoring signalling
Dec 30 15:54:49 WARNING[10297]: Ignoring switchtype
Dec 30 15:54:49 WARNING[10297]: Ignoring signalling
Dec 30 15:54:49 WARNING[10297]: Ignoring switchtype
Dec 30 15:54:49 WARNING[10297]: Ignoring signalling
Dec 30 15:55:45 WARNING[10297]: No audio available on SIP/6111-e7fa??
Dec 30 15:55:45 WARNING[10297]: Can't change device with no technology!

The entries from 15:54:49 tend to suggest something caused the card to
restart. The log entries are the result of me being to lazy to remove:
 switchtype=national
from zapata.conf file for these four channels. But, they certainly 
indicate a symptom!!!

When the TDM card failed, the T100P (isdn pri) continued to function.
After a user report of the problem, I dialed two of the four pstn
lines attached to the TDM card. Received Ring-No-Answer and absolutely
no CLI indication that a call was coming in.

As a non-programmer, that would appear to be an issue likely related
to the TDM driver (wctdm) and/or hardware. (Asterisk had been running
for about two weeks since the drivers were last reload.)

I'm posting this to both -user and -dev in 'hopes' that maybe someone
would help identify the TDM issues. There have been several people that
have called Digium Support on this same issue over the last several
months, yet the problem continues. And, I think we've all seen the 
recommendations from many on the list not to purchase the TDM card 
due to its problems; that's certainly not in the best interest of 
Digium nor the Asterisk project overall.

(And, no its not a shared interrupt problem, etc, etc, etc; been
there, done that, tried about everything a non-programmer can try
since the TDM card was first sold.)

So, can we get some attention from somebody on this?  Maybe add some
code into wctdm that at least might generate some log hints as to
what might be happening? (Again, I'm a non-programmer but trying to
learn.)

Respectfully,

Rich