Leaderboard

Another mid-quarter update featuring: New curated app installations Update to the qBittorrent installation Read more about it on our docsite here at Command Deck Update - November 25, 2025 NOTE: This update was applied automatically. You may need to clear your cache.

Let's Talk About Immich If you've been running Immich on HexOS this year, you know it hasn't been smooth sailing exactly. We want to talk about what's happened, why it was so challenging, and how we're working to handle these situations better in the future. What Happened? Earlier this year, Immich deprecated their old storage configuration and required all users to migrate to a new structure. For users running Immich through docker-compose or other manual setups, this meant updating some configuration files and running a few commands. Annoying, but manageable. For some HexOS users, the migration was more involved. Because of how TrueNAS SCALE structures application storage, moving to the new configuration required either reinstalling Immich fresh (the simplest solution) or manually migrating existing data between datasets (a process that involved SSH access, rsync commands, and careful attention to permissions). But if you're choosing between "reinstall the app" or "follow a 15-step guide," neither option feels great when you chose HexOS specifically to avoid that kind of complexity. Why Was This So Hard? When Immich made this change, we had a choice to make. We could have built a comprehensive rsync-based migration tool using the TrueNAS API. It has those capabilities. But that would have meant dropping everything else we were working on to build what amounts to using a cannon to kill a mosquito: a massive, complex solution for what we hope won't be a regularly recurring problem with this particular app. Instead, our community stepped up in a huge way. Users like @forsaken and @G-M0N3Y-2503 created detailed guides (to move or rsync your data). These guides walked through the manual migration process to preserve existing data in Immich. They focused on helping users through the immediate problem, while we continue building the platform we need to handle situations like this properly. That platform is HexOS Local: a locally-hosted management application that will let us perform complex operations without being bottlenecked by the engineering overhead of building one-off solutions through the SCALE API every time an application throws us a curveball. This reduces the technical burden on our team and, more importantly, gives us the flexibility to automate maintenance tasks that previously would have required manual intervention or massive engineering investments. This same platform will serve the Local UI/UX feature we've committed to delivering as part of our 1.0 release. We'll be talking a lot more about HexOS Local in an upcoming blog post, but the key takeaway is this: we're building HexOS to handle whatever the open-source ecosystem throws at it, without having to choose between "drop everything and build a custom tool" or "make users SSH into their servers." What About Right Now? If you're currently running Immich on the old storage configuration and haven't migrated yet, you have options: The simple path: Reinstall Immich fresh with the new configuration. Your photos will need to be re-uploaded, but the setup is clean and straightforward. The preservation path: Follow one of the community migration guides to keep your existing data in place. These guides are more technical and require command-line access, but they work. Our recommendation depends on your situation. If you have a manageable photo library and good backups, the fresh install is probably your best bet. If you have years of photos, carefully organized albums, and user configurations you don't want to recreate, the migration guides are there for you. And if this seems to daunting, email support@hexos.com so we can schedule a time to assist you directly. Moving Forward The Immich situation showed us exactly where we need to invest engineering effort. We can't keep facing the choice between building massive one-off solutions or asking users to break out the terminal. That's not sustainable, and it's not the HexOS we're building. Immich is an incredible project. It's exactly the kind of self-hosted solution we want to make accessible to everyone. The team behind it recently released v2.0, marking their stable release with better upgrade paths going forward. We're committed to making sure that when the next complex maintenance task comes up, whether it's Immich or any other application, we have the infrastructure in place to handle it gracefully. That's the HexOS we're building. Thanks for your patience while we get there.

OMG, I totally see the confusion now. I have edited that post and reworded it to be more clear. TrueNAS has no built-in means to do backups as part of their OS unless you consider rsync cron jobs/replication tasks as "backup". They don't facilitate the connection between servers or even setup the SSH permissions for you (you have to do all of that yourself). This is where our solution takes things a step further. We intend to make the experience much simpler. First you "friend" someone else that is a HexOS user (if you're doing the "buddy" approach). Then you will grant that user "permission" to backup to your system (and set a quota). Then they accept the request and authorize the backup in return (if you're doing that). Then we do the rest (establishing a secure VPN tunnel between systems, creating appropriate SSH users and handling the key exchange, etc.). You'll definitely be able to leverage the buddy backup system between two systems licensed under the same account. That would be silly of us to prohibit or unnecessarily complicate. I get the confusion. Buddy backup is just the marquee feature name for our backup solution between two servers. Every server needs a license. So if you and a buddy want to backup, you both have to have your own licenses under your own accounts. If you have two licenses for two systems under 1 account, you can use the same "buddy backup" system to handle that job as well. Apologies for muddying the water on my initial response. I can see how that had some folks confused, but I hope this clarifies.

We know about this problem all-too-well. The VM features will come in layers, but our plan is to try and get ahead of those kinds of problems by detecting hardware changes and automatically adjusting device ID mappings to your VMs before they start. This still needs a lot of R&D work and there are lots of potential gotchas, but I bet we can address a lot of common scenarios like what you've described. It gets more tricky when you have multiple GPUs in a system of the same make/model, as it's hard to know which one is which for assignment to VMs. I'd say that's a problem for after the initial VMs feature release for us to investigate further. Ultimately there will be some limits to what we can do, but we're going to do this as best as is humanly possible.

That thought did occur to us, but to be perfectly honest we'd rather anyone not comfortable following the guides or having difficulty to just contact us directly for support. That's what you all paid for and we are gonna provide it. I don't want to start asking our users who aren't comfortable to navigate the TN interface.

For detailed installation instructions, please refer to this thread: Illustrated Installation Guide - START HERE! =)