Disabling Snaps in Ubuntu 20.04
kevin-custer.com
kevin-custer.com
Since 16.04, Snaps have been a huge pain for me with running LXC in production environments.
By default, Snap applies updates and restarts Systemd services anytime it likes, and there's no way to turn this behavior off! The only way to get around it is to download the Snap package binary and install that directly. Then Snap won't "know" where to get updates.
(Caveat emptor: "Workarounds" like this can easily lead to a bad security scenario, since any critical security patches won't be installed by any standard system update process)
Did I mention that a fair percentage of the time the Snap updates would leave LXC in a completely broken state? In production (and development, too)!
The final nail in the coffin in this scenario comes in the form of Snap being the official recommended way to install LXC. I don't know if Stéphane and friends even publish Debian packages anymore.
I get the idea behind snap and appreciate it, but the lack of configurability and no clear definition of what stable really even means . . .
Even if people should be updating regularly, forcing them feels completely antithetical to the Linux ethos of users having control over their devices.
Ubuntu users may be more tech proficient than average users but even for developer machines, I think it takes a lot of concern out of the equation if the software is up-to-date.
As far as I'm concerned this doesn't deserve an '/s'.
With Ubuntu I just wait until the next LTS.
Since 2017: https://bugs.launchpad.net/ubuntu/+source/snapd/+bug/1575053
I am on the rolling releases already but I'm not going to install an LTS version that is so deeply integrated with the distributor while at the same time ignores the directory standards so incredible blatantly.
I'll also admit that I can't for the life of me get Arch to properly install with all the tools and features that are normally built-in to desktop environments to work on first boot. I'm sure I can make it work if I give it another try now that I've gotten more used to using Manjaro, but the Arch installation experience is not something I want to go through again any time soon.
I understand that most Arch users will want to decide exactly how they set up their system and all, but for my personal (non-work) machine I just want an operating system that works, allows me to mess with the standard Linux stuff and manages to install itself without me holding its hand. I'd much rather tick a box that says "enable full disk encryption" than manually configure cryptsetup and LUKS parameters. It's just too much effort for what I get out of it.
I think the (not that bad) manual installation process might keep some folks from installing it, but it makes you learn to choose from the very beginning.
After the initial install, update it on your schedule, and install things as you decide they are needed.
Ubuntu is the opposite -- install the world, then go madly trying to disable and remove the phone-home, forced-update, bloat.
# ps
I feel bad for the maintainers; it must be difficult to deal with all the rude and borderline disrespectful comments -- and particularly for something you're donating your time for free after all. But I must say that the architectural decision of creating a ~/snap was a colossal mistake.
It's not an argument. It's freeness is a completely irrelevant property.
I really have no idea what people are thinking sometimes.
The automatic startup of newly installed services (before you even have the opportunity to configure them) on Debian (and derivatives) has always bothered me and is a (small) factor in my decision.
Fortunately, there are a few "workarounds" to prevent newly installed services from being automatically started!
First, there's the brute-force / heavy-handed approach -- override the systemd presets [0] to set all services to disabled by default:
$ cat /etc/systemd/system-preset/00-disable-all-services.preset
disable *
$ ln -s /dev/null /etc/systemd/system/nginx.service
$ apt install nginx
# ... configure nginx as desired ...
$ systemctl enable nginx
$ systemctl start nginx
(I'd recommend avoiding the latter entirely and going with whichever of the first two options better suits your needs.)
---
EDIT: Just remembered another method but I'm not sure if it still works:
$ RUNLEVEL=1 apt install nginx
[0]: https://www.freedesktop.org/software/systemd/man/systemd.pre...
[1]: https://people.debian.org/~hmh/invokerc.d-policyrc.d-specifi...
A cultural acceptance of WIP software seems common, in free software communities, in my experience.
On the one hand, yeah it's harder to complain when something is only built in someone's spare time for free. On the other hand... is non-FOSS really any better? Has Windows yet managed to re-merge the old and new control panels?
The first is a fact of life. The second is bad judgement.
Due to LXC now being a snap, the file is simply not there. I guess it's 5 layers deep in namespaces, overlayfs and other stuff. I was so fed up with this that I removed Ubuntu (and replaced it with Gentoo).
I can't understand why they would break such fundamental thing and seemingly don't care about it.
I don't know how must be installed on your version, but on 19.10, snap can be completely removed: the only draw back is that pulseaudio depends on it, but I guess that for a server it's not that much of a problem.
Dear snap developers, don’t make false assumptions that every user in Linux has to have home directory as /home/username just because standard Ubuntu installer does this. Only thing that you should care about here is $HOME
I uninstalled it on all my Ubuntu systems and put it on hold. No new Ubuntu installs, switched to Debian.
I think I am going to have to start figuring out a plan to migrate them all over to RHEL...
1. prevent machines in the fleet from pulling the broken LXD update
2. rollback broken machines to the previously working LXD version on the same channel, since it no longer existed in the Snap Store™.
What a joke! Now we're burnt on snap _and_ LXD.
Some developers are trying to package for Debian. The work is in progress at https://wiki.debian.org/LXD
The following distributions package LXD (that I know of):
* Void Linux
* Alpine Linux
* Arch Linux
* Gentoo
Some of those are more suitable for production installs than others, but if you know what you are doing and manage your deployments well, all of them could work.
