Hi,
I've just committed major update for rtpproxy/nathelper, which adds support for proxying RTCP and also make RTP proxy behave much better for non-NATed clients by pre-loading remote addresses[1] of each party from the SIP request/reply. Please note that proxy's command protocol has been extended to support new functionality, so that you need both new rtp proxy and new nathelper (old nathelper will not work with new rtp proxy and vice versa). Both of them can be obtained from the ser's cvs repository:
http://cvs.berlios.de/cgi-bin/viewcvs.cgi/ser/rtpproxy/ http://cvs.berlios.de/cgi-bin/viewcvs.cgi/ser/sip_router/modules/nathelper/
Please let me know if there are any problems with the new version.
-Maxim [1] currently only IPv4 addresses can be pre-loaded, though it should be trivial to extend proxy, nathelper and command protocol to accomodate IPv6 as well. Proxy's core supports IPv4<->IPv4, IPv4<->IPv6 and IPv6<->IPv6 relaying already both for RTP and RTCP.
Maxim Sobolev wrote:
Hi,
I've just committed major update for rtpproxy/nathelper, which adds support for proxying RTCP and also make RTP proxy behave much better for non-NATed clients by pre-loading remote addresses[1] of each party from the SIP request/reply. Please note that proxy's command protocol has been extended to support new functionality, so that you need both new rtp proxy and new nathelper (old nathelper will not work with new rtp proxy and vice versa). Both of them can be obtained from the ser's cvs repository:
http://cvs.berlios.de/cgi-bin/viewcvs.cgi/ser/rtpproxy/ http://cvs.berlios.de/cgi-bin/viewcvs.cgi/ser/sip_router/modules/nathelper/
Please let me know if there are any problems with the new version.
-Maxim [1] currently only IPv4 addresses can be pre-loaded, though it should be trivial to extend proxy, nathelper and command protocol to accomodate IPv6 as well. Proxy's core supports IPv4<->IPv4, IPv4<->IPv6 and IPv6<->IPv6 relaying already both for RTP and RTCP.
P.S. Forgot to mention: nathelper now inserts special flag into SDP body which indicates that the session is already forced to go through one rtp proxy and do not apply rtp proxy to messages with such flag already set.
Hi Maxim,
I am in the process of testing this new version in our lab with 0.8.13. We have been using the older versions with great success for many months now. But the new version does not work. We are testing with Grandstream and Sipura units. When a Sipura calls another Sipura, the nathelper/rtpproxy fails to insert the proper "Connection Information (c)" in the SDP. Instead of filling in the IP Address of the RTPProxy it just leaves the same address and adds these four characters "\000" to the end which seem to make the other Sipura unhappy because it terminates the call right away with a "488- Not Acceptable" Message.
When a Grandstream is making the call, the same thing happens, with the exception of the four characters. (IP Address in Connection Information (c) is not updated)
The Ports do seem to get changed appropiately by the nathelper/rtpproxy in both cases. But since the IP is not substituted there is no chance of audio being setup properly.
I can send the Ethereal traces if you want.
Let me know what we can do to fix this issue.
Thanks,
Update...
I have now tested multiple versions of nathelper from January. The problem appears after the changes made by Tristan Colgate on 2004-01-16. Nathelper/rtpproxy works fine on the version from 2004-01-15.
Can you take a look at this Tristan? Maxim?
Thanks,
Andres wrote:
Update...
I have now tested multiple versions of nathelper from January. The problem appears after the changes made by Tristan Colgate on 2004-01-16. Nathelper/rtpproxy works fine on the version from 2004-01-15.
Can you take a look at this Tristan? Maxim?
I'll be travelling next 5 days. I will take a look at the problem upon my return if somebody doesn't beat me on it.
-Maxim
What change do you mean ? I reviewed and commited some changes on behalf of Tristan, so please blame me (and provide me with more details if possible) :-).
Could you make sure that the version before my commit works ?
Jan.
