Re: [OpenSIPS-Users] Opensips Multi instance issue on SIP TLS
Hi Solved the issue. Actually what it was doing the location table is storing the ip and port of the server where register request lands in socket field so when B-Party register request lands on different server it saves the ip and port of that server in location table . what I was doing in my code on invite request i was calling the lookup() function of registrar.so module to check if the B-Party is registered. It is also fetching the socket information through where opensips has to relay the invite request to B-Party. So in my case it was reading the wrong information of socket due to which this error was occurring because it cannot find the socket through which invite has to relay to B-Party What I did I just simply forcefully relayed the invite request to B-Party through that server where invite request lands by calling the force_send_socket(udp:1.2.3.4:5060) right below the lookup() function Thank You Benjamin and Gohar Ahmed for your assistance regards Feroze On Fri, Apr 8, 2016 at 8:27 PM, feroze waris wrote: > Ben > > are you saying log lines printed above listen parameter or above error. > if you are talking about listen parameter then no and if you are talking > about error then yes. This error occurs when proxy tries to relay invite > request to B party > > regards > Feroze > > > > > On Fri, Apr 8, 2016 at 8:04 PM, Benjamin Cropley < > benjamin.crop...@gmail.com> wrote: > >> Feroze, >> >> Are any other lines of logging printed just above it? >> >> Ben >> >> On Fri, Apr 8, 2016 at 3:48 PM, feroze waris >> wrote: >> >>> Ben >>> >>> I have put the address of my interface >>> >>> Regards >>> Feroze >>> >>> On Fri, Apr 8, 2016 at 7:39 PM, Benjamin Cropley < >>> benjamin.crop...@gmail.com> wrote: >>> Feroze Can I just check.. in the listen variable have you put the address of your interface or 1.2.3.4? You need to put your address.. not 1.2.3.4 On Fri, Apr 8, 2016 at 3:30 PM, feroze waris >>> > wrote: > Hi ben > > As you said i have enabled the listen=tls:1.2.3.4:5061 on receiving > proxy as well. Now i am getting the following error in this case > > Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:msg_send: send() for > proto 3 failed > Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:t_forward_nonack: > sending request failed > Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:w_t_relay: > t_forward_nonack failed > > > regards > Feroze > > On Fri, Apr 8, 2016 at 1:15 PM, Benjamin Cropley < > benjamin.crop...@gmail.com> wrote: > >> Feroze, >> >> I think you misunderstood :) >> >> I was saying, the URI has ;transport=tls, but the receiving server >> probably doesn't have a port open to listen for it.. Don't remove this if >> you want to use TLS, as by using rewrite() you're effectively breaking >> TLS >> ;) >> >> In the documentation it says if you dont specify a listen= variable, >> then it will listen on 'all interfaces'. It doesn't say whether it will >> listen to UDP TCP and TLS. But I would guess it's possibly only UDP and >> TCP. >> >> Therefore, make sure on the servers receiving the packets that you >> specifically set a tls listening interface with listen = tls: >> 1.2.3.4:5061 >> >> Good luck >> Ben >> >> >> On Fri, Apr 8, 2016 at 7:29 AM, feroze waris < >> feroze.chaud...@gmail.com> wrote: >> >>> Hi ben >>> >>> I have checked the traces and you were right that there was >>> transport = tls in RURI when request forwards from load balancer to 2 >>> opensips instances . I have replaced the RURI using rewrite() function >>> and >>> removed the transport parameter but the issue remains and i am getting >>> the >>> same error >>> >>> >>> regards >>> Feroze >>> >>> On Thu, Apr 7, 2016 at 1:09 PM, Benjamin Cropley < >>> benjamin.crop...@gmail.com> wrote: >>> Wild guess here, based on what I've read.. It might be your request URI is explicitly declaring which protocol to use, and OpenSIPS has no interface configured for that protocol.. eg. ;transport=tls is added to the request URI but there's no listen = tls: 1.2.3.4:5061 I know it's tricky to look at a trace for TLS. But have a look at http://wiki.snom.com/FAQ/How_to_decode_TLS_calls_using_wireshark Good luck Ben On Wed, Apr 6, 2016 at 6:24 PM, feroze waris < feroze.chaud...@gmail.com> wrote: > Hi ben > > yes i am load balancing the Register and Invite request to two > proxy servers using dispatcher module but all three servers (i.e load > balncer and two opensips servers) are sharing same opensips database. > This issue is not occurring in U
Re: [OpenSIPS-Users] Opensips Multi instance issue on SIP TLS
