Video Screencast Help

Backup and restore of VM's

Created: 13 Nov 2012 | 8 comments

I know that to a lot of people this is going to seem like a simple question, but for the life of me with this new interface I am on the verge of insanity!!

Instructions have been very difficult to come accross for this product - Each time I search I get directed to best practices, and as this is the first time I have used Backup Exec in about 10 years knowing where things are is difficult.

My Environment which is standalone and has no interfaces to other networks or the internet.  The reasoning behind this is that we have some legacy NT databases and apps written in VB that refuse to work in newer versions of windows, so our thought would be to isolate it and let it run its course until the book of business expires!

The configuration is as follows:

HP ML350 configured as an ESX server containing LTO4 tape drive hosting 4 virtual machines:
Windows 2003 Domain Controller
Windows NT4 Member Server hosting a SQL7 database
Windows 2008 R2 Server hosting Backup Exec 2012 (The above mentioned backup device is mapped to this)
Windows XP workstation (This is the sysprep'd clone that we build the physical workstations from)

As well as the Virtual infrastructure there are several Physical XP workstations, One of which we have created a shared area on to allow a backup to disk

What I would like to know is 'What would the best method of backing up the Virtual Machines?'  We have purchased the VMware agent and have managed to perform single backups.

I would guess that Incremental and differential backups would be no good to me as backups of VM's are basically flat files, so all files would be changed every time a backup was taken.  So I guess Full Backup is my only option.  I have looked at the different backup strategies and concluded that the 'Grandfather' strategy would suit best Monday - Friday with a retained backup every 4th Saturday.  This would give me all backups from the last 4 weeks and the option of historic backup of VM's from the retained tapes.

It is highly unlikely that we would ever need to perform a GRT restore, basically all we would need is to recover a complete Virtual Machine as the data on these machines is mainly static - the DC for instance will have no AD updates, no patches will be installed on any machines as they are isolated!  

My questions therefore are?
Do I need to backup to the Shared Network drive (which is the Windows share on the workstation) and then commit to tape, or can I backup direct to tape?
If I do backup to the network share first, does Backup Exec overwrite the previous days backup or do I need to create a command to do this first?
Is my best option to just do a recurring daily full backup job Mon-Fri to tape with an expiration time of 4 weeks?  Additionally a recurring job set to every 4th saturday?
Finally does anybody have instructions on how I would restore a VM?

All Help appreciated

Comments 8 CommentsJump to latest comment

VJware's picture

Do I need to backup to the Shared Network drive (which is the Windows share on the workstation) and then commit to tape, or can I backup direct to tape?

- Either ways, both are fine..

If I do backup to the network share first, does Backup Exec overwrite the previous days backup or do I need to create a command to do this first?

- In the backup job properties, you have an option to specify how many days you want to keep this backup set for..

instructions on how I would restore a VM?

If you choose the vCenter / ESX to initiate the restore, you can restore the complete vmdk...else if you choose the VM itself from the Backup & Restore tab, you can restore individual files/folders (GRT restore)...Depending upon where you intitate the restore from, select the files which you want to restore, select the destination & bunch of other options (such as notifications) appear in the guided restore wizard..

Have a look at this KB too ~ http://www.symantec.com/business/support/index?pag...

VillanOfAston's picture

Thanks,  I will try direct to backup tape and have a look at the docs!

Colin Weaver's picture

You can't officially  use the tape drive when it is connected (or physically part of) the ESX host. It has to be connected to a Physical Windows server running Backup Exec to be supported.

 

Also whilst we probably can backup the NT 4 machine as a complete VM, it is unlikely any indidividual file or SQL 7 database capability will function as we stopped supporting both NT4 and SQL 7 some time ago. This lack of support will affect VMware backups and traditional RAWS based backups.

Westholme's picture

Hi - I've searched a lot on this which is why I'm at this thread.

Can anyone tell me if you do a FULL VM restore (VMWARE) does it stage the file somewhere before overwritting the exiting VM please?

The reason I ask is, we tried to restore a full VM to a new location (a NAS drive elsewhere). We clicked the "different location" option and put the IP address in and a file path.

When the restore was set off we saw a byte count but no files appeared in this different loaction. When we looked at the job the location said it was restoring the VM environment (!).

The job was cancelled sharp'ish! The existing VM works fine and we have no idea where the bytes where sent to??? The Staging location is default but there are no files in there?? Teh job details were checked and all appears well but I am not sure what happened to the restore.

Can anyone please help explain?

Thanks

VJware's picture

Redirection of the complete VM (vmdk) will work only to the local media server. It cannot be redirected to any other location. (Unless you are redirecting to another VM or ESX and not as a flat file (vmdk)

Staging occurs only when you are attempting GRT restore from tape.

Colin Weaver's picture

Just to add some meat to your answer

It is acknowledged as part of a known issue - see the info mentioned in the workaround listed in http://www.symantec.com/docs/TECH188010

 

Westholme's picture

Thanks - bearing in mind there was no warning of this, where did these bytes get written to then?

We told the restor to go to \\IP.address\sharename\VMRestores\Server-SQL - (the VMRestores\Server-SQL part was a newly created folder sturcture specifically for this restore on a NAS).

The "job log" shows the restore as "Restoring data to VMVCB::\\VCENTER\VCGuestVm\(DC)PM DC(DC)\vm\Server-SQL\IT Software\VMRestores\Server-SQL."

Note it seems to have used part of the file path in the form of the part show in bold however this doesn;t seem to have been created at all???
 

Westholme's picture

Thanks, bearing in mind this is the first I've heard of this inability to redirect (and it lets you go through the process) - where will my full VM have started to write to please?

I had no other warnings etc.

 

Thanks