Change NTFS permission = Massive replication times

[JohnGG]

Gday folks, i have Veeam 9.5 and just changed NTFS permissions on my archive folder in Windows server to allow 2 people to write to it.
The size of the folder is +- 800 GIG and now my replication has gone from 30 gig a night to 400 + on that one drive.

Is this normal because it class's the change of NTFS permissions as a changed block?
Its been going for 50+ hours now and i have had to delete smaller VM replicated servers on my disaster site to to compensate for the huge amount of data!

[JohnGG]

This thread here veeam-backup-replication-f2/windows-sha ... 12874.html is whats happening to me but im replication my VM's to office site disaster.
I changed these permissions on a physical server located 400 kilometers away, then DFRS replicated to my virtual over night and im getting these massive change block changes.

I have stopped DFRS replication for the time being while these block changes are replicated to my office site

[soncscy]

Hey John,

It's absolutely normal.

Think of it this way:

Each file or group of files resides in a block on your hypervisor. (VMware in this case)

If even a single bit of the block is change, the entire block is considered "new". This is just how blocks work.

Since Veeam does its backups from the Hypervisor POV, this means each change to the blocks is new data and needs to be processed.

As you done, you basically changed potentially hundreds of thousands of blocks. The actual data change may be tiny in the __GuestOS__, but the Hypervisor has no knowledge of this -- all VMware sees is that a data block is different than before.

Without extensive magic from VMware, there's no way to avoid this as I get it. Maybe Agent based backups will help, but then you have to deal with Agent based backups, and "yuck". Once you hit a certain threshold, it's just impossible to manage without a competent backup software.

For my purposes, we just eat the change cost. It's not a "huge" deal usually for our clients. You have to understand such changes in the guest aren't free, and this is part of then cost of agentless.

[JohnGG]

No worries mate and thanks for the reply.
I would have thought that just the $MFT would have changed because it was just NTFS permissions but i had a sneaking suspicion thats why my replication size blew up lol.
The one replication is still going thanks to the slow NBN we have in australia!

[DaveWatkins]

To get around this specific issue (and sometimes it's not possible) use groups to give permissions. Adding a user to a group that has pemrission means no ACL changes to the files themselves

[JohnGG]

YeP I added the users to the group who had write permission but still here I am lol