Ben are you saying log lines printed above listen parameter or above error. if you are talking about listen parameter then no and if you are talking about error then yes. This error occurs when proxy tries to relay invite request to B party regards Feroze On Fri, Apr 8, 2016 at 8:04 PM, Benjamin Cropley wrote: > Feroze, > > Are any other lines of logging printed just above it? > > Ben > > On Fri, Apr 8, 2016 at 3:48 PM, feroze waris > wrote: > >> Ben >> >> I have put the address of my interface >> >> Regards >> Feroze >> >> On Fri, Apr 8, 2016 at 7:39 PM, Benjamin Cropley < >> benjamin.crop...@gmail.com> wrote: >> >>> Feroze >>> >>> Can I just check.. in the listen variable have you put the address of >>> your interface or 1.2.3.4? >>> >>> You need to put your address.. not 1.2.3.4 >>> >>> On Fri, Apr 8, 2016 at 3:30 PM, feroze waris >>> wrote: >>> Hi ben As you said i have enabled the listen=tls:1.2.3.4:5061 on receiving proxy as well. Now i am getting the following error in this case Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:msg_send: send() for proto 3 failed Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:t_forward_nonack: sending request failed Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:w_t_relay: t_forward_nonack failed regards Feroze On Fri, Apr 8, 2016 at 1:15 PM, Benjamin Cropley < benjamin.crop...@gmail.com> wrote: > Feroze, > > I think you misunderstood :) > > I was saying, the URI has ;transport=tls, but the receiving server > probably doesn't have a port open to listen for it.. Don't remove this if > you want to use TLS, as by using rewrite() you're effectively breaking TLS > ;) > > In the documentation it says if you dont specify a listen= variable, > then it will listen on 'all interfaces'. It doesn't say whether it will > listen to UDP TCP and TLS. But I would guess it's possibly only UDP and > TCP. > > Therefore, make sure on the servers receiving the packets that you > specifically set a tls listening interface with listen = tls: > 1.2.3.4:5061 > > Good luck > Ben > > > On Fri, Apr 8, 2016 at 7:29 AM, feroze waris < > feroze.chaud...@gmail.com> wrote: > >> Hi ben >> >> I have checked the traces and you were right that there was transport >> = tls in RURI when request forwards from load balancer to 2 opensips >> instances . I have replaced the RURI using rewrite() function and removed >> the transport parameter but the issue remains and i am getting the same >> error >> >> >> regards >> Feroze >> >> On Thu, Apr 7, 2016 at 1:09 PM, Benjamin Cropley < >> benjamin.crop...@gmail.com> wrote: >> >>> Wild guess here, based on what I've read.. >>> >>> It might be your request URI is explicitly declaring which protocol >>> to use, and OpenSIPS has no interface configured for that protocol.. eg. >>> ;transport=tls is added to the request URI but there's no listen = tls: >>> 1.2.3.4:5061 >>> >>> I know it's tricky to look at a trace for TLS. But have a look at >>> http://wiki.snom.com/FAQ/How_to_decode_TLS_calls_using_wireshark >>> >>> Good luck >>> Ben >>> >>> On Wed, Apr 6, 2016 at 6:24 PM, feroze waris < >>> feroze.chaud...@gmail.com> wrote: >>> Hi ben yes i am load balancing the Register and Invite request to two proxy servers using dispatcher module but all three servers (i.e load balncer and two opensips servers) are sharing same opensips database. This issue is not occurring in UDP call but on TLS this error occurs. Regards Feroze On Wed, Apr 6, 2016 at 7:34 PM, Benjamin Cropley < benjamin.crop...@gmail.com> wrote: > Feroze, > > Are you load balancing REGISTER *and *INVITE requests to > different nodes? > > If so, then unless you share location information between the > Registrars, then they wont all know where to route INVITEs. > > Ben > > > On Wed, Apr 6, 2016 at 3:20 PM, feroze waris < > feroze.chaud...@gmail.com> wrote: > >> Hello >> >> I have one load balancer connected to 2 opensips server >> Request is balancing through dispatcher module on load balancer >> the problem is when Register request of B party lands on >> different server than that server at which A Party sends invite >> request the >> call drops and i am getting following error >> Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:update_uac_dst: >> failed to fwd to af 2, proto 3 (no corresponding listening socket) >> Apr 6 07:58:31 localhost OPENSIP[1793]:
Re: [OpenSIPS-Users] Opensips Multi instance issue on SIP TLS
Feroze, Are any other lines of logging printed just above it? Ben On Fri, Apr 8, 2016 at 3:48 PM, feroze waris wrote: > Ben > > I have put the address of my interface > > Regards > Feroze > > On Fri, Apr 8, 2016 at 7:39 PM, Benjamin Cropley < > benjamin.crop...@gmail.com> wrote: > >> Feroze >> >> Can I just check.. in the listen variable have you put the address of >> your interface or 1.2.3.4? >> >> You need to put your address.. not 1.2.3.4 >> >> On Fri, Apr 8, 2016 at 3:30 PM, feroze waris >> wrote: >> >>> Hi ben >>> >>> As you said i have enabled the listen=tls:1.2.3.4:5061 on receiving >>> proxy as well. Now i am getting the following error in this case >>> >>> Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:msg_send: send() for >>> proto 3 failed >>> Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:t_forward_nonack: >>> sending request failed >>> Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:w_t_relay: >>> t_forward_nonack failed >>> >>> >>> regards >>> Feroze >>> >>> On Fri, Apr 8, 2016 at 1:15 PM, Benjamin Cropley < >>> benjamin.crop...