Space usage report

[georgy.goshin]

Hello!

How can I get backup datastore space usage by VM or by namspace?

Thanks

 

[Hannes Laimer]

Hey,

you mean the summed size of all used chunks? If that's the case, currently not, at least not easily.

 

[georgy.goshin]

Hey,

you mean the summed size of all used chunks? If that's the case, currently not, at least not easily.

Yes, the total weight of each VM of each namespace. Maybe some ZFS commands I'm not familiar with? Usual du does not shows any real info

 

[Hannes Laimer]

There are index files(.fixd and .didx) that reference the chunks used for a backup, those are basically lists of chunks, you can get a list of the referenced chunks with

proxmox-backup-debug inspect file --output-format json drive-<...>.img.fidx

with this list of digests you can find the referenced chunk files in the .chunks/ directory and sum up their sizes.

Note: different VMs/snapshots do reference the same chunks, so keep that in mind when summing up the sizes.

 

[georgy.goshin]

There are index files(.fixd and .didx) that reference the chunks used for a backup, those are basically lists of chunks, you can get a list of the referenced chunks with

proxmox-backup-debug inspect file --output-format json drive-<...>.img.fidx

with this list of digests you can find the referenced chunk files in the .chunks/ directory and sum up their sizes.

Note: different VMs/snapshots do reference the same chunks, so keep that in mind when summing up the sizes.

This will give me total image size but noting about how many disk space I'm spending for storing each VM

 

[Hannes Laimer]

No it won't. If lets a index file references 6 chunks, then the size of these 6 chunks is how much the backup of the disk this index file represents takes up in space. Ofc, this is always a little weird since the same chunks may be referenced by different index files, so summing up will yield a higher amount than what is actually used because chunks are counted as often as they are referenced.