| Author |
|
jmatthewsr
Junior
Joined: April 09 2008
Posts: 40
|
| Posted: May 22 2008 at 12:51pm | IP Logged
|
|
|
YES, I have read the doc notes on multithreaded usage of Terminate Call :-). There are no mutexes in my app, notifications are pushed into a worker thread from the lanscape engine to the worker thread via a lockless queue. After a call is in the SipInCall state, the app calls TerminateCall() but never returns from that call. I've tried both asynchronous version and synchronous versions of the TerminateCall method.
any ideas?
thanks,
-justin
|
| Back to Top |
|
| |
support
Administrator
Joined: January 26 2005
Location: United States
Posts: 1666
|
| Posted: May 22 2008 at 3:57pm | IP Logged
|
|
|
Hi Justin,
Hmmm….. What trial version are you running?
Support
|
| Back to Top |
|
| |
jmatthewsr
Junior
Joined: April 09 2008
Posts: 40
|
| Posted: May 22 2008 at 5:01pm | IP Logged
|
|
|
Code:
LME pMediaEngineVersion: 5.12.8.1
pDspMicroCodeVersion: 5.12.8.1
pAudioEngineVersion: 608.0032.0109
pAudioMixerVersion: 500.0000.0147
pIvrStreamingMediaVersion: 302.0005.0021
pSipStackVersion: 301.0010.0026
pRtpStackVersion: 304.0015.0162
|
|
|
|
| Back to Top |
|
| |
support
Administrator
Joined: January 26 2005
Location: United States
Posts: 1666
|
| Posted: May 23 2008 at 7:01am | IP Logged
|
|
|
Hi Justin,
Thanks for the version info.
A while back we had a report of a TerminateCall() API deadlock that was fixed in a past release. Its been quite a while since then too. Is it possible we missed something? Sure.
Since we received your post yesterday, we ran test code here that specifically hammers on the TerminateCall() API procedure in an effort to try and dead lock that API. The test code ran all night. Hmmm…..
The trial media engine is v5.12.8.1 (your current image). Internally, we are at 5.12.8.4. Its possible that what you are seeing no longer exists. Can’t be sure unless we can reproduce it here. Give us as much of a description as to what you think is going on to cause the deadlock. Better yet, if you can get a test app together that experienced the issue, then we will be able to fix it immediately.
Support
|
| Back to Top |
|
| |
jmatthewsr
Junior
Joined: April 09 2008
Posts: 40
|
| Posted: May 23 2008 at 7:27am | IP Logged
|
|
|
I've uploaded some logs "lanscape-logs-deadlock.zip" that shows this problem occurring. It seems to be happening when I shut down the bulk call. A call to ConferenceLine fails and then when TerminateCall is called it locks. See ivrcaconf.log entry 7964. Thread 2540 is the only worker thread in my app that asynchronously processes non-immediate events. 192.168.2.21 is the load testers, 192.168.2.41 is where lanscape is running. 192.168.2.6 is a cisco 7960 handset that I use to test to see if the system is running. After the deadlock occurs I just hear rining from the 7960. If I break into the app using the debugger, the call to TerminateCall never returns.
-justin
|
| Back to Top |
|
| |
support
Administrator
Joined: January 26 2005
Location: United States
Posts: 1666
|
| Posted: May 23 2008 at 7:37am | IP Logged
|
|
|
Good job. We will take a look....
|
| Back to Top |
|
| |
jmatthewsr
Junior
Joined: April 09 2008
Posts: 40
|
| Posted: May 23 2008 at 7:42am | IP Logged
|
|
|
Also, there is another device in there on the 2.41 address (a sip server). So the call flow is:
simulator -> our sip server -> Lanscape conference engine
192.168.2.21 -> 192.168.2.41:5060 -> 192.168.2.41:5061
|
| Back to Top |
|
| |
support
Administrator
Joined: January 26 2005
Location: United States
Posts: 1666
|
| Posted: May 23 2008 at 11:02am | IP Logged
|
|
|
Justin,
Don't see the 7964 entry in the ivrcaconf.log log. are we not loking properly????
Support
|
| Back to Top |
|
| |
jmatthewsr
Junior
Joined: April 09 2008
Posts: 40
|
| Posted: May 23 2008 at 11:05am | IP Logged
|
|
|
sorry, 7164. After the call to ConferenceLine fails at that line, the next call is TerminateCall (not explicetly shown in the logs)
|
| Back to Top |
|
| |
support
Administrator
Joined: January 26 2005
Location: United States
Posts: 1666
|
| Posted: May 24 2008 at 6:53am | IP Logged
|
|
|
Justin,
Is the machine you used for testing a multi-core?