@gmail.com> wrote: >>> Feroze, I think you misunderstood :) I was saying, the URI has ;transport=tls, but the receiving server probably doesn't have a port open to listen for it.. Don't remove this if you want to use TLS, as by using rewrite() you're effectively breaking TLS ;) In the documentation it says if you dont specify a listen= variable, then it will listen on 'all interfaces'. It doesn't say whether it will listen to UDP TCP and TLS. But I would guess it's possibly only UDP and TCP. Therefore, make sure on the servers receiving the packets that you specifically set a tls listening interface with listen = tls: 1.2.3.4:5061 Good luck Ben On Fri, Apr 8, 2016 at 7:29 AM, feroze waris >>> > wrote: > Hi ben > > I have checked the traces and you were right that there was transport > = tls in RURI when request forwards from load balancer to 2 opensips > instances . I have replaced the RURI using rewrite() function and removed > the transport parameter but the issue remains and i am getting the same > error > > > regards > Feroze > > On Thu, Apr 7, 2016 at 1:09 PM, Benjamin Cropley < > benjamin.crop...@gmail.com> wrote: > >> Wild guess here, based on what I've read.. >> >> It might be your request URI is explicitly declaring which protocol >> to use, and OpenSIPS has no interface configured for that protocol.. eg. >> ;transport=tls is added to the request URI but there's no listen = tls: >> 1.2.3.4:5061 >> >> I know it's tricky to look at a trace for TLS. But have a look at >> http://wiki.snom.com/FAQ/How_to_decode_TLS_calls_using_wireshark >> >> Good luck >> Ben >> >> On Wed, Apr 6, 2016 at 6:24 PM, feroze waris < >> feroze.chaud...@gmail.com> wrote: >> >>> Hi ben >>> >>> yes i am load balancing the Register and Invite request to two proxy >>> servers using dispatcher module but all three servers (i.e load balncer >>> and >>> two opensips servers) are sharing same opensips database. >>> This issue is not occurring in UDP call but on TLS this error >>> occurs. >>> >>> Regards >>> Feroze >>> >>> On Wed, Apr 6, 2016 at 7:34 PM, Benjamin Cropley < >>> benjamin.crop...@gmail.com> wrote: >>> Feroze, Are you load balancing REGISTER *and *INVITE requests to different nodes? If so, then unless you share location information between the Registrars, then they wont all know where to route INVITEs. Ben On Wed, Apr 6, 2016 at 3:20 PM, feroze waris < feroze.chaud...@gmail.com> wrote: > Hello > > I have one load balancer connected to 2 opensips server > Request is balancing through dispatcher module on load balancer > the problem is when Register request of B party lands on different > server than that server at which A Party sends invite request the call > drops and i am getting following error > Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:update_uac_dst: > failed to fwd to af 2, proto 3 (no corresponding listening socket) > Apr 6 07:58:31 localhost OPENSIP[1793]: > ERROR:tm:t_forward_nonack: failure to add branches > Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:w_t_relay: > t_forward_nonack failed > > I am using Openisps version 2.1.1 and using TLS on it > > > Regards > Feroze > > ___ > Users mailing list > Users@lists.opensips.org > http://lists.opensips.org/cgi-bin/mailman/listinfo/users >>>
Re: [OpenSIPS-Users] Opensips Multi instance issue on SIP TLS
Ben I have put the address of my interface Regards Feroze On Fri, Apr 8, 2016 at 7:39 PM, Benjamin Cropley wrote: > Feroze > > Can I just check.. in the listen variable have you put the address of your > interface or 1.2.3.4? > > You need to put your address.. not 1.2.3.4 > > On Fri, Apr 8, 2016 at 3:30 PM, feroze waris > wrote: > >> Hi ben >> >> As you said i have enabled the listen=tls:1.2.3.4:5061 on receiving >> proxy as well. Now i am getting the following error in this case >> >> Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:msg_send: send() for >> proto 3 failed >> Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:t_forward_nonack: >> sending request failed >> Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:w_t_relay: >> t_forward_nonack failed >> >> >> regards >> Feroze >> >> On Fri, Apr 8, 2016 at 1:15 PM, Benjamin Cropley < >> benjamin.crop...@gmail.com> wrote: >> >>> Feroze, >>> >>> I think you misunderstood :) >>> >>> I was saying, the URI has ;transport=tls, but the receiving server >>> probably doesn't have a port open to listen for it.. Don't remove this if >>> you want to use TLS, as by using rewrite() you're effectively breaking TLS >>> ;) >>> >>> In the documentation it says if you dont specify a listen= variable, >>> then it will listen on 'all interfaces'. It doesn't say whether it will >>> listen to UDP TCP and TLS. But I would guess it's possibly only UDP and TCP. >>> >>> Therefore, make sure on the servers receiving the packets that you >>> specifically set a tls listening interface with listen = tls: >>> 1.2.3.4:5061 >>> >>> Good luck >>> Ben >>> >>> >>> On Fri, Apr 8, 2016 at 7:29 AM, feroze waris >>> wrote: >>> Hi ben I have checked the traces and you were right that there was transport = tls in RURI when request forwards from load balancer to 2 opensips instances . I have replaced the RURI using rewrite() function and removed the transport parameter but the issue remains and i am getting the same error regards Feroze On Thu, Apr 7, 2016 at 1:09 PM, Benjamin Cropley < benjamin.crop...