Link to home
Start Free TrialLog in
Avatar of anthonypham
anthonypham

asked on

VMWare HA Failed - All VMs are in unknown state. This includes VCenter. How to fix?

We have two ESXi Hosts and a Netapp filer.  Our VCenter server is on one of the hosts.  The network went down for one host and HA was initiated.  However, the network went down for the second host. We tried rebooting the hosts and the storage, but this did not fix the issue.

All VMs are now in unknown state.  How can I get VCenter working? Can I move it to a desktop and run it on VMWare Server? Would this be the fix?

What would be the best solution?

Thanks
Avatar of jmcilhargey
jmcilhargey
Flag of United States of America image

Login to the hosts individually and check to make sure that your storage adapters are connected. I frequently have to reset connections in my SAN and then rescan from vsphere to reestablish links for the guests.
Avatar of anthonypham
anthonypham

ASKER

I have tried resetting the filer connection and connecting to the hosts, but when I add them to the inventory, it says unknown (invalid) as the VM.  This includes the vSphere.  
Avatar of Paul Solovyovsky
Had a Netapp panic this morning as well...fun stuff.

1.  First make sure that you can browse the datastores or nfs mounts.
2.  Login to the host and do the same (using VI client)

If you can than that's good, all you need to do is remove from inventory and browse the nfs mount or datastore and add it back in.



1.  We can browse our datastores and NFS mounts
2.  We can log into the host using VI client.
We tried removing it from inventory and browsing the nfs amount to add it.  However, when it's added, it says Unknown (invalid).  We have older VMs on direct storage and it doesn't say unknown on those.  This seems to only occur on VMs on the filer.  These VMs were setup with Vsphere 4.  The direct storage VMs weren't associated with Vsphere before.

Thanks!
see if you can unmount the nfs mount and re-add back in.  Since it's NFS you don't have any re-signature issues.

If you have another ESX/ESXi host sharing the same mount you should remove from inventory and re-register to the second host.  Try to run, if it works then you can just vmotion over to he first host
I'm still waiting for VMWare support.  But is it possible to recreate the vxm file?  Would this fix the issue?
It is.   What I normally do, just in case

1.  Create a VM with no hard disks or a dummy hard disk
2.  Copy a vmdk file and vmdk-flat file to the new folder.
3.  Add exisiting hard drive ot the VM

Not sure if it will resolve issue but give a small VM a try.  

Have you tried to remove and add the VM from inventory localy on one of the hosts or on vCenter only?
ASKER CERTIFIED SOLUTION
Avatar of SunshineVK
SunshineVK

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
We had two ESXi hosts and vCenter was running on one of them.  We tried adding and removing VMs locally, but it was still invalid...  This was a good lesson to not have vCenter or the vCenter DB reside with all your other VMs!

Though we wanted to have the latest data, we ended up just restoring our database and vCenter by copying it from the .snapshot in Netapp.  It was from 5PM from the earlier day.  After I moved it, I reconnected our VMs.

Thanks for the help everyone!
SOLUTION
Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
You can browse the storage, excxellent, try a rescan a few times more, it often needs this.. But see the following as well:

You stated "However, when it's added, it says Unknown (invalid).  We have older VMs on direct storage and it doesn't say unknown on those."

What client are you using?  If you attempt to access vSPhere VMs with an older client you will recieve this message while trying to add them to the inventory, verify you are using the corrent client, connect the ESX web page http://ESXHOSTNAME and install it.