[Asterisk-Users] RE:PRI issues with TE410P
Scott Stingel
scott at evtmedia.com
Mon Mar 22 11:28:34 MST 2004
Hi Azher-
I experience many of the same problems, especially the frame
retransmissions, which cause channel hangups from time to time. These
channel hangups clear themselves after a few minutes normally but do
interfere with calls.
I'm not the expert on Linux internals, so I'm not sure what can be done to
improve things. If you have a SCSI disk instead of an ide, you could try it
instead, in case the IDE interrupts are blocking the servicing of the
TE410P. Of course you can also take steps to minimize the ETH0 interrupts
during operation as well.
Finally, I believe there are deficiencies in the zaptel driver in that it
does not handle overrun and underrun errors correctly. I'm not certain of
this, but it appeared that way to a programmer that I had look at the
driver. Mark Spencer at Digium has expressed interest in looking at this
problem, but has not thus far.
You might report the problems to Digium customer service, and tell them that
the PRI problems are preventing you from buying additional TE410p boards,
etc. Maybe if enough of us report the problems we'll get some attention.
Be sure to copy sales (Greg Vance) on everything (gvance at digium.com)
Sorry I don't have the magic answer.
Regards
Scott
Scott M. Stingel
Emerging Voice Technology Inc.
Palo Alto, California and London, England
Email: scott at evtmedia.com
URL: www.evtmedia.com
-----Original Message-----
From: asterisk-users-admin at lists.digium.com
[mailto:asterisk-users-admin at lists.digium.com] On Behalf Of Azher Amin
Sent: Monday, March 22, 2004 9:03 AM
To: asterisk-users at lists.digium.com
Subject: RE: [Asterisk-Users] PRI issues with TE410P
Dear Scott,
I am using kernel 2.4.25. RedHat Linux 9.0. libpri-0.5.2, zaptel-0.9.0,
and asterisk from CVS of 20th March.
About the IRQ misses, may be I am wrong, but from below, CPU0 has lot
load from t4xxp, ide0 and eth0. Is it possible that we can keep the ide0
and eth0 to cpu1 ? and just let the cpu0 handle t4xxp ? Any suggestion
??
CPU0 CPU1
0: 2295 8857853 IO-APIC-edge timer
1: 2 0 IO-APIC-edge keyboard
2: 0 0 XT-PIC cascade
14: 98854 0 IO-APIC-edge ide0
15: 5 5 IO-APIC-edge ide1
22: 91359 0 IO-APIC-level eth0
28: 88531746 50053 IO-APIC-level t4xxp
NMI: 0 0
LOC: 8860069 8860067
ERR: 0
MIS: 0
Also on load, I am getting lot frame retransmissions. I have attached
the messages file, plz take a look at that as well for some suggestions.
I don't care about Frame retransmissions as far as my PRI's stays stable
and the users don't get dropped :(
Regards
Azher
-------------------------------
http://www.consulttech.com.pk
-----Original Message-----
From: asterisk-users-admin at lists.digium.com
[mailto:asterisk-users-admin at lists.digium.com] On Behalf Of Scott
Stingel
Sent: Sunday, March 21, 2004 9:16 PM
To: asterisk-users at lists.digium.com
Subject: RE: [Asterisk-Users] PRI issues with TE410P
Hello Azher-
I have a similar setup in hardware, ie: TE410P running on dual-xeon
system,
however I'm running IVR only. I start getting the I-frame errors above
about 80 simultaneous calls. I do not get IRQ misses at all. Also I do
not
get the startup error messages. The errors I get the most under load
are
the frame retransmission messages in /var/log/asterisk/messages. Do you
get
those as well?
Since you are getting IRQ misses, you may have some basic problem, ie:
something keeping the zaptel driver from getting around to servicing the
TE410 interrupts. I think that they have to be serviced without fail
every
1 msec, or errors start occurring.
Which kernel are you running?
What does your Perl script do?
Did you try disabling mysql logging to lower the disk load? Maybe the
disk
driver is interfering with the zaptel driver interrupts.
Regards
Scott
Scott M. Stingel
Emerging Voice Technology Inc.
Palo Alto, California and London, England
Email: scott "at " evtmedia.com
URL: www.evtmedia.com
>-----Original Message-----
>From: asterisk-users-admin at lists.digium.com
>[mailto:asterisk-users-admin at lists.digium.com] On Behalf Of Azher Amin
>Sent: Sunday, March 21, 2004 3:41 PM
>To: asterisk-users at lists.digium.com
>Subject: [Asterisk-Users] PRI issues with TE410P
>
>Hi,
>
>I am having some problems mentioned below, the box is in
>production live
>environment with traffic around 30 - 100 calls.
>
>I am running T/E410P in a Dual P4 xeon with HT disabled. I am using
>zaptel 0.9.0 and asterisk stable 1 release. There is no gui,
>just mysql,
>perl (small script) and asterisk.
>
>System runs very smoothly if the calls are around 40-50 and
>comes one by
>one , however sometimes at immediate load of around 30 more calls ... I
>get the following processes in the ps -ax, and asterisk starts droping
>the calls, irq misses rise and console shows lot of pri errors (which
>donot occur in a smooth load of around 50 calls).
>
>Can someone explain why this happens .. however these get cleared once
>the channels are handled.