@gmail.com> wrote: > Wild guess here, based on what I've read.. > > It might be your request URI is explicitly declaring which protocol to > use, and OpenSIPS has no interface configured for that protocol.. eg. > ;transport=tls is added to the request URI but there's no listen = tls: > 1.2.3.4:5061 > > I know it's tricky to look at a trace for TLS. But have a look at > http://wiki.snom.com/FAQ/How_to_decode_TLS_calls_using_wireshark > > Good luck > Ben > > On Wed, Apr 6, 2016 at 6:24 PM, feroze waris < > feroze.chaud...@gmail.com> wrote: > >> Hi ben >> >> yes i am load balancing the Register and Invite request to two proxy >> servers using dispatcher module but all three servers (i.e load balncer >> and >> two opensips servers) are sharing same opensips database. >> This issue is not occurring in UDP call but on TLS this error occurs. >> >> Regards >> Feroze >> >> On Wed, Apr 6, 2016 at 7:34 PM, Benjamin Cropley < >> benjamin.crop...@gmail.com> wrote: >> >>> Feroze, >>> >>> Are you load balancing REGISTER *and *INVITE requests to different >>> nodes? >>> >>> If so, then unless you share location information between the >>> Registrars, then they wont all know where to route INVITEs. >>> >>> Ben >>> >>> >>> On Wed, Apr 6, 2016 at 3:20 PM, feroze waris < >>> feroze.chaud...@gmail.com> wrote: >>> Hello I have one load balancer connected to 2 opensips server Request is balancing through dispatcher module on load balancer the problem is when Register request of B party lands on different server than that server at which A Party sends invite request the call drops and i am getting following error Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:update_uac_dst: failed to fwd to af 2, proto 3 (no corresponding listening socket) Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:t_forward_nonack: failure to add branches Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:w_t_relay: t_forward_nonack failed I am using Openisps version 2.1.1 and using TLS on it Regards Feroze ___ Users mailing list Users@lists.opensips.org http://lists.opensips.org/cgi-bin/mailman/listinfo/users >>> >>> ___ >>> Users mailing list >>> Users@lists.opensips.org >>> http://lists.opensips.org/cgi-bin/mailman/listinfo/users >>> >>> >> >> ___ >> Users mailing list >> Use
Re: [OpenSIPS-Users] Opensips Multi instance issue on SIP TLS
Feroze Can I just check.. in the listen variable have you put the address of your interface or 1.2.3.4? You need to put your address.. not 1.2.3.4 On Fri, Apr 8, 2016 at 3:30 PM, feroze waris wrote: > Hi ben > > As you said i have enabled the listen=tls:1.2.3.4:5061 on receiving proxy > as well. Now i am getting the following error in this case > > Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:msg_send: send() for > proto 3 failed > Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:t_forward_nonack: > sending request failed > Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:w_t_relay: > t_forward_nonack failed > > > regards > Feroze > > On Fri, Apr 8, 2016 at 1:15 PM, Benjamin Cropley < > benjamin.crop...@gmail.com> wrote: > >> Feroze, >> >> I think you misunderstood :) >> >> I was saying, the URI has ;transport=tls, but the receiving server >> probably doesn't have a port open to listen for it.. Don't remove this if >> you want to use TLS, as by using rewrite() you're effectively breaking TLS >> ;) >> >> In the documentation it says if you dont specify a listen= variable, then >> it will listen on 'all interfaces'. It doesn't say whether it will listen >> to UDP TCP and TLS. But I would guess it's possibly only UDP and TCP. >> >> Therefore, make sure on the servers receiving the packets that you >> specifically set a tls listening interface with listen = tls:1.2.3.4:5061 >> >> Good luck >> Ben >> >> >> On Fri, Apr 8, 2016 at 7:29 AM, feroze waris >> wrote: >> >>> Hi ben >>> >>> I have checked the traces and you were right that there was transport = >>> tls in RURI when request forwards from load balancer to 2 opensips >>> instances . I have replaced the RURI using rewrite() function and removed >>> the transport parameter but the issue remains and i am getting the same >>> error >>> >>> >>> regards >>> Feroze >>> >>> On Thu, Apr 7, 2016 at 1:09 PM, Benjamin Cropley < >>> benjamin.crop...