How to Investigate a core dump

When investigating a functional problem with the Sangoma SBC (appliance and/or software version), it is a good idea to check if a 'core' file was created around the time of issue.

How to check if the SBC created a 'core' file

To check if your SBC has created a 'core' file, open an SSH session to the SBC.

An SSH session can be created by using either of the two methods:

1. use a terminal program (such as 'putty' on a windows box) and use your log in credentials with the IP address of the SBC

   Open

   

2.use the SBC'swebSSH feature which is built into thewebgui.

*note, you must use httpsto use the webssh

Open

If a 'core' file was created, it will be located in the /var/core directory.
To navigate to this directory simply type the following in your SSH session:
->cd /var/core

Now type 'ls -al' to see if any 'core' files are present. This will show you any core files plus the time the core file was created.
If the time of the core file was weeks or months before the time of the issue, it is most likely not of interestto your case which means you can stopreadign this wiki page :)

The following is a screenshot of an SBC that had a core file created on December 8 2015:

Now lets go ahead and open this 'core' file for details

Opening a core file to investigatedetails

1. To investigate the previous core file created copy the complete core file name.
   i.e. "core.nsc.11.19604.1449583109"
   
   

2. Type the following string in your SSH session. (You can be in any directory when you type this)
   -> gdb /usr/local/nsc/bin/nsc <complete core file name>
   -> press <enter>

   i.e

3. Now type 'backtrace' at the end of the output.
   i.e.

    

4. You should see more output now. Below is an example output after typing 'backtrace'

    

5. The output from the above 'backtrace' is what needs to be investigated. This is what Sangoma Support and Sangoma R&D want to see when investigating a 'core' file.
   Please send this output along with trouble ticket with Sangoma.
   If more output is required, you can type the following command:
   -> thread apply all bt

   This output will provide the details of all threads, not just the thread that caused the 'core' file.

    

   Note: core file investigation is required to be performed on the SBC. If you try to copy the core fileoff the SBC andinvesgitate on another system, the investigation will not be useful.