Failover cluster cannot be created, Please help

mandy7777

I was setting up a Lab of" creating failover cluster node" on windows server 2012 R2 via Vmware, an error occur when it forming cluster 'cluster1'
Please see the attachements, i uploaded all the step that i have perform on this lab.

cluster node:
server3 : 192.168.100.103
server4 : 192.168.100.104

cluster server:
sever2: 192.168.100.102

All of the server are in test.com any solution , will be really appreicated!!

Arkrain

Is server3 a Domain controller?

https://support.microsoft.com/en-us/kb/2795523

mandy7777

all of them are domain controller

Arkrain

When you build a Windows Server 2012 failover cluster environment, you cannot add a server that has the Active Directory Domain Services (AD DS) role as a node.

mandy7777

thanks , Arkain. i set back the server 3 and server 4 to just the plain server, not domain controller ( snapshot in vmware), then i got this error: The operation has failed: an eror was encountered while loading the network topology. error code : 0xx80070005 access is denied"

if you could , please explain me a little more about this message .

because i'm using the Administrator of domain to log in, and all server 2, 3, 4 are able to connected together by using ping .

mandy7777

not Quite sure what kind of access should i get , since i'm using the administrator of test.com

sthomas

Setting the servers back from a snapshot isn't the way to go. You need to demote them as DCs the proper way.

mandy7777

thanks for resply sthmomas, i was not demote them , i snapshot every lab that i have completed, include the base install (fresh install), so i the snapshot that i set is the exactly fresh install server 20012 r2 state, did not promote as DC yet.

poolmanjim

We don't have a clear idea of your network so we have to make a few assumptions. I assume you have another domain controller running around somewhere. When you promoted Server3 and Server4 to be DCs that other Domain Controller knows that they were DCs. If you just reverted the snapshots on 3 and 4 the remaining Domain Controller is confused (and may not have all the FSMO roles it needs).

If you cannot go back to the version where AD was configured on 3 and 4, you will either need to rebuild your lab using proper configurations or you would need to seize roles from the old DCs to the existing one and do a metadata cleanup. I would recommend starting from scratch (forgoing the failover cluster for now).

Lastly, I would suggest you do some googling and study on how Active Directory works and how it is supposed to be deployed. The best way to learn this stuff is to break it and then try to figure out why it broke, fix it, and repeat. Google the topics, read some books, and watch some videos before trying to lab it all like this so you have a jumping off point.

Enticles

Whilst i have little experience dealing with failover clustering, i feel the need to weigh in a touch about the way you reverted your 2 other DC's back to member servers, and the issues it creates.

when they were domain controllers they were hosting Active Directory databases (NTDS.dit) and most likely DNS Server and Global Catelog server roles also. these roles all replicate between one another and rely on each other when it comes to KCC and ensuring they hold the most up to date information there is available. when you go through the process of DC demotion it updates active directory with the demotion changes, effectively shouting to the other remaining domain controllers "hey im not a DC/DNS server/GC server anymore!" and during this process if there are any forest or domain master roles on the server it will let you know this and allow you to take corrective action before proceeding further.

by just reverting to a previous snapshot you leave the one remaining domain controller in a state where it "thinks" the other 2 servers are domain controllers, Global Catelog Servers, DNS server, and potentially a domain or forest master for one or more of the 5 FSMO roles they can host. The server will continue to attempt to replicate all Active directory data to and from the now-member servers, and when it can't it will start throwing errors into multiple parts of event viewer and the problem will snowball into DNS replication errors, and DFS errors if that is configured also.

Like poolmanjim suggested, in this situation it will most likely be easiest to just squash the entire lab and rebuild it again from scratch. It sucks, but look on the bright side - you learnt something you may not have known before. Isn't that the whole idea behind labbing?

I can't tell you how many times i have broken my lab trying out new things, or doing things differently, or trying to be a smartass and automate something and getting the scripting wrong. I have actually done the exact same thing you have done here with a server 2008 R2 lab. It drove me round the bend before i eventually admitted to myself i borked the lab and began rebuilding it.

Hopefully this helps

mandy7777

Thank you so much for the advice, i restarted the Lab today, it works. i am able to create the cluster node without any problem .