0

On a CentOS 6.5 server I am running a very simple HTTP listener with Sinatra that exists simply to accept POST requests and pass the payload content of the POST request to another service.

I often find that the process with the PID of the Sinatra service no longer exists, with no explanation in the logs.

I have debug-level logging in the Sinatra code and still have no clues towards what is happening on the box to cause the process to die.

What can I do on the Linux side of this problem to monitor this particular process and determine what is causing it be killed?

Andrew Schulman
  • 8,811
  • 21
  • 32
  • 47
asdoylejr
  • 281
  • 1
  • 2
  • 7
  • `/var/log/messages` is a good start – Wesley Jun 19 '15 at 16:12
  • 3
    This question is being voted for closure because the author does not show a level of technical understanding or appropriate due diligence in researching the topic that the community judges as being a minimum barrier to participate. – Wesley Jun 19 '15 at 16:12
  • 1
    `/var/log/messages` was my first stop in trying to solve this issue and there was nothing regarding the Sinatra service's particular PID. – asdoylejr Jun 19 '15 at 16:15

1 Answers1

-1

You could try to do

strace -p pid-number >& /tmp/process.txt

Once the process has died you can study the contents of /tmp/process.txt

pid-number should be replaced with the numeric pid if the process you want to study.

Henrik Carlqvist
  • 492
  • 2
  • 5