Windows Physical Backup includes Swap File and Too much Data

[John Mallard]

I am using Nakivo to backup Windows 10 Workstations (laptops) and for 2 backups each day of 1 laptop, each backup is 16 - 17 GB !!!

Now I remember seeing in Release Notes that a physical Windows backup is still not ignoring the Windows swap file, which certainly does not help and should be rectified in an upcoming software release. However, in my case, that is still only 1.3 GB, so i am sure that there is still an awful lot more being backup up that is not necessary. Could this be addressed son, please ?

[Official Moderator]

Hi @John Mallard,

The backup is still not ignoring the Windows swap file because the swap file serves as additional memory available to hold different data from physical memory, which is more complicated than the VM system. Some modern Windows apps won't run — others might run for a while before crashing. The Windows OS shouldn't disable the swap file unless you're an experienced Windows user, and the backup shouldn't skip them.

We can't affect the tracking algorithm, but this is what you can do in your situation:
- Reduce the data volume.
- Reduce System temporary files/updated files, adjust the pagefile/fiberfill if you have enough hardware resource.
- Disable automated disk defragmentation maintenance.

Please feel free to contact us if you need any further information.

[DaniAvni]

@Official Moderator, I think that while it is true that the page file serves as additional memory, it should not be backed up on a physical machine. the file contents are reset when the physical machine turns on and the file itself is not usable as you cannot restore it or use it's contents in any way (windows prevents access to pagefile.sys so you cannot restore it anyway).

I would say the same applies for hiberfil.sys which is the hibernation file for windows

 

On my machine (32GB ram) the page file is about 13GB which is a waste of backup space and time as it is not usable and even if you skip the file during restore then windows will create it again

[Mario]

Just an Idea for an ugly hack:

Put pagefile on a diffrent partition and exclude this from backup?

Not nice, but maybe effective

I think the change tracking of nakivo is not very effective, I've seen mucht better solutions by other vendors.

(but I still like Nakivo)

[DaniAvni]

@Mario, I thought about it (thanks!) but the problem is that once I start excluding drives I need to do it over all physical machines and if I add a drive it will not be included automatically. A much better solution is for the backup software to exclude these files.

[DaniAvni]

I took a look on what Arcserve is doing and they are not backup up the pagefile and hiberfil contents but they do backup their location on the MFT so that the file returns to the same place (only content is replaced). Any solution is valid as long as the contents of the file do not waste space in my backups. My machine with almost nothing changes on disk can still produce a 100GB backup just because page file moves, antivirus changes and the fact that my SSD is fragmented (making a lot of disk blocks appear as changed)

[Official Moderator]

@DaniAvni and @Mario, hi!

Currently, the exclusion of specific files from the backup of the Physical machines is not yet supported. However, this feature is in our backlog. We do not have ETA.

We would recommend sending us a support bundle with a detailed description of the feature request so we can pass it to our Product Team.

Thank you a lot for your suggestions! They let us improve. 

[anm06965]

Any progress here with excluding specific files from physical backups? It's still seems to be a case - pagefile changes all the time so all my daily backups contain these changes for no obvious reason -> waste of backup space...

[Leslie Lei]

I think Nakivo should change the way the software gathers the data from the server.  If the software allows you to exclude the files from the file system, that means the software is looking at the data changes in each file.  This is the slowest method when doing data backup because the backup process needs to walk through every file in the file system to determine the changes.  I think that is the main reason the physical server backup process takes a very long time even if only a few files get changed.  It spends the most of the time walking the file with no changes.  

The best way to backup the data on the disk is the block-level backup.  The software should just take a snapshot of the data blocks on the disk level instead of the individual file level.  Slide the disk with multiple (million or billion blocks).  Look for the change in each block and only backup the changed blocks.  This method will not care what files has changed.  Every time a file gets changed, it will change data in the blocks in which the file resides.  The backup software just backs up the changed block.  This should be the best way for Nakivo to speed up the backup process.

I normally put the page file and the printer spooler folder or any file that I don't want to backup in another volume or disk in the Windows system.  Skip that disk/volume in the backup job to reduce the change that needs to be backed up.

 

[The Official Moderator]

17 hours ago, Leslie Lei said:

I think Nakivo should change the way the software gathers the data from the server.  If the software allows you to exclude the files from the file system, that means the software is looking at the data changes in each file.  This is the slowest method when doing data backup because the backup process needs to walk through every file in the file system to determine the changes.  I think that is the main reason the physical server backup process takes a very long time even if only a few files get changed.  It spends the most of the time walking the file with no changes.  

The best way to backup the data on the disk is the block-level backup.  The software should just take a snapshot of the data blocks on the disk level instead of the individual file level.  Slide the disk with multiple (million or billion blocks).  Look for the change in each block and only backup the changed blocks.  This method will not care what files has changed.  Every time a file gets changed, it will change data in the blocks in which the file resides.  The backup software just backs up the changed block.  This should be the best way for Nakivo to speed up the backup process.

I normally put the page file and the printer spooler folder or any file that I don't want to backup in another volume or disk in the Windows system.  Skip that disk/volume in the backup job to reduce the change that needs to be backed up.

 

Hello, @Leslie Lei,

Thank you for your post. We acknowledge the importance of optimizing the backup process for efficiency. Currently, we leverage VSS/LVM-based snapshots to read data from the source physical host whenever possible. Our approach involves comparing data with existing backups using hash sums to determine changes at the block level. We do not use file-level comparisons. 

In instances where snapshotting is not possible, we resort to low-level reading by sectors on the source disk.

The existing change tracking mechanism is not ideal, and we are working on improvements to speed incremental backups. This improvement has been added to our development roadmap, although a specific ETA is not available at this stage.

We have also forwarded your request to our development team. To help our team understand your environment and solution plan, consider sending us a support bundle. For information on how to create and send a support bundle, please refer to: https://helpcenter.nakivo.com/User-Guide/Content/Settings/Support-Bundles.htm

We value your input and look forward to any further insights you may have.