" Failed to connect to the service manager on 'serverThree' " - validate a cluster
- I'm trying to validate a cluster configuration between two BL460c blades.
If I try and run the verification from serverThree, everything progressse smoothly, I can select the tests to run against the servers. All is well.
If I try and run the verification from serverOne, it gives the error " Failed to connect to the service manager on 'serverThree' ", and obviously proceeds to not add it to the list of servers slated to be validated.
There is no firewall on serverThree so I dont know why this seems to be a one way problem.
If this isn't a concern and will not cause problems in managing the cluster in the future, I'm fine to just ignore the error. But I would like to ideally get this solved.
Thank you for your help,
-Aaron
All Replies
- Hi Aaron Any details on OS and SP level?
Chalkie - Sorry, you're quite right,
W2K8 R2 build 7100 Hi,
Was any error logged in Event Viewer and the FailoverClustering log?
1. Open Event Viewer.
2. Expand Applications and Services Logs.
3. Expand Microsoft and then expand Windows.
4. Locate FailoverClustering and expand it.
5. Observe the Operational log.
6. Close the Event Viewer.
7. Under Administrative Tools or inside Server Manager (expand Diagnostics), select Reliability and Performance Monitor.
8. Expand Reliability and Performance.
9. Expand Data Collector Sets.
10. Select Event Trace Sessions and observe FailoverClustering in the list.
11. Select Startup Event Trace Sessions and observer FailoverClustering in the list.
12. Right-click on FailoverClustering and select Properties.
13. Select the File tab and observe the location of the log file at the bottom of the interface.
14. Open Explorer and navigate to the log location at %systemroot%\system32\winevt\logs.
15. Observer the trace log which is the cluster log. An example of such a log is clusterlog.etl.001.
16. Also observe the cluster Operational log – Microsoft-Windows-FailoverClustering%4Operational.
Tim Quan - MSFT
- Hi Tim,
Alright, well, the main event logs hold nothing other than informational entries
FailoverClustering under winevt has about 18 entires about failed network connections, removing systems from the cluster, virtuals going down, virtuals coming up, virtuals failing to come up or down, more network problems. (on which note I am using HP' network util for network teaming)
Nothing though directly pertaining to why systemOne does not recognize Hyper-V manager on systemThree within the clustering feature, as within the Hyper-V manager pluggin, it recognizes it just fine, and allows me to configure VMs running on systemThree while logged into the server manager of systemOne. - I've since torn down all features and services from these two boxes.
I've added a third box (serverTwo).
serverOne and serverTwo communicate fine, but for whatever reason they both keep saying that serverThree either doesn't have the service manager installed, or it doesn't have the failover cluster role installed. (I havn't added hyper-v yet as I'm taking my time in getting the clustering working)
In the event logs on ServerThree, there are many audit failures event id 4656 all referencing the service manager.
Any further ideas?
-Aaron Hi Aaron,
What’s the exact wording of the error message?
Can you successfully ping serverThree from serverOne or serverTwo?
Tim Quan - MSFT
- Here is the main body of the error message:
A handle to an object was requested. Subject: Security ID: NETWORK SERVICE Account Name: serverThree
It has the associated event ID 4656.
Account Domain: TESTING Logon ID: 0x3e4 Object: Object Server: SC Manager Object Type: SC_MANAGER OBJECT Object Name: ServicesActive Handle ID: 0x0 Process Information: Process ID: 0x28c Process Name: C:\Windows\System32\services.exe Access Request Information: Transaction ID: {00000000-0000-0000-0000-000000000000} Accesses: DELETE READ_CONTROL WRITE_DAC WRITE_OWNER Connect to service controller Create a new service Enumerate services Lock service database for exclusive access Query service database lock state Set last-known-good state of service database Access Reasons: - Access Mask: 0xf003f Privileges Used for Access Check: - Restricted SID Count: 0
Any ideas?
-Aaron - To me it looks like there is some sort of problem around some NT credentials the serverOne and serverTwo system are throwing at the serverThree service manager? Or is it something is wrong with how the service manager on serverThree is configured that's causing the refusal?
There is no firewall configured on these systems.
The only other thing in between them is HP network utility which manages the NIC teaming.
Thank you,
-Aaron - Upgrade to final RTM release and report back.
Chuck Timon Senior, Support Escalation Engineer (SEE) Microsoft Corporation - Still havn't been able to get my hands on an RTM, but, I have worked around the problem in an awkward manner which resolves the issue... but doesn't explain anything.
Apparently, changing the system name resolved all issues I was having? I dont understand why, but any system assigned as 'serverThree' aquires the same WMI problems.
So now it all works fine with my set of 'serverOne' 'serverTwo' 'serverFour'. Mostly I'd like to understand why 'serverThree' seems to have gone sideways on me. Deleting 'serverThree' from the AD and recreating it didn't solve anything, and it has always resided in the some OU as the other servers. - You're right...that does not make any sense.
Chuck Timon Senior, Support Escalation Engineer (SEE) Microsoft Corporation
