NET LOGONSERVICE START FAILED

[klugetor]

hello everbody,

i have a serious problem at the moment. yesterday i backed up my inactive node from the mscs. after that i was rebooting it, and the boot failed. so i restored the server with the backup but it didn't work. so i used an older backup to restore the clusternode, that worked. but now i have the problem that the clusterservice will not start, because the acess to the netlogonservice is denied (error: 0005 access denied). please help me!!! tnx in advance

greetings
torsten

 

[vwing]

Hi Torsten

First that comes in mind is that to check the account that's running the clusterservice, is it the right one and if so, has it the right rights?
Your account need at least:
* Act as part of the operating system.
* Logon as a service.
* Logon locally.

And make sure the cluster service account is a member of both nodes. There are more rights that might be handy like increase quotas but they won't make the cluster service fail.

Easiest way is to check the working nodes serviceaccount and confirm those settings in Local computer policy> Computer Config.>Windows Settings> Local Policies> "User Rights Assignment.

Just a thought

Regards
Johan

 

[brayboy]

Did you ever get a solution to the issue you posted on Apr 4th of this yr?

I just ran into the same problem yesterday.

thanks for you help.

 

[fabiolander]

I get the same problem when the second node try to join the cluster...

All privileges are granted.

Nslookup is ok.

could you help me ?

 

[fabiolander]

My config is 2 HP DL380 with MSA 4100.
Domain controller run on W2k3.

Thanks

 

[xmsre]

If the backup was an old one, the machine account passowrd may have changed [every thirty days by default] or the cluster service account password may have changed.

If it's the machine account password then reset the secure channel using nltest or netdom.

If it's the cluster account password, then you have to take both nodes down at the same time and brign them back up because the token is cached by the surviving node.

 

[vwing]

With the backup-thing. Uninstall the cluster-service, reboot and then install it again into the cluster since the other node worked and had the cluster up and running.

 

[fabiolander]

This trick doesn’t work.
But I found some sclecli error on application events:
Policy object propagated with error... format invalid..

I had to reset group policy on domain controller to clean this event. Policy applied with success.

However second refuse to join cluster.

If I uninstall first node and install second one as primary.
The second node creates and start cluster with success.
But the first one doesn't want to join the cluster anymore...

I formatted, reinstalled and joined domain again but I got the same error.

any idea ?

 

[fabiolander]

Finally I found a solution :

I built another domain, move the two nodes inside and everything works fine.
I built a approbation between two domain and people from domain 1 can connect to cluster on domain 2.

It's works but its a dirty way

Now, I want to know what is wrong with the first domain ?

How could I compare the both to find the bug ?

Thanks in advance for your help.

 

[xmsre]

What was in the group policy? You get a similar issue if you muck with certain settings. Check out paragraph 10.

http://support.microsoft.com/default.aspx?scid=kb;en-us;823659