Ok, so I'm building templates for what I'll refer to here as "application groups"; the items are processes (proc.num[yada]) and ports (net.tcp.listen[yada]) that define the application group. Some of these application groups include an instance of mysql (which may be listening on a different port), and/or an instance of httpd (might be a different port). But these are also Linux boxes, so they're already using the Template_Linux which also has entries for these processes & services. On my test host, I have disabled those triggers that are part of Template_Linux, so I won't get spurious triggers firing.
To my mind, I don't ever want to see such a conflict as an error (although a separate command line utility to detect such issues would be welcome). I want those items and triggers and whatever else to simply be added. Think of it this way: if I define a trigger (and enable it), it is only because it is imperative that it fires when its conditions are met.
I'm guessing the "proper" solution is to copy Template_Linux to T_Linux, hack out the bits I don't want, and delete Template_Linux from the host(s) in question, and add T_Linux in its place. I don't like it, though.
To my mind, I don't ever want to see such a conflict as an error (although a separate command line utility to detect such issues would be welcome). I want those items and triggers and whatever else to simply be added. Think of it this way: if I define a trigger (and enable it), it is only because it is imperative that it fires when its conditions are met.
I'm guessing the "proper" solution is to copy Template_Linux to T_Linux, hack out the bits I don't want, and delete Template_Linux from the host(s) in question, and add T_Linux in its place. I don't like it, though.
Comment