Solved: What data does SNMP return and what is the best way to explore it?

LondonCisco · ‎10-28-2011

Hi All

I have some queries about using SNMP on Cisco devices.

1. What is the best way to get an idea of what data MIBs can return?

I have entered the commmand "show snmp mib". However, the number of MIBs generated is huge!

2. Can you enter a command on the Cisco device itself that returns snmp data?

e.g. if I want to see what data a MIB entry returns, can I interrogate it on the Cisco device itself rather than having to rely on network management software?

The reason I ask is that we are currently looking for Network Management Software. However, I have no idea what data SNMP can return, therefore don't know if what the Management Software offers is comprehensive or not.

Any suggestions on how best to get an idea the data SNMP can generate plus any other suggestions are very welcome!

Thanks

John

Marvin Rhoads · ‎10-28-2011

Generally speaking, snmp can query about any parameter of the system. An enterprise class device has almost all conceivable parameters instrumented thus. A device's inventory (chassis, cards, ports, power supplies, software image, etc.) and the state of its interfaces and traffic they are carrying are among the primary ones.

You can query a device's snmp variables remotely "by hand" using an open source distribution of a tool like snmpwalk. You need to know what you're looking for - i.e., the exact oid (object ID) string or else you get a long reply like you alluded to. It also helps to have the actual MIB you are querying against locally so that the output you get will be human-readable. Otherwise you'll get long numeric strings (the snmp oid) followed by alphanumeric values which may or may not make sense at first glance depending on the oid being queried. A MIB's purpose is to add that abstraction layer / interpretation to make the output more useful to the operator.

The purpose of an NMS (very generally speaking) is to give an even higher layer of abstraction to organize the querying and reporting of all these various data into a useful system with dashboards, graphs, reports etc.

View solution in original post

simionov.adrian · ‎10-30-2011

You should take the MIB file, you can find them on Cisco Website, and look inside this file with a text editor. Here you will also find some description for every oid. After you have the oid, is like knowing what to ask the device.

User snmpwalk to ask the device before finding the exact OID for your device. The oid from the MIB file usually will miss the last portion, which is specific to device/value. Can be a pointer, an interface id, but snmpwalk will help you to find the oid.

.1.3.6.1.4.1.9.2.1.57.0 this oid for example will give you processor utilization.

Cacti is a good NMS which can help you to setup some graphs if you know the exact OID value. For mass deploiment is not so convenient to use it.

For mass deploiment I used:

rrdtools (for graphs generation)

perl (processing the output)

braa (mass scanner)

For a professional solution I would definitely recommand creating specific tools, it will be easy to integrate them with sms gateways for example, ticketing tools.

View solution in original post

Marvin Rhoads · ‎10-28-2011

Generally speaking, snmp can query about any parameter of the system. An enterprise class device has almost all conceivable parameters instrumented thus. A device's inventory (chassis, cards, ports, power supplies, software image, etc.) and the state of its interfaces and traffic they are carrying are among the primary ones.

You can query a device's snmp variables remotely "by hand" using an open source distribution of a tool like snmpwalk. You need to know what you're looking for - i.e., the exact oid (object ID) string or else you get a long reply like you alluded to. It also helps to have the actual MIB you are querying against locally so that the output you get will be human-readable. Otherwise you'll get long numeric strings (the snmp oid) followed by alphanumeric values which may or may not make sense at first glance depending on the oid being queried. A MIB's purpose is to add that abstraction layer / interpretation to make the output more useful to the operator.

The purpose of an NMS (very generally speaking) is to give an even higher layer of abstraction to organize the querying and reporting of all these various data into a useful system with dashboards, graphs, reports etc.

simionov.adrian · ‎10-30-2011

You should take the MIB file, you can find them on Cisco Website, and look inside this file with a text editor. Here you will also find some description for every oid. After you have the oid, is like knowing what to ask the device.

User snmpwalk to ask the device before finding the exact OID for your device. The oid from the MIB file usually will miss the last portion, which is specific to device/value. Can be a pointer, an interface id, but snmpwalk will help you to find the oid.

.1.3.6.1.4.1.9.2.1.57.0 this oid for example will give you processor utilization.

Cacti is a good NMS which can help you to setup some graphs if you know the exact OID value. For mass deploiment is not so convenient to use it.

For mass deploiment I used:

rrdtools (for graphs generation)

perl (processing the output)

braa (mass scanner)

For a professional solution I would definitely recommand creating specific tools, it will be easy to integrate them with sms gateways for example, ticketing tools.

LondonCisco · ‎11-15-2011

Thanks Guys. Useful info.

John