Hi everyone,
in my Proxmox I've created a storage dir to store all kinds of data and made that one a ZFS dataset on its own. I want to be able to configure individual auto-snapshot retention policies this way. Everything works fine, I'm able to create VMs etc. in that dataset and the web-UI shows the usage of the pool and the amount of data within that dataset.
Though, I've decided to put one VM into an additional child dataset to apply individual retention policy on that as well. That additional dataset is
Is that behaviour designed that way and as expected?
I expected that Proxmox looks at my datasets just as directories and takes the child directories and files into account when calculating the size. But it seems to be dataset-aware in this case, even though I just created "dir".
in my Proxmox I've created a storage dir to store all kinds of data and made that one a ZFS dataset on its own. I want to be able to configure individual auto-snapshot retention policies this way. Everything works fine, I'm able to create VMs etc. in that dataset and the web-UI shows the usage of the pool and the amount of data within that dataset.
Bash:
pvesm add dir 'zpool-hdd' --content 'backup,images,iso,rootdir,snippets,vztmpl' --is_mountpoint 'yes' --prune-backups 'keep-all=1' --path '/zpool-hdd/pve'
Though, I've decided to put one VM into an additional child dataset to apply individual retention policy on that as well. That additional dataset is
zpool-hdd/pve/images/100
and really not otherwise known to Proxmox, but just a plain ZFS dataset. While that still works as expected, I just recognized that the size calculation in the web-UI doesn't work properly anymore. Before my change it showed the size of all the files in all the directories within zpool-hdd/pve
, nowe it only show the size of the files directly in the dataset zpool-hdd/pve
without taking child datasets into account. So the size calculation is bound to the boundaries of datasets.Is that behaviour designed that way and as expected?
I expected that Proxmox looks at my datasets just as directories and takes the child directories and files into account when calculating the size. But it seems to be dataset-aware in this case, even though I just created "dir".