Hello. After lurking for some time, this is my first post.
Background:
Inspired by the TinyMiniMicro Project and apalard's youtube videos, I have, over the past several weeks created a 3-node Hyper-converged, ceph-enabled, HA Cluster.
I plan to add a true, dedicated NAS with appropriate disk redundancy at some point to the homelab. Until then, I have a 4TB USB External HDD looking for a job and an interesting quandry to think about and solve.
For this discussion we can assume (1) achieving maximum I/O performance is not important to me (2) I am willing to fully accept the risk of data loss due to disk failure for anything I put on this device.
The objectives and outline of a plan (open to critique)
I think I'd like to use this HDD for back-ups and general "network sharing". I would like the set-up to accommodate large local file transfers, storing back-ups of all kinds including windows, and proxmox back-ups from all nodes if possible, but definitely from pve1.
I'm leaning towards setting up an HA LXC that will act as as a networked fileserver, call it "Fileserver LXC" (inspired by this video that leverages a Debian container + Cockpit to manage SMB) on my primary node, pve 1, and then plugging in the USB HDD device into pve1.
In the event of PVE1 failure, I will setup and expect HA rules dictate that the FileServer LXC be brought up on node 2 or 3.
Then, the admin user experience I desire for myself would be to unplug that USB device from pve1, plug it into the node that has the FileServer LXC up and running and in as few, uncomplicated, steps as possible (e.g. LXC reboot) be back to normal operations.
My main question is: How would you set-up this device+container+cluster combo given the information above? Assume The ocntainer is ready and waiting and the USB disk is wiped and initialized with GPT... also waiting.
Some things I've been asking myself:
Thank you in advance! I will be responsive though I may not be able to reply until after work hours East Coast time.
Background:
Inspired by the TinyMiniMicro Project and apalard's youtube videos, I have, over the past several weeks created a 3-node Hyper-converged, ceph-enabled, HA Cluster.
I plan to add a true, dedicated NAS with appropriate disk redundancy at some point to the homelab. Until then, I have a 4TB USB External HDD looking for a job and an interesting quandry to think about and solve.
For this discussion we can assume (1) achieving maximum I/O performance is not important to me (2) I am willing to fully accept the risk of data loss due to disk failure for anything I put on this device.
The objectives and outline of a plan (open to critique)
I think I'd like to use this HDD for back-ups and general "network sharing". I would like the set-up to accommodate large local file transfers, storing back-ups of all kinds including windows, and proxmox back-ups from all nodes if possible, but definitely from pve1.
I'm leaning towards setting up an HA LXC that will act as as a networked fileserver, call it "Fileserver LXC" (inspired by this video that leverages a Debian container + Cockpit to manage SMB) on my primary node, pve 1, and then plugging in the USB HDD device into pve1.
In the event of PVE1 failure, I will setup and expect HA rules dictate that the FileServer LXC be brought up on node 2 or 3.
Then, the admin user experience I desire for myself would be to unplug that USB device from pve1, plug it into the node that has the FileServer LXC up and running and in as few, uncomplicated, steps as possible (e.g. LXC reboot) be back to normal operations.
My main question is: How would you set-up this device+container+cluster combo given the information above? Assume The ocntainer is ready and waiting and the USB disk is wiped and initialized with GPT... also waiting.
Some things I've been asking myself:
- Should I just use USB pass through and rely completely on the container to interact with the device? (I think this is the right answer for this interim state)
- Should I instead set-up the device as "Storage" (if so what file system?) on pve1, then plug it into pve2, do the same, and then on pve3 and do the same to "stage" the drive?
- Is there some obvious, simple answer I'm overlooking like "just set it up as XYZ on PVE1, configure this or that, and boom..."
Thank you in advance! I will be responsive though I may not be able to reply until after work hours East Coast time.