>
>20992 pts/3 S 0:01 zttool
>21412 ? S 0:00 asterisk
>21413 ? R 0:00 asterisk
>21418 ? S 0:00 asterisk
>21419 ? S 0:00 asterisk
>21420 ? S 0:00 asterisk
>21421 ? S 0:00 asterisk
>21422 ? S 0:00 asterisk
>21423 ? S 0:00 asterisk
>21424 ? S 0:00 asterisk
>21425 ? S 0:00 asterisk
>21426 ? S 0:00 asterisk
>21427 ? S 0:00 asterisk
>21429 ? S 0:00 asterisk
>21430 ? S 0:00 asterisk
>21431 ? S 0:00 asterisk
>21432 ? S 0:00 asterisk
>21433 ? S 0:00 asterisk
>21434 ? S 0:00 asterisk
>21435 ? S 0:00 asterisk
>21436 ? S 0:00 asterisk
>21437 ? S 0:00 asterisk
>21438 ? S 0:00 asterisk
>21439 ? S 0:00 asterisk
>21440 ? S 0:00 asterisk
>21441 ? S 0:00 asterisk
>21442 ? S 0:00 asterisk
>21443 ? S 0:00 asterisk
>21444 ? S 0:00 asterisk
>21445 ? S 0:00 asterisk
>21446 ? S 0:00 asterisk
>21447 ? S 0:00 asterisk
>21448 ? S 0:00 asterisk
>21449 ? S 0:00 asterisk
>21451 ? S 0:00 asterisk
>21452 ? S 0:00 asterisk
>21453 ? S 0:00 asterisk
>21454 ? S 0:00 asterisk
>21455 ? S 0:00 asterisk
>21456 ? S 0:00 asterisk
>21457 ? S 0:00 asterisk
>21458 ? S 0:00 asterisk
>21459 pts/2 R 0:00 ps -ax
>21460 ? S 0:00 asterisk
>
>
>
>Further I am also getting
>
>Mar 22 01:33:09 WARNING[114696]: chan_zap.c:5970 zt_pri_error: PRI: !!
>Got I-frame while link state 2
>Mar 22 01:33:09 WARNING[114696]: chan_zap.c:5970 zt_pri_error: PRI: !!
>Got I-frame while link state 2
>Mar 22 01:33:09 WARNING[114696]: chan_zap.c:5970 zt_pri_error: PRI: !!
>Got I-frame while link state 2
>Mar 22 01:33:09 WARNING[114696]: chan_zap.c:5970 zt_pri_error: PRI: !!
>Got I-frame while link state 2
>Mar 22 01:33:09 WARNING[114696]: chan_zap.c:5970 zt_pri_error: PRI: !!
>Got I-frame while link state 2
>Mar 22 01:33:09 WARNING[114696]: chan_zap.c:5970 zt_pri_error: PRI: !!
>Got S-frame while link down
>Mar 22 01:33:09 WARNING[114696]: chan_zap.c:5970 zt_pri_error: PRI: !!
>Got S-frame while link down
> == D-Channel on span 1 up
>
>and certain IRQ misses like around 270 after an operation of 2-3 hours.
>
>Further when I load the T/E410p card I get this error ...
>
>Mar 22 00:33:48 VoiceOne kernel: TE410P: Launching card: 0
>Mar 22 00:33:48 VoiceOne kernel: TE410P: Setting up global serial
>parameters
>Mar 22 00:33:48 VoiceOne kernel: TE410P: Timing from source 0
>Mar 22 00:33:48 VoiceOne kernel: Found a Wildcard: Wildcard
>TE410P-Xilinx
>Mar 22 00:33:48 VoiceOne kernel: Registered tone zone 0
>(United States /
>North America)
>Mar 22 00:33:48 VoiceOne kernel: TE410P: Span 1 configured for
>CCS/HDB3/CRC4
>Mar 22 00:33:48 VoiceOne kernel: SPAN 1: Primary Sync Source
>Mar 22 00:33:48 VoiceOne kernel: Uhhuh. NMI received for unknown reason
>20.
>Mar 22 00:33:48 VoiceOne kernel: Dazed and confused, but trying to
>continue
>Mar 22 00:33:48 VoiceOne kernel: Do you have a strange power
>saving mode
>enabled?
>Mar 22 00:33:48 VoiceOne kernel: TE410P: Span 2 configured for
>CCS/HDB3/CRC4
>Mar 22 00:33:48 VoiceOne kernel: TE410P: Span 3 configured for
>CCS/HDB3/CRC4
>Mar 22 00:33:48 VoiceOne kernel: TE410P: Span 4 configured for
>CCS/HDB3/CRC4
>
>I have another card and above also come with that.
>
>I look forward for resolution on these issues with some suggestions to
>improve the stability. Coz otherwise system is not stable to be sold to
>client.
>
>Regards
>Azher Amin
>-------------------------------
>http://www.consulttech.com.pk
>
>
>
>
>_______________________________________________
>Asterisk-Users mailing list
>Asterisk-Users at lists.digium.com
>http://lists.digium.com/mailman/listinfo/asterisk-users
>To UNSUBSCRIBE or update options visit:
> http://lists.digium.com/mailman/listinfo/asterisk-users
>
>
_______________________________________________
Asterisk-Users mailing list
Asterisk-Users at lists.digium.com
http://lists.digium.com/mailman/listinfo/asterisk-users
To UNSUBSCRIBE or update options visit:
http://lists.digium.com/mailman/listinfo/asterisk-users
More information about the asterisk-users
mailing list