Manjaro LTS (or the idea of slow mirrors).

Based on these existing conversations,

I thought an interesting idea would be to have Manjaro mirrors that are updated only once or twice per year. They'd work the same way, but they simply wouldn't receive updates for a while.

I'm a user that, for example, sometimes likes to add or remove software that I need or don't need, and in some cases I would prefer not to have to do a full system upgrade (it can be inconvenient sometimes, or sometimes people have limited network connectivity and want to download only one package, or etc).

I think this would be great! "Manjaro Slower" (or some similarly named release) would be set up so that /etc/pacman.d/mirrorlist points only to mirrors updated every 6 months.

Perhaps a release called "Manjaro LTS" (or something) would have a /etc/pacman.d/mirrorlist that points to mirrors that are updated only every year, or two years, or similar; longer than "Manjaro Slower".

@realmain mentioned in the other thread,

This has already been discussed, and will not happen. You'd need a whole new team to do that.

I'm wondering, do we really need a whole new team for that? Or can we simply just choose to update the "slower" mirrors to whatever the state of the rolling mirrors are, every 6 months?

What makes it necessary to have a whole team for this?

1 Like

...not again. :fearful:

5 Likes

Haha... Oh... so it has already been discussed that much? (Sorry, I missed those threads!)

Yes, it has. :man_facepalming:

Look, Manjaro is a curated rolling-release distribution. It doesn't receive updates as frantically as Arch does, because the updates are more properly tested, and they are bundled together. That makes it a very stable recipe ─ at least, if you stick to the Stable branch.

When you install Manjaro, then you know what you're committing to. If you don't like that, then why would you want to change it into something it was never meant to be ─ namely a fixed-point distro ─ when there are so many other distributions already that cater to that very need?

Besides, there are fixed-point distributions that have more issues than you'd ever get in Manjaro. So why change the recipe? :man_shrugging:

4 Likes

Fine you know what ..

I can think of a way of doing this ... sorta.

(mirrors as drop-in-replacement)

But I would need server space.

If there is enough interest .. and I can bang out a working scenario without draining my meager funds .. then I might give it a shot.

But there is a big caveat here. . running a system this way will make the AUR less stable/useful than it is now.

No promises :wink:

It's not the intention of manjaro to be as stable as possible. It's a rolling release if not a bleeding edge distro. there are other distros which deliver the stability. Using manjaro/arch means that you have a good working backup/restore strategy in case the system crashes and the discipline to create regular backups.

Mirrors are not maintained by Manjaro but universities, ISP's or dedicated users. You cannot tell a mirror - hey - you should only update twice a year - they would laugh in your face.

If you want to do something like that - you can easily do this your self.

It only requires you to create your own mirror - it don't have to be official - just a local mirror - you can then update it when you you deem it necessary.

Then edit your local /etc/pacman.d/mirrorlist and insert your mirror server.

If you don't feel like running a server for it - you can still create a file based mirror.

Then you can install software using that mirror - it will only change when you update it.

3 Likes

And, when you then update it after a whole year, you can rejoice over all the breakage you'll be getting because of certain changes in configuration, package bundling and dependencies that took place during the year that you didn't update. :stuck_out_tongue:

4 Likes

I see, so you're saying it would be the same as a pacman -Syyu a year later on normal mirrors. :thinking:

..I mean .. that just lumps them together at once for issues that might exist. And you might even get lucky and simply skip ones that have since been fixed.

1 Like

Yes, because building a point-release distribution is a very different thing. With a point-release distribution, you collect all of your software as source code, and you build it into binaries, and then you start ironing out the bugs before you can release it to the end-users.

So what that basically means is that you create a snapshot of the evolution of a GNU/Linux system, and then you make sure that you get rid of the vast majority of the bugs in that snapshot, while in the meantime and in the background, the evolution of the software continues.

By the time you've gone through an initial alpha release, several beta releases, several release candidates and then onto the official final release, the only thing that might still be up-to-date with the evolution of the GNU/Linux ecosystem would be your kernel, because it's easier to grab a newer kernel than, say, to grab a new glibc, recompile the whole system against that and then find that it conflicts with the versions of certain other libraries and/or packages you've got on your system.

And even then still, you've just put out that shiny new release full of outdated software, and then you're going to have to keep that up-to-date with new updates because of security flaws that have been discovered in the meantime, and so on. So in the end, your long-term-support release is going to require even more work to keep it in a stable and functional state, and you'll still be left with outdated software.

Now take Manjaro instead. New software is bundled into snapshots and tested, and these new snapshots will contain most of the security fixes ─ the exception being a critical flaw that was only just discovered and that requires that certain components be updated immediately.

The snapshots you're creating are also fairly consistent, because after all, Arch ─ which is Manjaro's upstream ─ has already gone through some testing of its own before those same new versions of the software appear in Manjaro, and then Manjaro applies more testing via its Unstable, Testing and Stable-Staging branches.

