Craig Guy | 1 Feb 16:25 2007
Picon

Re: ooh323c stack calling ATA problem

Thanks Pavek,

It was a codec mismatch which has been sorted out now.  Two other issues I 
have found are that when initiating the call from Asterisk (ooh323 being the 
endpoint), ooh323 won't negotiate the framesPerPkt down, chan_ooh323 offers 
4, however the ATA wants 3.  Examination of the ooh323 stack source reveals 
an empty stub to reduce the endpoint transmission framesPerPkt rate 
(ooCapability.c, line 974 of 0.8.2).  This was worked around by changing the 
default framesPerPkt from 4 to 3 in the g723.1 capabilities function of 
ooh323cDriver.c.  There is no problem when the call originates from the ATA 
as ooh323 will happily match the reduced framesPerPkt in this instance.

One final issue, a different ATA model from Hughes will only support G729AB, 
which ooh323 interprets as G729B.  Searching around a bit on google shows 
that G729AB seems to be different to G729B and there are suggestions that 
G729AB may be compatible with G729A.  Is this true?, if so then I could 
possibly get a successful connection by modifying the codec compare code in 
ooh323 (ooCapability.c) to match G729B to G729A and ensuring that VAD is 
disabled on the ata.

The reason I am stuck with the particular ata's is because I am trying to 
have asterisk work across a satellite link provided by Hughes Satellite 
Broadband ( www.hns.com ), and the ata is provided by Hughes.  Other ata's 
have been tried, however the Hughes one has some proprietary way of 
requesting and receiving traffic priority across the satellite.  Normal 
ata's are subject to massive jitter.

Craig

----- Original Message ----- 
(Continue reading)

Craig Guy | 1 Feb 23:38 2007
Picon

Re: ooh323c stack calling ATA problem

Hi Avin,

This is the code that that is causing the call to fail when originating from 
ooh323 and the called ata has a capabiliity less than ours, it is lines 964 
through 978 in ooCapability.c (version 0.8.3)

   /* Can we transmit compatible stream */
   if(dir & OOTX)
   {
      OOTRACEDBGC3("Comparing TX frame rate: channel's=%d, requested=%d\n",
         ((OOCapParams*)epCap->params)->txframes, noofframes);
      if(((OOCapParams*)epCap->params)->txframes <= noofframes) {
         return TRUE;
      }
      //else {
      //   TODO: reduce our ep transmission rate, as peer EP has low receive
      //   cap, than return TRUE
      //}
   }
   return FALSE;

Craig

----- Original Message ----- 
From: "Avin Patel" <apatel <at> obj-sys.com>
To: "Craig Guy" <cguy <at> bigpond.net.au>
Cc: "Pavel Jezek" <pavel.jezek <at> i.cz>; <ooh323c-devel <at> lists.sourceforge.net>
Sent: Friday, February 02, 2007 12:55 AM
Subject: Re: [ooh323c-devel] ooh323c stack calling ATA problem

(Continue reading)

Avin Patel | 2 Feb 00:02 2007

Re: ooh323c stack calling ATA problem

Hi Craig,

Craig Guy wrote:
> Hi Avin,
> 
> This is the code that that is causing the call to fail when originating 
> from ooh323 and the called ata has a capabiliity less than ours, it is 
> lines 964 through 978 in ooCapability.c (version 0.8.3)
> 
>   /* Can we transmit compatible stream */
>   if(dir & OOTX)
>   {
>      OOTRACEDBGC3("Comparing TX frame rate: channel's=%d, requested=%d\n",
>         ((OOCapParams*)epCap->params)->txframes, noofframes);
>      if(((OOCapParams*)epCap->params)->txframes <= noofframes) {
>         return TRUE;
>      }
>      //else {
>      //   TODO: reduce our ep transmission rate, as peer EP has low receive
>      //   cap, than return TRUE
>      //}
>   }
>   return FALSE;
> 

I would need the tcpdump file to find out, what is the case?

I know this part, as I have added this comment to reduce the rate as 
todo task.

(Continue reading)

Avin Patel | 1 Feb 16:55 2007

Re: ooh323c stack calling ATA problem

Hi Craig,

Craig Guy wrote:
> Thanks Pavek,
> 
> It was a codec mismatch which has been sorted out now.  Two other issues I 
> have found are that when initiating the call from Asterisk (ooh323 being the 
> endpoint), ooh323 won't negotiate the framesPerPkt down, chan_ooh323 offers 
> 4, however the ATA wants 3.  Examination of the ooh323 stack source reveals 
> an empty stub to reduce the endpoint transmission framesPerPkt rate 
> (ooCapability.c, line 974 of 0.8.2).  This was worked around by changing the 
> default framesPerPkt from 4 to 3 in the g723.1 capabilities function of 
> ooh323cDriver.c.  There is no problem when the call originates from the ATA 
> as ooh323 will happily match the reduced framesPerPkt in this instance.
> 
ATA -> Asterisk( receiving side): Packet rate matches reduced rate.
Asterisk -> ATA (sending side): Matches the by reducing, But can't 
increase the rate. I think I have the receiver increase the rate for 
g.723.1 codec, than suggested. That's why it is failing.

