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.

VIM polling appears broken with SAM 6.2 and NPM 11.5

mharvey over 8 years ago

Been dealing with an ongoing issue where hardware health polling wasn't correct with some of my ESX hosts. Working with support and have been unable to determine why disk polling seems to be failing. Today I noticed that several VMs that were migrated between ESX hosts weren't showing on the correct hosts. These guest VMs were migrated weeks ago. In order to correct I had to change the polling method on the host to poll esx directly (which lost disk polling), then moving back to through VCenter updated the disk polling, but now shows disk off that are actually on. In another instance, the ESX host re-appeared in italics as if it was not being monitored. I had to correct this by disabling ESX polling on the monitored server, and re-enable it. I also noticed I had to hit poll now in order to get some of the guests to update. It would appear as if the VIM polling is no longer working on it's own since the upgrade.

0 robert.a over 8 years ago

Same issue here, looks like its also being reported in another thread.
Virtual Machines in italics/Already being managed by NPM
It's also filing to link up nodes that have just been added to the virtual host that's running them.
Cancel
Vote Up 0 Vote Down

Cancel
0 chrispaap over 8 years ago

Hey Matthew do you have an ongoing support case number I can check out. On the surface it seems similiar to a couple of different problems I have seen before but i want to touch base with the support rep to see what we can cross of first.
Cancel
Vote Up 0 Vote Down

Cancel
0 shack over 8 years ago in reply to robert.a

In regards to the link here(a thread I started). I believe it has always been this way.
Cancel
Vote Up 0 Vote Down

Cancel
0 chrispaap over 8 years ago

Does the appliance health show that configuration collection completed successfully?
Cancel
Vote Up 0 Vote Down

Cancel
0 cbunge over 8 years ago

I'm having a similar issue and have been working with support since the update back in June/July. It has to be a failure with IVIM. If I look at VMAN everything is polling fine. If I configure NPM to poll the vCenters directly, everything polls fine. When I sync NPM with VMAN, I get errors about incorrect credentials, access denieds, failure to poll, etc. Considering the two sides work individually and only break when they communicate to each other there is definitely a bug. I'm about to have a service call in a couple of minutes so hopefully I'll have some more suggestions for a fix.
Cancel
Vote Up 0 Vote Down

Cancel
0 stevenwhunt over 8 years ago in reply to cbunge

Chris, can you provide me with the support case number?
Cancel
Vote Up 0 Vote Down

Cancel
0 vispetto over 8 years ago

We had a similar issue where we were monitoring the VM's but when you looked at the ESX host is was italicized (as not being monitored). Ultimately the issue was with the mapping of the VM to the ESX host. What I discovered was a NULL value in the HostID column in the VIM_VirtualMachines table. Once I populated the HostID with the correct data the servers showed under the Host properly.
We also discovered this anomaly occurs if you add the node from the Add Nodes in Settings. If you add the VM from the Host server (when it's italicized) then everything works properly.
Cancel
Vote Up 0 Vote Down

Cancel
0 cbunge over 8 years ago in reply to stevenwhunt

‌860584 stevenwhunt‌
To troubleshoot we moved all vCenters and ESX in orion to the same poller. If that doesn't work on Tuesday if will modify the VIM table to increase the time outs of the api
Cancel
Vote Up 0 Vote Down

Cancel
0 stevenwhunt over 8 years ago in reply to cbunge

Chris, I just read through the case. Looking forward to hear the outcome of the troubleshooting. Please feel free to send me an email regarding how this process flow could be different, from your perspective, to ensure your environment does not experience any of these issues.
Cancel
Vote Up 0 Vote Down

Cancel
0 cbunge over 8 years ago in reply to stevenwhunt

This seems to be working properly now. -Knock on wood- . I'm still curious with a few blips missing in data if the timeout is also an issue, but it's odd that all ESX hosts and vCenters have to be on the primary poller for this to successfully stay connected. It sounded like from the tech support that this was a known bug a while ago and should have already been resolved.
Cancel
Vote Up 0 Vote Down

Cancel