I have a policy called cfengine_automated_execution.cf that verifies cf-agent doesn't pile up, bad things don¹t happen, etc..
If I were to move to a long-running daemon mode, I wouldn't use a supervisory daemon like runit or s6. I'd just drop a cron entry like this: 0 * * * * /var/cfengine/bin/cf-agent -b cfengine_automated_execution Have cfengine fire up cf-exced and put all of your infrastructure in Cfengine itself, instead of depending on other stuff. You can do a lot more process matching and have more flexibility in what you want to administrate if you go the route of using Cfengine to fire up / maintain daemons. On 5/11/12 11:37 PM, "Taj Khattra" <taj.khat...@gmail.com> wrote: >On Fri, May 11, 2012 at 7:31 PM, Mike Svoboda <msvob...@linkedin.com> >wrote: >> Process supervision? What are you trying to use? Runit? > >s6, a close cousin of runit (http://www.skarnet.org/software/s6) > >> You should be able to spawn a process and verify the pid still is >>online. > >yep, that's what s6-supervise tries to do. but in fork-mode, original >cf-execd child process (started by s6-supervise) exits after spawning >its own long-running child cf-execd. _______________________________________________ Help-cfengine mailing list Help-cfengine@cfengine.org https://cfengine.org/mailman/listinfo/help-cfengine
