Problems with ISDN PRI Lines

Hi all,

I have a TE410P in an ibm xSeries 345 with two Intel Pentium Xeon 2.8 and 2 GB RAM. I have Debian with kernel 2.4.27 and Asterisk 1.0.10. I have connected one ISDN E1 Line fiber to First Port, one ISDN E1 Radio Line connected to Second port and ISDN PTP Line with an AVAYA Definity in de 4th Port.

We have about 200 users with SwissVoice IP10 and Polycom IP500. Switches are Cisco Catalyst with two VLAN per Port, one for Voice and one ofr the rest.

This equipment has been in production for a Year, but from some weeks we have next problems:

Sometimes Users can’t make call to out through TE410P but they can between SIP Users. If we stop asterisk, unload modules, load again, and start asterisk, I solve the problem.

Sometimes Users can’t make calls to out and in. Restart Asterisk could solve the problem.

I have seen too much messages like

"Stopping retransmission… for request 102"
“Echo canellation already on”
“Dropping duplicate answer”
“Didn’t get a frame from channel…”
“Driver for channel … doesn support indication 3, emulating it”

I’m crazy with this problem and any help could be much appreciated.

but from some weeks we have next problems

What was exactly changed ago a few weeks ?

My customer says that they haven’t made any change.

I would like to know if this kind of problem could be produced by changes in Switches or by bad Card(But I don’t think because we have two machines in this site with linux-ha, and tow machines reports same problems), or by a big amount of calls, or by a misconfiguration.

I updated asterisk from 1.0.3 to 1.0.10 when the problems start but this will not better the problem. And the problem is being more frequent in the time.

I need any help to know how can I test where is the problem. My customer are a little estrict and let me do few things out of the machien.


[quote=“srsergio”]My customer says that they haven’t made any change.

Yep, no1 ever does: Its all the nightelfes and little dwarves doing that on their very own :unamused: :unamused:

Well, that is a NASTY problem, cuz it happens only under load, no “sheme” in it, not reproducable for sure (when needed).
So you have no other chance then firing up the extensive debug functions and reading LONG LONG logs when this happens.

But at first, you want to update to - there are MANY bugs fixed to the version you are using.
Read the UPDATE.TXT notes, there were changes making it nessecary to check the diaplans and the “old” command format you are using now.

I would like to know if this kind of problem could be produced by >>>changes in Switches or by bad Card

Sure, everything is possible - it could be even the new elevator in the house, where the powerline is close to the telephoneline of this company with a bad shielding… :open_mouth: (i had this very situation with a customer)…we/you need logs !