Backing up 500GB Incremental nightly

[norrin-rad]

Hi

We have netbackup 7.1.0.3 installed on a windows 2008 cluster, which we are setting up and is not yet in use.

There are two SAN drives, drive T and U and a quorum drive, which is Q drive.

There are also two virtual servers configured, which point to each san drive.

Virtual server 1 = T drive (1Tb of data)

Virtual server 2 = U drive (1Tb data)

We have three policies configured 

1st policy - backup all drives on nodes 1 and 2, excluding san drives - incremental daily, fuill weekend

2nd policy - backup virtual server 1 drive (T drive) and Quorum (Q drive) - incremental daily, fuill weekend

3rd Policy - backup virtual server 2 drive (U drive) - incremental daily, fuill weekend

The first two policies work fine, no issues, however the 3rd policy seems to backup 500gb incremental every night. The servers are not in use, so there should be no changes (or very little i.e megabytes).

We've tried several settings, adding the U drive to 2nd policy and changing the T drive to 3rd policy, changing client settings from timestamp to archive bit and vice versa.

This has been ongoing for a while and we really need to get it dorted, but we can't work out why it would be doing such large incrementals

Any help would be appreciated. Thanks

[Michael_G_Ander]

Could be caused by the U drive failing over from one node to another between backups, have seen this unfortunate behaviour especially with TIR

Other things I can think of is

- an application that changes something on a lot of files some example attributes, as for example some antivirus programs

- open database files or their backup files

- temporary files

One thing that is interesting is to see what the OS finds if you do search for files modified in the last day on the U: drive

[Genericus]

1. Sometimes, you need to do a valid full, before you can do an incremental.

2. Sometimes, the way you set up your systems means that the archive bit cannot be set, so the data looks new every day.

Can you set the client atributes to use timestamp instead of archive bit? That has helped me in the past...

[norrin-rad]

Thanks Michael\Genericus

I'll check to see what files I can find modified in the last day (good shout, didn't think of that) :)

As for the timestamp, I've tried that and unfortunately it made no difference.

Thanks for the replies. I'll see what changes have been made in the last days. 

[norrin-rad]

Zero files found modified in the last 2 days :(

[quebek]

Hi

See in backups what files were backed up with INC backup... then compare on the live drive.

[norrin-rad]

Thanks for the reply, Sorry new to NBU, how do I check what files were backed up as part of the incremental backup? 

[norrin-rad]

We had the same thing last night .. 568GB incremental backup :(

[quebek]

Hmm

You need to do some reading - about Backup, Archive and Restore GUI... Available in NBU admin guides...

[norrin-rad]

Hi Thanks for that, up until now been looking in the admin console, just checked the backup, archive and restore console. There are two folders on the drive and it seems to be backing up one of the folders daily, as part of the incremental changes (although none of these files have been accessed\changed). Any idea why one directory would be constantly backing up.. also would you know how to reset the archive bit on that folder?

[quebek]

btw is your backup based on archive bit or timestamp? take a look

I do advice to user timestamp along with change journal - change journal has to be created on the given volume you do backup - newer releases of NBU are creating this automatically.

Once changed to change journal - you should restart client services and perfrom full backup.

[Genericus]

Please run a FULL backup of that policy. That should reset everything.