Increment_flatten can be executed using qemu-img commit instead of qemu-img convert

[nachowork90]

Description
The goal is shorten the time and I/O required to merge the full and the next incremental on the chain when the keep_last value reached.

Use case
When we execute the qemu-img commit on the top file that point to the base image (backing_file) or full image in this case. Only the changes of the incremental goes to override the clusters in the base image, even if there are more than 1 elements in the chain.

Interface Changes
remotes/datastore/rsync/increment_flatten to modify the function merge_chain call
remotes/tm/lib/backup.rb to add a new function the handle this scenario.

Additional Context
Operations performed in the rsync host:

qemu-img commit -f qcow2 '/var/lib/one/datastores/100/7/768671/disk.0.3'                                                                                                                │·
mv '/var/lib/one/datastores/100/7/2757f7/disk.0.0' '/var/lib/one/datastores/100/7/768671/disk.0.3'

Progress Status

• Code committed
• Testing - QA
• Documentation (Release notes - resolved issues, compatibility, known issues)

[Franco-Sparrow]

This has a storage and backup time saving approach. Awsome nacho :)