We have a production machine that keeps crashing on regular basis. Our staging environments run fine, but when the application is deployed to production, the production machine crashes after two or three days.

We are unable to find any scenario on staging environment which causes the machine to crash. Is there any way to determine which processes were running when the machine crashed?

This might give an indication to issue with application or machine?

  • do you have a linux install made from scratch or are you using a specific distro? – Lorenzo Von Matterhorn Apr 10 '13 at 15:10
  • 1
    What is the nature of the application; Web? DB? Have you checked the respective logs under /var/log – Steve N Apr 10 '13 at 15:19
  • as @SteveN said, check the common locations for logs for any occurencies that may match your needs. (/var/log , /var/log/messages ) - for logging you have syslogd and klogd packages as the most used. – Lorenzo Von Matterhorn Apr 10 '13 at 15:21
  • I've had gone through similar issue in the past, the way I've solved it, was by running a cronjob which executed periodically commands that would have their output persisted to a logfile, like the output of dmesg -c, ps aux, cat /proc/net/sockstat, among others. Then, everyday I'd compare outputs to look for clues on where to start troubleshooting. That maybe one starting point. – Marcel Apr 10 '13 at 16:06
  • Restart the new application every 24 hours and check if this significantly increases the time-to-crash for production machine. – nik Apr 10 '13 at 16:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.