@gmail.com> wrote: >>> Wild guess here, based on what I've read.. It might be your request URI is explicitly declaring which protocol to use, and OpenSIPS has no interface configured for that protocol.. eg. ;transport=tls is added to the request URI but there's no listen = tls: 1.2.3.4:5061 I know it's tricky to look at a trace for TLS. But have a look at http://wiki.snom.com/FAQ/How_to_decode_TLS_calls_using_wireshark Good luck Ben On Wed, Apr 6, 2016 at 6:24 PM, feroze waris >>> > wrote: > Hi ben > > yes i am load balancing the Register and Invite request to two proxy > servers using dispatcher module but all three servers (i.e load balncer > and > two opensips servers) are sharing same opensips database. > This issue is not occurring in UDP call but on TLS this error occurs. > > Regards > Feroze > > On Wed, Apr 6, 2016 at 7:34 PM, Benjamin Cropley < > benjamin.crop...@gmail.com> wrote: > >> Feroze, >> >> Are you load balancing REGISTER *and *INVITE requests to different >> nodes? >> >> If so, then unless you share location information between the >> Registrars, then they wont all know where to route INVITEs. >> >> Ben >> >> >> On Wed, Apr 6, 2016 at 3:20 PM, feroze waris < >> feroze.chaud...@gmail.com> wrote: >> >>> Hello >>> >>> I have one load balancer connected to 2 opensips server >>> Request is balancing through dispatcher module on load balancer >>> the problem is when Register request of B party lands on different >>> server than that server at which A Party sends invite request the call >>> drops and i am getting following error >>> Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:update_uac_dst: >>> failed to fwd to af 2, proto 3 (no corresponding listening socket) >>> Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:t_forward_nonack: >>> failure to add branches >>> Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:w_t_relay: >>> t_forward_nonack failed >>> >>> I am using Openisps version 2.1.1 and using TLS on it >>> >>> >>> Regards >>> Feroze >>> >>> ___ >>> Users mailing list >>> Users@lists.opensips.org >>> http://lists.opensips.org/cgi-bin/mailman/listinfo/users >>> >>> >> >> ___ >> Users mailing list >> Users@lists.opensips.org >> http://lists.opensips.org/cgi-bin/mailman/listinfo/users >> >> > > ___ > Users mailing list > Users@lists.opensips.org > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > ___ Users mailing list Users@lists.opensips.org http://lists.opensips.org/cgi-bin/mailman/listinfo/users >>> >>> _
Re: [OpenSIPS-Users] Opensips Multi instance issue on SIP TLS
Hi ben As you said i have enabled the listen=tls:1.2.3.4:5061 on receiving proxy as well. Now i am getting the following error in this case Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:msg_send: send() for proto 3 failed Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:t_forward_nonack: sending request failed Apr 8 14:08:12 localhost OPENSIPS[623]: ERROR:tm:w_t_relay: t_forward_nonack failed regards Feroze On Fri, Apr 8, 2016 at 1:15 PM, Benjamin Cropley wrote: > Feroze, > > I think you misunderstood :) > > I was saying, the URI has ;transport=tls, but the receiving server > probably doesn't have a port open to listen for it.. Don't remove this if > you want to use TLS, as by using rewrite() you're effectively breaking TLS > ;) > > In the documentation it says if you dont specify a listen= variable, then > it will listen on 'all interfaces'. It doesn't say whether it will listen > to UDP TCP and TLS. But I would guess it's possibly only UDP and TCP. > > Therefore, make sure on the servers receiving the packets that you > specifically set a tls listening interface with listen = tls:1.2.3.4:5061 > > Good luck > Ben > > > On Fri, Apr 8, 2016 at 7:29 AM, feroze waris > wrote: > >> Hi ben >> >> I have checked the traces and you were right that there was transport = >> tls in RURI when request forwards from load balancer to 2 opensips >> instances . I have replaced the RURI using rewrite() function and removed >> the transport parameter but the issue remains and i am getting the same >> error >> >> >> regards >> Feroze >> >> On Thu, Apr 7, 2016 at 1:09 PM, Benjamin Cropley < >> benjamin.crop...@gmail.com> wrote: >> >>> Wild guess here, based on what I've read.. >>> >>> It might be your request URI is explicitly declaring which protocol to >>> use, and OpenSIPS has no interface configured for that protocol.. eg. >>> ;transport=tls is added to the request URI but there's no listen = tls: >>> 1.2.3.4:5061 >>> >>> I know it's tricky to look at a trace for TLS. But have a look at >>> http://wiki.snom.com/FAQ/How_to_decode_TLS_calls_using_wireshark >>> >>> Good luck >>> Ben >>> >>> On Wed, Apr 6, 2016 at 6:24 PM, feroze waris >>> wrote: >>> Hi ben yes i am load balancing the Register and Invite request to two proxy servers using dispatcher module but all three servers (i.e load balncer and two opensips servers) are sharing same opensips database. This issue is not occurring in UDP call but on TLS this error occurs. Regards Feroze On Wed, Apr 6, 2016 at 7:34 PM, Benjamin Cropley < benjamin.crop...