Yesterday I had some time on my hands, noticed 6.5.0b for vCenter appliance was released, and 6.1.4 vdp was available.
Upgraded vCenter appliance no problem (from 6.5.0a). Then tried to upgrade 2 vdp instances to 6.1.4 from 6.1.3. Both upgrades failed (no errors, just failed around 80% stating I had to recover from snapshot). Also, after reverting, the vDP plugin is no longer loaded in the vCenter web interface.
I restored the vdp appliance from snapshot, restarted, still no vdp icon to access. I also removed the vdp from the mob interface, and tried to re-register the vdp appliance using vdp-configure. It failed on re-registering to the vcenter instance.
At this point, I decided to try to just re-create my vdp instances. I downloaded the 6.1.4 ovf file, deployed, and during vdp-configure it fails (tried 3 times, on two different datastore types iscsi/nfs).
I still had the 6.1.3 ovf, so I tried to deploy the older version, fails with the same errors as 6.1.4 is failing with.
Keep in mind I deployed 6.1.3 a month ago with absolutely NO issues. I'm wondering if this has something to do with vCenter 6.5.0b.
The errors when deploying brand new instances (both 6.1.3 and 6.1.4) are attached a screenshots.
As mentioned above, I tried both on NFS and iSCSI datastores (iSCSI datastores running VMFS6).
Anyone know anything about this?
So, just an update.
I tried and re-deployed probably 10+ times trying different things (upgrading the hardware, vmware-tools, etc...). Anyways, I FINALLY got it to work on deployment today... I hope this helps someone else...
Some of these steps may not be needed, but I'm listing exactly what I did to get it to work.
1. Deploy OVA
2. Right click Appliance, upgrade VM hardware version to latest (6.5)
3. Go in to settings of VM, configure vmware tools to update on boot (probably not needed, but just listing what I did)
4. Boot VM
5. Load up vdp-configure web interface, during process
6. Using the VM console, log in as root, set timezone (I'm using Mountian) using "cp /usr/share/zoneinfo/America/Edmonton /etc/localtime"
7. Restart vmware-tools (for time update, and just for the heck of it) using "service vmware-tools restart"
8. Go back to vdp-configure interface, on vcenter configuration used "vsphere.local\user" format, instead of "user@vsphere.local"
9. Few errors on showing storage information, but vdp-continued
10. Few errors on showing CPU/MEM info, but vdp-continued
11. Configuration completed, booted!
Major things I did different than normal
1. use vSphere.local\User instead of user@vsphere.local
2. Upgraded the VM hardware version
3. Set time zone of VDP linux guest OS, restart vmware-tools service
Just another update...
I deployed a second VDP appliance doing everything identically, and it failed with the original errors. I have no idea why this is failing...
Well, I finally found a re-producible fix (forget about my first draft of fixes):
During the vdp-configure web interface, when it gets to the storage part, and errors present re: storage configuration (it won't show any datastores), simply hit back, and next, and keep going back and next until the datastore info populates, presents and shows properly.
When they show properly, continue with the vdp-configure web interface. Chances are the CPU/Mem config will error (showing 0//0, however sometimes it is accurate), but you can continue as this won't cause the configuration to fail.
I've tested, and re-produced this multiple times now.