The new Linuxserver.io docker image at the very least has solved the annoying update cycle NextCloud has and seems to have fixed the need to do that every few months. I haven’t ever had it die but I don’t push it hard and I keep the plugins to a minimum because I just don’t trust it and it doesn’t run all that well.
I gave up on owncloud just before it became nextcloud because it kept breaking every time I updated it.
Wallabag is similar for me now. I’m stuck on a slightly out of date version because I can’t get newer ones to run. Everything else I self host is painless though.
I wish there were an alternative in a sane programming language that I could actually contribute to. For some reason PHP is extremely sparse in its logging and errors mostly only pop up on the frontend. Having to debug errors after an update and following some guide to edit a file in the live env that sets a debugging variable, puts the system in maintenance mode and stores additional state in the DB is scary.
Plus PHP is so friggin slow. Nextcloud takes noticeable time to load nearly anything. Even instances hosted by pros that only host nextcloud are just slow.
I run a k3s Kubernetes cluster on a single KVM host(multiple VMs). Honestly I do not care a single f*ck about that machine nor k3s itself. I update once a year, do not have any documentation written nor IaC somewhere. I always forget how I configured the networking stuff for example. But that machine runs my critical services flawlessly without a single crash in like 3 years. So no I cannot relate.
It is fine, but then again I update it often too late which is actually pretty bad. The problem is Nextcloud pushes new features and a high frequency schedule of releases with those at an alarming rate of speed. Perhaps for corporate environments it is not as big of a deal as a professional team can fix obscure bugs with their knowledge and experience on their mirrored test servers, but home users don’t have these resources available and public community knowledge and bug fixes need time which that release schedule hinders.
I still wouldn’t say it is bad by default, simply because somehow it runs pretty stable for me since a decade. Updates are a pain though with many breaking changes and little bugs.
Yep. Got such a service as well. I’ve got this one docker container that’s supposed to connect to a VPN and provide access from the outside to another one. The bitch keeps just crashing to a point where even “restart policy: always” will give up on it. Doesn’t matter too much usually, since I can start the container before I need it, and it will usually run for half a day or so, yet still
I disagree–a system (even Arch!) should be able to update after a couple months and not break! I recently booted an EndeavourOS image after 6 months and was able to update it properly, although I needed to completely rebuild the keyring first
Arch and EndeavourOS are the same thing. There is no functional difference between using one or the other. They both use pacman and have the same repos.
Very true–the specific EOS repo has given me a bit of trouble in the past, but it takes like 3 commands to remove it and then you’ve got just arch (although some purests may disagree 🤣)
I know this is how it’s supposed to be and how it should be but sadly it doesn’t always go this way and arch is notoriously known for this exact problem, the wiki itself tells you to check what’s being upgrades before doing because it might break. Arch is not stable if you don’t expect it to be unstable.
I’m using opensuse tumbleweed a lot - this summer I’ve found an installation not touched for 2 years. Was about to reinstall when I decided to give updating it a try. I needed to manually force in a few packages related to zypper, and make choices for conflicts in a bit over 20 packages - but much to my surprise the rest went smoothly.
I regularly “deep freeze” or make read-only systems from Raspberry Pi, Ubuntu, Linux Mint LMDE and others Linux Distros whereas I disable automatic updates everywhere (except for some obvious config/network/hardware/subsystem changes I control separately).
I have had systems running 24/7 (no internet, WiFi) for 2-3 years before I got around to update/upgrade them. Almost never had an issue. I always expected some serious issues but the Linux package management and upgrade system is surprisingly robust. Obviously, I don’t install new software on a old system before updating/upgrading (learned that early on empirically).
Automatic updates are generally beneficial and helps avoid future compatibility/dependency issues on active systems with frequent user interaction.
However, on embedded/single purpose/long distance/dedicated or ephemeral application, (unsupervised) automatic updates may break how the custom/main software may interact with the platform. Causing irreversible issues with the purpose it was built for or negatively impact other parts of closed circuit systems (for example: longitudinal environmental monitoring, fauna and flora observation studies, climate monitoring stations, etc.)
Generally, any kind of update imply some level of supervision and testing, otherwise things could break silently without anyone noticing. Until a critical situation arises and everything break loose and it is too late/too demanding/too costly to try to fix or recover within a impossibly short window of time.
Bad stories about nextcloud scare me 😂 I hope Im not gonna jinx myself, but my nextcloud runs super stable for almost a year. I get some errors while updating, but service doesnt stop working and its usually simple fix by following the message it shows.
I removed apps that I dont use (most of them) and web ui became super fast on my budget server
Actually all services are so smooth and almost no issues, maybe beginner luck 😉
I’ve hosted mine for years on my own bare metal Debian/Apache install and 28 is the first update that has been a major pain. I’ve had the occasional need to install a new package to enable a new feature, or needed to add new/missing indices to the database, but the web interface literally tells you how to do those things, so they’re not hard.
28 though broke several of the “featured” apps that I use regularly, like “Retention”. It also introduced some questionable UI changes that they had to fix with the recent .1 update. I’ll get occasional errors when trying to move or delete files in the web interface and everything. 28 really feels like beta software, even though we’re a point release in and I got it from the “stable” update channel.