@gmail.com> wrote: > Feroze, > > Are you load balancing REGISTER *and *INVITE requests to different > nodes? > > If so, then unless you share location information between the > Registrars, then they wont all know where to route INVITEs. > > Ben > > > On Wed, Apr 6, 2016 at 3:20 PM, feroze waris < > feroze.chaud...@gmail.com> wrote: > >> Hello >> >> I have one load balancer connected to 2 opensips server >> Request is balancing through dispatcher module on load balancer >> the problem is when Register request of B party lands on different >> server than that server at which A Party sends invite request the call >> drops and i am getting following error >> Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:update_uac_dst: >> failed to fwd to af 2, proto 3 (no corresponding listening socket) >> Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:t_forward_nonack: >> failure to add branches >> Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:w_t_relay: >> t_forward_nonack failed >> >> I am using Openisps version 2.1.1 and using TLS on it >> >> >> Regards >> Feroze >> >> ___ >> Users mailing list >> Users@lists.opensips.org >> http://lists.opensips.org/cgi-bin/mailman/listinfo/users >> >> > > ___ > Users mailing list > Users@lists.opensips.org > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > ___ Users mailing list Users@lists.opensips.org http://lists.opensips.org/cgi-bin/mailman/listinfo/users >>> >>> ___ >>> Users mailing list >>> Users@lists.opensips.org >>> http://lists.opensips.org/cgi-bin/mailman/listinfo/users >>> >>> >> >> ___ >> Users mailing list >> Users@lists.opensips.org >> http://lists.opensips.org/cgi-bin/mailman/listinfo/users >> >> > > ___ > Users mailing list > Users@lists.opensips.org > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > ___ Users mailing list Users@lists.opensips.org http
Re: [OpenSIPS-Users] Opensips Multi instance issue on SIP TLS
Feroze, I think you misunderstood :) I was saying, the URI has ;transport=tls, but the receiving server probably doesn't have a port open to listen for it.. Don't remove this if you want to use TLS, as by using rewrite() you're effectively breaking TLS ;) In the documentation it says if you dont specify a listen= variable, then it will listen on 'all interfaces'. It doesn't say whether it will listen to UDP TCP and TLS. But I would guess it's possibly only UDP and TCP. Therefore, make sure on the servers receiving the packets that you specifically set a tls listening interface with listen = tls:1.2.3.4:5061 Good luck Ben On Fri, Apr 8, 2016 at 7:29 AM, feroze waris wrote: > Hi ben > > I have checked the traces and you were right that there was transport = > tls in RURI when request forwards from load balancer to 2 opensips > instances . I have replaced the RURI using rewrite() function and removed > the transport parameter but the issue remains and i am getting the same > error > > > regards > Feroze > > On Thu, Apr 7, 2016 at 1:09 PM, Benjamin Cropley < > benjamin.crop...@gmail.com> wrote: > >> Wild guess here, based on what I've read.. >> >> It might be your request URI is explicitly declaring which protocol to >> use, and OpenSIPS has no interface configured for that protocol.. eg. >> ;transport=tls is added to the request URI but there's no listen = tls: >> 1.2.3.4:5061 >> >> I know it's tricky to look at a trace for TLS. But have a look at >> http://wiki.snom.com/FAQ/How_to_decode_TLS_calls_using_wireshark >> >> Good luck >> Ben >> >> On Wed, Apr 6, 2016 at 6:24 PM, feroze waris >> wrote: >> >>> Hi ben >>> >>> yes i am load balancing the Register and Invite request to two proxy >>> servers using dispatcher module but all three servers (i.e load balncer and >>> two opensips servers) are sharing same opensips database. >>> This issue is not occurring in UDP call but on TLS this error occurs. >>> >>> Regards >>> Feroze >>> >>> On Wed, Apr 6, 2016 at 7:34 PM, Benjamin Cropley < >>> benjamin.crop...@gmail.com> wrote: >>> Feroze, Are you load balancing REGISTER *and *INVITE requests to different nodes? If so, then unless you share location information between the Registrars, then they wont all know where to route INVITEs. Ben On Wed, Apr 6, 2016 at 3:20 PM, feroze waris >>> > wrote: > Hello > > I have one load balancer connected to 2 opensips server > Request is balancing through dispatcher module on load balancer > the problem is when Register request of B party lands on different > server than that server at which A Party sends invite request the call > drops and i am getting following error > Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:update_uac_dst: > failed to fwd to af 2, proto 3 (no corresponding listening socket) > Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:t_forward_nonack: > failure to add branches > Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:w_t_relay: > t_forward_nonack failed > > I am using Openisps version 2.1.1 and using TLS on it > > > Regards > Feroze > > ___ > Users mailing list > Users@lists.opensips.org > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > ___ Users mailing list Users@lists.opensips.org http://lists.opensips.org/cgi-bin/mailman/listinfo/users >>> >>> ___ >>> Users mailing list >>> Users@lists.opensips.org >>> http://lists.opensips.org/cgi-bin/mailman/listinfo/users >>> >>> >> >> ___ >> Users mailing list >> Users@lists.opensips.org >> http://lists.opensips.org/cgi-bin/mailman/listinfo/users >> >> > > ___ > Users mailing list > Users@lists.opensips.org > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > ___ Users mailing list Users@lists.opensips.org http://lists.opensips.org/cgi-bin/mailman/listinfo/users
