I've had a Ciena node running happily for several months now and when I updated the custom fields on it, it decided that it is now a Cisco UCS??? WTH? How do I get this back to its right configuration?
How did you update the custom properties? Is there a possibility that there's a Cisco UCS in your environment that is using the same IP addresses as your Ciena?
I just updated them on the node custom properties section. When it saved it switched to a UCS for some reason.I can log into this node and it's NOT a Cisco UCS.
Edited... human error and a feature. no bug.
I saw this same feature with a Palo Alto firewall, that was suddenly changed to a Cisco UCS by someone selecting "poll for ucs". That is a one-way setting. I repaired it by going to add a node, SolarWinds complained it was already monitored. I selected the option to re-add it because I was having problems with the node. It resolved the problem without screwing up all of the dependencies like Orion Maps.
But this is a genuine SolarWinds feature. Normally you let node details collect it's info, but in the case of Cisco UCS there isn't anything for node details to detect, so the scan for UCS options are like an override to force UCS. To undo this re-onboard the node using the same poller engine, and select the "i'm having trouble do it anyway" option. It will put you back to normal but the UCS option may be checked and greyed-out (depending on which of the two UCS options you checked, it is in 2 places).
I spotted this because the node was being displayed with the default node details summary page instead of the custom firewall node detail page. The node was correct in all ways except the SYSOBJECTID and MACHINE TYPE was wrong. The machine's "description" which is populated with data from SNMP was correctly recorded, so when node details were collected it was all correct. Scanning for UCS hard-codes the node to UCS.
Ouch! 2 days ago I had one node doing this, yesterday I found a second. Now I'm wondering if there are more... Looking for a way to easily find them, manage nodes search does not include machine type in it's search, so I'll probably have to use swql...
Hi @Lyphosaurus ,
Check the Node Details poller to make sure it hasn't changed from what it's supposed to be:(From managing the node, select List Resources)
It might not be the specific problem you're experiencing, but this happened to me once.
This will find them:
SELECT NodeID, Caption, NodeDescription, Description, Vendor, SysObjectID, MachineType, EngineIDFROM Orion.NodesWHERE MachineType LIKE '%UCS%'
Replying to myself because a now deleted reply from someone else asked a question that might be useful for others who stumble across this.
In my case, the Nodes Details poller was changed by what I believe should be considered a bug, but was technically triggered by me. I was seeing nodes switch back and forth intermittently from the expected Machine Type (Cisco ISR routers) to a statically defined Machine Type. What happened: I was attempting to ensure that all child entities on all of the routers in my environment were being monitored, so used the API to perform the ScheduleListResources followed by ImportListResourcesResult. This also changed the assigned Node Details poller from the Node Details by SolarWinds to one of our in-house pollers, even though the Automatically poll nodes during network discovery, add node, and list resources to determine if this poller is a match setting is disabled on our custom pollers.
ok, it turns out this was an erroir caused by a human. The Cisco UCS detection is not done by snmp, it is done by selecting the UCS credentials. And that cannot be unselected, but you can work around it with a forced "i'm having problems" re-add. The bug I called out was not a bug. it is a feature .
turns out there are two scan for UCS options, one was not selected the other was selected sometime later. I updated my post.
So "scan for UCS" is an option in two places. I think there is no "scan" for UCS, I think UCS cannot be detected. If you set "scan for UCS" what it seems to do is hard code the node for UCS. To remove that you have to re-add the node without deleting it. When you re-ad it select the same poller engine used before. It will tell you that the node is already being monitored. But if you picked the original poller, you will have an option "I'm having trouble with the node and want to re ad it anyway" or something like that. Select that option, and it will "fix" your node without losing history.
Tested on 2023.3.1
