Update of nextcloud package in linux distribution independently from updating the server?

ℹ️ Support
,
1

Suppose I’m running nextcloud 29 on a debian machine, but I’m thinking about switching to openSuse. Something bothers me since I started with nextcloud:

Question: If Debian or openSuse offers an updated nextcloud package, can I simply accept this update, because the new package (let’s say version 30) includes all that’s necessary to go on running nextcloud 29 on my server? So in Debian do: apt update, apt upgrade, even if there is a new major version of nextcloud going to be installed?

In other words, can I update the package of my distribution independently from updating my running nextcloud server to the next version?

1 Like
2

1. Automatic updates via package manager are risky

Nextcloud requires manual steps during updates, such as running the web installer or the occ upgrade command. If you allow the package manager to automatically update Nextcloud, the code might update, but critical steps like database migrations may be skipped. This could result in your installation breaking or becoming unstable.

Another common issue is that updates via the package manager can cause incompatibilities between the database, configuration, and Nextcloud code. This is a frequent source of problems in such setups.

2. Updates depend on the distribution

  • Debian/Ubuntu: These distributions provide stable packages, but they tend to be older. On a production server, this can be fine, but don’t expect to have the latest features right after release.
  • openSUSE: If you’re using a rolling release like Tumbleweed, packages are usually more up-to-date, but there’s a higher risk of something breaking.

3. My personal experience

Over the last three years, I’ve tried every method of installing Nextcloud – manual installation, package managers, Snap packages, and Docker containers. The best solution for me is undoubtedly Nextcloud AIO (All-In-One), which I’ve been running for the past two years.

Here’s why I recommend AIO:

  • Fewest issues with updates: All necessary steps (code, database, apps) are handled automatically without any manual intervention.
  • Stability: I haven’t encountered any major issues with functionality since installation.
  • App functionality: Office and other Nextcloud apps work flawlessly.
  • Independence from distribution: Since AIO runs on Docker, you can use it on any server that supports Docker. Personally, I use Ubuntu Server with the xfs filesystem because I handle large multimedia files.
  • Modularity for weaker hardware: With AIO, you can enable only the modules you actually need. This is especially helpful if your server hardware is not very powerful.

4. Answer to the question

No, it’s not safe to update Nextcloud via the package manager without manual intervention. This could lead to issues with the database, configuration, or apps. If you want full control:

  • Use the official installation method, Docker, or better yet, Nextcloud AIO.
  • If you rely on distribution packages, always test updates on a staging environment first.
  • Follow the documentation, especially for major updates.

In my opinion, if you want reliability, I highly recommend Nextcloud AIO. It works independently of the Linux distribution, minimizes potential issues, and offers the flexibility to run only the modules you need, making it ideal for systems with limited hardware resources. For a production server, it’s the ideal solution.

2 Likes
3

I’m running Ubuntu 24.04.1 LTS and nextcloud 30.0.4 bare metal. Collabora is running in Docker. Yes i had issues with f.e. office, but that’s solved now.
Also Adguard/wireguard on the same machine.
On a 2/2Gbit fibre line.

Some issues i didn’t have with Docker/snap showed up during BM installation but way less.

I thought i only had to use the occ upgrade command to upgrade nextcloud?

I tried docker and snap but always ran into editing .php files and/or locating them. Switching to BM made tons of errors go away, configuration seems simpler and you learn stuff by doing so. The AIO one i want to try on a different machine to see how that goes but for now i’m happy.

And what do you mean by :

For what needs? i’ll have maybe 10 users on a i5 3570/16GB/240GB SSD (can add more storage later) . Once i figured out (someone pointed out) the built-in CODE server wasn’t good at all and better to run colabora separately i’m basically done (just want to add previews for certain files now). It’s running great, great speeds, office opens fast, system boots in 10 seconds after a reboot.

Steaming 4GB movies over lan-vpn / mobile-vpn works great, talk works (only tested 2 users) If i really need power i can move it to a dual Xeon 32c64t/192GB server (but electricity is expensive). And i would probably make this a VM.

4

Hey,
based on your configuration and what you’ve shared, I thought I’d chime in with some suggestions from my own experience running a self-hosted setup at home. Like you, I’ve been self-hosting for years, and here’s why I think you might want to consider using Proxmox for your server setup.

Why Proxmox?

Using your entire hardware setup (i5 3570/16GB/240GB SSD) for just one Ubuntu server is, in my opinion, an inefficient use of resources. Proxmox allows you to:

  1. Optimize Resource Utilization:
    You can divide your hardware into multiple virtual machines (VMs) and allocate resources dynamically. This means your server can handle multiple services simultaneously without wasting unused capacity.

  2. Simplified Management:
    Proxmox provides an excellent interface for managing VMs, containers, and backups. One of the best features is automated VM backups, which require zero intervention once set up.

  3. Better Fault Isolation:
    If one service crashes or faces issues, it doesn’t affect the entire system. For example, I run my Nextcloud AIO in a dedicated VM, completely isolated from other services like Plex or Jellyfin.

  4. Flexibility to Experiment:
    Want to try new setups or software like Nextcloud AIO? With Proxmox, you can spin up a new VM or container without affecting your main server environment.