On 30-01 11:14, Andres wrote:
Update...
I have now tested multiple versions of nathelper from January. The problem appears after the changes made by Tristan Colgate on 2004-01-16. Nathelper/rtpproxy works fine on the version from 2004-01-15.
Can you take a look at this Tristan? Maxim?
Thanks,
-- Andres Network Admin http://www.telesip.net
Andres wrote:
Hi Maxim,
I am in the process of testing this new version in our lab with 0.8.13. We have been using the older versions with great success for many months now. But the new version does not work. We are testing with Grandstream and Sipura units. When a Sipura calls another Sipura, the nathelper/rtpproxy fails to insert the proper "Connection Information (c)" in the SDP. Instead of filling in the IP Address of the RTPProxy it just leaves the same address and adds these four characters "\000" to the end which seem to make the other Sipura unhappy because it terminates the call right away with a "488- Not Acceptable" Message.
When a Grandstream is making the call, the same thing happens, with the exception of the four characters. (IP Address in Connection Information (c) is not updated)
The Ports do seem to get changed appropiately by the nathelper/rtpproxy in both cases. But since the IP is not substituted there is no chance of audio being setup properly.
I can send the Ethereal traces if you want.
Let me know what we can do to fix this issue.
Thanks,
Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers
Jan Janak wrote:
What change do you mean ? I reviewed and commited some changes on behalf of Tristan, so please blame me (and provide me with more details if possible) :-).
Could you make sure that the version before my commit works ?
Jan.
Right now we are running the nathelper version from 2004-1-15 and it works fine. We tried the version from yesterday and also from 2004-1-22 and it fails to work (presumably after the patch made by Tristan). See the attached Ethereal trace and examine frame #3 (the INVITE after nathelper processes it). Take a look at the "Connection Information" in the SDP. The IP was not substituted and extra characters were added that makes Sipura units unhappy.
If you need anything else please let me know.
Andres.
On 30-01 11:14, Andres wrote:
Update...
I have now tested multiple versions of nathelper from January. The problem appears after the changes made by Tristan Colgate on 2004-01-16. Nathelper/rtpproxy works fine on the version from 2004-01-15.
Can you take a look at this Tristan? Maxim?
Thanks,
-- Andres Network Admin http://www.telesip.net
Andres wrote:
Hi Maxim,
I am in the process of testing this new version in our lab with 0.8.13. We have been using the older versions with great success for many months now. But the new version does not work. We are testing with Grandstream and Sipura units. When a Sipura calls another Sipura, the nathelper/rtpproxy fails to insert the proper "Connection Information (c)" in the SDP. Instead of filling in the IP Address of the RTPProxy it just leaves the same address and adds these four characters "\000" to the end which seem to make the other Sipura unhappy because it terminates the call right away with a "488- Not Acceptable" Message.
When a Grandstream is making the call, the same thing happens, with the exception of the four characters. (IP Address in Connection Information (c) is not updated)
The Ports do seem to get changed appropiately by the nathelper/rtpproxy in both cases. But since the IP is not substituted there is no chance of audio being setup properly.
I can send the Ethereal traces if you want.
Let me know what we can do to fix this issue.
Thanks,
Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers
Yes, indeed, there was a problem with force_rtp_proxy(). I've just committed a fix (1.38). The problem was that you were trying to use results of one call to ip_addr2a() after another call to that function. Since ip_addr2a() returns pointer to a static internal buffer, it was leading to incorrect results.
-Maxim
Jan Janak wrote:
What change do you mean ? I reviewed and commited some changes on behalf of Tristan, so please blame me (and provide me with more details if possible) :-).
Could you make sure that the version before my commit works ?
Jan.
On 30-01 11:14, Andres wrote:
Update...
I have now tested multiple versions of nathelper from January. The problem appears after the changes made by Tristan Colgate on 2004-01-16. Nathelper/rtpproxy works fine on the version from 2004-01-15.
Can you take a look at this Tristan? Maxim?
