Immutable Nixos. My entire server deployment from partitioning to config is stored in git on all my machines.

Every time I boot all runtime changes are “wiped”, which is really just BTRFS subvolume swapping.

Persistence is possible, but I’m forced to deal with it otherwise it will get wiped on boot.

I use LVM for mirrored volumes for local redundancy.

My persisted volumes are backed up automatically to B2 Backblaze using rclone. I don’t backup everything. Stuff I can download again are skipped for example. I don’t have anything currently that requires putting a process in “maint mode” like a database getting corrupt if I backup while its being written to. When I did, I’d either script gracefully shutting down the process or use any export functionality if the process supported it.

I haven’t tested in Windows, but this is my setup Linux to Linux using rclone which the docs say works with Windows.

Server

• LUKS
• LVM
• Volgroup with a mishmash of drives in a mirror configuration
• Cache volume with SSD
• BTRFS /w Snapshots (or ZFS or any other snapshotting FS)
• (optional) Rclone local “remote” with Crypt if you want runtime encryption at rest and the ability to decrypt files on the server. You can skip this and do client side only if you don’t want the decryption key on the server.
• SFTP (or any other self-hosted protocol from https://rclone.org/docs/)

Client

• Rclone Config /w SFTP (or chosen protocol)
• (optional) Rclone Config /w Crypt
• Rclone mount with VFS.

I use this setup for my local files and a similar setup to my Backblaze B2 off site backups.

The VFS implementation has been pretty good. You can also manually sync. Their bisync I don’t fully trust though.

I can access everything through android using https://github.com/newhinton/Round-Sync. Not great for photos though as thumbnails weren’t loading without pulling the whole file last I tested a year ago.

One method depends on your storage provider. Rsync may have incremental snapshots, but I haven’t looked because my storage provider has it.

Sometimes a separate tool like rsnapshot (but probably not rsnapshot itself as I dont think its hard links interact well with rsync) might be used to manage snapshots locally that are then rsynced.

On to storage providers or back ends. I use B2 Backblaze configured to never delete. When a file changes it uploads the new version and renames the old version with a timestamp and hides it. Rsync has tools to recover the old file versions or delete any history. Again, it only uploads the changed files so its not full snapshots.

Yes. You compose a crypted vault over your storage vault. I pay about $1/mo for B2 Backblaze. Around 150G last I checked.

Important stuff (about 150G) is synced to all my machines and a b2 Backblaze bucket.

I have a rented seed box for those low seeder torrents.

The stuff I can download again is only on a mirrored lvm pool with an lvmcache. I don’t have any redundancy for my monerod data which is on an nvme.

I’m moving towards an immutable OS with 30 days of snapshots. While not the main reason, it does push one to practicing better sync habits.

I believe you will see OOM errors in journalctl. Also considered a systemd service or something to restart the process?

Glad you figured it out though!

I use EteSync to sync my contacts. It can do calendars also and has a self-hosted option. Personally I just use Proton for my calendar at the moment.

I haven’t tested any desktop syncing with EteSync, but maybe it will work for you.

I use DavMail to proxy an Exchange account so I don’t have to install Outlook on my phone.

Never used ic before, but there is an example in the README on GitHub.

>>> import logging
>>> from icecream import ic
>>>
>>> def warn(s):
>>>     logging.warning(s)
>>>
>>> ic.configureOutput(outputFunction=warn)
>>> ic('eep')
WARNING:root:ic| 'eep': 'eep'

5x work shirts, 5x work pants, bagged/black t-shirts and fisherman pants for home, all black socks. L1 cache is the drier. L2 is the shelf next to the drier. There is nothing beyond L2.

Blows my mind, lol. Usually means no redundancy that allows one set to be done while the other set handles the traffic.

Can try installing Avahi on the RPi (may come on the default image). It will advertise .local over mDNS / DNS-SD. I believe Avahi will advertise on link local if there is no default route to the internet.

Your system may automatically resolve the domain if its able to pickup the mDNS records to SSH in. Been a couple years since I’ve done it, so I could be forgetting a nuanced detail, but I vaguely remember just ‘plug and play’ if internet for the RPi wasn’t required.

