BE 2014, USB Drive Rotation, Deduplication? Full/Incr?

[Anchor_Consulti]

Hi all

Looking for solution options for a small client who will be running BE 2014.  They want to do nightly backups that will goto 2tb USB 3.0 drives that will be rotated out each day and brought offsite.  We're looking at having 5 drives total, one for each weekday.  Unfortunatly, they can not backup to AWS or something else over the cloud due to the system that they're backing up being completely offline (long story).

I know this was possible with BE2012 - we've done it before with a storage pool.  Essentially, BE didn't care which drive was inserted, it would backup to that.  If they skipped a day of rotation, they'd get 2 backup sets on the drive if there was space.  We've done this before with Western Digitial My Passport disks, setting the disk ID to unique ID's before inserting into the BE system.

Questions:

1. Is there a new / better way to do this with BE2014?  I can't seem to find any formal guide.
    
2. This client has almost 2tb of data that needs backing up and unfotunately only has usb 2.0 ports.  We're looking at adding USB 3.0, but if that's not an option is there a way to safely implement a strategy that does a full backup over the weekend to the connected drive and then incrementals throughout the week so there's not huge amounts of data that needs to be copies during the week?  FYI, this is a hyper-v w2k8r2 evnironment, there will be some complete VM's being backed up, others where we just backup critical data.  There's an Exchange 2014 server too that we absolutely want to protect both in full and with granular message recovery.
   
   a) I worry about a recovery, figuring it which bits are on which disk for recovery.  Is this an issue or does BE2014 indicate which disk is needed?
   
   b) I was also thinking of having some sort of deduplication strategy in place as an alternative, but I don't know how this would work with rotating disks - especially since we can't guarantee that the client will rotate correctly.  Thoughts there?

Any guidance would be apprecited.  Seems like something that a lot of end users would want to implement, and I'm hopefully that there's a way....

[pkh]

There is no change in the way disks are rotated in BE 2014.  You still need to use a storage pool.  See my article below

https://www-secure.symantec.com/connect/articles/how-rotate-external-harddisks-be-2012

For BE 2014, you can now specify that BE use the disk with the most free space, but this does not apply in your case because you would only have 1 online disk at one time.

If you do full and incremental backups, you should keep the entire backup chain on one disk because BE would require all the backup sets to be present when there is a restore.  I am not sure whether you would be prompted for the correct disk if it is not online.  Regardless, you would have a problem because your disks are rotated offsite.

You can only have one dedup folder per media server so I think this solution is out for you.

[Colin_Weaver]

Just for info the requirement to keep the full and incremental backup sets on an online disk (or the same disk if the disk will ever be detached) relates to GRT enabled backup sets. In thoery standard remote agent Full and Incremental backup sets can be on different detached disks (although it is possible that restore jobs created from the new style 'synthetic' selections view might have some issues with media requests)

BTW this GRT limitaion is true for all versions of Backup Exec that offer forms of GRT as such is not a new limitation just for 2014. Also the 'synthetic' restore view (if it does have limitations in this area) was introduced with BE 2012

[Anchor_Consulti]

Thank you both.  

Some more information on how I'm now planning on doing this for the client:

1) We'll have a large drive (array) stay onsite and have scheduled full backups there.  Dedup will be used.  This will be the primary source for restorations if necessary.

2) The USB drives that will go offsite will be used for restoration only if the main drive (#1 above) fails or is lost in fire, etc.

So, scratch deduplication on the usb drives.

"If you do full and incremental backups, you should keep the entire backup chain on one disk because BE would require all the backup sets to be present when there is a restore."

Besides the annoyance of needing all offsite drives brought in in the event that recovery is needed from them, would incrementals to them work?  If so: my new thoughts / qutestions

1) We'll use more than 9 usb drives, not 5.  

2) Full backups will be done Friday night (probably into Saturday) onto the one of the multiple "Friday" USB drives that we'll have, #1-#5 for the 1st through 5th Friday of the month - that way we have weekly ful backups offsite going back a month

3) Incremental backups will be done Mon-Thu nights to the other USB drives, with a retention of of 2 weeks

So, would this work??  If we'd need to restore from the media that goes offsite, we'd need all of the media there, but would BE be able to do the restoration?  If the BE machine was lost to disaster, could the portable media be indexed by a new install and restored from - even if it means there's a bunch of disk swapping?  Would plugging all of the drives up at once to to a USB 3.0 hub back a DR recovery work better?

Moreover, am I missing something here? Shouldn't this be a pretty standard setup, essentially replacing tapes with large USB drives?  Is there a better way?

Thanks again.

[RvStenis]

i'd like to bump this thead, seems to be quiet since november.. I'm in pretty much the same situation, vm backup to dedup storage and wanting some offsite sets of backups while having limited internet bandwidth on a remote office location.

So is there a guide on this allready?

Thanks,

Ron.