Re: [OpenSIPS-Users] Opensips Multi instance issue on SIP TLS
Hi ben I have checked the traces and you were right that there was transport = tls in RURI when request forwards from load balancer to 2 opensips instances . I have replaced the RURI using rewrite() function and removed the transport parameter but the issue remains and i am getting the same error regards Feroze On Thu, Apr 7, 2016 at 1:09 PM, Benjamin Cropley wrote: > Wild guess here, based on what I've read.. > > It might be your request URI is explicitly declaring which protocol to > use, and OpenSIPS has no interface configured for that protocol.. eg. > ;transport=tls is added to the request URI but there's no listen = tls: > 1.2.3.4:5061 > > I know it's tricky to look at a trace for TLS. But have a look at > http://wiki.snom.com/FAQ/How_to_decode_TLS_calls_using_wireshark > > Good luck > Ben > > On Wed, Apr 6, 2016 at 6:24 PM, feroze waris > wrote: > >> Hi ben >> >> yes i am load balancing the Register and Invite request to two proxy >> servers using dispatcher module but all three servers (i.e load balncer and >> two opensips servers) are sharing same opensips database. >> This issue is not occurring in UDP call but on TLS this error occurs. >> >> Regards >> Feroze >> >> On Wed, Apr 6, 2016 at 7:34 PM, Benjamin Cropley < >> benjamin.crop...@gmail.com> wrote: >> >>> Feroze, >>> >>> Are you load balancing REGISTER *and *INVITE requests to different >>> nodes? >>> >>> If so, then unless you share location information between the >>> Registrars, then they wont all know where to route INVITEs. >>> >>> Ben >>> >>> >>> On Wed, Apr 6, 2016 at 3:20 PM, feroze waris >>> wrote: >>> Hello I have one load balancer connected to 2 opensips server Request is balancing through dispatcher module on load balancer the problem is when Register request of B party lands on different server than that server at which A Party sends invite request the call drops and i am getting following error Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:update_uac_dst: failed to fwd to af 2, proto 3 (no corresponding listening socket) Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:t_forward_nonack: failure to add branches Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:w_t_relay: t_forward_nonack failed I am using Openisps version 2.1.1 and using TLS on it Regards Feroze ___ Users mailing list Users@lists.opensips.org http://lists.opensips.org/cgi-bin/mailman/listinfo/users >>> >>> ___ >>> Users mailing list >>> Users@lists.opensips.org >>> http://lists.opensips.org/cgi-bin/mailman/listinfo/users >>> >>> >> >> ___ >> Users mailing list >> Users@lists.opensips.org >> http://lists.opensips.org/cgi-bin/mailman/listinfo/users >> >> > > ___ > Users mailing list > Users@lists.opensips.org > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > ___ Users mailing list Users@lists.opensips.org http://lists.opensips.org/cgi-bin/mailman/listinfo/users
Re: [OpenSIPS-Users] Opensips Multi instance issue on SIP TLS
Wild guess here, based on what I've read.. It might be your request URI is explicitly declaring which protocol to use, and OpenSIPS has no interface configured for that protocol.. eg. ;transport=tls is added to the request URI but there's no listen = tls: 1.2.3.4:5061 I know it's tricky to look at a trace for TLS. But have a look at http://wiki.snom.com/FAQ/How_to_decode_TLS_calls_using_wireshark Good luck Ben On Wed, Apr 6, 2016 at 6:24 PM, feroze waris wrote: > Hi ben > > yes i am load balancing the Register and Invite request to two proxy > servers using dispatcher module but all three servers (i.e load balncer and > two opensips servers) are sharing same opensips database. > This issue is not occurring in UDP call but on TLS this error occurs. > > Regards > Feroze > > On Wed, Apr 6, 2016 at 7:34 PM, Benjamin Cropley < > benjamin.crop...