Now, for G.273.1, I have read, that reducing rate, actually increase the 
bandwidth, as more packet need to be send. I couldn't find it where. So 
it might need to be compared reverse than other codecs. ..???

You can provide the tcpdump file for failing case, we can check it 
further. I have seen this bug, with others also. We can fix this, if you 
provide needed logs/dump.

Regards,
Avin Patel
(Continue reading)

Avin Patel | 1 Feb 16:38 2007

Re: Timing fix

Hi Tim,
This change leads to deadlock. I will add the owner lock with some changes.

Regards,
Avin Patel
Objective Systems, Inc.

Tim King wrote:
> I have been making the following change and I don't quite understand 
> what I am doing, but without this change Asterisk crashes regularly on a 
> dual-Xeon machine, and with it Asterisk is stable. Note this only 
> affects dual/processor/core machines AFIK. If anyone who understands the 
> locking can verify this then please add it to the next release.
> 
> In file chan_h323.c in directory 
> asterisk-addons-x.y.z/asterisk-ooh323c/src line 808 change:
> 
>       if(p->owner)
>       {
>          p->owner->tech_pvt = NULL;
>          p->owner = NULL;
>       }
> to
>       if(p->owner)
>       {
>          ast_mutex_lock(&p->owner->lock);
>          p->owner->tech_pvt = NULL;
>          ast_mutex_unlock(&p->owner->lock);
>          p->owner = NULL;
>       }
(Continue reading)

George Melika | 3 Feb 00:02 2007
Picon

ooh323: Faking Ring Problem

Hi,

I have been looking around online for a fix for my problem but I have not been able to find one.  My problem is
that when I dial a destination on a GW using OOH323 I get a fake ring tone instead of the actual ringtone
generated from the GW.  This is very easy to spot specially when most of the GW's I am calling are in
international locations with very different ringtones.  This is also evident when calling a US number
that supports a custom ringtone.

I am not an expert at all, but my suspicion is that OOH323 is not opening an RTP channel to receive the progress
indicator and instead simulating the ringtone locally.  So is this a configuration option that I have
missed? A known problem? Or, is it something that should work?

Thank you,
George

 
____________________________________________________________________________________
Do you Yahoo!?
Everyone is raving about the all-new Yahoo! Mail beta.
http://new.mail.yahoo.com

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier.
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
Mike Tubby | 7 Feb 01:13 2007

ooh323 drops registration with Cisco IOS GateKeeper - bug or config issue?

All,
 
I'm running (attempting to) ooh323 with Asterisk and a Cisco 2621XM router operating as a H.323 GateKeeper, however when I bring the Asterisk box up it registers successfully with the GateKeeper (exchanges GRQ/GCF, then RRQ/RCF) it notes the GateKeeper supports keepalive at 300 seconds, when it gets to time to re-register its sends an RRQ again and gets rejected with RRJ (unspecified reason) and so closes the session.
 
At this point the H.323 session is lost and never retried.
 
The set up is as follows:
 
    Asterisk Box: RedHat/Fedora Core 3, Asterisk 1.2.14 (built from source), ooh323 rev 0.8.2 from Asterisk-Addons-1.2.5
 
    Cisco GateKeeper: 2621XM router with c2600-jsx-mz.123-22.bin
 
Config on Asterisk box:
 
 
; Objective System's H323 Configuration example for Asterisk
; ooh323c driver configuration
;
; [general] section defines global parameters
;
; This is followed by profiles which can be of three types - user/peer/friend
; Name of the user profile should match with the h323id of the user device.
; For peer/friend profiles, host ip address must be provided as "dynamic" is
; not supported as of now.
;
; Syntax for specifying a H323 device in extensions.conf is
; For Registered peers/friends profiles:
;        OOH323/name where name is the name of the peer/friend profile.
;
; For unregistered H.323 phones:
;        OOH323/ip[:port] OR if gk is used OOH323/alias where alias can be any H323
;                          alias
;
; For dialing into another asterisk peer at a specific exten
;       OOH323/exten/peer OR OOH323/exten <at> ip
;
; Domain name resolution is not yet supported.
;
; When a H.323 user calls into asterisk, his H323ID is matched with the profile
; name and context is determined to route the call
;
; The channel driver will register all global aliases and aliases defined in
; peer profiles with the gatekeeper, if one exists. So, that when someone
; outside our pbx (non-user) calls an extension, gatekeeper will route that
; call to our asterisk box, from where it will be routed as per dial plan.
 

