[Linux-HA] Confusion about MailTo RA and monitoring
dk at in-telegence.net
Mon Jul 16 03:47:24 MDT 2007
> This is incorrect, or at least incomplete.
> Even if no "monitor" operation is configured, the cluster will do
> startup probes to find out whether the resource is running somewhere and
> in what state.
> If those calls blindly return "no error" (ie, 0, success), the cluster
> will go into multiple active recovery mode, which is not usually what
> you want.
> You _do_ want monitor to return the proper state of your resource.
> The safe return code for those pseudo-resources like MailTo is
> "OCF_NOT_RUNNING", but that should be done sparingly. It also means that
> they'll not pass the ocf-tester script, but such is life.
Thank you very much. Good to know as I will have to write my first
custom OCF scripts soon.
More information about the Linux-HA