Support
|
| Back to Top |
|
| |
jmatthewsr
Junior
Joined: April 09 2008
Posts: 40
|
| Posted: May 24 2008 at 6:55am | IP Logged
|
|
|
I reproduced this on 2 machines, both are dual core:
Pentium D 2.8Ghz
Core 2 Duo T7500 2.2Ghz (laptop)
|
| Back to Top |
|
| |
support
Administrator
Joined: January 26 2005
Location: United States
Posts: 1666
|
| Posted: May 24 2008 at 7:02am | IP Logged
|
|
|
Justin,
Thanks for the response. We are suspecting a possible multi-core threading related issue. Not sure yet. Will post when we get more info.
Are you working this morning?? I though we were the only ones crazy/stupid enough to be working this early on a Saturday morning. :)
Support
|
| Back to Top |
|
| |
jmatthewsr
Junior
Joined: April 09 2008
Posts: 40
|
| Posted: May 28 2008 at 2:55pm | IP Logged
|
|
|
Any update to this issue? We are anxious to wrap up the testing and delpoy the software. Have you been able to reproduce this with your updated 5.12 release? Is it possible for us to get the latest 5.12 release to see if the problem is fixed in that release?
thanks,
-justin
|
| Back to Top |
|
| |
support
Administrator
Joined: January 26 2005
Location: United States
Posts: 1666
|
| Posted: May 29 2008 at 9:14am | IP Logged
|
|
|
Hi Justin,
We have not seen the deadlock here using the latest release. We could get you an “engineering release” of the current version to test but that will have to be later this week.
Unfortunately we are in the process of replacing network infrastructure and development machines this week which will hamper our normal workflow. We are trying to make this changeover have as little impact as possible but you know how it goes… something benign and simple seems to take days not hours.
Support
|
| Back to Top |
|
| |
jmatthewsr
Junior
Joined: April 09 2008
Posts: 40
|
| Posted: May 29 2008 at 9:33am | IP Logged
|
|
|
Ok, please send me the engineering release as soon as you can.
thanks,
-justin
|
| Back to Top |
|
| |
support
Administrator
Joined: January 26 2005
Location: United States
Posts: 1666
|
| Posted: June 02 2008 at 2:46pm | IP Logged
|
|
|
Justin,
Thanks for waiting for this post.
We just completed the hardware and software updates from last week. As usual, the updates took a few days.
We just put into place some new Intel quad core development and test machines. We want to conduct a bit more deadlock testing before we have you access an updated version. With these faster quad core hosts, we will be able to really stress the thread timing of the media engine. Our hope is two fold: Either we see the deadlocks you have described or they are not experienced (due to already being fixed). Once we run our tests, we will get you a test image. That should be possible this week.
We will repost to this thread when we have completed running all our normal QA tests.
Support
|
| Back to Top |
|
| |
support
Administrator
Joined: January 26 2005
Location: United States
Posts: 1666
|
| Posted: July 11 2008 at 9:12am | IP Logged
|
|
|
Hi Justin,
We deeply regret the long delay in getting back to this post. Testing and QAing of the media engine always seems to take way more time than what we want. Its just the way it is. We hope you are still interested in our products.
We have v5.12.8.6 “engineering release” of the media engine ready for you if you still want to evaluate and test against. This version of the media engine has gone through the highest degree of multi-core testing to this date. The good news is that we have uncovered multiple and very subtle multi-core issues as the result of our testing on quad and octal core based host machines. In addition to that, multiple multi-core related performance improvements are included in this release that improve call handling performance.
It is important that you continue your eval and test using this new product image. Please let us know if you want FTP access to this engineering release.
Thanks Justin,
Support
|
| Back to Top |
|
| |
jmatthewsr
Junior
Joined: April 09 2008
Posts: 40
|
| Posted: July 11 2008 at 1:17pm | IP Logged
|
|
|
We are no longer considering using the Lanscape products.
Thank you for your efforts and help during the evaluation. Please close any open requests.
Thanks,
-justin
|
| Back to Top |
|
| |
support
Administrator
Joined: January 26 2005
Location: United States
Posts: 1666
|
| Posted: July 14 2008 at 3:28pm | IP Logged
|
|
|
Hi Justin,
Thanks for the final feedback. Maybe next time…
Kind regards,
Support
|
| Back to Top |
|
| |
Active Topics 
Member List 
Search 
Help 
Register 
Login
LanScape Support Forum -> LanScape VOIP Media Engine™ - Pre-Sales Technical Support
Topic: TerminateCall not returning (deadlock?)
