Hi,

I hope the correct solution to the problem is to add RR parameters on sequential requests.

The attached patch restores vsf=, vst= and did=.

Dragons ahead, so far it works for me.

Regards,

Kristian Høgh

Uni-tel A/S

On Friday 27 November 2015 14:07:08 Kristian F. Høgh wrote:

> Hi again,

>

> I looked at the wrong request, when I wrote it was a patch we applied in-house.

> (I looked at Route: header received from UA, not RR send by kamailio.)

>

> RR on sequential requests have no did/vsf, using vanilla kamailio version 4.2.5 as written below.

> If required I can test later using master. (I tested ealier, but it included our patches, which doesn't make any difference on 4.2.5)

>

> The questions still remain.

> Should kamailio add did/vsf on sequential requests, when I call record_route()?

> Should I set dlg_match_mode to 1? (I use 0)

> Is it wrong to call record_route() from sequential requests?

>

> Regards,

> Kristian.

>

>

> On Friday 27 November 2015 12:54:50 Kristian F. Høgh wrote:

> Hi list,

>

> Record-route on sequential requests doesn't have did/vsf parm.

>

> On initial INVITE, I call record_route() and set dialog flag.

> kamailio adds the following RR

> Record-Route: <sip:ww.xx.yy.zz;lr;ftag=15af612df;vsf=AAAA....;did=4a8.3ca2>

>

> UAC sends a re-INVITE, containing the following route:

> Route: <sip:178.21.251.54;lr;ftag=15af612df;vsf=AAAA....;did=4a8.3ca2>

> I call loose_route() followed by record-route(), and kamailio adds the following RR:

> Record-Route: <sip:ww.xx.yy.zz;lr;ftag=15af612df>

>

> I've got a phone which updates the routeset on sequential requests (which I don't think it should) and the did-matching fails.

> Should kamailio add did/vsf on sequential requests, when I call record_route()?

> Should I set dlg_match_mode to 1? (I use 0)

> Is it wrong to call record_route() from sequential requests?

>

> Which way is the best to resolve the problem?

>

> Regards,

> Kristian Høgh

> Uni-tel A/S