[general]
;Define the asetrisk server h323 endpoint
 
;The port asterisk should listen for incoming H323 connections.
;Default - 1720
;port=1720
 
;The dotted IP address asterisk should listen on for incoming H323
;connections
;Default - tries to find out local ip address on it's own
bindaddr=0.0.0.0
 
;This parameter indicates whether channel driver should register with
;gatekeeper as a gateway or an endpoint.
;Default - no
gateway=yes

;Whether asterisk should use fast-start and tunneling for H323 connections.
;Default - yes
;faststart=no
;h245tunneling=no
 
;H323-ID to be used for asterisk server
;Default - Asterisk PBX
;h323id=ObjSysAsterisk

h323id=ASTERISK
e164=100
 
;CallerID to use for calls
;Default - Same as h323id
callerid=ASTERISK
 
;Whether this asterisk server will use gatekeeper.
;Default - DISABLE
;gatekeeper = DISCOVER
gatekeeper = 192.168.1.6
;gatekeeper = DISABLE
 
;Location for H323 log file
;Default - /var/log/asterisk/h323_log
;logfile=/var/log/asterisk/h323_log

;Following values apply to all users/peers/friends defined below, unless
;overridden within their client definition
 
;Sets default context all clients will be placed in.
;Default - default
context=default
 
;Sets rtptimeout for all clients, unless overridden
;Default - 60 seconds
;rtptimeout=60      ; Terminate call if 60 seconds of no RTP activity
                    ; when we're not on hold
 
;Type of Service
;Default - none (lowdelay, thoughput, reliability, mincost, none)
;tos=lowdelay
 
;amaflags = default
;The account code used by default for all clients.
;accountcode=h3230101
 
;The codecs to be used for all clients.Only ulaw and gsm supported as of now.
disallow=all     ;Note order of disallow/allow is important.
allow=alaw
allow=ulaw
allow=gsm

; dtmf mode to be used by default for all clients. Supports rfc2833, q931keypad
; h245alphanumeric, h245signal.
;Default - rfc 2833
dtmfmode=rfc2833

[cisco-gk]
type=peer
ip=192.168.1.6
port=1720
context=from-h323
disallow=all
allow=alaw
allow=ulaw
allow=gsm
rtptimeout=60
dtmfmode=rfc2833
h323id=PABX
 
 
trace of it disconnecting:
 
 
[root <at> pabx asterisk]# cat h323_log
---------Date 02/06/07---------
09:53:08:742  Signalling IP address is set to 0.0.0.0
09:53:08:742  Listen port number is set to 1720
09:53:08:742  Using local RAS Ip address 192.168.1.5
09:53:08:742  Gatekeeper Mode - RasUseSpecificGatekeeper
09:53:08:742  Gatekeeper IP:port set to - 192.168.1.6:1719
09:53:08:742  Enabled RFC2833 DTMF capability for end-point
09:53:08:742  H323 listener creation - successful
09:53:08:742  Creating CMD listener at 0.0.0.0:7575
09:53:08:742  CMD listener creation - successful
09:53:08:742  H.323 Endpoint Configuration is as follows:
09:53:08:742    Trace File: /var/log/asterisk/h323_log
09:53:08:742    FastStart - enabled
09:53:08:742    H245 Tunneling - enabled
09:53:08:742    MediaWaitForConnect - disabled
09:53:08:742    AutoAnswer - disabled
09:53:08:742    Terminal Type - 50
09:53:08:742    T35 CountryCode - 1
09:53:08:742    T35 Extension - 0
09:53:08:742    Manufacturer Code - 71
09:53:08:742    ProductID - objsys
09:53:08:742    VersionID - v0.8.2
09:53:08:742    Local signalling IP address - 0.0.0.0
09:53:08:742    H225 ListenPort - 1720
09:53:08:743    CallerID - ASTERISK
09:53:08:743    Call Establishment Timeout - 60 seconds
09:53:08:743    MasterSlaveDetermination Timeout - 30 seconds
09:53:08:743    TerminalCapabilityExchange Timeout - 30 seconds
09:53:08:743    LogicalChannel  Timeout - 30 seconds
09:53:08:743    Session Timeout - 15 seconds
09:53:08:743  Gatekeeper Client Configuration:
09:53:08:743    Gatekeeper mode - UseSpecificGatekeeper
09:53:08:743    Gatekeeper To Use - 192.168.1.6:1719
09:53:08:743  H323 RAS channel creation - successful
09:53:08:743  Sent GRQ message
09:53:08:750  Gatekeeper Confirmed (GCF) message received.
09:53:08:750  Gatekeeper Confirmed
09:53:08:750  Sent RRQ message
09:53:08:760  Registration Confirm (RCF) message received
09:53:08:760  Gatekeeper supports KeepAlive, Registration TTL is 300
09:57:48:761  Sent RRQ message
09:57:48:766  Registration Reject (RRJ) message received.
09:57:48:766  RRQ Rejected - Undefined Reason
09:57:48:766  Error: Gatekeeper error. Either Gk not responding or Gk sending in valid messages
09:57:48:766  Error: Gatekeeper error detected. Closing GkClient as Gk mode is UseSpecifcGatekeeper
09:57:48:766  Destroying Gatekeeper Client
09:57:48:766  Closed RAS channel
[root <at> pabx asterisk]#
 
Any ideas whether this is a bug or a configuration issue?
 
