Question about VMWare ESXi traps

[Sean Murray_1]

Have the latest MIB set in SIM. I usually download from HP as I see any new ones.

I can look at the SNMP traps within the console.

However all I keep getting are "Generic Unregistered" traps.

Some servers are also sending said traps constantly.

Do I need to need to get the MIBs from VMWare instead?

If I compile and add will they affect the ones that came with the HP MIB download?

Thanks

[LGentile]

What type of events are you expecting to see in SIM from ESXi?  Hardware events?  If you are using the embedded HP install for ESXi you should not have to do anyting at all with SIM.  Just discover your ESXi hosts, then run the Subscribe to WBEM events task so SIM will accept them and show hardware status and you should be good to go.  Getting ESXi set up is one of the easier things to do.  There are a few docs on this on both VMware's site and HP's I believe.

If you are not using the HP ESXi install, you can install the components separately, it's just a little more of a hassle.  This will put the CIM providers on the box so SIM can interpret hardware status (and vCenter as well if you use that).

[Sean Murray_1]

Hardware alerts. I can see the status of the systems no problem. All green and good to go.

Just noticed at one point that one had a accelerator issue but I never saw an alert come in for it.

Thanks.

[SwisspostIT]

Hi,

remember that ESXi with HP Image installes some CIM Providers for the events. However these events are (as far as I know) not SNMP based, they're WBEM based!

So check that port 5989 is open between HP SIM Server and the managed systems.

Also check that you have configured subscription of WBEM Events of the managed ESXi Hosts on the HP SIM Server!

Regards,

Ville

[LGentile]

If the issue occurred prior to the discovery in SIM, I do not think the condition will regenerate (meaning after discovery, if the HW status showed the failed status, you won't see an WBEM event for it).

If you have a test system, discover it, then pull out a power supply and see if the event comes in.  I never got a real answer on this - some servers tend to spam events constantly until fixed, where others will simply send one event (SNMP or WBEM) and be done with it until the status of the component changes again.