Plymouth as an opt-in feature

For consideration:

I’m aware it’s been suggested more than once to remove Plymouth from the default install; or, at least, have it disabled by default.

How feasible is it to have Plymouth as an opt-in feature instead (via Calamares), with a simple warning that it may cause boot failure for some users?

10 Likes

I am afraid I’ll have to respond with this “Me Too” post. :wink:

3 Likes

Generally speaking - if one is not satisfied with how the ISO is laid out - why not build a custom ISO.

On a reasonably powered hardware the build time is neglible.

It will take some time to learn - but once you got it - you get addicted.

While that’s a valid suggestion, it does nothing to enhance the new user experience; a user follows the recommended installation procedure only to be met with a black screen and/or boot failure.

One could also argue that Linux comes with no guarantees; and to the extent that’s true, the argument still fails to address the potential negative impacts on new users and wider adoption.

In my view, if Plymouth (a known bad player) were given the status of an opt-in feature rather than being a mandatory inclusion in the default install - if it then causes a failure to boot during initial setup - at least there is an (unspoken) “…we told you so…”.

Rather than removing Plymouth completely from the default install; a topic that often seems to arise; this at least seems like a fair and practical compromise. Is it feasible to have Plymouth be an opt-in feature? I don’t know; at least, not from every perspective; but, I wonder if it might not be better than the constant tweaking with hopes that Plymouth issues might magically go away.

In any case, this was intended to promote some conversation on the topic, rather than being a feature request, so if you’re reading this, you’re invited.

Don’t you have anything more to add? :wink:

1 Like

The size of the Manjaro userbase is largely unknown but I reckon the user base is not counted tens of thousand more likely hundreds of thousands.

Seen in that perspective - the relatively few that actually opens an issue - only a subset is actually related to plymouth where the rest lands withing Nvidia driver realm, Gnome extensions or badly maintained Plasma customization.

That may have been true - and given the rolling nature of Manjaro - it is a valid suggestion - I just don’t see it happen.

I happen to appreciate Plymouth - even more with the 24.x in unstable branch - as this gives me a localized encryptiion phrase input - and that is something I have come to appreciate.

Not really except IMHO if it’s known to cause issues and isn’t (in many cases) needed, it should be optional.

I haven’t had any issues but then I don’t have nvidia hardware.

1 Like

I only had problem with black screen on boot once, caused by an update. Uninstalled plymouth und everything was fine again. And my hardware is pure AMD (ryzen 7735u).
Imo everything should be optional, when it shortens the list of “probably no problems booting when buying following hardware” to “Intel processor with iGPU, but not something with that old crappy Celeron which is known to have other problems displaying a graphical desktop”.

2 Likes

Ah, yes… Celeron…

…that’s one vegetable I was never tempted to take a byte from.

I considered starting a thread focused on dumping Nvidia…

…but then reconsidered, as it was likely too much like standing in the middle of Manchester United turf, and shouting “Man U sucks royal d…”

1 Like

Though really part of the issue is even ISOs not starting in some cases due to plymouth.
While I would welcome this choice over nothing at all, it would not fix that problem.

(Building ISOs without it on the other hand does. Ill post mine when Plasma updates are out.)

Calamares in general would be extra lovely with any number of selection screens.

This topic was hot when one was created for office suites. Still we have not had similar developments.

And just in case anyone stumbles here;

1 Like

I think you hit the nail on the head there.

Even though Manjaro isn’t aimed at complete newbies, if that’s the first ISO someone tries it might put them off trying others e.g. Mint or whatever, or later being able to ask for help and learn something (although, we know how that often goes).

Maybe that’s worth consideration in the “grand scheme of things”?

The issues with Plymouth are relatively few given that we don’t actually know how many systems is running Manjaro Linux.

But as Manjaro Linux appears to be very popular - judging from all the hate speech - several 100.000s system must be in play.

Given that - I don’t think Plymouth is that big an issue - and not an issue that warrants a complete removal.

I agree there may be systems which has issues - but the team has gone a long way to pave the road so the vast majority will boot and install and work.

If a user runs into black screen and is not able to solve it - either on their own or by asking in the forum - the user is likely a :shopping_cart: user and those tend to cause unspeakable horrors on the forum.

A short while ago I acquired a Thinkpad X13 AMD gen.4 (blank as in no OS) - with AMD 7840U - Manjaro default booted OOB - no issues what-so-ever.

By removing Plymouth and it all works - you imply that Plymouth is the cause.

Given the fact that Plymouth actually works - the issue stems from the configuration of the system - configuration which make Plymouth fail.

Fix the configuration instead of removing Plymouth - the system continues to work.

