I would like to monitor Oracle asm . I have two nodes, how can I achive, that monitor works always on the one active node?. ther are A/B nodes, first I want to monitor asm on A host, after A host fail, I would like automatic monitor on B.
I’m not familiar with “Oracle asm” specifically, but from your description it
sounds like it is some sort of high-availability resource which may be running
on one machine, or another, but is (almost, hopefully) guaranteed to be
running somewhere.
Most HA resources have a floating IP address, which is how the clients which
actually want to use them connect (without caring which actual machine they
are running on), and this is what you want to tell Icinga to do as well.
For HA resources I always have two types of checks:
-
one set of service and host checks which look at the real machines and make
sure they are running whatever it is they are supposed to be running (such as
the HA management software, for example), but specifically excluding any
floating resources which might be running one machine, or might be on the other -
another set of both service and host checks for the HA resources
themselves, which Icinga is told to look for on the floating IP address (which
would generally have a DNS name as well, so you can also use that)
Then you get a series of results - one set for each “real” machine in the HA
cluster group, plus one set for the floating resources which could be running
anywhere , but you want to make sure is/are running somewhere
So, the basic point is - point Icinga at the same floating IP address as you
would point any client wishing to use the HA resources at.
Antony.