Hi Paul,
When doing an Automatic Upgrade on the same hardware, you need to ensure a lot of different prerequisites are fulfilled:
1. Available disk space (at least 990GB of free space on the ESXi host where Admin VM is residing).
2. Ensure all the ESXi hosts in your deployment have NTP configured to “Start and Stop with host”, that NTP is started, and that time is in sync on all hosts. If not, and if you need to make adjustments to time, this has to be done when all the VMs are gracefully shut down, as DB data is very dependent on time stamps.
To check this do the following:
a) Use vSphere client to connect to vCenter.
b) Click on ESXi host.
c) Click on Configuration tab
d) Click on Time Configuration link in the Software section
e) Click on Properties in the top right
f) Verify the time and that NTP Client is Enabled – if time is incorrect. Don’t adjust it why your VMs are running. Find the time to put the system into Maintenance Mode, gracefully power them off (Power > Shut down Guest), and then adjust the time. (The time should be adjusted when NTP Client is started and reconnected to NTP server. If that doesn’t happen automatically, you may need to put ESXi host into Maintenance and then Restart it, bring it out of Maintenance, and then start all CWMS VMs)
g) If the time is OK, click on Options button
h) Ensure Start and Stop with Host is selected (if not, correct it). If the time is correct and NTP Client is running, but this setting is wrong, you can make the change without powering off the VMs.
i) Repeat this on all ESXi hosts in your CWMS deployment.
3. When doing an Automatic upgrade, you will need to dedicate a temporary IP address and hostname to CWMS 2.5 Admin Upgrade VM. When dedicating a temporary IP address and hostname to the CWMS 2.5 Admin upgrade VM (per documentation), ensure that your firewall between the IRP server and internal network applies all the needed rules to this temporary IP address and hostname as well. Hence, all the rules that apply to the currently running Admin VM IP address and hostname, should be applied to the temporary IP address and hostname of the CWMS 2.5 Admin upgrade VM. This is needed, so the Automatic Upgrade process can automatically create 2.x IRP VM on the ESXi hosts in the DMZ.
4. When doing an Automatic upgrade, Ensure you are deploying CWMS 2.5 Admin upgrade VM on the same ESXi host that is running your existing Primary Admin VM, and make sure your 1.5 CWMS VMs are all up and running.
5. Ensure your vCenter is reachable via hostname and not only via IP address. If you run a vSphere client, in the IP Address / Name of the server, enter the FQDN of the vCenter server and ensure that is resolvable in your network and that you can actually connect to it. This is important for Step 6 of the Automatic upgrade instructions where you need to enter vCenter hostname and credentials for the Upgrade VM to connect to.
6. When doing an Automatic upgrade, use Standard vSwitch on your VMs instead of Distributed vSwitch. If you are using Distributed vSwitch, you will HAVE TO do a MANUAL upgrade.
To check this do the following:
a) Use vSphere client to connect to vCenter.
b) Click on ESXi host.
c) Click on Configuration tab
d) Click on Networking link in Hardware section
e) Click View: vSphere Standard Switch button. If you see your NICs configured here, it means you are using Standard vSwitch.
f) Still, click on View: vSphere Distributed Switch button, and ensure there is nothing configured there.
7. Don’t create any VM snapshots on the existing CWMS 1. 5 VMs. CWMS 2.5 upgrade won’t let you proceed if any snapshots are detected. Also, snapshots aren’t really needed since there is no change being done to the original system. Only data will be copied from the original system. At a certain point in the upgrade, the Automatic upgrade process will power off the original system, so it can create all the new VMs and assign the same IP addresses and hostname as the original system had. If for some reason the upgrade is not good, all you will need to do is power OFF 2.5 VMs and power ON 1.5 VMs and the system is back up the way it was before the upgrade.
Finally, Automatic upgrade can take time as you first need to deploy 250 Upgrade Admin Vm. If I understood that is where it seems you are having an issue as the VM is not being deployed and it stays stuck. That piece of the upgrade should actually be rather fast as it is just deploying the VM from OVA (I would say 20min normally if not faster).
Are you using vCenter to deploy this VM, or you connect vSphere client directly to ESXi host and then try to deploy OVA? If yes, that won't work. You need to use vSphere client to connect to vCenter server, and then deploy VM from OVA to a specific ESXi host.
I hope this will help.
-Dejan
