2.0 Shaper not recognizing ports to shape voip, custom definitions



  • I do this same thing with 1.2.3 with no problem, in that I write my own port assignments for bandwidth shaping in 1.2.3 by altering the traffic shaper inc/xml file to replace the name and the port ranges for rtp for my sip deployment.

    I did the same thing with 2.0, (tried multiple versions through 24th nov build) and find that the queue status no longer shows my rtp in the voip queue, but rather in the default queue. It does however show a minimal amount of traffic in the queue (sip signalling probably). A packet capture shows the media on the ports specified by me, so I am confused why this does not get observed properly.

    Is there a way to add a specific provider to the shaper script and have it included in the distribution?


  • Rebel Alliance Developer Netgate

    Are your phones going out directly, or do you have a local PBX?

    Sometimes the way Asterisk and other PBX systems flag outgoing packets they can fall into outgoing ACK queues instead of the VOIP queue, but I haven't seen them fall into the default queue before.

    Not unless another rule was matching the traffic first, or the traffic really wasn't matching the VOIP rules in some way.



  • The sip server is local, but it declines to match packets whether i use the local IP address or not. I simply "restated" the sip server name (replace asterisk) and changed the RTP range. When a call goes to the trunk provider, it matches the RTP range but still goes in the default queue.

    Since all of my media is anchored by my sip server, I simply added a LAN rule to match all traffic from it and place it in the voip queue, but I've never had to do this before.

    I also think some of the queue's are not being created using the wizard (which i defaulted the customization by updating and removing the shaper and running the default asterisk shaper script).



  • @grazman:

    The sip server is local, but it declines to match packets whether i use the local IP address or not. I simply "restated" the sip server name (replace asterisk) and changed the RTP range. When a call goes to the trunk provider, it matches the RTP range but still goes in the default queue.

    Since all of my media is anchored by my sip server, I simply added a LAN rule to match all traffic from it and place it in the voip queue, but I've never had to do this before.

    I also think some of the queue's are not being created using the wizard (which i defaulted the customization by updating and removing the shaper and running the default asterisk shaper script).

    This is still happening through the Dec. 19 snapshot. Also, the custom port definitions that were in the original VOIP shaping rules are no longer in there (i.e. udp 10000-20000 for asterisk rtp), and matching by ip address does not work. If adding a lan rule to match the traffic, it only recognizes one direction, adding a second rule for the other direction does not work and ceases to recognize any traffic in the desired queue.


Locked