Offline Primary server recovery

[KeirL]

Hi

I would like to test the recovery of a production NetBackup 10.3 Primary Server but do this 'offline' so as not to interrupt the production environment. I plan to build a new Windows VM on an isolated network to allow me to create it with the same server name and IP address as the original. I will then take a catalog backup to a basic disk storage unit onto a local drive on the production NBU server..

I'm hoping I can then copy the catalog backup files, associated .drpkg file and catalog DR file to my test VM which will give me everything I need to recover the NBU primary server - should this work? There will not be any other servers on this network (no AD, DNS, or Internet access) but I can use 'host' files If necessary. Our environment is small with just this single NBU Primary server that also acts as a media server.

Once I have configured the VM with the same name and IP address, I will install the NetBackup binaries to the same location (D:\Program Files in our environment) and also select the installation in DR mode and use the .drpkg file. The nbwebsvc user is local to the server and not an AD account - so I will create this as a local account on the test server.

I will also ensure the volume used by the Disk storage unit (on the production NBU server) also exists on the NBU recovery VM and that the catalog backup files are copied over to the same path and folder.

Next I will, run the catalog recovery and select the DR file that will be copied locally with the catalog backup files.

Question - Do I first need to create the disk storage unit manually and also run an 'Import' on the catalog backup files before the start the catalog recovery process? Or will it simply work if I just make sure the backup image files exist in the same path as detailed within the DR file?

I'm just doing this to 'practice' the process should it need to be done for real, but I just want to make sure it 'should' work using this approach so I can troubleshoot any errors.

many thanks 

[Nicolai]

Hi @KeirL 

It look like a good plan. Even thru is purely for training. Even a failed attempt is a lesson learned.

The basic disk stu is very simple , it just a "pointer" to a disk area. Create the STU in Netbackup and place the files from the catalogue in it and you should be ready to run the catalogue recovery wizard.  Remember to bounce Netbackup daemons after catalogue recovery.

From my experience , you will likely have to run the recovery process a couple of times to iron out issues encountered. Write down the experiences in a recovery instruction / checklist. 

[KeirL]

Thanks @Nicolai 

I might try this tomorrow as I'm pretty much set to test.

So for the Disk STU, all I need to do is to create the STU with a pointer to a volume that is the exact same path as the STU on the original Primary server. I place the catalog files into this location, BUT I don't need to run an 'Import' media task?? 

I see the  DR files includes the path to the catalog files so I guess it get's it from the path in the DR file rather than having to check it's own catalog for the location.

Am I also correct in thinking I can use this exact same procedure to migrate the server to a new version of Windows? The current server is Windows 2012 and we're getting rid of these in favor of Windows 2019 or 2022. I can also use this process to migrate Windows platforms can't I?

thanks

[Nicolai]

hi @KeirL 

Well - The basic stu is the pointer to the disk location.

On top of my head, no, you don't need to import media task when using the basic disk storage unit, that is why i prefer using it for DR backup /restores. It is so much more simple. If using MSDP/Tape or storage servers, then media import is required, because Netbackup doesn't know where data would be residing. And from my experience, import in a DR situation is always troublesome.  

In theory, yes, that is the support way replace hardware. It really comes down to the size of the Netbackup image catalogue. If size is in multiple terabytes, then other techniques is needed.

Have fun running the DR drill. Remember firemean and airline pilots also do drills, so why not in IT. 

[Nicolai]

The trouble shooting guide also contain a good handful of recommendation for catalogue recovery. Especially around certificates, as they are not part of the recovery procedure.

https://sort.veritas.com/DocPortal/pdf/15179611-160934780-1

[Tape_Archived]

The media import is not required as mentioned by @Nicolai 

I followed similar steps to refresh my OS from 2008 to 2016. I had created basic local stu on a file share to store the catalog image and then create the basic stu on new server providing the same path.

@KeirL  have you thought about alternative option to take a VMware backup of your Master and restore the VM as new/offline master server?

[Michal_Mikulik1]

Hello guys,

just one note - you don't need storage unit definition(s) for any restores including Catalog. You need accessible folder/file with catalog backup (for disk based catalog backup), or accessible library device+tapes for a tape based backup, bot not necessarily a storage unit as a logical entity on the top of it.

regards

M.