Replies: 3 comments 2 replies
-
It makes no sense to tune the borg chunker which uses a hash to determine individual chunk sizes that way Unless there was a custom chunker which is aware of the exact format boundaries its likely not to be a good match |
Beta Was this translation helpful? Give feedback.
-
borg 1.2 has a
So guess this is too simple for VMA and it would require a very special custom chunker, as @RonnyPfannschmidt said. |
Beta Was this translation helpful? Give feedback.
-
Another problem with the VMA backup is, that all data blocks can be stored out-of-order, varying between backup runs (so in a first backup blocks could be stored in the order 4,1,3,2 and in a second run they could be stored in the order 3,1,2,4). Would it even be possible to write a custom chunker to reorder the contained data blocks? |
Beta Was this translation helpful? Give feedback.
-
Hello,
I'm asking for help figuring out what would be the optimal chunker parameters for dedup'ing asynchronously written Proxmox VMA images.
The VMA format is described here: https://git.proxmox.com/?p=pve-qemu.git;a=blob;f=vma_spec.txt
According to them, they use 65k cluster size with a maximum of 4M extent size, also adding a 512b header at the beginning of each extent.
Does this mean I'd need these for successful deduplication?
Beta Was this translation helpful? Give feedback.
All reactions