I when back and looked at some of my old
debug logs and came across one of them where the issue with cdrs
appearing in the debug log occurr. Here is the trace.
2004/08/27 10:24:38.955
6
RasTbl.cxx(1901) GK Removing callptr: 8e 58
d0 12 39 65 11 d4 a4 1d a4 db bb 45 0b 9b
2004/08/27 10:24:39.156
5
yasocket.cxx(699) ProxyH(1) 1 sockets selected from 22, total 22/2
2004/08/27 10:24:39.156 5
ProxyChannel.cxx(472) Q931d Reading from xxx.xxx.xxx.xxx:1720
2004/08/27 10:24:39.156 3
ProxyChannel.cxx(681) Q931d Received: Alerting
CRV=25833 from xxx.xxx.xxx.xxx:1720
2004/08/27 10:24:39.157 4
ProxyChannel.cxx(634) Q931 Received: {
q931pdu = {
protocolDiscriminator = 8
callReference = 25833
from = destination
messageType = Alerting
IE: Progress-Indicator = {
80
88
..
}
IE: Signal = {
01
.
}
IE: User-User = {
23 80 06 00 08 91 4a 00 04 28 00 b5 00 00
12 40 #.....J..(.....@
02 3c 05 18 04 01 40 99 09 90 99 09 90 0c
83 83 .<....@.........
34 40 0c 83 93 34 40 0c 83 c3 34 40 28 0d
86 00 4@...4@...4@(...
11 00 ab 51 30 c1 39 65 11 d4 a4 86 a4 db
bb 45 ...Q0.9e.......E
0b 9b 01 00 01 00 10 80 01
80
..........
}
}
h225pdu = {
h323_uu_pdu = {
h323_message_body = alerting {
protocolIdentifier = 0.0.8.2250.0.4
destinationInfo = {
vendor = {
vendor = {
t35CountryCode = 181
t35Extension = 0
manufacturerCode = 18
}
}
gateway = {
protocol = 2
entries {
[0]=voice {
supportedPrefixes = 4 entries {
[0]={
prefix = dialedDigits "xx#xx#xx#xx"
}
[1]={
prefix = dialedDigits "xxxxxx"
}
[2]={
prefix = dialedDigits "xxxxxx"
}
[3]={
prefix = dialedDigits "xxxxxx"
}
}
}
[1]=h323 {
supportedPrefixes = 0 entries {
}
}
}
}
mc = FALSE
undefinedNode = FALSE
}
}
callIdentifier = {
guid = 16 octets {
ab 51 30 c1
39 65 11 d4 a4 86 a4 db bb 45 0b 9b .Q0.9e.......E..
}
}
multipleCalls = FALSE
maintainConnection = FALSE
}
h245Tunneling = TRUE
}
}
}
2004/08/27 10:24:39.158 6
yasocket.cxx(610) xxx.xxx.xxx.xxx:60949 94 bytes sent
2004/08/27 10:24:39.168
5
yasocket.cxx(699) GkStatus 1 sockets selected from 3, total 3/0
2004/08/27 10:24:39.168
6
yasocket.cxx(610) xxx.xxx.xxx.xxx:55729 2 bytes sent
2004/08/27 10:24:39.168
5
job.cxx(360) JOB Worker threads: 26 total -
20 busy, 6 idle
2004/08/27 10:24:39.168
5
job.cxx(188) JOB Starting Job StatusCmd
debug trc 0 at Worker thread 2475572144
2004/08/27 10:24:39.168
5
GkStatus.cxx(1014) STATUS Got command debug trc 0 from client
xxx.xxx.xxx.xxx:55729
NUMINDGK231|262053|18|16|||412780c80003ffa5|xxx.xxx.xxx.xxx|5a 81 aa 73 93 07
11 d4 9d ab dc db 11 3d 26 47|5a 81 aa 73 93 07 11 d4 9d a9 dc db 11 3d 26 47|10:24:01.000
EDT Fri Aug 27 2004|10:24:21.000 EDT Fri Aug 27 2004|10:24:39.000 EDT Fri Aug
27
2004|xxx.xxx.xxx.xxx|45368|xxx.xxx.xxx.xxx|1720||525557528798:dialedDigits|xxx.xxx.xxx.xxx:45368|525557528798
NUMINDGK231|262093|0|16|||412780c80003ffcd|xxx.xxx.xxx.xxx|6b bb 74 aa 93 07 11
d4 9d de dc db 11 3d 26 47|6b bb 74 aa 93 07 11 d4 9d dc dc db 11 3d 26
47|10:24:30.000 EDT Fri Aug 27 2004||10:24:40.000 EDT Fri Aug 27
2004|xxx.xxx.xxx.xxx|45385|xxx.xxx.xxx.xxx|1720||525557524994:dialedDigits|xxx.xxx.xxx.xxx:45385|525557524994
.
