kupjones

The vSphere Basic Admin indicates that the HOSTD process on ESX 4 exposes a new set of MIBs that appear to be fully populated with the data that was originally populated into the HOST-RESOURCES MIB. (see page 55 in the doc www.vmware.com/.../vsp_40_admin_guide.pdf). An alternative given by VMware is to enable Net-SNMP and…

Casey, on your last statement about ESX using net-SNMP, this is correct for 3.5 but incorrect for 4.0. On 3.5 VMware leveraged net-SNMP running in the COS and populated the MIBs that are part of the standard net-SNMP agent. With 4.0 that all changed. The actual data is being populated via the SNMP support built into HOSTd…

My, you certainly did - my apologies. All I can say at this point since you are running the beta is -- it's a beta. The facts are and remain unchanged, these being: * VMWare broke SNMP support * VMWare documentation provides incorrect information, that being ALL previous SNMP data was brought forward into HOSTd * VMWare is…

Correct, the current Orion is an SNMP only NMS and the SNMP interface as it is implemented in ESX/ESXi HOSTd *does not* support the original MIBs as did 3.5 -- despite the fact that VMware indicates otherwise. The SNMP support is bare bones and reports the underlying hypervisor OS and some stats. On the current version of…

Choosing the API road is unfortunate, as I consider programming API's to be vastly inferior mechanisms for gathering system performance data -- not to mention being much more susceptible and sensitive to infrastructure variance. Question: Why not deliver both? Build vSphere models that use the new SNMP MIBs, then stay on…

I can only tell you what I got from Denny and what I got from VMware. From Denny (paraphrased) "VMware has dropped supporting SNMP from ESX4 therefore we (SW) are not updating Orion to support ESX4 via SNMP out-of-the-box. Vmware is not fully populating the VMware MIBs so there is no point. SW is building in support for…

VMWare has made the management of both ESX and ESXi hypervisors identical -- via the API. There is an SNMP component now built into the HOSTd daemon (SNMPd is no longer used) that provides *some* basic ESX/i information, but management data is via the API. Clearly there are some things that ESXi will never report on --…

Brandon, sorry I was not clear -- so I cobbled together a screen shot of NPM with what I *think* the Presentation should look like. Operationally, this is how I think it should work: - SNMP will give you both the interface table (L2) and the IP table (L3). - We are interested in both. -- The L2 table is good as it allows…

BTW, "hokey" is a technical term. Yes, they no longer expose the vnic interface statistics. VMware really mucked this one up.

Unfortunately, the above is not a solution. On a virgin ESX4 install that directory and library is not available -- I suspect it is a holdover from an in-place upgrade? Nevertheless, the bottom-line is neither SW nor VMware are making much headway in providing a solution **FOR THE IMMEDIATE NEEDS** of the customer base.…

Terry, your choices for Windows SNMP agents are (as far as I know): - Stay with the one that comes from Microsoft - Implement Net-SNMP according to either of the two methods provided by NET-SNMP.ORG and relish in the complexity - Purchase a 3rd party SNMP agent like the one from SNMP Informant(which is still only v1 and v2…

Brandon, the vSphere Basic Admin indicates that the HOSTD process on ESX 4 exposes a new set of MIBs that appear to be fully populated with the data that was originally populated into the HOST-RESOURCES MIB. (see page 55 in the doc www.vmware.com/.../vsp_40_admin_guide.pdf). So the question is -- is it really necessary to…

I hope that decision works out - sadly, I'm not sure Juniper gets it. I just got through reading an interview (www.networkworld.com/.../020810-juniper-stratus.html on the two Juniper visionaries and frankly their explanation of what they believe drives a datacenter is odd and the timeframe for new Stratus product is really…

This is a common problem with most NMS' -- they do not allow for the fact that monitoring SNMP is a requirement for determining monitored system health. The thought seems to be that an ICMP ping is sufficient. The answer is, there really is no good answer. I've setup simple monitors to test that something is responding on…

Brandon, something else to consider for features: When performing a discovery of a network we will run into network devices that are multi-homed (MH) and in cases of fairly large networks these multi-homed devices will have multiple Layer3 paths. The issue we run into is that during the discovery each IP address, although…

Brandon, all good stuff -- we look forward to the continued improvements. The topology aware feature is very much needed! Thanks!