So my understanding is that in order to be able to do Hardlinks the folders need to be inside a dataset. But they cant be other datasets as I did. So I fucked up when setting it up.
Is there any easy way to fix this because I already have all my apps setup and working with the *arr suite and Jellyfin running. Can I just make some folders under /data move all the files (is this a good idea or bad because of intense I/O) and afterwards delete the media and torrents datasets? And just point the apps to the new directories?
I am having great difficulties understanding the details of this issue in order to give detailed advice.
Any chance of some screen shots of the datasets screen and details of what apps are using what datasets and of what links have attempted to be created?
And output of ls -l for some of the created hard-links (or whatever they are - my guess is soft-links) so we can see what was actually created?
P.S. As an explanation of why hard links don’t work, unlike most file systems which have a single mount point for the entire file system, ZFS mounts each dataset individually and so from a Linux perspective each dataset looks like a separate file system - and so hard links are not possible across datasets.
And I would like it to be possible that files inside media and torrents to be Hard linked. And in order to do that my question was if just creating simple directories under /data Dataset move all the contents of /media dataset and /torrent dataset to those new directories is a good idea? Or is there a better approach?
I doubt that putting everything as subdirectories of the root dataset for the pool will work because you cannot have an SMB share on either a root dataset or a subdirectory, only on sub-datasets.
I still have no idea what you are trying to achieve and why, and so no idea whether hard links are the correct solution (which is doubtful) or whether a single dataset and subdirectories is really appropriate or not and I cannot help until I understand these.
To restate (yet again):
to help you remove the links you have already created we need still details of these links.
to help you set up your datasets and apps in the best way we need details of datasets (which you have given) and details of app configurations (not yet given).