I’m on my laptop so I thought I would elaborate on my first comment to give you things to watch out for if/when you update. I’ve been hosting mine with the zip file manually installed with my own Apache/PHP/MySQL/MariaDB setup for ages now without issue. It’s been rock solid except for, like I said, the occasional changes required to take advantage of new features such as adding new indices to the database or installing an additional php addon. Here’s the things that I noticed with updating to 28.
The 3 dot/ellipses menu was missing in the web interface and was replaced with dedicated buttons for “Download”, “Add to Favorites” and “Delete”. Shift clicking was also broken. This meant that when I, for example, take a lot of photos for a holiday, I can’t use the web interface to select a large range of multiple files and then move them all from “InstantUpload” into a more permanent album. I either had to use the mobile app, or do them one at a time. The ellipses menu, along with the options to bulk “move/copy” have been added back since then with the *.1 update, but shift clicking in the web interface to select a range of files is still broken.
The “Retention” app, which is listed as a “Featured” app doesn’t function any more. I used it to automatically delete backups of my Signal messenger, files in the “InstantUpload” folder that were over a year old, etc. You can enable it, but it doesn’t actually work and just throws errors in the log file, which is now reported in the “Overview” portion of the “Administration” page with a note of “X number of errors since somedate”, and prevents you getting the green checkmark. It’s probably safe to assume that other apps will also have issues because I had half a dozen get automatically disabled with the update.
Occasionally when I use the web interface to move or copy a file, I’ll get an error message that the operation failed. Sometimes this is true, sometimes it’s not and the operation actually succeeded. If it ends up being true and the move did actually fail, doing it again results in a successful move.
It seems like they’ve made some substantial under-the-hood changes to the user interface that shouldn’t have been shipped to the “stable” channel. It’s not completely broken, it “is” usable, especially after they restored my bulk move/copy button, but I still can’t use the Retention app, at least last time I looked, so I’ve literally got daily cron scripts to check those folders for old files and delete them, then trigger an occ files:scan of the affected directories to keep the Nextcloud database in sync with the changes. This however, bypasses the built-in trash bin so I can’t recover the files in the event of an issue. I actually considered rolling back to 27 for a bit, but decided against it, so if I were you, I would stick with 27 for a while and keep an ear to the ground regarding any issues people are having that are or aren’t getting fixed in 28.
I haven’t had any issues with Nextcloud yet. But any torrent client refuses to work. I’ve tried various qbittorrent containers, transmission, deluge briefly, they all work for a while but eventual refuse to do anything.
Good call. I do some backups now but I should formalize that process. Any recommendations on selfhost packages that can handle the append only functionality?
I use and love Kopia for all my backups: local, LAN, and cloud.
Kopia creates snapshots of the files and directories you designate, then encrypts these snapshots before they leave your computer, and finally uploads these encrypted snapshots to cloud/network/local storage called a repository. Snapshots are maintained as a set of historical point-in-time records based on policies that you define.
Kopia uses content-addressable storage for snapshots, which has many benefits:
Each snapshot is always incremental. This means that all data is uploaded once to the repository based on file content, and a file is only re-uploaded to the repository if the file is modified. Kopia uses file splitting based on rolling hash, which allows efficient handling of changes to very large files: any file that gets modified is efficiently snapshotted by only uploading the changed parts and not the entire file.
Multiple copies of the same file will be stored once. This is known as deduplication and saves you a lot of storage space (i.e., saves you money).
After moving or renaming even large files, Kopia can recognize that they have the same content and won’t need to upload them again.
Multiple users or computers can share the same repository: if different users have the same files, the files are uploaded only once as Kopia deduplicates content across the entire repository.
There’s a ton of other great features but that’s most relevant to what you asked.
I’ve used rclone with backblaze B2 very successfully. rclone is easy to configure and can encrypt everything locally before uploading, and B2 is dirt cheap and has retention policies so I can easily manage (per storage pool) how long deleted/changed files should be retained. works well.
also once you get something set up. make sure to test run a restore! a backup solution is only good if you make sure it works :)
As a person who used to be “the backup guy” at a company, truer words are rarely spoken. Always test the backups otherwise it’s an exercise in futility.
Keep your Apple TV and use it as a streaming client for whatever you stand up on the backend. Personally I have a Synology NAS that I love and I use the net to get all my content. Use the net. 😉
Appreciate your comment, and that seems like a common setup. If you didn’t have the ATV, what would you front end the Plex server with? I have a Synology router and would probably buy a Synology NAS, if I went that route.
Actually with a Synology NAS you don’t need Plex, they have a built in equivalent called DS Video with apps for Apple TV, iOS, Android, etc!
I’ve had an Nvidia shield in the past as well and it works reasonably well, but the video experience is definitely better on the Apple TV. The Android boxes make more sense if you want a place to install emulators that also occasionally streams.
Thank you for this! I’ll look more at the Synology NAS devices and see what that’s all about. I’m probably the other way around, stream more, and emulate once in a while.
selfhosted
Newest
This magazine is from a federated server and may be incomplete. Browse more on the original instance.