Outgoing VoIP traffic ends up in wrong queue …
-
I used the shaper wizard in 1.2.3 to setup traffic shaping. It's a very very simple setup i'd made inside the wizard. I dont understand how to embedd Screenshots inside here, so i try to explain the problem via text:
1. Priorize VoIP Traffic on IP 192.168.3.188 guaranteed bandwith 128 Kbit/s (this overrides provider settings in wizard)
2. No Penalty
3. Activate lower priority settings with CatchAll (no specific protocols and no bandwith limit)
4. No priorization of network game traffic
5. No priorization of other network protocolsAfter that the wizard generates me queues for in/out VoIP traffic, and in/out for p2p CatchAll.
During a phone call the incoming packets from outside are in the proper queue … but my outgoing packets (upload) ends up in the p2pCatchAll Upload queue. I can't understand why? The ruless are configured to the ip of the phone ... no tos/qos or other weird settings! Maybe it's of interest or not, but this is my setup:
Local LAN with VoIP Phone on 192.168.3.188 --- [PFSENSE/IPSEC] … IPSec tunnel in internet ... [CISCO-ASA/IPSEC] –- Local LAN with VoIP PBX
My problem is, that the upload of the IP 192.168.3.188 is in the wrong CatchAll p2p queue ... and is therefor not priorized!!! I'd read several Postings but no ones gave me the right hint.
Thx for your help in advance.
-
Just a guess - the phone may be setting ToS bit(s) that cause this. Try putting the VOIP phone's IP as the target in the wizard's rule?
-
I disabled the tos bit 3 on the phone. But my shaper rule is set up anyway to "Tos: don't care" and only restrictet to the IP address of my phone. I woul'd say: normally a supereasy setup … but it don't works ... nothing changes.
-
Hmmm, odd, I have seen VOIP traffic go to the ACK queue because of ToS bits, but if you put the phone's internal IP address in that field it should work. What do your shaper rules look like?
-
I'd figured out how to insert screenshots … here are my rules:
And here are my queues:
I can't understand why my phone upstream isn't feeded in the proper queue but in qwandef …
The VoIP downstream is in the right qVOIPDown queue … but this isn't much important than in other direction ...
Here are is the edit rule screen:I don't know. For me, this is looking good, but unfortunaly it doesn't work :/
I tried also to modify the outgoing rule using destination 192.168.1.32 (VoIP Server) instead of source 192.168.3.188. Nothing changes. Anyway wrong queue.
Could there be a issue cause it is going through a ipsec tunnel? But i cant setup the rule from LAN to IPSEC … only LAN to WAN is possible!?!? -
IPSEC? Hmmm, I think that may be the issue. You may want to search the forum for ipsec shaping. I'm not sure this will work as is, but I don't use ipsec, so I can't comment there.
-
I'd "solved" it with a workaround. I configured a IP alias which contains all clients. Then i had set up the shaper only with the penalize box pointing to the alias. The allowed bandwith was [MeasuredBandwith] - [150 Kbit per Phone]. The remaining bandwith is for the phones …
... not a really charming solution, and it's working only with my very easy setup!