pve-zsync vs storage replication: differences in the internals

[Twissell]

Hello Proxmox Community!

I'm digging into Proxmox internals (literally Perl scripts) a bit, but I can't

find answer for my question: what's is the differences in internal implementation
between pve-zsync and full storage replication task.
As far as I can see both mentioned methods use ZFS snapshots under the hood.

So, one question is in subject and the second is where I mistaken ontopic.

 

[wolfgang]

Hi,

pve-zsync is written to operate on Linux ZFS hosts and has no dependency on Proxmox VE.

 

[Twissell]

Hi,

pve-zsync is written to operate on Linux ZFS hosts and has no dependency on Proxmox VE.

Got it!
But, how does full ZFS pool replication has been implemented in Proxmox VE ?

 

[wolfgang]

But, how does full ZFS pool replication has been implemented in Proxmox VE ?

What do you mean? pvesr only replicate single VM's. Also, pve-zsync replicates only single datasets.

 

[Twissell]

@wolfgang Thanks I see where was my misunderstanding of the process.
Is there only one script pvesr that replicate VM from one node to another?
As for pve-zsync I've already seen how it works.

 

[wolfgang]

Is there only one script pvesr that replicate VM from one node to another?

No, it completely integrated into the Proxmox VE framework.

 

[Twissell]

No, it completely integrated into the Proxmox VE framework.

Ok, I've another question in that case: is implementation itself binary file or script written in Perl/Python etc ?
Or it's couple of files which are integrated part of Proxmox VE?

 

[wolfgang]

We use Perl in Proxmox VE management framework, so it is a script and not a binary file.