Thanks,
-- Andres Network Admin http://www.telesip.net
Andres wrote:
Hi Maxim,
I am in the process of testing this new version in our lab with 0.8.13. We have been using the older versions with great success for many months now. But the new version does not work. We are testing with Grandstream and Sipura units. When a Sipura calls another Sipura, the nathelper/rtpproxy fails to insert the proper "Connection Information (c)" in the SDP. Instead of filling in the IP Address of the RTPProxy it just leaves the same address and adds these four characters "\000" to the end which seem to make the other Sipura unhappy because it terminates the call right away with a "488- Not Acceptable" Message.
When a Grandstream is making the call, the same thing happens, with the exception of the four characters. (IP Address in Connection Information (c) is not updated)
The Ports do seem to get changed appropiately by the nathelper/rtpproxy in both cases. But since the IP is not substituted there is no chance of audio being setup properly.
I can send the Ethereal traces if you want.
Let me know what we can do to fix this issue.
Thanks,
Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers
Sorry for this, I should have been more careful.
Jan.
On 31-01 11:56, Maxim Sobolev wrote:
Yes, indeed, there was a problem with force_rtp_proxy(). I've just committed a fix (1.38). The problem was that you were trying to use results of one call to ip_addr2a() after another call to that function. Since ip_addr2a() returns pointer to a static internal buffer, it was leading to incorrect results.
-Maxim
Jan Janak wrote:
What change do you mean ? I reviewed and commited some changes on behalf of Tristan, so please blame me (and provide me with more details if possible) :-).
Could you make sure that the version before my commit works ?
Jan.
On 30-01 11:14, Andres wrote:
Update...
I have now tested multiple versions of nathelper from January. The problem appears after the changes made by Tristan Colgate on 2004-01-16. Nathelper/rtpproxy works fine on the version from 2004-01-15.
Can you take a look at this Tristan? Maxim?
Thanks,
-- Andres Network Admin http://www.telesip.net
Andres wrote:
Hi Maxim,
I am in the process of testing this new version in our lab with 0.8.13. We have been using the older versions with great success for many months now. But the new version does not work. We are testing with Grandstream and Sipura units. When a Sipura calls another Sipura, the nathelper/rtpproxy fails to insert the proper "Connection Information (c)" in the SDP. Instead of filling in the IP Address of the RTPProxy it just leaves the same address and adds these four characters "\000" to the end which seem to make the other Sipura unhappy because it terminates the call right away with a "488- Not Acceptable" Message.
When a Grandstream is making the call, the same thing happens, with the exception of the four characters. (IP Address in Connection Information (c) is not updated)
The Ports do seem to get changed appropiately by the nathelper/rtpproxy in both cases. But since the IP is not substituted there is no chance of audio being setup properly.
I can send the Ethereal traces if you want.
Let me know what we can do to fix this issue.
Thanks,
Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers
Pls can you tell me where to get the update version. I have just downloaded the latest tarball from CVS 1.4 but it does not appear to be the one updated today.
Thanks, Dinesh
Dinesh Mahbubani The International Marketing Exchange Ltd. Email: dinesh@imelhk.com mailto:dinesh@imelhk.com Tel: (852) 2541-2617 Fax (852) 2543-4537
-----Original Message----- From: serusers-bounces@iptel.org [mailto:serusers-bounces@lists.iptel.org] On Behalf Of Maxim Sobolev Sent: Saturday, January 31, 2004 5:56 PM To: Jan Janak Cc: serusers@lists.iptel.org Subject: Re: [Serusers] New versions of RTP proxy/nathelper commited
Yes, indeed, there was a problem with force_rtp_proxy(). I've just committed a fix (1.38). The problem was that you were trying to use results of one call to ip_addr2a() after another call to that function. Since ip_addr2a() returns pointer to a static internal buffer, it was leading to incorrect results.
-Maxim
Jan Janak wrote:
What change do you mean ? I reviewed and commited some changes on
behalf
of Tristan, so please blame me (and provide me with more details if possible) :-).