.
.
.
Notice that this happen when I when from a debug trace of 6 to 0.
There are like a few hundred more cdrs after the first two I posted here. This
issue happen on more then one occasion.
I had one other case where gnugk became non responsive it would not
accept calls and if one logged into the status port one would get a blank screen.
If I remember correctly, this case happen when I did the following steps:
1.) rotatelog
2.) debug
trc 6
3.) Status Port
Froze
4.) I
Tried to telnet again to status port. Got blank screen.
5.) Gnugk
process was still running but CPU Process usage was 0 and no calls were being
accepted.
I think this trace helps us see the problem a little better now.
Regards,
Freddy.
From:
openh323gk-users-admin@xxxxxxxxxxxxxxxxxxxxx on behalf of Freddy Parra
Sent: Tue 9/14/2004 10:12 AM
To: openh323gk-users@xxxxxxxxxxxxxxxxxxxxx
Subject: Re:
Monitoring gnugk
Hi
Komnet,
I'm currently running Gnugk2.2b5 using Pandora RC1. My
Kernel is 2.4.21-4.ELsmp #1 SMP Fri Oct 3 17:52:56 EDT
2003 i686 i686 i386 GNU/Linux.
I'm running with 2 gigs of ram and 4 3.2 gig cpus.
What I've notice was that when I was running less then
400 calls everything seem to work fine. I could do
everything correctly from the status port.
I had no problems doing rotation and setting debug
traces. Never once did it crash.
The status port problems started happening when I had
more then 500 calls. Currently, I am running 3
DS3s(~1800 calls) in routed mode with 30 to 40 percent
cpu utilization. Gnugk2.2b5 is very stable and rock
solid, the best build I'v tested by far. Its been
running for over 3 weeks with very heavy traffic
without crashing. But before this, I manage to crash
and make gnugk non-responsive by setting debug traces
when I needed to do debug for customers. Then with
rotation debug logs I came across the weird behavior
of CDRs being logged into the the debug log. When
running this much traffic there is no room for error
which is why I only use the status port when I need to
reload or make a critical change. Maybe it is the OS
that I am using, I don't know, I wish I knew. I can
only tell you what happen and how it happen. I hope
this helps.
The following commands I ran were:
debug trc 0
debug trc 6
rotatelog
Regards,
Freddy
-----Original Message-----
From: openh323gk-users-admin@xxxxxxxxxxxxxxxxxxxxx
[mailto:openh323gk-users-admin@xxxxxxxxxxxxxxxxxxxxx]On
Behalf Of
Kompnet
Sent: Tuesday, September 14, 2004 1:56 AM
To: openh323gk-users@xxxxxxxxxxxxxxxxxxxxx
Subject: Re: Monitoring gnugk
But I don't know.
I am connecting by a telnet to the status port and
within all working day I
have an opportunity to look active calls, to reload a
configuration, to look
the table of registration, to register or on the
contrary to disconnect an
endpointpoint, to break off suspending (now almost
never it happens) or
unnecessary session, to change a tracelevel.
There is no problems since summer or autumn 2003, but
then the bug has been
found and eliminated instantly.
Maybe your problems are OS-depended problems?
Or separate commands which I do not use do not work?
Version 2.2.5b
Regards.
__________________________________
Do you Yahoo!?
New and Improved Yahoo! Mail - Send 10MB messages!
http://promotions.yahoo.com/new_mail
-------------------------------------------------------
This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
Project Admins to receive an Apple iPod Mini FREE for your judgement on
who ports your project to Linux PPC the best. Sponsored by IBM.
Deadline: Sept. 13. Go here: http://sf.net/ppc_contest.php
_______________________________________________________
List: Openh323gk-users@xxxxxxxxxxxxxxxxxxxxx
Archive: http://sourceforge.net/mailarchive/forum.php?forum_id=8549
Homepage: http://www.gnugk.org/
