Making Duplicate copies of a B2D folder

[Rick_Thering]

I am trying to work out a disaster recovery situation.  We are currently using a single disk connected to the server as a B2D device.  I would like to take a copy of this disk over to a rotating off site disk.   What I am proposing is that we would run our normal backups to the B2D folder.  Then as a post backup command we would run a RoboCopy of the on-site disk to the off-site disk that is connected to the server at that time.   Then the next day the off-site disk would get removed and replaced with a 2ND off-site disk for that nights backup. 

In my mind I would then have 1 on-site disk that would have everything.  1 off-site disk that would have everything, and 1 off-site disk that would be a day behind in it's content. 

In my reading on this I have  gotten the impression that this is not going to work properly, that the files would not be able to be recovered. 

Has anyone done this?  Will it work?  If not does anyone have any suggestions for alternate methods?

Thanks


Rick.

[RahulG]

You can set up a duplicate backup job in backup exec for duplicating the backup set from the onsite b2d to offsite b2d . You can click  on file | new | Duplicate backup set job

select the option to duplicate backup for an existing job .

So as per your question it work ...  well I am not sure about this statement   "Then as a post backup command we would run a RoboCopy of the on-site disk to the off-site disk that is connected to the server at that time".... 

Instead of running any post backup commad it would be easier to run a duplicate  backup Job
 

[Rick_Thering]

Thank you for your response. 

I have made an attempt at using the Duplicate Backup for an existing job.  The problem that we run into with this is directing it to the rotating off-site media.  We need to be able to go to 1 of 2 different off site disks.  I thought that this might be accomplished using a device pool, but when they swap disks, it still wants to see the first disk it had used. 

Also, unless we do a full backup daily, if I use the duplicate job, I will not get everything needed to restore the server on each of the off-site disks. 

As an example, If I want to do a weekly full backup with Daily Differentials.

Day 1 I would get the Full backup to the on-site Disk, and a Duplicate to the off-site disk 1. 
Day 2 I would get the Differential to the on-site disk, and a duplicate to the off-site disk 2.
Day 3 I would get the differential to the on-site disk, and a duplicate to the off-site disk 1
Day 4 I would get the differential to the on-site disk, and a duplicate to the off-site disk 2
ETC...

At the end of the week, I would have everything I needed on the On-site disk, but a big mess figuring out what I needed from the off-site disks to get a server back up and running. 

[RahulG]

When you run the duplicate backup set the catalog is also written in the catalog folder , so you can know the data which you  want to restore is on which b2d file . if the drive to which you are going to backup is going to be changed daily you can create a removable backup to disk folder in your backup exec .
you can set up proper media set so the bakup file does not get overwritten accidently ..
Also if you are trying to restore on the offiste location from different backup exec install you would need to just catalog the both the disk and when you run the restore it would ask you for the file which contain the data to be restored ....

[Rick_Thering]

Rahul, 

If I understand what you are saying, Setup the 2 Removable B2D folders, Set the Duplication job to backup to each one alternating for each day.  And make sure the human factor puts the correct disk in place.  (I don't see a way to have it just use which ever media is in place correct?)  

Now if I need to restore, I will need both off-site disks.  Catalog Each, and do a restore.  Is this correct?

Since this backup is for disaster recovery, and one of the two disks is always going to be connected to the server.  My concern is that I have not protected the server.   In the event of a physical disaster, the disk that was connected to the server when the disaster happened, is likely to be unavailable/damaged leaving me with only part of what is needed to restore (Unless every backup is a full backup) 

[RahulG]

Media set defines the overwrite prootection and append period for the bkf files . You define a media set in the backup job which determines for how much time the backupset should not be overwritten

Yup if you are going to restore any data you would need both the disk connected to the server . and need to catalog both of them

In this case i Guess you need to run Full backup daily .as if you run differential it would not accomplish the startergy u want as oe of the disk might will be always connected and if its a drive containing full backup then u cannot restore the differential as well because during the restore u restore the full and the last differential bkup .

[Todd_Payne]

We do a very similar thing here.  I run full backups every Friday night to Friday backup to disk folders.  Then I copy these Friday folders with a batch file along with other files to a 400GB removable disk and send them off site.  I have successfully restored files from the off site removables drives but it is a little tricky. It seems that backup exec thinks the off site file is already cataloged.  There must be something in this old Friday backup file that tricks backup exec into thinking it is the current Friday backup file. This is probably because I just copied the file instead of running a duplicate backup sets job.  Here is how I work around the issue.

Before I attempt the restore, I "delete"  the back up to disk file from my catalog.  To delete the file from the catalog I first associate the file with the retired media set and then using backup exec  "delete"  it.  This removes it from the catalog but does not remove the actual file from the file system. 

Then I copy the off site file to a special backup to disk folder I have  named Restore.  I Inventory the backup to disk folder and  catalog the "imported media ".  Now every thing is synced up and I run my restore job. 

After restoring  the files I needed, I have to put things right.  I remove the off site file that I used from the catalog much the same as I did for the current Friday back up to disk file .  Then I Inventory my Friday folder to get the current Friday back to disk file back in the catalog.   Finally I associate the current Friday backup to disk file with the scratch media set so that it will can be used on Friday again.  I don't bother to catalog the media unless I need it for a restore as it is going to be overwritten on Friday any how.

The longest part of this process is cataloging the imported media.  We  run separate backup jobs for each server to separate backup to disk folders and that helps reduce the time it takes to catalog the media.  Our largest backup to disk file is 137Gigabytes.  It takes about 30 minutes to catalog the file.  Considering that it usually takes a day to get the off site disk back on site, it is not so bad. 

It is not the most Ideal plan.  But the primary focus of this strategy is a disaster recovery situation in which our servers were lost so I would be installing Backup Exec and then copying the off site files to the backup to disk folders.    

Hope it helps.

[Rick_Thering]

Todd,

Thank you very much for your post.  It appeared to me that this was the case, and I am glad to get some confirmation.  I did do some test by moving the data from the on site disk that I was using to an separate disk, then connecting that to a separate server and cataloging and restoring data. 

I agree, it is not an ideal plan as it does cause you to do the catalog, but as you said this is for disaster recovery, not restoration of a user's oops. 



Rick.