Could you make sure that the version before my commit works ?
Jan.
On 30-01 11:14, Andres wrote:
Update...
I have now tested multiple versions of nathelper from January. The problem appears after the changes made by Tristan Colgate on 2004-01-16. Nathelper/rtpproxy works fine on the version from
2004-01-15.
Can you take a look at this Tristan? Maxim?
Thanks,
-- Andres Network Admin http://www.telesip.net
Andres wrote:
Hi Maxim,
I am in the process of testing this new version in our lab with 0.8.13. We have been using the older versions with great success for
many months now. But the new version does not work. We are testing
with Grandstream and Sipura units. When a Sipura calls another Sipura, the nathelper/rtpproxy fails to insert the proper
"Connection
Information (c)" in the SDP. Instead of filling in the IP Address of the RTPProxy it just leaves the same address and adds these four characters "\000" to the end which seem to make the other Sipura unhappy because it terminates the call right away with a "488- Not Acceptable" Message.
When a Grandstream is making the call, the same thing happens, with the exception of the four characters. (IP Address in Connection Information (c) is not updated)
The Ports do seem to get changed appropiately by the nathelper/rtpproxy in both cases. But since the IP is not
substituted
there is no chance of audio being setup properly.
I can send the Ethereal traces if you want.
Let me know what we can do to fix this issue.
Thanks,
Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers
_______________________________________________ Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers
The unstable CVS branch has been updated.
Jan.
On 31-01 19:17, Dinesh wrote:
Pls can you tell me where to get the update version. I have just downloaded the latest tarball from CVS 1.4 but it does not appear to be the one updated today.
Thanks, Dinesh
Dinesh Mahbubani The International Marketing Exchange Ltd. Email: dinesh@imelhk.com mailto:dinesh@imelhk.com Tel: (852) 2541-2617 Fax (852) 2543-4537
-----Original Message----- From: serusers-bounces@iptel.org [mailto:serusers-bounces@lists.iptel.org] On Behalf Of Maxim Sobolev Sent: Saturday, January 31, 2004 5:56 PM To: Jan Janak Cc: serusers@lists.iptel.org Subject: Re: [Serusers] New versions of RTP proxy/nathelper commited
Yes, indeed, there was a problem with force_rtp_proxy(). I've just committed a fix (1.38). The problem was that you were trying to use results of one call to ip_addr2a() after another call to that function. Since ip_addr2a() returns pointer to a static internal buffer, it was leading to incorrect results.
-Maxim
Jan Janak wrote:
What change do you mean ? I reviewed and commited some changes on
behalf
of Tristan, so please blame me (and provide me with more details if possible) :-).
Could you make sure that the version before my commit works ?
Jan.
On 30-01 11:14, Andres wrote:
Update...
I have now tested multiple versions of nathelper from January. The problem appears after the changes made by Tristan Colgate on 2004-01-16. Nathelper/rtpproxy works fine on the version from
2004-01-15.
Can you take a look at this Tristan? Maxim?
Thanks,
-- Andres Network Admin http://www.telesip.net
Andres wrote:
Hi Maxim,
I am in the process of testing this new version in our lab with 0.8.13. We have been using the older versions with great success for
many months now. But the new version does not work. We are testing
with Grandstream and Sipura units. When a Sipura calls another Sipura, the nathelper/rtpproxy fails to insert the proper
"Connection
Information (c)" in the SDP. Instead of filling in the IP Address of the RTPProxy it just leaves the same address and adds these four characters "\000" to the end which seem to make the other Sipura unhappy because it terminates the call right away with a "488- Not Acceptable" Message.
When a Grandstream is making the call, the same thing happens, with the exception of the four characters. (IP Address in Connection Information (c) is not updated)
The Ports do seem to get changed appropiately by the nathelper/rtpproxy in both cases. But since the IP is not
substituted
there is no chance of audio being setup properly.
I can send the Ethereal traces if you want.
