Summary:ASTERISK-05145: libpri clocking errors
Reporter:Robert Christian (thetatag)Labels:
Date Opened:2005-09-25 12:54:21Date Closed:2011-06-07 14:02:51
Versions:Frequency of
Description:I have noticed that on my PRI connection (ESF/B8ZS, dms100) there are constantly slipping frames.  When clocked with the telco, they occur every twenty seconds.  The CLI states:

Sep 25 09:44:36 NOTICE[18432]: chan_zap.c:8089 pri_dchannel: PRI got event: HDLC Abort (6) on Primary D-channel of span 1

If I run pritest, the output is:

   Read on 3 failed: Unknown error 500
   Got event on PRI interface: 6

On internal timing the problem is less frequent, but still substantial.  Each time one of these messages pops up, there is a little pop sound on the phone.  Periodically, it causes the D-channel to reset and drops all calls.

For the record, my TE110P is not missing or sharing IRQ's, I am not running graphics, and I have gone through all the suggestions on the mailing lists.  I am using a Digium-recommended motherboard and not using the 3GHz/2GB-RAM machine for anything else.  I have run exhaustive testing with the telco, and their timing appears to be perfect.

The general concensus on the mailing lists is that this problem can be corrected by rolling back to a significantly earlier version of libpri (and, I assume, zaptel).  This would indicate that there is a problem with the timing in the current version of libpri.  (Although I cannot substantiate this, I wonder if it happened around the time configurable clocks were added.)

I don't know what causes this, or why it only happens to some people, but if an eariler version works fine, it seems libpri has an issue that should be addressed.

Thank you.
Comments:By: Mark Spencer (markster) 2005-09-25 12:56:36

This is a technical support issue and should be pursued through Digium technical support (support@digium.com).  Feel free to e-mail me if you have any trouble.