Strings are just a pain… Common solution is a runtime guard or doc comment. Can maybe try an overload and annotate a string param with NoReturn and an @throws doc comment.

Been awhile since doing python, but even that might not work since string satisfies both overloads. I recall having issues with unions where types were not discrete. I may have solved by ordering the overloads differently so string is considered first, but again, its been awhile.

https://stackoverflow.com/questions/44912374/python-type-annotation-for-sequences-of-strings-but-not-for-strings

Look at the official docs. There is a table part way down stating which methods are available for each. I pick the one closest to how I use it. So if I’m not mutating I’ll use Sequence over List to inform the caller I’m treating as immutable and to safe guard myself from mutating it in my implementation via static type analysis.

https://docs.python.org/3/library/collections.abc.html

My NAS is an mATX mobo with an i5, 64G RAM, 8 disk drives, 3 nvme drives, and an ARC GPU for video transcoding.

Disk drives are all mirrored. One nvme runs NixOS which is easy enough to redeploy if the drive dies. One nvme is cache on top of the disk drives. Last nvme I use for temp fast storage like Jellyfin transcoding.

Its more of a combo NAS/server as I run most self hosted apps on it (tor node, monero node, jellyfin, *arr stack, etc).

Seconding wireguard. I just have PiVPN manage my configs on an RPi, port forwarding on my router, and a script to do dynamic DNS via API on IP change.

I don’t have CGNAT though and have full control over my work machine.

Similar to what I do. I just have a script that triggers on IP change directly on my router.

EOL support. I have a 11-12 year old System76 laptop. Works perfectly on the latest Ubuntu version.

Their shitty walled garden for both software (iOS) and hardware (soldered components that don’t need to be).

Overpriced.

Fake sense of privacy.

I used Mac OS 6.x through 10.4. When I was in college and couldn’t afford to replace my aging G4, I triple booted Fedora, Mac OS X, and Windows on a hackintosh where I gravitated towards mostly Linux and Windows for a couple games. Owned a couple iPhones but decided to role Android when the nexus 6 came out to save some money when I had my first child on the way and my current phone was dying.

I don’t miss anything I left behind. Had a short stint at work during COVID where I was given a MacBook. While not horrible, I ran into enough nuances I was able to justify to my work using a Linux laptop instead. I just don’t find anything appealing to give them my business.

Jellyfin recommends not using SBCs. I was in the same boat as you a month ago. Started on an RPi. Works fine for raw (no transcoding). Poor performance if you do any scrubbing or try to watch something while new content is processing. Got a mini PC. It was better but its basically a laptop chipset, so still not the best experience. Had other things I wanted to do on my self-hosted setup so decided to just bite the bullet and make a proper build: 12th gen i5, Intel Arc GPU, 4+8 SATA ports with PCI card, 3xNVME, 10xHDD/SSD case. Can’t speak to the performance yet. Learning Ansible to automate managing it including installing the OS.

I would stay away from NAS systems like QNAP or Synology. They tend to not be much better than a SBC.

For the budget constraints I would just echo getting the cheapest desktop-class PC you can get your hands on in a suitable form factor.

https://jellyfin.org/docs/general/administration/hardware-acceleration/#hardware-acceleration-on-docker-linux

While hardware acceleration is supported on Raspberry Pi hardware, it is recommended that Jellyfin NOT be hosted on Raspberry Pis or other SBCs. Many hardware acceleration features are not supported and will fallback to software. In addition, they are generally too slow to provide a good experience when transcoding is needed. Please consider getting a more powerful system to host Jellyfin.

https://github.com/geekau/media-stack/blob/main/full-vpn_single-yaml/docker-compose-media-stack.yaml

Planning to base my next build on this stack. Still waiting on a couple parts.

Have not dug through the whole thing yet, but looks like they use Swag. Haven’t looked into it yet, I normally just go nginx.

I already have wireguard with dynamic dns on my router if I want to access it remotely.