OvenMediaEngine | signals.cpp:119 | OME received signal 11 (SIGSEGV), interrupt.

[alex-jhones]

Hello. @getroot
Again the problem has been reproduced. I will attach the error log.

crash_20200805.dump.txt
ovenmediaengine.log.txt

Originally posted by @alex-jhones in #164 (comment)

[getroot]

Thanks for the bug report.
This seems to be a different cause from the previously resolved crash. It seems to be a problem caused by certain transcoding options.
Please share your Server.xml.

[getroot]

I've tested this with an automatic script player for hours (repeat connect, terminate) but haven't been able to reproduce the same problem. Please provide the following information:

1. Server.xml
2. OBS Settings
3. How to reproduce
4. Can you test again after clean build of the current master branch?

[alex-jhones]

1. Server.xml
   Server.xml.txt

2. OBS Settings
   

3. How to reproduce
   After setting the attached obs and Server.xml, the problem is reproduced within 5 minutes.
   And obs publishes 5 different streams simultaneously.

crash_20200811.dump.txt
ovenmediaengine.log.txt

4. Can you test again after clean build of the current master branch?
   The test was done on the last master branch.

[getroot]

Unfortunately I've been testing very repeatedly for a long time using your Server.xml and six OBSs, but I haven't been able to reproduce the crash.

But I could find in your logs that our kernel version is different. (Other than that, everything was the same.)

[mine]
OvenMediaEngine v0.10.7 (v0.10.6-6-g9516cfc) is started on [OME-dev] (Linux x86_64-5.3.0-59-generic, #53~18.04.1-Ubuntu SMP Thu Jun 4 14:58:26 UTC 2020)

[Your log]
OvenMediaEngine v0.10.7 (v0.10.6-6-g9516cfc) is started on [hicys76] (Linux x86_64-4.15.0-109-generic, #110-Ubuntu SMP Tue Jun 23 02:39:32 UTC 2020)

Could you please check if the problem reproduces after updating the Linux kernel? Thanks for your contribution.

[Keukhan]

Hello, @alex-jhones .

I tried testing in the same kernel environment, configuration and concurrent publishing. but, the crash doesn't occur.

Did you modify the code? Or the same as the master branch source code.

If there is a modified code, can you send me the 'diff' information?

I hope the problemis solved quickly. thank you.

[Testing Kernel Version]
Linux ubuntu-6cpu-16gb-sg-sin1 4.15.0-109-generic #110-Ubuntu SMP Tue Jun 23 02:39:32 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux

[alex-jhones]

@getroot,@kwonkeukhan I ran it on the latest version of ubuntu desktop and didn't modify ome's code either.

The problem is reproduced even in the oven player that has not been modified. In order to exclude environmental differences in the experiment, the server/client experimented without modifying the code.

The failure was reproduced by configuring another server to quickly identify the problem. We have sent you the temporarily configured server connection information by e-mail. I hope it helps to reproduce the issue by connecting obs.

[getroot]

We've done a lot of tests so far.
And I could see that it works well with 10 OBS for 5 days when jemalloc is excluded. So I temporarily removed jemalloc from the default library. (We will do more tests on this.)

Please test with the latest master branch(clean build required) and let me know if your problem is resolved.

[alex-jhones]

@getroot The current modified version doesn't crash as before. However, after about 1 to 2 hours, there is no audio, and this symptom is intermittent. (It is not easy to reproduce.)

[getroot]

Audio output issues are not related to this patch. Because this patch just doesn't use JEMALLOC
If the audio interruption is a bug, please register this as a separate issue. (But I think this is most likely an encoder problem.)

I will close this issue. Thanks for the contribution.