 
Regards
 
 
Mike
 
 
 
 
-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier.
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
ooh323c-devel mailing list
ooh323c-devel <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ooh323c-devel
Avin Patel | 5 Feb 18:37 2007

Re: ooh323: Faking Ring Problem

Hi George,
This is NOT a problem. That is how H.323 works. I would tell that PSTN 
is fake ring. If there is a reason for call can't be established, than 
that reason field should be used to specify, NOT RTP channel.

Subject name should be, dealing with PSTN line.

The early audio establishment is adding now in ALERTING message. So this 
should not be a problem any more. Are you using updated version?

Regards,
Avin Patel
Objective Systems, Inc.

George Melika wrote:
> Hi,
> 
> I have been looking around online for a fix for my problem but I have not been able to find one.  My problem is
that when I dial a destination on a GW using OOH323 I get a fake ring tone instead of the actual ringtone
generated from the GW.  This is very easy to spot specially when most of the GW's I am calling are in
international locations with very different ringtones.  This is also evident when calling a US number
that supports a custom ringtone.
> 
> I am not an expert at all, but my suspicion is that OOH323 is not opening an RTP channel to receive the
progress indicator and instead simulating the ringtone locally.  So is this a configuration option that I
have missed? A known problem? Or, is it something that should work?
> 
> Thank you,
> George
> 
> 
>  
> ____________________________________________________________________________________
> Do you Yahoo!?
> Everyone is raving about the all-new Yahoo! Mail beta.
> http://new.mail.yahoo.com
> 
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier.
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> ooh323c-devel mailing list
> ooh323c-devel <at> lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/ooh323c-devel
> 

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier.
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
George Melika | 7 Feb 19:42 2007
Picon

Re: ooh323: Faking Ring Problem

I'm using the version that is shipped with 1.4.  Is there a newer one?

----- Original Message ----
From: Avin Patel <apatel <at> obj-sys.com>
Cc: ooh323c-devel <at> lists.sourceforge.net
Sent: Monday, February 5, 2007 9:37:08 AM
Subject: Re: [ooh323c-devel] ooh323: Faking Ring Problem

Hi George,
This is NOT a problem. That is how H.323 works. I would tell that PSTN 
is fake ring. If there is a reason for call can't be established, than 
that reason field should be used to specify, NOT RTP channel.

Subject name should be, dealing with PSTN line.

The early audio establishment is adding now in ALERTING message. So this 
should not be a problem any more. Are you using updated version?

Regards,
Avin Patel
Objective Systems, Inc.

George Melika wrote:
> Hi,
> 
> I have been looking around online for a fix for my problem but I have not been able to find one.  My problem is
that when I dial a destination on a GW using OOH323 I get a fake ring tone instead of the actual ringtone
generated from the GW.  This is very easy to spot specially when most of the GW's I am calling are in
international locations with very different ringtones.  This is also evident when calling a US number
that supports a custom ringtone.
> 
> I am not an expert at all, but my suspicion is that OOH323 is not opening an RTP channel to receive the
progress indicator and instead simulating the ringtone locally.  So is this a configuration option that I
have missed? A known problem? Or, is it something that should work?
> 
> Thank you,
> George
> 
> 
>  
> ____________________________________________________________________________________
> Do you Yahoo!?
> Everyone is raving about the all-new Yahoo! Mail beta.
> http://new.mail.yahoo.com
> 
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier.
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> ooh323c-devel mailing list
> ooh323c-devel <at> lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/ooh323c-devel
> 

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier.
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
ooh323c-devel mailing list
ooh323c-devel <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ooh323c-devel

 
____________________________________________________________________________________
Sucker-punch spam with award-winning protection. 
Try the free Yahoo! Mail Beta.
http://advision.webevents.yahoo.com/mailbeta/features_spam.html

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier.
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
Avin Patel | 7 Feb 19:57 2007

Re: ooh323: Faking Ring Problem

Hi George,
You can synchronize with svn 1.4 branch to get latest changes. I don't 
know which 1.4 release will have this change.

I think he have solved this problem.

Regards,
Avin Patel
Objective Systems, Inc.

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier.
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

Gmane