Guys,
sorry for being boneheaded, but even after a lot of reading the docs and postings here, I still fail to fully understand the "best" (security, stability, performance - in that sequence) solution to what I would like to do:
I have some TB in personal files on a USB-HDD (LUKS with btrfs-fs) and would like to make that available to all clients (Linux PC and Android smartphones) on my home network. There is nothing else on this HDD, so it could be fully dedicated to whatever I set up for this purpose.
Normally, I guess, the answer would be something like OMV-VM and SMB.
Im my case, though, I would like to simulate a setup comparable to a cloud-storagebox, because that is where my files might at a later stage find themselves.
So the current local setup should serve as testing ground for what I intend for later. For that reason I decided to use Rclone to access the files, because that allows chunking and encrypting on the client side, with only chunked and encrypted files on the HDD. (Which means I could even forgo the LUKS part above and do plain btrfs on the HDD.)
Multi-user could also become an issue later, which I imagine could be done by setting up users (and ACL if necessary) in the guest and connecting with Rclone as the relevant user.
Question 1:
What kind of guest should I use?
- Debian LXC with Rclone? (Most likely not going to use Rclone Mount, so FUSE on unprivileged LXC should not become an issue.)
- OMV-VM with Rclone in addition to everything else it offers?
Question 2:
(somewhat depending on the answer above)
How to make the files from the HDD available in the guest?
I have read https://pve.proxmox.com/pve-docs/pve-admin-guide.html#pct_settings, but fail to grasp which of the three versions of handling mount points my use case would most benefit from.
(In case the answer to Q1 is unprivileged LXC, the warning "With unprivileged containers you might run into permission problems caused by the user mapping and cannot use ACLs." might rule out Bind Mount Points or require user mapping, if I understand this correctly?)
My impression is, that the standard would be option 1, the Storage Backed Mount Points, and passing through the mounted filesystem.
Are there any significant disadvantages to doing it that way, if it is indeed "the Proxmox standard"? (Again: Security, stability, performance)
Thanks a lot for your help.
sorry for being boneheaded, but even after a lot of reading the docs and postings here, I still fail to fully understand the "best" (security, stability, performance - in that sequence) solution to what I would like to do:
I have some TB in personal files on a USB-HDD (LUKS with btrfs-fs) and would like to make that available to all clients (Linux PC and Android smartphones) on my home network. There is nothing else on this HDD, so it could be fully dedicated to whatever I set up for this purpose.
Normally, I guess, the answer would be something like OMV-VM and SMB.
Im my case, though, I would like to simulate a setup comparable to a cloud-storagebox, because that is where my files might at a later stage find themselves.
So the current local setup should serve as testing ground for what I intend for later. For that reason I decided to use Rclone to access the files, because that allows chunking and encrypting on the client side, with only chunked and encrypted files on the HDD. (Which means I could even forgo the LUKS part above and do plain btrfs on the HDD.)
Multi-user could also become an issue later, which I imagine could be done by setting up users (and ACL if necessary) in the guest and connecting with Rclone as the relevant user.
Question 1:
What kind of guest should I use?
- Debian LXC with Rclone? (Most likely not going to use Rclone Mount, so FUSE on unprivileged LXC should not become an issue.)
- OMV-VM with Rclone in addition to everything else it offers?
Question 2:
(somewhat depending on the answer above)
How to make the files from the HDD available in the guest?
I have read https://pve.proxmox.com/pve-docs/pve-admin-guide.html#pct_settings, but fail to grasp which of the three versions of handling mount points my use case would most benefit from.
(In case the answer to Q1 is unprivileged LXC, the warning "With unprivileged containers you might run into permission problems caused by the user mapping and cannot use ACLs." might rule out Bind Mount Points or require user mapping, if I understand this correctly?)
My impression is, that the standard would be option 1, the Storage Backed Mount Points, and passing through the mounted filesystem.
Are there any significant disadvantages to doing it that way, if it is indeed "the Proxmox standard"? (Again: Security, stability, performance)
Thanks a lot for your help.