We only want to monitor and show status on critical interfaces (links to other switches and non-transient network devices) on our switches; so we list the resources on the switch and deselect all but the critical links which supposedly turns off monitoring of those interfaces. But if one of those non-monitored interfaces goes down, we sometimes get a red icon on the node icon. If you expand (from Managed Node View, for instance) the interfaces on the device, they all show green (because all monitored interfaces are up), but the red icon is still overlayed on the node icon.
Any ideas why this is happening and maybe how to stop that behavior?
I need you to run the following SQL query and record the results.
SELECT * FROM Nodes WHERE GroupStatus like 'Up-%.gif' and not exists(SELECT 1 FROM Interfaces WHERE Interfaces.NodeID=Nodes.NodeID and Interfaces.Status<>'1')
I then need you to open a support case with the results.
Support had me run a script which cleaned it up. Since I've seen it pop up again, I surmise this is a bug. Not sure if it was cleared up in the latest release.
Well I can tell you that they are running 10.2.1 so it has not been fixed. I would like to know what script cleaned it up. I discovered (through an idea I got reading through this post ;)) that if you list resources on a node in question and add an interface that is up then go back in and remove it, it will stop showing with a down child status. However, it would be extremely tedious and time consuming to go through all of them one by one and do this if there is an easier way of cleaning this up.
To be honest, Chrystal, I would open a support ticket... If I remember, they had me run one script to determine what the next script would look like, and your situation may be different. It's disappointing to learn that this issue hasn't been resolved yet. They seem to have a real problem with interface discovery.
We resolved this by running a couple SQL queries to get lists of the nodes with the issue, and then unmanaging for 5 minutes. When they remanaged themselves, they all showed up with no child status of down. Yay. Not an ideal solution, but it did work.
I know this is an old thread, but I'm on 10.3 and having this issue as well. Oddly, though, the down child objects only show up on Network Atlas Maps. Node Details show everything green with no issues or child object problems, but I can't get my map to clear up . And I've tried the unmanage thing as well as removing all interfaces and adding them back--no dice. I even shutdown all interfaces that weren't up, so nothing should be operationally down w/o being admin down. Hasn't changed a thing.
A couple of things:
1) See above. There's a SQL query which can clean up existing situations.
2) We've not had the problem since I ran the query.
3) I've made extensive use of the "unplugged" feature for interfaces that are not uplinks/downlinks/links-to-critical-nodes. Eliminates a lot of this.
Doesn't look like the same issue.
1) No results from the query SW gave you
2) Yours seemed to show up in Node Details as well; mine is all-clear there, just not in Network Atlas
3) The only issues are on core switches, so the unpluggable feature isn't appropriate (I want to know if something comes up/down)
I'll prolly open up another thread or a SW case when I get time. Thanks for the feedback.
Also: don't have time to figure out how to find the ticket (Portal is broken, won't display existing tickets, I've been told it's my problem even though it happens with all browsers including IE 7) and attach the results. Clues would be appreciated.
Just to see if this is what you are doing,
Pick up a switch X and navigate to the "Node Details" page
Then select "list resources" and say pick one of them and deselecting it.
Then you submit and get back to the Node details page for the same Switch X.
There you see the "Node is Down"(red color).
If this is exactly the problem then I would recommend to open a support ticket to investigate further.If not please explain the steps involved .At the same time are you able to ping the switch from the NPM server.
Correct except the last line: it's not "node down", it's "child" (interface) down (if the interface de-selected is in fact down).
In other words: Why is the node showing interface down when the interface that is down is not being monitored?
Got it.I tried to replicate this on my lab.Here is what I observed.
I followed the steps mentioned in my last thread response.Just a recap,I selected a switch and moved to Node details page.I did list resources wherein I went on and deselected one of its child interface.Then I logged into the database quickly and looked up for the table "Interface".I found the child interface was no more in the interface table.
Coming back to the GUI and also referring to the statement in your last thread " Why is the node showing interface down when the interface that is down is not being monitored?".I did a list resource and found the interface which I unselected appeared unselected itself.Coming to the exact problem,I unselected a child interface which was up(green) at the first place.And then when I came back after this interface disappeared from the "Interface" table in the database,it still showed unselected Up(green) in the GUI.
So I am not sure if you had time to test just one interface or a few more interface before reporting the problem.I would like you to go-head and test a few more,probably try unselecting a down(red) and a up(green) interface and get back with the results.
I'm still not sure you have this. It's pretty simple:
SolarWinds solutions are rooted in our deep connection to our user base in the THWACK® online community. More than 150,000 members are here to solve problems, share technology and best practices, and directly contribute to our product development process. Learn more today by joining now.