  5. Energy Efficiency:
    Despite running multiple VMs and services, my server’s power consumption is only around 40–50 W, which is quite efficient for the workload it handles.

  6. Solving Overkill Memory Issues:
    I had a similar issue with overkill memory crashes on Ubuntu server. Moving to Proxmox resolved this problem completely by using memory management and fine-tuned resource allocation.

How I Use My Server with Proxmox

Here’s how my setup looks, so you can get an idea of what’s possible:

  • Hardware:
    32 GB RAM, check my home scheme
    server_scheme_color
    server_scheme_color1974×1342 412 KB

    .
  • VMs and Services Running:
    1. Nextcloud AIO:
      Dedicated VM with 16 GB RAM and 2 TB disk (XFS filesystem) for large files. Modules are enabled selectively (see screenshot below).
    2. Plex & Jellyfin:
      For multimedia streaming, with media libraries hosted on a NAS.
    3. Audiobookshelf:
      A personal audiobook library hosted as a self-contained service.
    4. NGINX Proxy Manager:
      Manages all subdomains and routes requests to the correct VMs/services.
    5. Other services:
      Lightweight services like Vaultwarden, Home Assistant setup, Wordpress pages, Calibre web…

Addressing Your Questions

Modularity for Weaker Hardware

Nextcloud AIO allows you to enable only the modules you actually need. For example, if your hardware is limited (like your i5/16GB setup), you can skip resource-intensive modules and run just the basics.
Here’s a screenshot showing how I’ve selectively enabled modules in AIO.
modules

Collabora/OnlyOffice Issues

The problems you mentioned with Collabora and OnlyOffice on Snap or bare metal installations are very familiar to me. These were a constant headache until I switched to Nextcloud AIO. It takes care of everything automatically—just enable the module, and it works out of the box.

A Recent Success Story

I recently helped a friend with a very similar hardware setup to yours. He also made the switch to Proxmox, and he’d definitely agree it was the right move. He’s now running multiple VMs efficiently, and his hardware is being utilized to its full potential.

At the same time, he switched to Nextcloud AIO, and he’s been pleasantly surprised at how well everything works—without the slightest issue. Even tasks like Collabora setup, which used to be a hassle, now work seamlessly.

A Few Final Thoughts

If you’re happy with your current setup, that’s great! But I’d suggest trying Proxmox to see how it fits your workflow. For me, and for my friend, it’s been a game-changer in terms of flexibility, stability, energy efficiency, and automation.

Feel free to reach out if you want more details or have any specific questions about Proxmox, Nextcloud AIO, or my setup.
Just take this as something to think about. :blush:

1 Like
5

Hi,

Open Source software allows you the freedom to choose your deployment route or customization. There isn’t a single way which is the best way. Depends on your available resources, and need of the project along with your technical skills on which way you may prefer.

Some users have already given their way of doing things.

But coming to your specific question, for Debian, you can selectively update or upgrade packages. This isn’t exactly a nextcloud related thing or question.

For example, instead of upgrading all, you can selectively update things in the OS.

sudo apt upgrade packagename1 packagename2

Thanks.

6

Happy new year!

And thanks for the info.

I do have some more powerful and also power hungry and very loud systems here.
(i got 6 of those Dell Optiplex machines with similar specs).

My real server however is a dual node 2U Supermicro system.

2 x 1000watts psu (redundant)
1 node contains

  • 2 x Xeon 16core/32threads (model i forgot, 2.4ghz to 2.8ghz)
  • ~192GB ECC memory (if i recall correctly some modules died).
  • 3 x 240-500GB SSD’s
  • 3 x 4TB HDD’s
  • 2 x Gbit nic
  • 1 x ipmi
    1 node has a Quadro card for passthrough.

This all connects to a Juniper 48port switch, connecting it fully takes 6 to 8 ports.

It used to run Vmware ESXI 8.0 and a separate machine ran Vcenter.
Currently it runs Proxmox on node 1 and ESXI on node 2.

I tried proxmox but i couldn’t get the storage to work out as i wanted, i’m not exactly sure what the issue was, something with HBA/raid from the onboard controller and adding that to Proxmox - drive/storage pools/caching?
It’s been a while since i’ve been working on it due to sickness etc.

Maybe i’ll do a thorough check soon and see if i can configure it correctly.
Since it’s a dual node system VMware was able to load balance/HA VM’s across the nodes using Vcenter and at that time Proxmox didn’t have anything like that.

Also VMware ESXI was running from an onboard USB drive since it didn’t need much, Proxmox performed poorly that way and took away a drive bay/ssd reasons lots of technical stuff.

When i give it a go i’ll post to see what my findings are, webhosting has never been my forte and i’m unfamiliar with optimizing settings and such.
And i need to figure out the storage, in these cases i’d like to keep my OS drive (proxmox installation drive) separate from my storage pools but then i waste a storage bay.

7

You can’t update the Nextcloud package independently from the server upgrade. If you run apt upgrade or the equivalent on openSUSE, it might upgrade Nextcloud to the latest version (e.g., 30), which could require additional steps like database migrations. To avoid this, you should lock the Nextcloud version in your package manager or use Nextcloud’s own update system for version control.

8

This topic was automatically closed 8 days after the last reply. New replies are no longer allowed.