I'm trying to model this reachability chain (fixed width font better to 
view this):

               , - my-fritz-box -- 8.8.8.8 -- internet -- gmail.com
               |                `- 8.8.4.4 -'          `- 
my-web-server-via-some-proxy
my-web-server -'                                                   |
               `---------------------------------------------------'

Ideally I want to do this within WATO, not vi.

Looking first at the my-fritz-box-to-internet bit: the unreachability
rules should be:

R1) if *both* 8.8.8.8 and 8.8.4.4 are unreachable then everything to
   the right is unreachable

R2) for other hosts, simply: if one of them is unreachable then
   everything to the right is unreachable

I can see two solutions to do this:

S1) defining a cluster (cluster "internet" is unreachable
    if *all* its member nodes (8.8.8.8 and 8.8.4.4) are unreachable).

S2) using host-parent-child relationships (internet is unreachable if
    *all* its parents (8.8.8.8 and 8.8.4.4) are unreachable).

So my first questions:

Q1) is one of these solutions better, and if so, then why?

Looking next at the 
my-web-server-to-internet-and-to-my-web-server-via-some-proxy
bit: I check access to my website via a public http proxy: the unreachability 
logic rules should be:

R3) if *either* internet is unreachable *or* my-web-server is unreachable then
   my-web-server-via-some-proxy is unreachable.

But I don't find any way in WATO to describe this 
"if-either-parent-is-unreachable-
then-host-is-unreachable" logic.

So my next questions:

Q2) should my-web-server-via-some-proxy be a host or a service? Why?

Q3) is it possible to model these "unreachable if *either* A or B unreachable"
   relationships?

I did plenty of googling and reading Check_MK docs, but didn't manage to
find anything. Any hints appreciated!

Alexis
_______________________________________________
omd-users mailing list
[email protected]
Manage your subscription or unsubscribe
https://lists.mathias-kettner.de/cgi-bin/mailman/listinfo/omd-users

Reply via email to