How to get the smallest backup files?

[mensa]

Hello,

backing up VMs and Windows PC (via Agent) to a CIFS Share on a NAS.
The Compression setting "Extreme" brings the smallest files size, if I understood correctly.
But does the "Storage optimization" setting also affect the file size? Or is this only relevant for how much data has to be transferred?

Other question: Higher compression settings need more CPU usage. But of Which host?
In case of backing up VMs the CPU of the Veeam B&R server and in case of backing up Windows PC via Windos Agent the CPU of the Windows PC? Is my understanding correct?

Any other settings which affect file size at final target store?

[Shestakov]

Hello,

The Compression setting "Extreme" brings the smallest files size, if I understood correctly.
But does the "Storage optimization" setting also affect the file size? Or is this only relevant for how much data has to be transferred?

The smallest size will be achieved with Storage optimization = WAN target, Compression level = Extreme. First chooses the size of data block, second how hard backup proxy will try to find similar data blocks.

Other question: Higher compression settings need more CPU usage. But of Which host?
In case of backing up VMs the CPU of the Veeam B&R server and in case of backing up Windows PC via Windos Agent the CPU of the Windows PC? Is my understanding correct?

On backup proxy for VM backups, on PC running Veeam agent for physical servers.

Any other settings which affect file size at final target store?

Hardware deduplication on target repository.
I would suggest to read this and this topics.
Thanks!

[mcz]

Hi everybody!

Last week I did a full backup of a VM which had a size of 550 GB. I used the default settings for compression and of course I used the encryption. The VM itself contains three huge .DBS-files (gupta database) with a sum somewhere around 500 GB. To my big surprise the backup file had a size of 500 GB, which means a reduction of only 9,1 % which is definitely not the best compression ratio.

So I asked myself: How does compression work? e.g. if you would ZIP a big file (e.g. veeam's configuration db), you could see compression ratios of 80 % and more (because there are a lot of repetitions within the file) but how would veeam do this? AFAIK, veeam uses a block size of 512 kB and if it compresses each block for its own, then I could imagine why it didn't help a lot in my case.

Thanks very much for the clarification!

[PTide]

Hi,

ZIP works with files therefore it is able to find those repetitions. In case of VM backup compression algorithms work with data blocks that are read from snapshot (blocksize depends on storage optimization settings).

Thanks