Which do you think is easier to maintain? And perhaps even more crucially, which do you think will ultimately leave you with the most stable (and supported) operating system on your computer? The point-release distro that still needs to be kept up-to-date, or the tested and up-to-date curated rolling-release system?

There's a reason why Manjaro works the way it does. There are plenty of Debians and Debian spinoffs, RedHats and RedHat spinoffs already. And then there's always Slackware ─ still receiving security updates, but their software itself is archaic because there hasn't been any new release anymore since 2014. (Their KDE version still stands at 4.2 at installation time and at 4.14 after updating.)

6 Likes

Roughly yes.

I have played with older installations - how to successfully update an old package selection. When you install from an old ISO using Calamares you get a system with an old set of packages.

You can then test your update troubleshooting skills in a virtual machine.

Beforehand you already know a few of the issues you will be facing

  • new pacman hooks which requires pacman 5.2
  • hplip - python cache files
  • firewalld - python cache files
  • gnome shell extensions causing trouble

Start honing your skills

  • Download an old ISO from the archive.
  • Boot a virtual machine using the ISO and install using the GUI installer.
  • Shut down the vm
  • Make a snapshot using VirtualBox snapshotting feature.
  • Boot the system - set an optimal mirror and update
  • Make notes along the way
  • When you fail
    • close the virtual machine
    • restore the snapshot
    • learn from what went wrong the first time
    • try again
2 Likes

Please correct me if I am wrong. Since Manjaro has been supported for quite a long time on my machines, so I guess I could consider it as LTS. Something which is different from fixed point release.

2 Likes

Well, no. Technically, the term LTS applies to software that's receiving security updates and bug fixes, but without that the software's version gets updated. This means that the API and ABI will remain the same all along, and that is not the case with Manjaro. :wink:

3 Likes

Manjaro LTS is then kind of contradiction in "terms".

2 Likes

Philip experimented with pushing updates less frequently to Stable several years ago, but switched back to pushing them if no issues got reported and the reported ones fixed in Testing.

You got the explanation why it is not feasible to slow down the update cycle of the rolling Manjaro.

Philip mentioned a different possible approach which would effectively mean to create a different distro Manjaro Solid.

There have been no updates on this idea so far.

4 Likes

@trusktr - Rolling release will provide you with bug fixes which is already available upstream because they always will be up-to-date with the upstream meaning less bugs and a more stable system (Bonus :point_right::point_right: security patches for not just Firefox or Thunderbird.. but for every packages because some people still treat security bugs as just another bug and stick to normal releases. :roll_eyes: )

But YOU, won't get them! You will have to wait another 6months to get that fix because you are using a point release distribution (like Ubuntu or your Manjaro LTS :wink: ). But wait, you can use a PPA (if you're using Ubuntu) to get the fixes or the latest versions... but then the PPA breaks (Ohhhhooo)! Now you remove the PPA and go back to old version and tolerate the same old software that you are using with the same old bug until the next point release or if the fix doesn't arrive in the point release, wait for the next major release! Do you want Manjaro to be like that too? Many of us don't want our system to be this way! Hence this approach. :slightly_smiling_face: Hope this gives you an idea about the issue of point releases and the advantages of a rolling release. :slightly_smiling_face:

It can be overwhelming if you don't know what you're doing...

:point_right: Bandwidth wise for you because there will be a lot of updates and GB's to download.
:point_right: Also, system maintenance wise. More updates, how to fix... etc.

Hope this helps you to clear why Manjaro users won't be interested with Manjaro LTS. Manjaro is really stable and i've never broke my system. I've been using Manjaro for 4 years. I've reinstalled twice, but for hardware upgrades only. Once an SSD upgrade and then a new laptop. Stick to sane defaults if you're not confident (KISS philosophy). Learn from the community (we're awesome) and install Timeshift. You should be fine!!! :wink:


@eugen-b - Yeah. Still don't understand why snap!? There are enough threads in this forum who complaints about snap issues! I myself removed snapd recently from my system and saved 10 freaking seconds from my boot! I actually wanted to kickstart the conversation then when I removed snapd about this unbreakable system but then decided to not rant about it since there is no update on it.

But somehow snap is a great technology as per Philm without a good explanation. Snap is neither intuitive nor performance wise good. Hope he changes his mind about snap :crossed_fingers::slightly_smiling_face:

Also, hope there will be an opt out if it moves forward with snap, like the option to use the current way of installation.. the normal way (Philm said the current way is needed by the way. But that begs the question, for how long until it will be dropped?) . Or at the very least move to Flatpak which is better in every sense when compared with snap! :wink:

2 Likes

https://repo.kde-git.manjaro.org/ISO/

1 Like

Everything has already been said about why it's not happening, but i just wanted to make the following reaction:

People must understand that being rolling is among what makes Manjaro itself !

:cookie:

5 Likes

completely agree. if it aint broke don't fix it

Forum kindly sponsored by