SIP testing - Issues?? #9
Comments
|
@tchandler48 thanks for the detailed info! As you might have guessed, the SIP plugin is quite immature right now: video support in particular, for instance, is quite flaky. I only made some quick tests by having the browsers call some extensions on Asterisk, and a few where an external softphone called a registered browser, and it seemed to work fine mostly. I hope your tests will help us make it more stable. Have you any further detail on the segmentation faults? Could you try and replicate the crashing scenarios launching Janus with either gdb or valgrind, in order to understand where exactly the code is breaking? My guess is that it's some race condition, as proper locking is not involved yet in the plugins, but there may be something else hidden somewhere. |
|
attached gdb of segmentation fault.... hope it helps, let me know if you Cheers On Tue, Mar 11, 2014 at 9:57 AM, Lorenzo Miniero
Reading configuration from ./conf/janus.cfg
{ Going to negotiate audio...
[582006710] We have 2 candidates for Stream #1, Component #1 [582006710] Stream #1, Component #1 [582006710] We have 2 candidates for Stream #1, Component #2 [582006710] Stream #1, Component #2
[582006710] Done! Ready to setup remote candidates and send connectivity checks...
Program received signal SIGSEGV, Segmentation fault. (gdb) backtrace |
|
Yes, as I guessed it's some kind of race condition. The call is closed but the related session-with-plugin and/or gateway handle is not valid anymore. I'll investigate the issue and try to fix this in the next commit. Thanks! |
|
No problem. Thank You. I STILL THINK this is a GREAT PRODUCT, and look Cheers On Tue, Mar 11, 2014 at 10:27 AM, Lorenzo Miniero
|
|
Tom, I made some fixes to the code that should help with the segfaults issues: could you try and replicate your scenarios with the latest revision and let me know if that did actually at least fix those? Once I'm sure that's settled I'll try and dig into the SIP issues as well. |
|
Be glad to, may be later today or tonight, but I promise I will get it done. Thank You for all your hard work, and for a GREAT project. Cheers On Thu, Apr 3, 2014 at 10:50 AM, Lorenzo Miniero
|
|
@tchandler48 apologies if this took so long, but I finally found some time to work on the SIP plugin. It should be more stable and work better in general. Could you let me know if you keep on getting the same issues when you find some time to replicate those tests? Thanks! |
|
Will start on it this morning, and report back to you. Thank You On Wed, May 14, 2014 at 8:36 AM, Lorenzo Miniero
|
|
Thanks, keep me posted! |
|
I am having trying registering on the new updates. I have the tried the sip:192.168.1.171:5060 or In the past, the 1st entry worked... Suggestions? Thanks On Wed, May 14, 2014 at 9:13 AM, Lorenzo Miniero
|
|
Hi Tom, In the latest version of the plugin the web interface for the SIP plugin For both the server and username parts you now need the sip prefix, e.g., Server: sip:serverip:5060 since the plugin now doesn't give for granted that your user or the one Let me know if that fixes it,
|
|
Thank you. Got registered, now to do some testing..... Tom c On Wed, May 14, 2014 at 11:13 AM, Lorenzo Miniero
|
|
Update below: Test's of new updates 5/14/2014
SetRemoteDescription." clearing the call. If I 192.168.1.177:36922 for asterisk. Thank you for all the HARD work. Sorry to be bearer of bad news. It may Cheers On Wed, May 14, 2014 at 11:32 AM, Tom Chandler [email protected]:
|
|
2014-05-14 18:59 GMT+02:00 tchandler48 [email protected]:
Weird, I guess this is caused by the fact that Firefox ignores REMB One of the plans for the gateway is to start handling NACKs within the
Actually, the popup should not be on the allow/deny prompt: the Did this happen all the time, or just sometimes? Was it Chrome or Firefox?
