I mean…just because there isn’t an easy way to boot directly off a snapshot doesn’t mean you can’t still the snapshots to restore.
So far I haven’t found anything, and I don’t think anything’s going to happen with it, systemd-boot’s mission seems to be basic and simple. I understand that but I lose too much that I use. That’s why I’m looking into rEFInd, but it seems to be a distant third or fourth in popularity (depending on whether you consider EFISTUB a “bootloader.” So not much of a community comparitively.
Yeah, but you’d need to have some sort of external boot device (or a grub or rEFInd entry kept around as a backup EFI option) to do it, right?
To be fair, I think you need to look at this from their perspective as well.
If you use grub the way Arch ships the package, this issue wouldn’t occur yet some people are trying to place all the blame for this on the Arch team.
The situation is complicated which is why I have been trying to not to assign blame anywhere. It took a lot of things including EOS or the end-user automating grub-mkconfig
for it to happen.
So who’s fault is it?
- It is grub’s fault because they made the change(except they never released it)
- It is Arch’s fault because they pulled those commits into their package(except it works if you use the package as they intended)
- It EOS’s fault because we automated
grub-mkconfig
(except if you don’t automate that the grub menu won’t get updated when things change)
Ultimately, all this highlights is that something has to change and we need to figure out what that something is.
A post was merged into an existing topic: Grub 2:2.06.r322.gd9b4638c5-1 won’t boot and goes straight to the BIOS after update
I didn’t understand this until my laptop got this “grub issue”, in fact, I used to think that arch-chroot
was a very difficult task for a noob like me, but it’s pretty easy and I’ve got my laptop running EOS again in less than 3 minutes, well, almost 10 beause I’ve read instructions before executing anything.
I think “infamous” is a big word for a package that works very well in most cases.
Previously you said that you’re using Linux 20+ years ago. I’m pretty sure you’ve faced a lot of issues since then and I don’t get how you blame a package for generating an issue that is relatively easy to solve. Anyway, I can only recomend you to read this dalto’s post about who to blame for this issue.
I agree, it’s better to know how to deal with issues and learn something new while fixing them than to just run away looking for an alternative that no always be better.
I’m one of those, but I’m trying to learn and I’m on the greatest Linux community to do so!
EOS way it’s much better, because is easy to get an arch based installation and its community is the right place to find answers for any issue.
Sorry if it is a “bad” word! I do not mean any insult to anybody.
Maybe I misused the word… please consider English is not my mother tongue.
I understand the word infamous means famous of something bad.
I agree with you. Grub is almost the standard boot loader … and it is really good. It was just a painful issue…
I will edit my post and remove the word.
Something we can agree on. I personally do not think anyone is at fault for the issue. The “protections” in place to prevent this kind of issue failed. Trying to assign blame is both pointless and counterproductive.
I didn’t think of “infamous” as a bad nor offensive word, but it seems that you’re saying grub is something that everyone sloud stay away from.
My point is that one issue in a long time doesn’t make grub a bad bootloader.
Of course I am not saying that.
It is absolutely up to everybody to use whatever he wants.
Linux is all about freedom.
I can not tell others what to do of course.
I did not say this either. All what I can say it is a bad thing to have. It is still… long story short… I myself I am back to grub after I successfully installed (and loved) systemd-boot because of snapshot support.
Some transparency on this would be great.
If I were to guess, only three options exist for a bootloader: stay on GRUB, use rEFInd or use systemd-boot. Everything else (like Syslinux and LILO) is rather exotic and shouldn’t be seriously considered.
If it were my decision, I’d just get rid of the hook. If anything, this issue has demonstrated how dangerous it is to automate any part of the update process on Arch. Besides, there is no need to regenerate grub.cfg
on every update to package grub
and the user ought to know when something has changed in /etc/default/grub
.
I am not trying to be obtuse, there just isn’t anything to share. We discussed(before this incident) the desire to potentially switch bootloaders and talked through the pros and cons of the various bootloader options available. That was about it. There is no secret plan.
May we know the result of the discussions? Of course you have the right not to answer if it is still ongoing or if you see better not to answer now.
My 2 cents after I tried over the past few days Grub, rEFInd and systemd-boot, the latest is the best for simplicity and boot speed. I just hope it supports snapshots.
Maybe we need a specific thread to vote?
Grub is fine. No need to vote.
OK. Thank you. I am on Grub actually after tried the others.
I was just curious about discussions mentioned by @dalto
Thank you.
He said there is nothing to share at the moment
I updated my second computer today, no problem neither. Two commands sudo pacman -Syu
and sudo grub-install
. To my understanding the previous update messed something up with EFI, but subsequent grub update will not affect it. If it is proven recurring, then perhaps it’s worth have the discussion. But this and all the other grub threads are becoming so long, I might have missed something.
Edit: I understand the feeling of insecurity and willingness to jump ship, but maybe unecessary worry. The intend of grub devs is certainly not to brick our computers, maybe one off mistep by one commit.
This topic here has nothing to do with the discussion about options to grub for the install process… as mentioned already we start to discuss already a long time before this grub issue currently.
The ideal would be to make bootloader optional to choose one… but all that depends on a lot more research testing discussion and a working solution before we can go to change something.
…and nothing related to this topic… here we try to get the info and help about this specific issue that is causing users to be unable to boot into the system.
And help upstream arch and grub development.
Understanding. This is the point.
It is not about Grub is good or bad. It is about people scared of not being able to boot their system.
Again, after I tried rEFInd, and systemd-boot (and loved it) I am back to Grub as systemd-boot does not support booting to snapshots.
And again, I really wish systemd-boot will someday support snapshots. It will be amazing (for me at least).
But for now, it is Grub. Let’s hope things will get better with it.
What I shared was the result of the conversations. There isn’t any more to share. I was just trying to be as transparent as possible with the community.
There is no need for a vote, it isn’t a popularity contest.
Maybe so but the recent ussue has generated a lot of interest in the subject. Might not be a bad thing for the community as a whole to discuss as well, that would give the dev team a much better sense of what the community is looking for regarding features and use cases.