Monitoring servers with SIM using ilo only

[ehodges]

We would like to monitor our servers without using agents on the OS, and it seems that using SIM to poll iLO would be an ideal way to achieve this.

So, I discovered the system with SIM, which it properly recongnized as a management processor. To test it, I unplugged a power supply. The iLO homepage shows the server's health as degraded, but in SIM it still shows as green, and no events were registered.

I know you can set iLO to send SNMP traps to SIM, but that's not necessarily what we want - it only sends ilo-related alerts, like 3 unauthorized login attempts in a row, the ilo event log is full, etc. Instead, SIM to show the health of the system - e.g. a memory module is bad, a HD failed, a power supply died, etc. Can that be done?

[Martin Smoral]

did you set the SNMP trap destinations to the loop back address and enable the SNMP Pass-through in iLO so that SNMP traps are routed onto the iLO network instead of over the management network?

[ehodges]

I'm trying to do this without installing any agents on the operating system at all, and just rely on iLO to send me health information. So unless I misunderstand your post, I have no SNMP traps to set destinations for.

I did set the SNMP destination on iLO itself to point to the SIM server, but as I said, that only alerts us to iLO-related alerts. We would like health status alerts as well.

[Rob Buxton]

I don't think that is going to work. It's the agents that monitor the hardware. They then send the SNMP Trap. The ILO can be configured to pass through SNMP traps from the OS, but it is not a hardware monitoring device.
So, in order to do the HW monitoring you want you have to install the agents.

In Vmware ESXi environments you add the HP ESXi bundle to achieve the same thing.

[ehodges]

It would be disappointing to find this can't be done. And odd as well, since if I go to the iLO System Status tab, particularly the System Information section, I can see exactly what is wrong with the system.

So iLO can definitely monitor the hardware, it's just a matter of passing that information along to SIM.

HP, if you're listening, this would be a nice feature to have in the future if we can't do it right now...

[Sean Carmen]

any news on this? this is also exactly what I would like to do. it's got to be possible, given that ilo can access all the health information. any input would be welcome..

[BPE]

You need to think about the parts which are really monitored by ILO and which parts can only be monitored by Insight SNMP sub-agent or WBEM provider.
If you use ILO without the agents on the OS side you do not get all the information.
Best example is the storage controller or an HBA. How do you synchronize the read out of the Disk SMART Data?

Power supply failure should be no problem but disk failure would be a problem.

The more important question is do you really use only SIM to monitor the HW?
Many people would like to alarm HP by insight remote support. In this case remote support should support your configuration.

ILO supports a pass-trough function but this requires the insight agents. In this case it should be possible to monitor all the error types. I never found a complete document for the setup. ILO user manual pass-trough option is the best place to start.

[Sean Carmen]

thanks for you input. I am sure you are right that there is no way around the agents if you are looking to do complete monitoring. for a lot of our smaller customers it would be more than sufficient to know if a power supply, fan or disk fails (internal DAS disks show up in ilo from what I saw). all the info is there in ilo. just hard to believe there is no way to get it out of ilo..

[Change_happens]

its good to have monitoring based on iLO. i believe HP is thinking on it, lets see.