Discovery configuration in NNMi not only requires methods
for discovery boosting (seeds and auto-discovery rules), but
limitations as well. So NNM has boundaries where discovery should be
stopped. This section describes the discovery filters used in NNMi.
What are discovery filters?
Discovery filters are rules within NNMi, which define discovery
boundaries and control so that NNMi wouldn't discover anything outside
these boundaries. In other words, that's a tool to limit NNMi's
Why do we need discovery filters?
Proper discovery planning can save you a lot of:
- Money: Save expenses on purchasing licenses, not important nodes excluding from discovery.
This can be done by dealing with unimportant nodes later on, trying to
solve issues based on received alarms, or trying to delete unimportant
nodes later on.
- System performance: The fewer nodes you manage, the better is the performance.
If you choose list-based discovery, then the only way to control the amount of discovered devices are seeds.
If you choose rule-based auto-discovery, then dealing with rules
makes you a manager in selection of choosing which devices need to be
discovered. Mistakes made in rules can make you run out of license
limits very quickly, or can decrease system performance that is
overloading your system. On the other hand, mistakes can make you
discover too few nodes and leave a part of your network undiscovered.
The following screenshot represents the Auto-discovery configuration window:
Auto-discovery rule ordering number affects the discovery in the following ways:
- IP address ranges: If the device complies with two or more auto-discovery rules, the rule with the lowest ordering number is applied.
- System Object ID (sysObjID) ranges:
If no IP address has complied with the auto-discovery rules, then
system object ID setting applies to all auto-discovery rules with
higher numbers. In the case that the IP address is included in the
auto-discovery rule, then system objects ID is applied only within that
particular auto-discovery rule.
Starting from NNMi 8.10, you can use ping sweep to help the
auto-discovery process discover nodes on listed IP address ranges. It
is very handy on networks that NNMi does not have control of, for
example, ISP networks.
Ping sweep works only if your subnet mask is 16 bit or smaller, that is, 192.168.*.*.
It is good practice to enable ping sweep on small network ranges. If
you have a firewall between NNMi and a node which you plan to monitor,
please check that the firewall is not blocking ping sweep packets.
Discovery seeds may be used in auto-discovery rules. It's recommended to provide at least one seed per rule.
You can also set IP address ranges to exclude an addresses from discovery.
There might be overlapping auto-discovery rules. That is, range A is configured as the range to be discovered and range B is configured as the range not to be discovered:
Range C is covered by both rules, but will comply only with the rule which has the lowest rule-ordering number.