admin portal as a client?
I am wondering if there is a way to use a temporary instance of RHV-M to setup a host for local storage, etc. andthen to create a new production instance of RHV-M and continue to manage the same datacenter and host?
I notice adding a new RHV-H host to a fresh instance of RHV-M doesn't bring in any of the configurations for the network, local storage, etc. In vmware you'd just use the ESXi client, since RHV doesn't have any lightweight client, does losing the RHV-M vm lose the whole datacenter configuration unless you had a backup? I can't use hosted engine, that won't help in my case, and i'd like to be able to use a temporary instance of RHV-M if my production instance was down.
Thanks,
Responses
Could you please elaborate about your setup?
Do you use a physical RHV-M or a virtual?
What do you consider down for your production instance? Just RHV-M or also other hosts?
What storage are you using for templates and VM's? Is that the local storage on the hypervisors you are referring to or is it on networked storage?
Please indicate why you cannot use hosted engine?
There is a better way: hosted engine :) Everyone asking you about hosted engine should be an indication of the better way.
Considering your scenario. How are you going to switch from your VirtualBox RHV-M to your VM RHV-M? The RHV-M and the hypervisors are tightly coupled. For instance, the data of templates and VM's in the postgresql database on the RHV-M and the contents of the storage on the hypervisors should be identical. When you start the VirtualBox RHV-M, it is communicating with the hypervisors and especially with the SPM. When you start the VM RHV-M, all of a sudden those hypervisors should be "controlled" by another RHV-M.
The most valuable asset in your RHV setup is the storage as it contains the VM's and templates. We use NFS storage and with major upgrades like 3 to 4, we detach two hypervisors and install a new hosted engine with a temporary storage domain. Then we stop the old RHV-M and import the storage domain and add the remaining hypervisors to the new RHV-M.
Creating a new RHV-M with a small temporary storage domain and then attaching the real domain and importing all templates and VMs has worked well for us instead of an inplace upgrade. I am not aware whether this is an official procedure.
When you lose your RHV-M, all you need to do is install a new RHV-M, add the (NFS) path of the storage domain and add all hypervisors. This way you lose all events in the RHV database, but that is a minor point in my opinion.
Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.
