Alternate between testing and stable

I have read in the announcement thread that more people should be running the testing branch to support finding errors early. I’d be happy to support the testing efforts by switching over to testing and deal and report a glitch here and there, but I do have days where I need a rock solid system because I can’t afford a crash during a presentation or a meeting.

Hence I am wondering if there is a clever way to run testing and stable in parallel with the same /home?

I know I can install a separate OS onto my (encrypted) root partition while keeping /home but I am on btrfs so maybe all I need to do is the below?

  1. Create a copy of @ and call it let’s say @testing
  2. Somehow add an entry to grub so I can boot @testing instead of @
  3. Somehow switch the new @testing subvolume to the testing branch
  4. Alternate between the two roots as needed while working with the same @home?

Has anyone done this already and if so anything I should be aware off?

1 Like

I guess you can do that, but

is not IMHO

You may include @home in your new setup but not as /home(I don’t know how that is done with btrfs). You can symlink your actual data (at home) to your new installation’s /home/$USERNAME folder.
Another way is re-using @home as /home but using a different user name for the other user.

Was wondering why you wrote this but upon thinking about this some more I guess it’s because unstable will come with new settings, defaults, and options which would go into /home and potentialy confuse the hell out of the stable branch with the consequence that both branches become unstable.

Sounds like mixing stable and unstable isn’t as bright an idea as it looked like initially :smirk:

1 Like

Perhaps you could switch to testing and not update on the day (or day before, maybe) when you have critical demands.

Nothing says you must update as soon as it is available.

Correct, but that also means I may be stuck with issues (known and unknown) the current unstable build has.

How do others work with testing? Is testing stable enough to be your daily OS or is it more like inside a VM and you spend an hour or so every now and then?

What I do is just do not update when I need a 100% reliable system. Once I know I have some spare time I backup with Back in Time and do an update.

1 Like

I’ll be getting a new ssd for my laptop soon. I plan to do a double installation of Manjaro, one for stable and one for testing, each with it’s own home.

1 Like

That’s exactly how I handle things with my stable branch installation right now.

It has been my daily and only OS for over a year now. Yes there are a few more issues than stable but the severity of them tends to be a bit less. Generally they have been quickly fixed by further updates, community feedback or even occasionally by using my own thinking bits.

Really the difference is more frequent, but smaller, updates.

All else fails, roll back to the backup made before updating - or so I tells myself since I have not yet needed to…

1 Like

Why? All your user config options are in /home/user. Stick with either one or the other, no need for two bare metal systems.

If you really want to stay on stable but help test then create a testing system in a VM, set it up similar to your stable system. If you are really keen then clone that VM and change to the unstable branch.

Yes.

IMO testing is easier to manage than stable, more frequent but smaller updates, some issues may pop up from time to time … that is the purpose of testing.

If you are a little anxious then simply wait a day after a testing branch snap has happened and read the announcement thread for potential issues. Or update in a testing VM before bare metal updating.

1 Like

Ok, convinced…I’ll switch over to testing :slight_smile:

This is how, right?

sudo pacman-mirrors --api --set-branch testing
sudo pacman-mirrors --fasttrack 5 && sudo pacman -Syyu

This is the reason I duoboot with two separate partitions. I also run, as now, 18.0.4 testing and 18.0.3 stable, so things do get confusing. Keep your home in the same partition as the OS. IMHO

Switching branches is quick (about a minute) but I when something breaks on unstable, you may need days to fix something and if you need to work something important that is not working right… So this is indeed problematic. I don’t see how you could have installed different packages versions on the same system and then choose which one to use…

So the best bet is: use often backups (like timeshift) and if something really problematic shows up, roll back from backup and switch to stable for the time you need to do something and then switch back to the “fun mode” ;).

However comparing testing and unstable, testing is much safer. Unstable can be stressful, testing rarely is, because biggest problems are stopped on unstable.

Well, the original idea was to create a btrfs snapshot of the stable os =(@) and call it e.g @Stable, switch to testing and whenver stability is required create another snapshot of the os (=@testing) and switch back to the original @stable snapshot. A matter of seconds + a reboot.

I still think this would work but where this approach falls short is that testing and stable will also make changes to settings stored in /home shared between the two @ snapshots.

I guess I could still achieve what I wanted if I kept my data separate from /home, but so far testing works flawlessly for me. Or better said testing shows the same hickups stable did :grinning:

This topic was automatically closed after 90 days. New replies are no longer allowed.