Let me know what we can do to fix this issue.
Thanks,
Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers
Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers
Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers
I am sorry to trouble you with this but how do I get to the unstable cvs Branch. I have created an account and tried to browse for it but don't seem to be able to find it. Presently I do not have any CVS software installed on my sys.
If such software is required I would be grateful if you could email me the update.
Thanks, Dinesh
Dinesh Mahbubani The International Marketing Exchange Ltd. Email: dinesh@imelhk.com mailto:dinesh@imelhk.com Tel: (852) 2541-2617 Fax (852) 2543-4537
-----Original Message----- From: Jan Janak [mailto:jan@iptel.org] Sent: Saturday, January 31, 2004 7:21 PM To: Dinesh Cc: serusers@lists.iptel.org Subject: Re: [Serusers] New versions of RTP proxy/nathelper commited
The unstable CVS branch has been updated.
Jan.
On 31-01 19:17, Dinesh wrote:
Pls can you tell me where to get the update version. I have just downloaded the latest tarball from CVS 1.4 but it does not appear to be the one updated today.
Thanks, Dinesh
Dinesh Mahbubani The International Marketing Exchange Ltd. Email: dinesh@imelhk.com mailto:dinesh@imelhk.com Tel: (852) 2541-2617 Fax (852) 2543-4537
-----Original Message----- From: serusers-bounces@iptel.org [mailto:serusers-bounces@lists.iptel.org]
On
Behalf Of Maxim Sobolev Sent: Saturday, January 31, 2004 5:56 PM To: Jan Janak Cc: serusers@lists.iptel.org Subject: Re: [Serusers] New versions of RTP proxy/nathelper commited
Yes, indeed, there was a problem with force_rtp_proxy(). I've just committed a fix (1.38). The problem was that you were trying to use results of one call to ip_addr2a() after another call to that
function.
Since ip_addr2a() returns pointer to a static internal buffer, it was leading to incorrect results.
-Maxim
Jan Janak wrote:
What change do you mean ? I reviewed and commited some changes on
behalf
of Tristan, so please blame me (and provide me with more details if possible) :-).
Could you make sure that the version before my commit works ?
Jan.
On 30-01 11:14, Andres wrote:
Update...
I have now tested multiple versions of nathelper from January. The problem appears after the changes made by Tristan Colgate on 2004-01-16. Nathelper/rtpproxy works fine on the version from
2004-01-15.
Can you take a look at this Tristan? Maxim?
Thanks,
-- Andres Network Admin http://www.telesip.net
Andres wrote:
Hi Maxim,
I am in the process of testing this new version in our lab with 0.8.13. We have been using the older versions with great success
for
many months now. But the new version does not work. We are
testing
with Grandstream and Sipura units. When a Sipura calls another Sipura, the nathelper/rtpproxy fails to insert the proper
"Connection
Information (c)" in the SDP. Instead of filling in the IP Address
of
the RTPProxy it just leaves the same address and adds these four characters "\000" to the end which seem to make the other Sipura unhappy because it terminates the call right away with a "488- Not Acceptable" Message.
When a Grandstream is making the call, the same thing happens, with
the exception of the four characters. (IP Address in Connection Information (c) is not updated)
The Ports do seem to get changed appropiately by the nathelper/rtpproxy in both cases. But since the IP is not
substituted
there is no chance of audio being setup properly.
I can send the Ethereal traces if you want.
Let me know what we can do to fix this issue.
Thanks,
Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers
Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers
Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers
Maxim Sobolev wrote:
Yes, indeed, there was a problem with force_rtp_proxy(). I've just committed a fix (1.38). The problem was that you were trying to use results of one call to ip_addr2a() after another call to that function. Since ip_addr2a() returns pointer to a static internal buffer, it was leading to incorrect results.
-Maxim
Thank you Maxim! It is working properly now with today's nathelper version on CVS.
-
Andres -
Dinesh -
Jan Janak -
Maxim Sobolev