Recent searches
Search options
Interested to know how many of you guys use object store to house your files/movies/pictures and shit like that in the #homelab or we just file systems all of the time?
@train I just do zfs pools in Proxmox.
@train I keep thinking about learning object storage in a homelab context but it's just too much effort ATM and always stuck in low priority. Btrfs storage pool serves me well and I can mount it anywhere and use traditional tools since it's just a filesystem. Object storage I feel would wind up limited to how you can interact with it.
@train@hachyderm.io bulk storage like that is on zfs, and pods access it via NFS. None of the media apps that I use know how to use object storage so there just isn't any point. Also the object storage that I have is a part of the ceph cluster which is SSD-backed with a much smaller capacity
@train for media, bulk storage that doesn't need to be fast and is easy to recover if too many drives die, I use snapraid and mergerfs. For the VMs and their data, zfs via truenas/iscai
@train Wish I’d had a need, but not so far. NFS/SMB for most large mounts, and Longhorn for fast/critical storage. Unraid as the backing for the network shares.
@train I have a minio vm (with zfs on the backend) that I use as backup target for my desktop computer and to manually backup photos/videos from my phone. I also use this vm to store travel photos taken with a camera while I’m still traveling before I can copy them to my computers to edit it.
@diegolakatos can I ask why not just use zfs? Why did you choose to use object?
@train some of the use cases that I mentioned demands access using a vpn or WiFi and I’m not very fancy of using nfs/smb in these scenarios, I think using an API to send files in this scenario is better. And for the desktop backups I don’t want to have to mount a filesystem just to perform a backup or have access to the backup files unless I explicitly want it.
@train Yupp, whenever an app can use S3, I use it. I love the fact that it's just this gigantic lake of data. I don't have to worry about filesystems, partitions or anything else like that. Plus, being able to access the data via HTTP is another bonus.
@mmeier I’m starting to lean that way!