Query multiple SNMP OID's no index

[oxo-oxo]

Good it's working: I thought it was going to be an intresting thread (which is why I asked for more info to keep it alive).

While muli DS template magic was being discussed I was trying this non-indexed area:

Code: Select all

$ vi resource/snmp_queries/tcp.xml
<interface>
        <name>Get tcp Statistics</name>
        <oid_index>.1.3.6.1.2.1.6.1</oid_index>
        <index_order>tcpRtoAlgorithm</index_order>
        <index_order_type>numeric</index_order_type>
        <index_title_format>|chosen_order_field|</index_title_format>
        <fields>
                <tcpRtoAlgorithm>
                        <name>Index</name>
                        <method>walk</method>
                        <direction>input</direction>
                        <oid>.1.3.6.1.2.1.6.1</oid>
                        <source>value</source>
                </tcpRtoAlgorithm>
                <tcpInSegs>
                        <name>tcpInSegs</name>
                        <method>walk</method>
                        <direction>output</direction>
                        <oid>.1.3.6.1.2.1.6.10</oid>
                        <source>value</source>
                </tcpInSegs>
                <tcpOutSegs>
                        <name>tcpOutSegs</name>
                        <method>walk</method>
                        <direction>output</direction>
                        <oid>.1.3.6.1.2.1.6.11</oid>
                        <source>value</source>
                </tcpOutSegs>
        </fields>
</interface>

[TheWitness]

Owen,

You do have that effect on people. Big time segway.

TheWitness

[Michalis]

Now, considering i have an indexed query that returns many rows and values.
How can i make cacti put all those data into a single datasource and a single RRD file?
Is this even possible with indexed queries?

Thanks

[TheWitness]

Now I am confused. Do you want to SUM or AVG the data in all the rows? What is your intention?

TheWitness

[Michalis]

I want to run an SNMP query on an indexed OID.
Then i will get 5 different values on different OID's.
I want to put the different OID's in a single datasource so when i go to create graphs for a device and select a data query i dont have 5 distinct DS being created for a device.

For example i want to query the temperature of a server. The server has 5 sensors with each sensor having its OID.
I can run an indexed query at a specific OID and get the 5 values. When i run the query i get "5 Items 5 Rows" in Cacti which is expected. Then i can see all 5 sensors and can select one to graph for a device. However i dont want to have 5 datasources and 5 graphs per device. I want to have a single datasource with the 5 values and graph that.

This is the query i am using :

Code: Select all

<query>
	<name>Get Compaq Temps</name>
	<description>Queries a Compaq host for Temps and Thresholds (Health MIB)</description>
	<oid_index>.1.3.6.1.4.1.232.6.2.6.8.1.2</oid_index>
	<index_order_type>numeric</index_order_type> 	
	<fields>
		<cpqHeTempIndex>
			<name>Index</name>
			<method>walk</method>
			<source>value</source>
			<direction>input</direction>
			<oid>.1.3.6.1.4.1.232.6.2.6.8.1.4.0</oid>
		</cpqHeTempIndex>
		<tempCPU1>
			<name>CPU1 Temperature</name>
			<method>walk</method>
			<source>value</source>
			<direction>output</direction>
			<oid>.1.3.6.1.4.1.232.6.2.6.8.1.4.0.1</oid>
		</tempCPU1>
		<tempCPU2>
			<name>CPU2 Temperature</name>
			<method>walk</method>
			<source>value</source>
			<direction>output</direction>
			<oid>.1.3.6.1.4.1.232.6.2.6.8.1.4.0.2</oid>
		</tempCPU2>
		<tempBoard>
			<name>Motherboard Temperature</name>
			<method>walk</method>
			<source>value</source>
			<direction>output</direction>
			<oid>.1.3.6.1.4.1.232.6.2.6.8.1.4.0.4</oid>
		</tempBoard>		
		<tempPSU>
			<name>Powersupply Temperature</name>
			<method>walk</method>
			<source>value</source>
			<direction>output</direction>
			<oid>.1.3.6.1.4.1.232.6.2.6.8.1.4.0.5</oid>
		</tempPSU>				
	</fields>
</query>

SNMP Walk .1.3.6.1.4.1.232.6.2.6.8.1.4.0 gives me these temperatures

Code: Select all

SNMPv2-SMI::enterprises.232.6.2.6.8.1.4.0.1 = INTEGER: 32
SNMPv2-SMI::enterprises.232.6.2.6.8.1.4.0.2 = INTEGER: 44
SNMPv2-SMI::enterprises.232.6.2.6.8.1.4.0.3 = INTEGER: 40
SNMPv2-SMI::enterprises.232.6.2.6.8.1.4.0.4 = INTEGER: 38
SNMPv2-SMI::enterprises.232.6.2.6.8.1.4.0.5 = INTEGER: 67

And SNMP walk the index .1.3.6.1.4.1.232.6.2.6.8.1.2 gives :

Code: Select all

SNMPv2-SMI::enterprises.232.6.2.6.8.1.2.0.1 = INTEGER: 1
SNMPv2-SMI::enterprises.232.6.2.6.8.1.2.0.2 = INTEGER: 2
SNMPv2-SMI::enterprises.232.6.2.6.8.1.2.0.3 = INTEGER: 3
SNMPv2-SMI::enterprises.232.6.2.6.8.1.2.0.4 = INTEGER: 4
SNMPv2-SMI::enterprises.232.6.2.6.8.1.2.0.5 = INTEGER: 5

I hope it makes some sense

[gandalf]

Please check out the DAta Query "SNMP - Interface Statistics" and the Data Template for "Errors and Discards". That shows an example how to achieve your goal
Reinhard