[Asterisk-Users] PRI HDLC Abort (6) Errors

[Tom]

Hello,
I have searched and searched, and come up with nothing.  I am running Asterisk
with a wcte110p configured for t1.  Our PRI is staying up, and we can make
calls however our service provider's logs are flooding with errors and we are
getting lots of HDLC Abort (6) on Primary D-Channel Errors.

Our provider says it looks like our box is trying to be the master timer on the
circuit (which is not correct they are providing the timing) we have tried both
span=1,1,0,esf,b8zs and span=1,0,0,esf,b8zs in zaptel.conf both produce the
same problems.  The problem is not in "Asterisk" per se as the errors start
happening as soon as I modprobe the driver and run ztcfg.  As soon as the
circuit comes up the errors start on the provider's end.

I have checked everything I've seen in the list which basically amounts to
making sure the wildcard is on its own IRQ, and ours is:

[root at test asterisk]# cat /proc/interrupts
           CPU0       CPU1
  0:   28998010   28954370    IO-APIC-edge  timer
  1:         10          0    IO-APIC-edge  i8042
  8:          1          0    IO-APIC-edge  rtc
  9:          0          0   IO-APIC-level  acpi
 12:         67          0    IO-APIC-edge  i8042
 14:     155314     147819    IO-APIC-edge  libata
 15:     260058     260532    IO-APIC-edge  ide1
169:      79998      70983   IO-APIC-level  libata
177:         16    1441995   IO-APIC-level  eth0
185:    1909840          0   IO-APIC-level  uhci_hcd, eth1
193:          0          0   IO-APIC-level  ehci_hcd
201:          0          0   IO-APIC-level  uhci_hcd
209:   27757492   28439765   IO-APIC-level  t1xxp
NMI:          0          0
LOC:   57954744   57954745
ERR:          0
MIS:          0

The system is not under any sort of load:
[root at test ~]# uptime
 13:40:53 up 15:54,  2 users,  load average: 0.01, 0.05, 0.02

I tried nicing Astrisk up as far as it can go and running it with -p, however,
it isn't the actual asterisk process that is the problem, I am 100% certain it
is in the configuration or the wcte110p driver itself as the errors start as
soon as the driver is loaded and configured with no asterisk running at all.
Here is my config:

#zaptel.conf
loadzone = us

span=1,1,0,esf,b8zs,yellow
bchan=1-23
dchan=24

#zapata.conf
[channels]
switchtype=national
immediate=no
context=bell
signalling=pri_cpe
group=1
channel => 1-23

We are running CVS Asterisk/zaptel/libpri from March 2nd 2005 on Fedora Core 3
fully patched as of last night, I was thinking the problem was with the 2.6
kernel getting preempted and therefore the driver not being able to do its
timings right, however fc3's kernels have preemption disabled by default.  Does
Digium hardware really need/expect a real time OS to run properly?
Like I said previously I think the problem is in the driver itself not in
asterisk.  Any help would be appreciated, and I can code a bit in c so if
someone can point me in the right direction I might be able to fix it myself...

We have been using Asterisk for 8 months now, however this is the first PRI
we've terminated on it, we've never had any problems with voice
quality/jitter/anything previously.  And even now, the calls work fine, sound
fine, everything seems great, but our provider is up our butt about the errors
flooding them, and they are pretty annoying on our end flooding the asterisk
console.

Tom Christensen
VP Product Development
SinglePoint Networks