Something about the basics of data transfers

I am doing a full intensive test of the software to evaluate the possibility of using it in production environment in some of my customers.

One of my big doubts is how really works data transfers between client and server.

I have readed all manual and documentation, but still don´t understand how works the data transfer.

I am testing the software only for file backup (not image backup) at the moment.

In a normal way, I understand that the first copy is a complete copy, this means that all the files and folders are transferred from client to server.
After that, the incremental file backups, how exactly works?
I understand that client composes a file list database that send to server and it compares with the server side database to know what files has been changed from last backup, then start data transfer, but…

  • It transfer complete changed files from client to server?

  • It transfer only parts of changed files (at block level) from client to server? (Something like rsync does)

  • It transfer complete changed files, but later on the server it only stores the partial modified data of every file doing some block level hashing?

  • How affect the server setting “Local incremental file backup transfer mode” to the behavior of the transfer? or the behavior of the server post-process?

  • What exactly happens if I enable Internet Mode on client? The client checks at block level files that has been changed and thansfer only those blocks? If this is true, why not is best working always at “Internet mode” to prevent big data transfers?

I am sorry for making a lot of questions, but, after 2 weeks of testing, (with very high satisfaction in general) I don´t understand the underlying work of the backup process.

Thank you so much.


That should answer your questions:

Thanks @orogor,
I readed all documentation before post. But, this does not resolve my doubts. I still not know what is transferred from the client to the server, the complete files, or only part of them.
One of my big challenges is the backup of mail storage on clients (Outlook PST).
Those can be large files, 50-60 GB in one only file for example. This files are constatly modified while remains opened by Outlook.
In every incremental backup, the software will copy entire file or only some parts of the file?
If the client is running in “Internet mode”, is the same behavior?.
If I change the server settings to “Block difference - Hashed”, is the same behavior?

Other question, can I see the real amount of data transfered from client to server in a backup session somewhere?

Thanks a lot

internet mode use block difference hashed by default
But you could set that for local as well.
it will work fine for the purpose of large files transferts in général.

Because of the database like behavior of pst file , i am unsure that you can backup them properly using urbackup. Outlook would need to be closed and disk io flushed and whatnot.
So please read the doc and forums as i remember there’s something about it.

I don’t use windows too much, but something i d try
Backup the mails from the exchange server it should have special tools to do backup. The client whill just sync from that.
Backup the whole windows drive as file and/or image once per week or so.
Backup c:\users as file using a virtual subclient, that you can do daily or more, just ignore outlook pst.

Eventually run in debug the first few incrementals, so that you see which file are always transferred , like cookies and whatnot and build your ignore list from that.

If you usually setup computers with a d: drive for photos or code or whatnot , setup a different subclient for that.
Subclient backup at different frequencies, can be cleaned, and can fail independently from each others, also if you have large volumes, they can somehow complete faster, as each sublient won’t backup the whole drive.

we backup several outlook clients with hashed block difference and all is working perfect.


Thanks all for your answers.

Currently I am done some outlook backups to test the software and it is working.
I have changed the server setting to “Block Difference - hashed” and it is working but I still don´t know what is transfered from client to server. Anybody can answer this?

In every backup, complete outlook file is transfered? or only some parts of the file?

Where can I see how much data has been transfer really?, not the size of the full or incremental backup, the transfered data only.

Thanks a lot