We are using the community Chef apache2 cookbook with great success; my question revolves around using the "notifies" notification mechanism to tell Apache 2 of configuration changes.

We run a custom watchdog utility to restart the whole app, including Apache, if it become unresponsive. This watchdog utility queries a particular status URL to determine availability. The problem is, if the timing is bad and the watchdog does a health check just as Chef is restarting Apache, then the whole app stack is bounced.

Is there a way I can make the "notifies" mechanism shutdown the watchdog when it restarts Apache at the end of the run? Then start the watchdog up again once Apache has been restarted.




Another way would be to make watchdog take an interval of time before being sure that service is indeed down. So it should test the URL 10 times in span of 3 mins to say for sure the service is indeed down. This 10 fail results should be continuous.

| improve this answer | |

To answer my own question - what I did was make the watchdog application "Chef aware". That is, when Chef is running it must not restart anything even if it notices something is down. That prevents spurious restarts.

| improve this answer | |

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.