Host Not Monitored?

[felyjos]

Hellom,

SCOM 2007 R2
Veeam MP 6.5.0.1549

Why do I have host "Not Monitored" individually on Cluster which are monitored globally and other host are monitored?
Thanks,
Dom

[Alec King]

Hi Dom,

There could be several different reasons for that....

First of all, could you try Rebuild Topology link in the VEEAM UI? This causes SCOM to rebuild and re-initialize all the VMware objects.
Please wait 30 minutes or so for that to complete.

Second, you could also try restarting the System Center Health Service on the Collector server which monitors this host. (you can see in the VEEAM UI which Collector is monitoring which Cluster(s))
That will also cause SCOM to re-discover and re-initialize managed objects, including VMware hosts.
Again wait about 30 minutes for that to complete.

If those solutions don't work, there could be a more complex problem with vCenter, permissions, the host itself, etc...if you raise a support case we can dive in.

Thanks!
Alec

[felyjos]

Hi Alec,

All collectors have been restarted.
I am waiting the "Rebuild Topology"

Thanks,
DOm

[sergey.g]

Hi, if you still experience the issue, check if all of these hosts are from the same VMware Collector? If so, monitor Operations Manager log on this server, I bet Health Service is experiencing issues delivering the data, we need to understand why - maybe amount of objects/workflows is too high for this Collector. When hosts are discovered, SCOM should initialize monitors for them regardless of VES or Collector state. So, most of times when hosts aren't monitored it's a Health Service issue. Maybe MPs haven't arrived yet or HS cannot handle this amount of objects.

Let us know about your progress.
Thanks.

[felyjos]

Hi, I see some performance issues YES.
I have 5 Collectors for 62 Hosts, 678 VMs

The first collector is giving an error:
Time Sampled: 12/3/2014 10:53:07 AM
Object Name: Process
Counter Name: Handle Count
Instance Name: MonitoringHost
Last Sampled Value: 18286
Number of Samples: 10602

the second collector is giving an eror:
Time Sampled: 12/3/2014 10:52:56 AM
Object Name: Process
Counter Name: Handle Count
Instance Name: MonitoringHost
Last Sampled Value: 15375
Number of Samples: 7846

and I have a third one which is not communicating properly...

still investigating but do you think I have enough collectors for the number of VMs?
Thanks,
Dom

[felyjos]

I am still getting

Log Name: Operations Manager
Source: OpsMgr Connector
Date: 12/8/2014 10:46:32 AM
Event ID: 21023
Task Category: None
Level: Information
Keywords: Classic
User: N/A
Computer: VOPVEEAMCOL1.ad.medctr.ucla.edu
Description:
OpsMgr has no configuration for management group SCOM-MED and is requesting new configuration from the Configuration Service.
....
does this means there is issue with the collector to contact the Management Server?

[felyjos]

On System Center Operations Manager 2007 R2: Stop the System Center Management service.
Locate the installation directory for OpsMgr in the following path:
SystemDrive\Program Files\System Center Operations Manager 2007
Delete the Health Service State directory.
Restart the service.

I did it on the 3 Management servers as well as on the 5 collectors with reboot

[felyjos]

the "not -communicating Collector" has been rebuilt... and communicate now...
checking why the machine are overloaded suddendly after two years
the number of VMs or Hosts or Clusters do not seem to be the reason as these two collectors do not have the more VMs, Clusters or Hosts!!!
All Collectors have the same configuration..
Name - Clusters - Hosts - VMs - Datastores
OCol1 - 5 - 17 - 120 - 0 ==> Overload
OCol2 - 1 - 9 - 162 - 0 ===> Overload
Mgr1 - 3 - 7 - 125 - 70
RCol1 - 0 - 16 - 126 - 0
RCol2 - 1 - 13 - 146 - 0

Still working on the troubleshooting trying to get something!!!
Thanks,
"Dom

[Alec King]

Hello Dom,

Your Collectors should not be overloaded; less than 200 VMs per Collector is OK. Assuming that Collector meets minimum hardware spec (2 x CPU, 4GB RAM)

Are those Collectors performing any other tasks?

Note - the event about "requesting new configuration" is expected, after you have cleared the Health State cache on the server.

Do you still see errors in the Operations Manager log on those Collectors?

[felyjos]

Hello
Yes two collectors are still having the alerts for
All the collectors are:
4 vCPU
16 GB RAM
Veeam VMware Collector: Health Service (Monitor Host) Handle Count threshold

No other task they are dedicated to be Veeam MP Collectors.

Yes and trying to Reset the Health on the SCOM Console takes for ever .... 45 minutes...

Let me uninstall Veeam Collectors and reinstall them ...

Thanks,
Dom

[Alec King]

Hi Dom,

I think the problems might be on the SCOM side - not the Collector side. So, reinstall of Collectors might not help.
Note - the Handle Count alert might not be a serious issue. This is a 'soft' threshold, depending on your environment - it may be you can increase this threshold.
Do you see other data from VEEAM MP is delivered OK? (hosts, VMs, performance data, reports and dashboards and so on)

Perhaps you should open a support case? Then we'll get your log files, and we can do some detailed analysis.

