Cron Process Keeps Being Killed in EngineYard

Question

I keep getting "FAILURE Process cron: is down" alerts in my engineyard application. A few minutes later I get a follow-up alert mentioning that the process is back up again. Has anyone ever noticed this issue before?

Answer 1

It turns out that engineyard is expecting a cron entry to touch a file every minute:

# This and the remote_file for cron_nanny go together
# Cron touches a file every minute
cron 'touch cron-check' do
  minute  '*'
  hour    '*'
  day     '*'
  month   '*'
  weekday '*'
  command 'touch /tmp/cron-check'
end

A compannion script called cron_nanny (/engineyard/bin/cron_nanny) checks the modification time of the touched file and if it is older than 120 seconds kills the crond process and restarts.

I happend to be deleting all cron entries in my custom chef-recipes, which caused the touch cron job to be removed, so every 120 seconds or so the cron_nanny script would restart the process.