VMware Cloud Community
insearchof
Expert
Expert
Jump to solution

vCenter HA has an invalid configuration. Remove vCenter HA to destroy the current cluster configuration and set up vCenter HA again.

Today I  rebuilt my HA

VMWare 6.7 U3

Have 1 datacenter

1 cluster

8 ESXI hosts 6.7 U3 also

All three nodes are up and running   Active Passive and witness

I was just going to check on things and I noticed this error

vCenter HA has an invalid configuration. Remove vCenter HA to destroy the current cluster configuration and set up vCenter HA again.

Why did this happen? 

What can I do to fix this.?

This was a rebuild I lost my primary node to disk failure weeks ago and I was told Remove Ha and then redeploy and thats what I did

How to find what the invalid configuration is?

Thank you

Tom

Reply
0 Kudos
1 Solution

Accepted Solutions
brad_101
Contributor
Contributor
Jump to solution

Hello insearchof...

 

If your cluster still shows 'unavailable' after doing a Manual snap/deploy of rebuilding it.. but you can ping all 3 HA IPs you probably need to do 1 last thing.

 

On the current active vCenter, login via SSH..

Change the 'root' users default shell to bash.. (i use the interactive method)

 

# chsh

Prompt should come up like..

root@xxx [ ~ ]# chsh
Changing the login shell for root
Enter the new value, or press ENTER for the default
Login Shell [/bin/appliancesh]: /bin/bash

root@xxx [ ~ ]#

 

Then reboot both the witness and HA nodes.

 

It should then complete HA setup.

 

After HA is setup you can change back the 'default' root shell to be the appliance one of (if you prefer)

root@xxx [ ~ ]# chsh

Changing the login shell for root
Enter the new value, or press ENTER for the default
Login Shell [/bin/bash]: /bin/appliancesh

root@xxx [ ~ ]#

Hope that helps

View solution in original post

Reply
0 Kudos
18 Replies
Nawals
Expert
Expert
Jump to solution

Hi,

when you re-deployed VCHA make sure DRS should be manually/disable set. If not set this error should occur. disabling DRS and deploying Passive and Witness nodes to the same host that the VC was running on.

NKS Please Mark Helpful/correct if my answer resolve your query.
Reply
0 Kudos
insearchof
Expert
Expert
Jump to solution

Nawals

Thank you

So I should do this

1. Remove Vcenter HA

2. Disable DRS on My vcenter cluster           Note I only have one Vcenter   1 datacenter and 1 vcenter cluster

3. Then deploy. HA

              I always place the nodes on separate Hosts and datastores.

One other question here.   Is it possible to rename the Nodes   I did not like the naming method used on the rebuiild    my one node was vcsa-1-passive-passive 

I saw during the build you can change the names there is that the best place to do so?

Thanks

Reply
0 Kudos
insearchof
Expert
Expert
Jump to solution

Ok I turned of DRS on the Vcenter Cluster that all the ESXI Hosts are a member of

I removed VCenter HA

Then enabled HA    waited for a while   went out to see my grand daughter and when I came back the job finished

I deleted the passive node created and shows the same error

What logs can I look at to help resolve this issue?

vCenter HA has an invalid configuration. Remove vCenter HA to destroy the current cluster configuration and set up vCenter HA again.

Any thoughts Ideas or suggestions??????

Thank you

Reply
0 Kudos
Nawals
Expert
Expert
Jump to solution

Check vpxd and vcha logs.

NKS Please Mark Helpful/correct if my answer resolve your query.
Reply
0 Kudos
insearchof
Expert
Expert
Jump to solution

NKS,

I reviewed both logs and I could not find anything in the logs that would help me

Any other ideas to look for?

Reply
0 Kudos
Nawals
Expert
Expert
Jump to solution

Follow below as they face same issue. This may help you.

VCHA Setup - vCenter HA has an invalid configuration. Remove vCenter HA to destroy the current clu...

NKS Please Mark Helpful/correct if my answer resolve your query.
Reply
0 Kudos
insearchof
Expert
Expert
Jump to solution

NKS

Article was helpful   but still have the error

Here is what I did this time

1. Turned off DRS on the Cluster.

2. Turned off  HA on the cluster.

3. Removed the HA from vcenter

4. configured all nodes to be on same ESXI host as current VCSA and on same Datastore.

so after a few hours my Passive node was cloned     100% completed

The the HA process was at 44% completed at 1:00 AM today I went to bed and now I see the same error on Vcenter.

But one thing did change the Passive node remained this time in the past procedure the Passive node would be deleted or it never finished cloning not sure

No witness node created.  

So not sure what is the invalid configuration error.

I have one thought.

