ESXI 5.0.0, 62380 on a DL380 G7

[Jan Soska]

Hello Donald,

what is the destination for your snmp traps? do you use tool like HP SIM (System Insight Manger) or other similar?

Jan

[Donald J Wood]

HP SIM was never considered for our environment due to size. We tried it and it couldn't handle all the servers we have. Our enterprise is huge that reports to a BMC trap catcher. Us having SIM or not having SIM is out of scope for this discussion anyway.

As you may know, the sender of the traps would be the ESXI hosts and that's where the messages originate. So back to my original question, why send traps from both enterprise IDs?

[Matti_Kurkela]

My guess is that VMware wants to provide a uniform, hardware-vendor-neutral interface for basic hardware monitoring: "If it's running VMware ESX, you can always get basic hardware health information this way, no matter if it's HP, Fujitsu, IBM, Dell or something else."

But the old vendor-specific MIBs also exist because a) they may offer more specific information in some cases, and b) they allow the VMware systems to be monitored also by existing vendor-specific solutions, like HP SIM... because some people already have those and will want to keep using them.

Apparently VMware has understood that in large enterprises, hardware monitoring is often integrated with a complex enterprise monitoring/management system, and trying to force the customer to make major changes to that by vendor decision would usually have the customer looking for other, more cooperative vendors instead.

[Donald J Wood]

Corporations like the one I work for have computer and software system from many different companies that send alerts for many different reasons. In large companies an enterprise monitoring solution that provides a single view into alerting for all of our systems is a must have because most companies now have a centralized monitoring team. HP SIM doesn't provide that. End of story.

The redundancy crated by letting OS manufactures have hooks into the hardware alerting is very confusing. In our environment with our HP systems, it uses to be done one way, with the HP PSP and some SNMP configs. Now we get two traps for the same thing. I think we should be using the HP traps because we’re using HP servers. But, it is leaving other people at my company questioning which one we should be using and if we don't change will we be left behind the eight ball at some point.