SNMP and VMFS

I have a monitoring software running which monitors all sort of server / services of 100s of server (alerts via SMS etc.)

Does anyone know how I can monitor the VMFS partition via SNMP ? I don't mean sending traps (push) but rather pull the information for each header (also ones which aren't connected to a vcenter) ?

Find more posts tagged with

Comments

bertieb

Do you NEED to use SNMP for this? It's possible to poll the hosts for similar information using the API (e.g. in my case I have a perl script running in nagios, basically). You can also do this with powershell (vSphere PowerCLI). I'd suspect these may be easier as there is quite a lot of information out there now.

Also, theres a half decent perl script on the Nagios Exchange, 'check_esx3' I *think* that may be of use.

HTH!
MikeB.

jibbajabba

bertieb wrote: »

Do you NEED to use SNMP for this?

Yes, like I say - we run our own software doing this and we did run perl scripts on other test hosts and it was quite CPU extensive so I rather don't run any cron'ed scripts on the hosts.

bertieb

Gomjaba wrote: »

Yes, like I say - we run our own software doing this and we did run perl scripts on other test hosts and it was quite CPU extensive so I rather don't run any cron'ed scripts on the hosts.

Fair enough. I'm lucky to have a perl wizard who took the check_esx3 script (amongst others) and tidied it up considerably - it is pretty intensive and inefficient in its base state.

I can't help with the SNMP query though, good luck in resolving that one

jibbajabba

bertieb wrote: »

Fair enough. I'm lucky to have a perl wizard who took the check_esx3 script (amongst others) and tidied it up considerably - it is pretty intensive and inefficient in its base state.

I can't help with the SNMP query though, good luck in resolving that one

Thanks, the biggest challenge at the moment is to make the MIB work with our software and get the right OID

astorrs

You'll have to manually create shell scripts on the ESX servers to expose those kinds of results via SNMP and then edit /etc/snmp/snmpd.conf to export them as unused OID's.

I'm with bertieb though, I'd use the API, even if all you do is pull it all in and then expose it from a central server through SNMP.