Backup-to-disk Question

[David_Penrice]

Hopefully someone can help me with this question.

We have started doing differential and full backups to disk during the week and then at the weekend we transfer those to a set of tapes.

Now my question is, how do I find out which .bkf file contains what data… for example, I backup 10,000 files and the using 10GB backup-to-disk files, I now have say 10 .bkf files and they get backed up to tape.

Two weeks later, I’m asked to recover one document from that 10,000 files, so I go to the restore folder and it shows me that I have 10 .bkf for that weeks backup but I know that only one of those .bkf files has that one document.

So is there any way of finding out from my catalogs (or anywhere else for that matter) which one of those .bkf files has the data or will I need to restore the whole set just to get the 1 file?

I hope this makes sense.

Thank in advance.

[nicktf]

although you get excellent throughput when backing up the .bkf files to tape, you get this inconvenience when trying to restore.

The preferred method is to create a duplicate job in the policy that you use to backup your data. This will expand the .bkf files and write individual files back to tape, which makes restoring considerably simpler at the cost of speed.

Incidentally, for you B2D job set compression to software, and for the duplicate job, set it to none (as the data has already been compressed as it is written to disk. This seems to give the best performance.

[CoolPix]

"Incidentally, for you B2D job set compression to software, and for the duplicate job, set it to none (as the data has already been compressed as it is written to disk. This seems to give the best performance."

I'm not sure about the performance but I use hardware compression also during the D2T stage because it seems that the data will still be more compressed. For example I  have an Oracle DB server with about 43 GB of data. After D2D the full backup size is about 23 GB and when it is duplicated to tape the backup size seems to be about 16 GB. The Oracle instance is shut down during the backup and the databases seem to compress really efficiently.

Message Edited by CoolPix on 01-16-2008 09:51 PM

[David_Penrice]

Thank you for the replies, sorry for the delay in responding.

I used the policy set and setup the jobs correctly but I'm confused on something...

Let's say i'm backing up a server that is 100GB and when it backups up to disk, i get lets say 99GB, how come when i then duplicate the job back to tape, the size on the tape is 89GB?

My exchange server reported, 101GB to disc and then showed 89GB when duplicated to tape..

Is this normal?

[David_Penrice]

Also, do i need to setup a different job to backup the IMG folders that are made from the SP/SQL agents etc? i'm sure that they are not getting backed up in the duplicate policy.

[nicktf]

Yes, this is a result of compression. The backup to disk job will show total bytes backed up from source, and the amount of data will reflect the amount of compressed data written to tape. This should be roughly the same as the size of .bkf files written to disk, for example I have a job which backs up 450Gb of data, compressing this down to 15 10Gb .bkf files which then get written to tape as 145Gb.

Of course, you can't beat a full restore from tape to ensure all is well.

Not sure about .img files as I don't use them here, though.


Message Edited by nicktf on 01-24-2008 04:26 AM