Yes, that's an inconsistent state within the plugin, that answers with a
Bad news are way better than no news, and I can only thank you for the
|
|
The popup problem was when I was testing with Chrome on a window 7 box. I I enjoy trying to help, and feel like in the long run, this will be the Keep up the GREAT work!!! Tom c On Wed, May 14, 2014 at 2:42 PM, Lorenzo Miniero
|
|
Additional note: I just tested using Firefix 29.0.1, and it worked GREAT. No seg fault. I Cheers On Wed, May 14, 2014 at 2:49 PM, Tom Chandler [email protected] wrote:
|
|
Please forgive me. I went back and tested using Chrome and the following chrome -> pstn. No seg faults, great audio. So hold off on the bug report, Thank You On Wed, May 14, 2014 at 2:59 PM, Tom Chandler [email protected] wrote:
|
|
Tom, I've made some more bugfixing in the SIP plugin and demo page, which should have made it more robust with respect to the inconsistencies (486) you got in your tests. Hopefully they'll be harder to replicate from now on! |
|
Using Firefox 29.0.1, the sip gateway works correctly. However on Chrome, WebRTC error... "Failed to set remote offer sdp: Session error code: This only happens on Chrome, not Firefox. I did a git pull just before these test and saw the updates that you have Hope this helps....Sorry Tom C On Fri, May 16, 2014 at 8:28 AM, Lorenzo Miniero
|
|
Good catch, I should have fixed this now. We already call createAnswer after setRemoteDescription, but it looks like there was a potential race condition I was not handling in JavaScript, namely when createAnswer was executed and setRemoteDescription had not completed yet. The createAnswer is now in the success callback of setRemoteDescription, so it shouldn't happen anymore. Fingers crossed for good news, now :) |
|
sip phone to chrome is still giving the following error: Using Chrome the only error message that I am getting now is: WebRTC error... "Failed to set remote offer sdp: Session error code: The create answer popup error is now gone. But I am still getting the Thank you for all your hard work..... Tom c On Fri, May 16, 2014 at 9:23 AM, Lorenzo Miniero
|
|
This could be an issue with the negotiated codecs in the SDP, could you 2014-05-16 16:56 GMT+02:00 tchandler48 [email protected]:
|
|
Lorenzo, MY BAD. I had "screwed up" my asterisk. Fixed it and the sip gateway Works great on Chrome or Firefox......... GREAT WORK.........now recording of audio/video on the janus server..... JUST Kidding....... Cheers On Fri, May 16, 2014 at 10:08 AM, Lorenzo Miniero
|
|
Good to know, thanks!
|
I am posting this here. It should be more a mailing list post, but for the time....
I tested Janus-gateway, siptest.html. Any comments or questions, I will try to answer.
Configuration:
Asterisk-11.8.rc2
(with opus/vp8 patch applied)
Janus-gateway
(current svn as of 3/10/2014)
ATA Sip phones
Browser: Chrome Version 33.0.1750.146 m (windows 7)
Chromium ersion 30.0.1599.114 Ubuntu 12.04
Testing the Janus-gateway sip plug-in.
Test #1
All components on the save private network
(192.168.1.xxx)
Test #2
siptest.html -> PSTN (Hangup webpage)
RESULT: Good two way audio, using HANGUP button
on webpage disconnects the call.
Test #3
Chrome -> Chromium
RESULT: Call Connected - Asterisk shows call active with
two channels in use. NO AUDIO in either direction.
Hangup button on either browser has NO affect.
Closing Chromium broswer, Janus segmemtation fault....
Asterisk SIP.conf
[2250] ;ATA sip phone
type=friend
secret=1234
host=dynamic
context=local
qualify=yes
transport=udp
directmedia=no
videosupport=no
disallow=all
allow=ulaw
[8000] ;Chrome browser
secret=1234
context=local
host=dynamic
trustrpid=yes
sendrpid=no
type=friend
qualify=yes
transport=udp,ws
callcounter=yes
icesupport=yes
directmedia=no
disallow=all
allow=ulaw,vp8
[8001] ;Chromium browser
secret=1234
context=local
host=dynamic
trustrpid=yes
sendrpid=no
type=friend
qualify=yes
transport=ws
callcounter=yes
icesupport=yes
directmedia=no
transport=udp,ws
disallow=all
allow=ulaw,vp8
Hope this helps, I will continue to test..
Tom C
The text was updated successfully, but these errors were encountered: