©XSIBackup-Free: Free Backup Software for ©VMWare ©ESXi

Forum ©XSIBackup: ©VMWare ©ESXi Backup Software

You are not logged in.

#1 2021-09-30 09:36:05

Registered: 2021-02-07
Posts: 25

some CBT backup queries

Hi, Thanks for any assistance.

I have a replication running daily for a windows server and wish to implement an offsite backup.
Server size static at 1.2TB
Diff size if 10-40GB+ daily.
Probably averaging around high twenties.
While 50GB remote transfer is manageable overnight, I would prefer to optimise it for offsite backups.

1)The server is a windows server doing AD etc plus file sharing.
I would be surprised if the data changed for than 1GB a day.

I assume the rest of the Diff is created from things like:

Volume shadow copies.

I am correct in guessing the above and is there a document on how to minimise these Diff sizes (without taking a noticeable performance hit)?

2)To avoid a 1.2TB initial transfer can you seed a CBT backup?
By this I mean run an initial CBT backup via SSH to a USB drive on a LAN connected Syncology NAS (1Gb bandwidth). Then transfer the USB drive to the offsite location, copy the backup files into a directory on the remote NAS box. Then reconfigure the XSIBackup-DC job to use the IP and directory on the remote site. Will DC recognise the transferred backup and just transfer a diff?

3) I assume IP transfer (ssh) is the best configuration for a remote site than, say, NFS over vpn?

Thanks again


#2 2021-09-30 16:43:01

Registered: 2017-04-21
Posts: 2,055

Re: some CBT backup queries

The plain answer is: yes. Just as long as you don't modify the contents of a first transfer you can move it around as you whish.

Just one thing: please do not use any USB device to store your data especially when backing up 1.2 TB of it. You can use eSATA, Thunderbolt, NAS, etc...

(c)XSIBackup uses a 1MB block size. That means that even if just a single bit of information changes at a given block in your Windows server, the whole block will be considered to have changed. Depending on how spread that data is in the disk you may produce a different amount of differential data.

NTFS is probably not among the best FSs at the time to keep contiguous data. It keeps many copies of the superblock, it stores it at random offsets in the disk and it tends to fragment data to then reassemble it, not good in terms of data density.


Board footer