FAQ

Procuring a Reporter memory dump for Linux

FAQ ID:    FAQ243
Version:    8.0
Status:    Published
Published date:    07/15/2009
Updated:    10/03/2011
 

Answer

This should only be done at the request of Bluecoat support.

  1. Stop reporter.

    1. To do this you execute  a "./bcreporter stopin the etc/init.d  folder.
    2. To make sure it's stoped you execute a "./bcreporter status" in the same folder.
  2. Run reporter in a command window. 

    1. Open a command line window and navigate to the opt/bc/reporter folder, or your installed location.
    2. Execute the "ls -l " command to procure a directory listing. Along with other folders and files, you should see the bcreporter binary, a few journal_xxxx.tst file or two , and a isready.txt file.
    3. Run these commands
      1. ulimit -c unlimited
      2. ./bcreporter
    4. Reporter will now be running, and functional, and you should be seeing it log events to the same terminal window.
    5. If it crashes again, the dump file, called "core",  will be in this folder.  Please upload to this to your SR for analysis by the Bluecoat team.

 

Procuring a dump, when Reporter causes LINUX to go into a High Utilization state:

  1. Stop reporter.

    1. To do this you execute  a "./bcreporter stop"  in the etc/initi.d folder.
    2. To make sure it's stopped you execute a "./bcreporter status" in the same folder.
  2. Run reporter in a command window. 

    1. Open a command line window and navigate to the opt/bc/reporter folder.
    2. Execute the "ls -l " command to procure a directory listing. Along with other folders and files, you should see the bcreporter binary, a few files ending with the .out extension, and a isready.txt file.
    3. Run these commands
      1. ulimit -c unlimited
      2. ./bcreporter
    4. Reporter will now be running, and functional, and you should be seeing it log events to the same terminal window.
    5. If it goes into a high utilization situation again, you need to execute this command in another terminal window, to determine the PID number: 

                         –" ps –aux |grep reporter"   

                6: In the same terminal window, use the number found in the above command to execute this command which will kill the Reporter service, and create a core: 
                         –" kill -segv <PID NUMBER> "
 
General LINUX Notes on taking a core:
  • To setup your LINUX server to permantely provide Memory dumps for all applications, follow these  two steps.
    • 1: Add below 2 lines in "/etc/security/limits.conf"
    • *             hard    core            unlimited
      *             soft    core            unlimited
    • 2: comment out the below line in "/etc/profile" if it is not commented by symbol "#"
    • TIP: In other words, change
    • ulimit -S -c 0 > /dev/null 2>&1
    • to
    • #ulimit -S -c 0 > /dev/null 2>&1
  • The “/proc/sys/kernel/core_pattern” file contains a pattern for the name of the core dump file and defaults to the value “core”.
  • The “/proc/sys/kernel/core_uses_pid” file contains a flag designating whether or not to append a dot followed by the PID of the dumping process.
  • The core_pattern may contain certain % specifications that cause other elements to be included in the name of the core dump file.  
  • Here are the possible % specifications:

%% -- output one '%'

%p -- pid

%u -- uid

%g -- gid

%s -- signal number

%t -- UNIX time of dump

%h -- hostname

%e -- executable filename

 

NOTE1: If the core_pattern file does not include “%p” (the default does not) and the core_uses_pid file contains a non-zero value (something other than ‘0’), then a dot and PID will be appended to the name of the core dump file.

NOTE2: You must be root to configure these things. 

NOTE3: Here is an example to make the core file name to contain executable file name and time of the dump:

echo core.%e.%t > /proc/sys/kernel/core_pattern

NOTE4: More information about the -segv switch can we found here WIKI SITE 
 

NOTE5:Generic informatoin about taking a memory dump on LINUX, can be found here IBM INFO.

 
 
 
 
 


Internal Comments

To debug the dump yourself, you need to execute this command "yum install gdb" but this will only work on a Red Hat registered machine.


Rate this Page

Please take a moment to complete this form to help us better serve you.

Did this document help answer your question?
 
 
If you are finished providing feedback, please click the RATE CONTENT button. Otherwise, please add more detail in the following text box and then click RATE CONTENT.
 
 

Your response will be used to improve our document content.

Ask a Question