Experiences and fact gathering - like the topic on Plymouth in QA section - will help avoiding the issue - also part of why a stable snap has been delayed since mid March - we are now in mid May - that is the longest time I can remember between stable snaps.

That long time implies that it has not been easy - a lot of hard work has been put in to making - amongst other things - Plymouth, a stable transition.

2 Likes

The only things that have been done to ‘fix’ plymouth that I know of have been to

  • roll back plymouth to some version that works more often
  • enforce kms
    (which is a default anyways that should have a pacnew, but now theres some hook things too)

Neither of these has a 100% success rate, because plymouth does not have 100% success rate.

Somewhat tangentially, until now I also see no good reason for it besides ‘eye-candy’.

Besides this I suppose - is that somehow provided by plymouth?
And somehow not provided by anything else?

1 Like

When it happened I investigated the problem und found problems with plymouth and amdgpu. The solution with least afford was to deactivate/uninstall plymouth. And voila, next plymouth release did fine again. But I removed it.
You can criticize my decision, but I would do that again.
Despite I’m able to fix things, I do not like to spent my free time with fixing my daily driver. Thats why I try to change the installation only when absolutely necessary and try to stay on installation standards. With no problems despite this. And all for some more shiny seconds when booting.
While I understand the idea to make manjaro as shiny as possible, I like it as efficient as possible. My reason to use it.
So I dont think that “fix the configuration for plymouth working” is the best approach, when the configuration came with the installation and did work perfectly.
But ok. I removed plymouth and I wont install it again, despite knowing, that its working fine at the moment, if I would.

Nonetheless - a lot of time has been spent to understand why and how the distribution could best accomodate for the widespread use of Plymouth.

A team works together - discuss what to do and how to do it - and it may be difficult to reach a consensus - but you back it - work with it - your own opinion does not always need to be the one - it is true - you may not agree with all decisions - that is how it is on a team - you take the sweets - some times you bite the dust.

I don’t see it as eye candy - it is merely an abstraction - a layer above the messages flowing over the screen.

I believe that someone at some point measured the difference and concluded the measured system actually booted faster when it didn’t have to write all the messages to the screen.

Whether you want the messages or not is a matter of preference.

I remember back when plymouth was dropped around 2017 - at the time it was a hit and miss.

Then the members asked how can I get rid of all that text flowing by - and what should I do about all the warnings I see - not to mention error messages - what about them?

And what about the flickering ? can I remove that ?

Plymouth has the function of creating a flickerfree system start - nothing more.

Yes - it is provided by plymouth - and not something else.

I don’t criticize your decision - no way, buddy - you are the sysadmin - not me - what ever you decide for your system - your decision.

1 Like

If it boots quick enough … the text can be a bottleneck.

Thats why I opt for ‘silent boot’.

It will beat plymouth every time.

I was going to go through quoting and answering more … but all the other bits can be answered by ‘silent boot’.

If you say so. I’m not familiar with that part.
But … we arent saying plymouth is the only way, right?

1 Like

If one wants to input encryption phrase using the native keymap then plymouth is - to my knowledge - the uncomplicated way - with 24.x to have it work OOB.

I reckon it is possible to achieve without plymouth - I just didn’t find out how - yet.

Plymouth is only a drop in the sea with relation to the next snap.

Bear in mind - the team works in spare time - despite the fact that a Linux consulting company exist - most of us are unpaid hobbyists …

1 Like

This is the worst possible time to be discussing this. Only calamares maintainer(s) can respond about the feasibility of changing calamares and they are probably busy getting the new ISOs ready for release

For a Live ISO I prefer to see the text messages so I edit boot options in GRUB
For installed OS I prefer quiet + nosplash in /etc/default/grub
Most other users should be able to do something similar
Some users on here could create a tutorial or BASH script for new users to disable or enable plymouth

IMO it is always a better option for a user to implement their own changes rather than expecting Manjaro Team to sort it out

Manjaro Team should have Plymouth on the main ISOs unless it is a major time-sink to maintain
but not including plymouth on minimal ISOs might be better for users with older or low spec hardware

I reckon it would depend on the actual setup.



Apparently it has been. It took over a year to even gather/accept the idea of kms. And even then it was only really done at the same time it became default in Arch.
It seems it will have to take about the same amount of time to be convincing about the idea that it isnt the whole solution.
And some people have been making a point to mention how much work has gone into this round of updates - in particular in relation to plymouth.
I have no idea why its been such a time sink. But according to the people doing the time it has been.

As to the merit of the discussion. Its about the same as its always been.
But the personal impetus among users to do something about it seems to grow over time.

1 Like