Requesting a list of all backups, which files are used to create the list?

[Jeeves_]

PBS stores the backups using a number of files. I see:
* Chunks
* img.fidx
* json.blob
* log.blob
* conf.blob

If you have a lot of backups, the amount of time it takes to generate a list of available backups sometimes takes too long for PVE.

I want to use the ZFS special device to speed things up, but I need to be careful about the https://openzfs.github.io/openzfs-docs/man/master/7/zfsprops.7.html#special_small_blocks setting, to prevent the amount of datausage on the special devices. So I don't want to set the value too high.

So which files are used to generate the list of available backups?

 

[Chris]

Hi,
for the listing mainly the directories and their file contents are scanned by opening the directory and reading the file entries. Apart from these, the owner file for each backup group is accessed and the content read for the owner check and the manifest file (index.json.blob) for each backup snapshot to get the verify and encryption state.

Chunk files as well as index files (being .fidx and .didx, for fixed index and dynamic index files) are not involved in the listing directly, they are only accessed when fetching the snapshots contents, e.g. during sync or restore.