The current VCSA was a passive node before my one node died on a bad hard dirve.

It has two network adapters on it. 

Before I build the HA should I remove that network adapter?  I have been just changing the ip address when I go thru the configuration.

Thaanks

Reply
0 Kudos
Nawals
Expert
Expert
Jump to solution

Don’t change any configuration.Make sure have all network connectivity. Here is the vmware guide for VCHA requirements and configuration. https://docs.vmware.com/en/VMware-vSphere/6.7/vsphere-esxi-vcenter-server-671-availability-guide.pdf

NKS Please Mark Helpful/correct if my answer resolve your query.
Reply
0 Kudos
insearchof
Expert
Expert
Jump to solution

NKS,

Good news I was correct I removed the HA nic on my VCSA  and deployed HA and now I have a passive and witness node

No errors on deploying the Ha this time.

On the HA status both the passive and witness show as  ? unavailable

I see the ip addresses I assigned.

I turned on Ha on the cluster

I turned on DRS on the cluster

But still showing ? unavailable

Also I am unable to ping the passive or witness ip address the have HA Network Adapters and on the same subnet as the active

I can ping the active HA Network adapter address but not on the passive or witness

Any ideas or suggestions

Thank you

Reply
0 Kudos
insearchof
Expert
Expert
Jump to solution

NWS

Update

The HA is not showing status on Passive and Witness Nodes both are up and running.

also on the page the edit button and initiate failover button are greyed out.

Any thoughts on this?

Thank you

Tom

Reply
0 Kudos
Nawals
Expert
Expert
Jump to solution

Check the permission also check in vsphere Web client .

NKS Please Mark Helpful/correct if my answer resolve your query.
Reply
0 Kudos
Nawals
Expert
Expert
Jump to solution

Make sure all 3 IPs should reachable. If not you have to check the network connectivity.

NKS Please Mark Helpful/correct if my answer resolve your query.
Reply
0 Kudos
insearchof
Expert
Expert
Jump to solution

Network is fine I am able to ping other device on 10.2.11.x just not passive and witness

I believe it is because the status in unavailable     Also Edit and initiate failover buttons greyed out   Tells me that HA is not setup correctly.

Any ideas?

pastedImage_0.png

Reply
0 Kudos
Nawals
Expert
Expert
Jump to solution

Yes configuration is not correctly. You have you follow the vmware guide to configure VCHA properly.

NKS Please Mark Helpful/correct if my answer resolve your query.
Reply
0 Kudos
brad_101
Contributor
Contributor
Jump to solution

Hello insearchof...

 

If your cluster still shows 'unavailable' after doing a Manual snap/deploy of rebuilding it.. but you can ping all 3 HA IPs you probably need to do 1 last thing.

 

On the current active vCenter, login via SSH..

Change the 'root' users default shell to bash.. (i use the interactive method)

 

# chsh

Prompt should come up like..

root@xxx [ ~ ]# chsh
Changing the login shell for root
Enter the new value, or press ENTER for the default
Login Shell [/bin/appliancesh]: /bin/bash

root@xxx [ ~ ]#

 

Then reboot both the witness and HA nodes.

 

It should then complete HA setup.

 

After HA is setup you can change back the 'default' root shell to be the appliance one of (if you prefer)

root@xxx [ ~ ]# chsh

Changing the login shell for root
Enter the new value, or press ENTER for the default
Login Shell [/bin/bash]: /bin/appliancesh

root@xxx [ ~ ]#

Hope that helps

Reply
0 Kudos
insearchof
Expert
Expert
Jump to solution

Brad_101

 

I just rebuilt the HA again this weekend  Same results 

I stopped DRS and HA 

Ran the HA Setup

Put the passive and witness on same host as active.

All Nodes are now online but passive and witness show unavailable.

 

I tried your suggestion but did not get very far.

SSHd into the vcenter 

root@TGCSNET-VCSA-1 [ / ]# chsh /bin/bash
chsh: user '/bin/bash' does not exist
root@TGCSNET-VCSA-1 [ / ]#

any suggestions

 

 

 

Reply
0 Kudos
insearchof
Expert
Expert
Jump to solution

Update

 

I went and restarted the nodes anyway.   

insearchof_0-1604967304654.png

 

I restarted the passive node a few times with same results.

 

Any ideas?

 

Reply
0 Kudos
insearchof
Expert
Expert
Jump to solution

After a few rebuilds 

I then restarted all the nodes

The witness node came back online up and green

The passive node came back as down. redeploy 

 

I redeployed the passive node and now all nodes are up and green healthy 

 

Now I need to migrate the passive to different host and datastore

 

That's tomorrows project

 

Thank you

 

.

Reply
0 Kudos