Monday, 27 August 2012

Veeam nworks Management Pack 5.7 - How to Troubleshoot


Credit: the following information is strongly based on the Veeam ProPartner University “Veeam nworks MP 5.7 - Troubleshooting” course and updated for SCOM 2012 with a new selection of images.

1. Collector and EM server not displayed in OpsMgr

Sympton:
Collector and EM server are not displayed in:
Operations Console > Monitoring > nworks VMware > Enterprise Manager > _Enterprise Manager Server State
Operations Console > Monitoring > nworks VMware > Enterprise Manager > Collector Servers > _Collector State



Probable causes:
·         Uncompleted discovery process
·         Insufficient permissions of the OpsMgr Agent Action Account
·         Communication problems between OpsMgr agent and OpsMgr server

Solution:
·         Wait for discovery to complete - or -
·         Restart the OpsMgr agent, or System Center Management service, on the Collector server or Enterprise Manager server to initiate discovery


·         Check that the proper permissions are assigned to the OpsMgr Agent Action Account (recommended that SCOM Agent Local System Account permissions are used - must have full administrative access to the local server)

Check:
Operations Console > Administrator > Agent Managed > find the server hosting the Collector and/or Enterprise Manager, and check if the Operations Manager Agent has been deployed on that server and if it is reporting back with Green status



2. Collector and EM server cannot connect to VMware Host/vCenter Server

Probable causes:
·         Incorrect access credentials for the nworks service account
·         Firewall or network issues
·         DNS configuration errors

Solution 1:
Nworks Enterprise Manager Web UI > Select the Host/vCenter in question > Operations: ‘Test Connection’ button
Note: There is also a ‘Test All Connections’ button



From the Host/vCenter > Administration > Edit Connection > verify the Connection settings
Note: At least read-only role across a vCenter server is required

Solution 2:
Check that the traffic on the port used for API access to the VMware systems is not blocked (default 443.) Either open firewall for the required port, or configure the ‘Proxy settings’ in nworks Enterprise Manager Connection settings



Solution 3:
If no connection using an FQDN DNS name, try using IP address.

3. VMware objects do not appear in OpsMgr views

Sympton:
VMware objects do not appear in:
Operations Console > Monitoring > nworks VMware > _All VMware State
Note: This view includes _Cluster State, _ESX Host State, _VM State

Probable cause:
Proxying is not enabled for the Operations Manager agent on the Collector Server(s)

Solution 1:
Operations Console > Administration > Device Management > Agent Managed > Agent Properties > Security tab > verify “Allow this agent to act as a proxy and discover managed objects on other computers” is ticked



Solution 2:
If only certain VMware objects are missing or have not been updated:
Nworks Enterprise Manager Web UI > Enterprise Manager tab > Operations > Rebuild Full Topology
Note: This will take several collection cycles (default 5 minutes each) to complete



4. Unstable behaviour of the OpsMgr Health Service

Symptons:
·         Heartbeat failure
·         OpsMgr agent restarts
·         Gaps in reports/charts
·         Agent and/or VMware objects “grayed out”

Probable causes:
·         OpsMgr Health Service is overloaded
·         Incorrect agent configuration
·         Collector is overloaded

Solution 1:
1.       Remove nworks monitoring jobs from the Collector
2.       Stop the “System Center Management” service on the Collector server
3.       Delete the “Health Service State” folder at C:\Program Files\System Center Operations Manager\Agent
Note: The Health Service State folder is recreated when the “System Center Management” service is restarted
4.       Start the “System Center Management” service
5.       Observe the “Operations Manager” log in Event Viewer > Applications and Services Logs
6.       Reapply nworks monitoring jobs to the Collector

Solutions 2:
·         Run Configure Operations Manager Agent task (include in nworks MP) to reapply configuration to the agent
·         Relocate monitoring jobs to other Collectors
·         Apply tuning options - such as increasing collection intervals, or disabling certain data classes - to reduce the amount of data processed by the Collector

5. No events/alerts in OpsMgr console

Sympton:
Alerts on vCenter events are not received in the Operations Manager console

Probable causes:
·         Connection or configuration issue between Enterprise Manager/Collector server and VMware systems
·         Disabled Rules and/or Monitors on the OpsMgr
·         Modified thresholds for alerts on vCenter events

6. Performance data is not collected

Sympton:
No data in OpsMgr console

Probable cause:
Windows Event Logging facility is not working properly; that is - events are not being delivered to the custom “nworksEventLog” found in Event Viewer on both Enterprise Manager and Collector(s)



Solution:
·         Check the existence of the ”nworksEventLog” and that new events are being delivered (at the very least, EventID 117 should appear at 5-minute intervals)
·         If the log does not exist: reboot the system, open the Computer Management snap-in, check nworks Virtual Infrastructure Collector service is installed and running, and that the nworksEventLog has been created
·         If the above does not help: reinstall the collector
·         If log still not created: check the license in the nworks Enterprise Manager Dashboard - the license type must be ‘MOM/SCOM’



7. Performance data is not delivered

Sympton:
No data in OpsMgr performance graphs

Probable cause:
Windows Management Instrumentation (WMI) is not working properly

Solution (using the Windows Management Instrumentation Tester):
·         From the command prompt> type wbemtest
·         Click Connect…
·         Enter for Namespace root\nworks
·         Click Connect



·         Click Enum Instances…
·         Enter superclass name as Host Stats (this an example metric class for collection from ESX hosts)



·         The presence of instances confirms the Collector is posting collected host metrics into WMI for OpsMgr agent



·         If no instances shown or they do not update every 5 minutes, there is an issue with connection to vCenter. Check connection as above, and if okay then try re-installing the Collector.

No comments:

Post a Comment