Steps for Dissimilar System, dissimilar disk restore in BMR?

[someoldman]

Netbackup 6.5.4.   Client is Windows 2003, running on a physical server.  To test the functionality of BMR, we want to BMR just the system volumes (OS) to a Virtual Machine.  (VMWare).  I suppose the same steps would apply to BMR on completely different physical hardware as well.

At first, we thought the process was to "Prepare to Restore" the physical server; make a boot disk/iso, and then boot the VM from that.   Doesn't seem to work that way.
Is it instead:
1) Prepare to restore the physical.
2) Make a boot cd/iso using that physical server.
3) Boot the VM from that iso to get a "base" OS in order to do a "Prepare to Discover" the VM
4) Discover the VM.
5) Make a copy of the configurations from the physical server. Edit that to reflect info from the "Discovered" VM.
6) Prepare to restore the copied and editted  client configuration for VM server.
7) Boot vm from original iso again, and since the Prepare to restore now has the new vm configs, things should progress.


Am I making this too hard?  Too easy?    Adding unnecesary steps? Leaving something out?

We could bring up a VM with Win 2003 very quickly, but what we are more after is proof of concept if we had to do this with different physical hardware.

[Nick_Morris]

From my understanding (i have not tested yet but been assured by support it should work) is do the following:

Create BMR boot ISO and store in location ESX can see it
Create blank VM from template or find VM to use
Boot VM from BMR ISO and create a discover config.
In BMR restore clients create new config and initialise drivers etc from discovered config
Do BMR restore on VM and done using same IP/Name etc.


Alternatively if it is not in a DR situation, then you could just use the P2V (VMWare Converter) to make a VM from a physical. Just as quick and easy in my experience

[someoldman]

We are wanting to test BMR with dissimilar system.  We just happen to have VMs available for testing instead of physical servers.   In many of our DR scenarios (something less than Data Center destruction) we would replace some of our current  physical boxes with new hardware.  

Another department within our company has "ownership" of the physical box we are using as test, as well as their own VM environment.  So it might be awhile before everything gets tested.

[Nick_Morris]

If it's just dissimilar BMR you are doing then that's pretty straightforward from my experience (though untested for me onto a VM as of yet). I just follow pretty much similar things as before:

* Make sure you have a BMR boot CD/ISO with same version of NBU on (only if you want to automate it completely or get new equipment that needs a driver adding it will need to change the CD)
* Do discovery of where you want to move the machine to from BMR boot CD
* Use discovered config to create new config for BMR restore and set that running
* Use BMR restore option
* Occasionally i have found post restore you might need to re-apply some drivers

BMRs biggest issues usually come from incorrect DNS resolution from personal experience. Also i think it was 6.5 or 6.5.1 where discovery didn't work properly but thats now fixed. I would not recommend changing IP address/name during a restore either. It just doesn't seem to work really. Though you generally use a BMR in a disaster and not for cloning really, so shouldn't be an issue. Also make sure old machine is switched off so no IP conflicts etc during restore.

BMR is the only thing for Netbackup for me that just works now with no issues. Shame same can't be said for most things :)