@gmail.com> wrote: > >> Feroze, >> >> Are you load balancing REGISTER *and *INVITE requests to different >> nodes? >> >> If so, then unless you share location information between the Registrars, >> then they wont all know where to route INVITEs. >> >> Ben >> >> >> On Wed, Apr 6, 2016 at 3:20 PM, feroze waris >> wrote: >> >>> Hello >>> >>> I have one load balancer connected to 2 opensips server >>> Request is balancing through dispatcher module on load balancer >>> the problem is when Register request of B party lands on different >>> server than that server at which A Party sends invite request the call >>> drops and i am getting following error >>> Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:update_uac_dst: failed >>> to fwd to af 2, proto 3 (no corresponding listening socket) >>> Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:t_forward_nonack: >>> failure to add branches >>> Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:w_t_relay: >>> t_forward_nonack failed >>> >>> I am using Openisps version 2.1.1 and using TLS on it >>> >>> >>> Regards >>> Feroze >>> >>> ___ >>> Users mailing list >>> Users@lists.opensips.org >>> http://lists.opensips.org/cgi-bin/mailman/listinfo/users >>> >>> >> >> ___ >> Users mailing list >> Users@lists.opensips.org >> http://lists.opensips.org/cgi-bin/mailman/listinfo/users >> >> > > ___ > Users mailing list > Users@lists.opensips.org > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > ___ Users mailing list Users@lists.opensips.org http://lists.opensips.org/cgi-bin/mailman/listinfo/users
Re: [OpenSIPS-Users] Opensips Multi instance issue on SIP TLS
Hi ben yes i am load balancing the Register and Invite request to two proxy servers using dispatcher module but all three servers (i.e load balncer and two opensips servers) are sharing same opensips database. This issue is not occurring in UDP call but on TLS this error occurs. Regards Feroze On Wed, Apr 6, 2016 at 7:34 PM, Benjamin Cropley wrote: > Feroze, > > Are you load balancing REGISTER *and *INVITE requests to different nodes? > > If so, then unless you share location information between the Registrars, > then they wont all know where to route INVITEs. > > Ben > > > On Wed, Apr 6, 2016 at 3:20 PM, feroze waris > wrote: > >> Hello >> >> I have one load balancer connected to 2 opensips server >> Request is balancing through dispatcher module on load balancer >> the problem is when Register request of B party lands on different server >> than that server at which A Party sends invite request the call drops and i >> am getting following error >> Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:update_uac_dst: failed >> to fwd to af 2, proto 3 (no corresponding listening socket) >> Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:t_forward_nonack: >> failure to add branches >> Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:w_t_relay: >> t_forward_nonack failed >> >> I am using Openisps version 2.1.1 and using TLS on it >> >> >> Regards >> Feroze >> >> ___ >> Users mailing list >> Users@lists.opensips.org >> http://lists.opensips.org/cgi-bin/mailman/listinfo/users >> >> > > ___ > Users mailing list > Users@lists.opensips.org > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > ___ Users mailing list Users@lists.opensips.org http://lists.opensips.org/cgi-bin/mailman/listinfo/users
Re: [OpenSIPS-Users] Opensips Multi instance issue on SIP TLS
Feroze, Are you load balancing REGISTER *and *INVITE requests to different nodes? If so, then unless you share location information between the Registrars, then they wont all know where to route INVITEs. Ben On Wed, Apr 6, 2016 at 3:20 PM, feroze waris wrote: > Hello > > I have one load balancer connected to 2 opensips server > Request is balancing through dispatcher module on load balancer > the problem is when Register request of B party lands on different server > than that server at which A Party sends invite request the call drops and i > am getting following error > Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:update_uac_dst: failed > to fwd to af 2, proto 3 (no corresponding listening socket) > Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:t_forward_nonack: > failure to add branches > Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:w_t_relay: > t_forward_nonack failed > > I am using Openisps version 2.1.1 and using TLS on it > > > Regards > Feroze > > ___ > Users mailing list > Users@lists.opensips.org > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > ___ Users mailing list Users@lists.opensips.org http://lists.opensips.org/cgi-bin/mailman/listinfo/users
[OpenSIPS-Users] Opensips Multi instance issue on SIP TLS
Hello I have one load balancer connected to 2 opensips server Request is balancing through dispatcher module on load balancer the problem is when Register request of B party lands on different server than that server at which A Party sends invite request the call drops and i am getting following error Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:update_uac_dst: failed to fwd to af 2, proto 3 (no corresponding listening socket) Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:t_forward_nonack: failure to add branches Apr 6 07:58:31 localhost OPENSIP[1793]: ERROR:tm:w_t_relay: t_forward_nonack failed I am using Openisps version 2.1.1 and using TLS on it Regards Feroze ___ Users mailing list Users@lists.opensips.org http://lists.opensips.org/cgi-bin/mailman/listinfo/users