I've worked on solving this many times in the past, including spending hours on the phone with SolarWinds Support and have never really gotten anywhere, but I figured I'd give it a try yet again... So I have a few questions that would help me understand this: How does SolarWinds attempt to correlate a UCS blade to an already managed node (usually this is an ESX host, at least in our environment)? I ask because in my experience, it is very inconsistent in matching blades to hosts. On the very same chassis, Orion will match a few hosts to blades but not match the others at all. The only difference I could find between the ones that matched and the ones that didn't are that, on the ones that did match, the MAC address for the ESX host's vmk0 management ethernet adapter matched the MAC address for the UCS blade's adaptor-1/host-eth-1 interface (I discovered this by painstakingly pouring through a dump of the UCS device's API and comparing it to the data in the SolarWinds VIM_HostMACAddresses table in the Orion Database.). This leads me to believe that maybe Orion checks the MAC of the UCS blade against the MAC of the ESX host? But this can't be right because, while the MAC addresses for ESX hosts are recorded in the database, I couldn't find the MAC addresses for UCS blades recorded anywhere... What is the criteria for if a UCS blade is clickable or not? In our environment at least, it seems that for most of the blades, if Orion doesn't match it to an existing Orion Node then the blade isn't clickable at all. We do have a few blades that aren't matched to nodes that we can click on and it gives us the option to manage, but a vast majority of our blades are not clickable at all, even though in reality most of our blades are populated by ESX hosts in the real world. I just can't figure out how Orion decides whether a blade is clickable. I've seen some suggestions that the fiber interconnect needs to have an external IP set, but that doesn't seem to matter as blades that are attached to FI's with the external IP set seem to have the same amount of trouble being matched to Orion nodes as ones without. I also maybe thought it had something to do with blades having their IP set a certain way, but that doesn't seem to matter either. Some of the blades have IP's recorded in the SolarWinds database and plenty don't, and the ones that do have IP's recorded in the DB aren't matched at a better rate than the ones without So Ideally this post is aimed at a SolarWinds employee that is really familiar with the inner workings of the Orion UCS algorithm, but if any user out there has been able to get their Orion UCS blade's all clickable and properly matched to their respective ESX hosts, I'd love to hear how you did it. And here's a quick sample of how pretty much all of our UCS manager pages look:

This discussion has been locked. The information referenced herein may be inaccurate due to age, software updates, or external references.

You can no longer post new replies to this discussion. If you have a similar question you can start a new discussion in this forum.

UCS Blades Not Matched or Even Clickable (Or: Down the UCS Rabbit-Hole Again)

xtraspecialj over 3 years ago

I've worked on solving this many times in the past, including spending hours on the phone with SolarWinds Support and have never really gotten anywhere, but I figured I'd give it a try yet again...

So I have a few questions that would help me understand this:

How does SolarWinds attempt to correlate a UCS blade to an already managed node (usually this is an ESX host, at least in our environment)?
- I ask because in my experience, it is very inconsistent in matching blades to hosts. On the very same chassis, Orion will match a few hosts to blades but not match the others at all. The only difference I could find between the ones that matched and the ones that didn't are that, on the ones that did match, the MAC address for the ESX host's vmk0 management ethernet adapter matched the MAC address for the UCS blade's adaptor-1/host-eth-1 interface (I discovered this by painstakingly pouring through a dump of the UCS device's API and comparing it to the data in the SolarWinds VIM_HostMACAddresses table in the Orion Database.). This leads me to believe that maybe Orion checks the MAC of the UCS blade against the MAC of the ESX host? But this can't be right because, while the MAC addresses for ESX hosts are recorded in the database, I couldn't find the MAC addresses for UCS blades recorded anywhere...
What is the criteria for if a UCS blade is clickable or not?
- In our environment at least, it seems that for most of the blades, if Orion doesn't match it to an existing Orion Node then the blade isn't clickable at all. We do have a few blades that aren't matched to nodes that we can click on and it gives us the option to manage, but a vast majority of our blades are not clickable at all, even though in reality most of our blades are populated by ESX hosts in the real world. I just can't figure out how Orion decides whether a blade is clickable. I've seen some suggestions that the fiber interconnect needs to have an external IP set, but that doesn't seem to matter as blades that are attached to FI's with the external IP set seem to have the same amount of trouble being matched to Orion nodes as ones without. I also maybe thought it had something to do with blades having their IP set a certain way, but that doesn't seem to matter either. Some of the blades have IP's recorded in the SolarWinds database and plenty don't, and the ones that do have IP's recorded in the DB aren't matched at a better rate than the ones without

So Ideally this post is aimed at a SolarWinds employee that is really familiar with the inner workings of the Orion UCS algorithm, but if any user out there has been able to get their Orion UCS blade's all clickable and properly matched to their respective ESX hosts, I'd love to hear how you did it.

And here's a quick sample of how pretty much all of our UCS manager pages look: