From the docs it says to take backup of your /repo/blob/* folder and also to a snapshot of your index.
What is in the index that you cannot recreate from the blobs? Couldn’t you just restore your blob-folders and reindex the content? or am I missing something?
I know recreating alot of content would take quite a while. Just want to know how it all works 
The index contains all meta-data about the nodes; e.g path, the blobs are just the content of the node. So a blob has no notion of the tree-structure, which version is used in which branch etc.
Its much slower to write updates to the blob than updating a meta-data field in the index, since the data must be serialized, written to disk etc, so we try to avoid that as much as we can. There is also work to be done here to speed up bulk updates and creates further,
That said, it is possible to write data to the blobstore to be able to recreate the meta-data, but for the moment this is not done, since the snapshot-functions is a safe way to backup the data.
Ok, that sounds logical