low-level discovery

[herta]

With assistance of mib2template.pl (https://github.com/Akint/mib2template), I created a template for our SAN. The template defines a number of discovery rules e.g. to find the LUNs and collect info on hit rates and such.

Only, this works a bit too well. As the items use SNMP OIDs ending in {#SNMPINDEX}, Zabbix tries to collect close to 29,000 items, because it doesn't just collect information about the LUNs that are actually created on the SAN, but about every possible LUN that could be created.

So I tried to filter out the LUNs that we do use.

In Administration -> General -> Regular expression, I created a regexp with

Name: LUNs for discovery
Expression: ^(102|1810|1812|1813|1814|1815|1816|1817|1819)$
Expression type: Result is TRUE
Case sensitive: unticked

I then added a filter to the Discovery rule with

Macro: {#SNMPINDEX}
Regexp: @LUNs for discovery

And that effectively limits the SNMP queries to the list in the regular expression.
Which is nice, were it not that we have more than one SAN, and of course, they aren't exact replica's of each other.

The only idea I can come up with at this time is to add a second SNMP community string to each SAN (which is unique to that SAN) and create one discovery rule per SAN, using that second SNMP community string and a regexp specific for that SAN.

Is that the way to go, or does anyone have a better idea?

Kind regards,

Herta

[axlrod]

I don't know what MIB you have for that SAN, but is there no snmp entry listing the LUN as active or inactive? Or maybe a counter thats higher than 0..

You can discover with any key you want ..