vCenter 5.0 Rebuild

[dowdingit]

Setup: All vCenter/vSphere 5.0 products

VM01 - 1st ESXi host - currently added to the vCenter cluster

VM02 - 2nd ESXi host

New vCenter Server

iSCSI SAN

I'm a new Systems Admin, taking over for a recently quit (shape up or ship out situation - he shipped out) sysadmin.  The vCenter server crashed, took the DB with it with no backup, and a VM Host running esxi 5.0 (VM01) went down.  HA was setup, and so all virtual machines moved over to the other vmhost (VM02).

I have since spun up a new VM, installed vCenter Server 5.0 on a VM created on VM01 host.  I added that host to the cluster on vCenter, and am now planning on adding the second host VM02 to this as well.

My questions are pretty simple I'm sure (I'm coming from a Hyper-V shop previously)

I see all the other VMs on the vCenter server here, underneath VM01, as Unknown - when I migrate the VM02 to this cluster, will those show up again, or will I have to setup the DRS, vMotion, and/or HA again?

HA was previously setup, but from my reading I see that the config (or some of it at least) is on the hosts themselves, can I import that configration back, or am I better off recreating that?

I'm doing this after-hours in a maintenance window, but the shop is technically 24/7.  Should I expect any real downtime?

Thanks for any help/suggestions! 

[abhilashhb · Accepted solution]

You have all the VM's running on the second host right?

When you add the second host, it will bring all the VMs under the inventory. Whatever is unavailable/unknown, right-click and remove from inventory. These are just orphaned files that are no more needed.

All you need to do with HA is that you need to add the hosts into the cluster and turn on HA. It will do all the required configurations so you don't have to worry much about it.

As far as all VMs are on one of the host and you add that host and remove all the orphaned Vms from inventory, nothing will go wrong. I don't see any downtime that you'll require here.

[dowdingit]

Thank you very much for the quick reply.  I have marked your response as correct, and will update this thread with the final results to let everyone knows how it goes. 

[abhilashhb]

One more thing you need to do is, Login into individual hosts and make sure they are not showing up as unknown because of any other reasons

In your case, in the first host, you should not find any VMs and when you login directly to 2nd Host, all the VMs should be up and running on it.

[dowdingit]

That is correct   I have removed the unknown inventory items, and adding the second host now. 

[dowdingit]

Things went pretty well, however, I mistakenly added the hosts to the datacenter, and not my HA Cluster.  From what I read, it looks like I have to shut down all the VMs, remove the hosts from the Datacenter, and readd them directly to the cluster.  Then I can power on my VMs.  Does this procedure sound right, or is there a way to add them to that cluster without shutting down all the VMs on each host? 

[Senthil0281]

No.You can just remove the host from data center and add it again into cluster without shutdown the vm's.

[dowdingit]

Update:

I performed the following steps (approx time rounded to nearest 15mins)

Created Snapshots of all VMs, verified backups of VMs to backup server/tapes.  Did a couple of test restores after hours to verify backups.  (2.25hrs)

Removed Host VM01 from Datacenter. No VMs went down in the process, full functionality in place.  VCenter server was on this VM, and continued to work the entire time (15mins)

Added Host VM01 to HA Cluster in Datacenter (15mins)

Removed host VM02 from Datacenter (15mins)

Added host VM02 to HA Cluster (15mins)

Tweaked DRS a little and tested vMotion capabilities (15mins)

All in all, about 3hours in total for entire process, the vast majority of time being taken up by backups and precautions.  I did notice doing the snapshot of the file server, which hosts static files, as well as a couple of Access databases caused about a 60s hiccup in network functionality.  This caused some issues with the Access databases.  They were fixed that evening, and a couple of things handled in the morning.  Otherwise, I need to finish tweaking settings in the cluster once I research a bit more, but all seems to be good.  Thanks for the help/suggestions!  Rated helpful and correct answers accordingly.  Hopefully this helps future admins!