[felyjos]

The collectors have been re-installed.
As you said no effect...
The SCOM agents have been re-installed ....
still in progress but so far the collectors are still green in SCOM!!!

[felyjos]

The collectors have been re-installed.
As you said no effect...
The SCOM agents have been re-installed ....
still in progress but so far the collectors are still green in SCOM!!!

Still one collector missing !!!
What will be the impact on the collectors/VMs/Host of a change in the MOM Action password?

Thanks,
Dom

[Alec King]

Hi Dom,

Quick answer - there should be no effect to VEEAM MP if you change the MOM Action Account password. Assuming that the Action Account is still in local Administrators group for that server, that's all we need. We don't use that password anywhere in our Collector our UI.

When you say one Collector is 'missing', you mean that you cannot see it in SCOM?
If you can't see it in VEEAM MP views folder - can you see it in all Computers view in SCOM console? Is the agent itself correctly installed and healthy?
And BTW when you reinstalled the agents; did you set Proxy permission for them? This is a requirement for Veeam MP, and if you reinstall the agents you have to set Proxy again (SCOM console, Administration, Agent-managed, double-click each server and check Security tab)

Please do raise a support case any time you want - we will get detailed logs and will know better what is happening....

Cheers
Alec

[felyjos]

Hi Alec,

Yes it is still in the Local Administrators Group ... Any issue with nested Groups?

When you say one Collector is 'missing', you mean that you cannot see it in SCOM?
I can see in some views but not a specific one...
YES it is showing in Agent Managed, Collector Heartbeat Views but not under _Collector Service Status View

YES the act as proxy is checked on all collectors.
I will open a case with Veeam as it seems I am looping!!!
"Your support case number is 00700694"
Thanks,
Dom

[Alec King]

Hi Dom,

No issue with nested groups in Administrators. We just need the agent action account to have admin rights to the server.

The _Collector Service Status view is CRITICAL - if you see no Collector there, then there will be no Hosts, no VMs, no Datastores - no vSphere objects. Discovery of Collector Service is the first stage of all other discovery. The Collector service must appear in the view and it must be "monitored" status.

You can try restarting the SCOM Health Service (Microsoft Monitoring Agent) on that server - that can force discovery of the service.

[felyjos]

I did restart the HealthService service on the collector missing...
It appears well on the Veeam Server

Finally it is here !!!
let see if the host are coming also...

[Alec King]

If you can see the Collector now (and it is Monitored - Green state), then please click Rebuild Full Topology in the VEEAM UI - that will make all Collectors send the Hosts, VMs, etc. That's the fastest way to confirm everything is working now.
Let us know how it goes!

[felyjos]

I will check later but the hosts missing are coming from the 3 collectors having troubles yesterday and so far none have sent anything even I click the Rebuild Full Topology this morning...
let see
cross-fingers..

[felyjos]

Hello,

Veeam MP issues - Case # 00700694 [ ref:_00D30RWR._50060iRbPy:ref ]
working with Nick for the past 3 days (opened 12/9/2014) ...
We found already some items but nothing to get back on track.... We did an upgrade of the collectors to 6.5.0.1549 which was the MP applied but the collectors were still behind...
227 out of 681...
20 Host not monitored...

still in progress!!!
Dom

[felyjos]

Hello,

Still working with Veeam Support on the issue but due to the holidays the engineer changed and a new set of information was just provided on Friday, 19 December 2014.
No clue yet what is going on!!! after 10 days!!!

Thank you

Merry Christmas - Happy New Year all my best Wishes....
Happy Holidays

Dom

[felyjos]

Still working with Veeam Support but today one clue...
the Run As account from SCOM is rejected on two collectors:
The Health Service could not log on the RunAs account AD\xxxx for management group SCOM. The error is Logon failure: unknown user name or bad password.(1326L). This will prevent the health service from monitoring or performing actions using this RunAs account

but as this is the default Run As account I wonder where this information is stored on the collector or where is it picked as it concerns only two collectors out of 5!!!

Happy New Year

[felyjos]

Still working with Veeam Support as the numbers are growing but so slowly ...!!!

12/30/2014 06:00 a.m. (After complete reinstallation of SCOM Agent and Veeam Virtualization Software)
Collectors state all (5) are okay
Clusters 10 (Ok) 6 are not monitored (+1 more monitored after 15 hours!!! )
Hosts 62 (Ok) 12 are not monitored (+12 more monitored after 15 hours!!!)
VMs 361(out of 686!!) 496 are not monitored(+145 VMs +192 VMs not monitored)

It is populating but it is really really really slow!!! Not sure why … still a lot of objects not monitored!!!

12/29/2014 06:00 a.m.

Collectors state all (5) are okay
Clusters 10 (Ok) 7 are not monitored (Same as 12/27)
Hosts 62 (Ok) 24 are not monitored (Same as 12/27)
VMs 361(out of 686!!) 304 are not monitored(Same as 12/27)


Happy New Year

[felyjos]

still working with Veeam Support as the numbers are not changing anymore now since a week...

Strange...
thanks,
Dom