I've been looking into .deb packaging for Caddy but it really feels like they require us to jump through too many hoops to make it happen. I'd much rather just ship a prebuilt binary.
Security. Dynamically linking stuff is always going to be better than statically linking. Do you trust upstream to keep track of security issues and rebuild in the absurd dependency tree that is Golang software? Or the multi-year old effort which is the current Debian security team?
The reason why it's absurdly complicated is solely on the Golang team, not Debian.
It's clearly not production stuff, but it works very well for a quick test bed before production deployments.
(Same-but-different for enforcing non-interactive mode on installs.)
The idea with LTS is to be predictable and stable.
Put another way:
- desktop people want ALL UPDATES NOW.
- server people want NOTHING TO CHANGE EVER.
(yes, not exactly that, but kind of)
I actually really like snaps and that's the only bad experience I've had. I don't know why LXC would come in that form, very weird.
It is a seriously difficult thing to get right. I always applaud people to take on difficult problems and try to solve them, but I would not expect it to have any sort of robustness for another 5 years at least.
For context: https://forum.snapcraft.io/t/disabling-automatic-refresh-for...
Snap developer responses are hilarious. No matter what your use case is, Snap developers know better than you, you silly irresponsible sysadmin/user. Snap is basically just another App store.
Snaps break debain's stable release model. They allow upstream to ship updates outside of the normal 6 month ubuntu releases. There are times when you might want this, but it should be opt in not mandatory. I thinking specifically of lxd which is only shipped via snaps.
The snap store's trust model is confusing. Its hard to tell who is making the packages and how they are sandboxed. If I'm going to install a proprietary piece of software I want to know exactly what it can and can't do. Lately I've been using firejail when I need to run things like this.
And now for a minor complaint that also feels most user hostile to me: why do the snap developers think its ok to require a non hidden directory in $HOME? Seriously my home directory is MINE, if you have to store application state there at least have the decency to do it in a hidden directory.
Applications need a place to save state, and you don't really want it to be hidden, because sometimes I as a user do need to change it manually for any number of obscure reasons. So you make a subfolder that's designated for apps to write to.
There's no good way to put this into the existing Linux model though. All you'd do is add yet another standard a la https://xkcd.com/927/
Especially when a . is used to make the directory hidden, because I can't unhide that specific directory without changing the name apps expect.
I get into the Library folder by doing: cmd-shift-g -> ~/Library -> enter
In Catalina, it's now Cmd-shift-fn-.
$ ln -s .hidden not_hidden
It's that easy. Really.
I'm being picky, I know! But I still think the Library approach is cleaner. (Again, back when it wasn't hidden by default!)
This model is terrible. OS asks up-front: program wants permissions to do A, B, C, X, Y, Z: grant/refuse? You the user decide that the program should not be allowed to do X, so you refuse. Now the program will not run at all. That's about the worst design possible. A 10 year old can brainstorm a better design within a few minutes.
Investigate prior art:
man 7 apparmor
man 4 capsicum
man 2 pledgeIn practice, apps that want access to my contacts usually keep asking for access every time I try to do anything, until I eventually either relent to make the prompts go away or click allow by accident.
And that's me as a computer-enthusiast. I would bet money most normal people just hit allow always. Because it's easier.
Some apps even try to circumvent this system by showing a "help" screen with instructions on how to re-enable the permission manually, but I only saw this twice.
Still, this is not practical enough. A true user-centric strategy would be to offer "mock-permissions" to an app, so that if an app says that it needs to read your home dir, you grant mock permission to the app, and it sees an empty home dir, not yours. From the point of view of the app, it should be impossible to know if it has been granted the real permission or just a mock permission.
People still just click "allow" on everything. They get tired of getting questions asked, and just want the program to work, so they don't even read anything and just tap tap tap until it lets them through.
Access your files? Sure. Access your documents? Whatever. Send data off to our corporate data vacuum? Whatever, I need to see what my face will look like with an AR moustache!
Also, are you running Catan as root or something? How is it able to compromise your whole system?
After fighting with node, vlc and ffmpeg snaps not working reliably, I had to add a warning for my PhotoStructure users on Ubuntu to avoid those packages (and how to install them via apt).
I agree with a bunch of people here: snap promises some great features, but in practice the horrible performance, additional resource consumption, and spotty reliability led to me actively avoiding snap packages and looking for an alternative installation.
4 years old now.
https://bugs.launchpad.net/ubuntu/+source/snapd/+bug/1575053
https://docs.microsoft.com/en-us/archive/blogs/patricka/wher...
https://docs.microsoft.com/en-us/windows/win32/shell/csidl
.NET has its own set which is also very old and goes back to .NET 1.0 (as old as Win98) https://docs.microsoft.com/en-us/dotnet/api/system.environme...
I have still been using `snap` for the past year, and have mostly been happy with it. It works well for apps like `spotify` and `zotero`, it's a better solution than `ppa` for third-party software, and I like that it tries to introduce a permission model. But not following the dotfile convention is annoying, and not prioritizing a search-and-replace of `~/snap` to e.g. `~/.snap` in the code after four years is strange when this is their most upvoted bug.
The actual heritage is when `ls` was written to exclude .. (dot dot) directories from it's listing the code literally just had the logic to exclude any files which had a dot/period as its first character. Later operators exploited that bug/feature to create items with a dot prefix to make those files and folders hidden. Eventually this became convention.
Whereas systems released more recently understood the need to have a hidden file attribute so made it a first class property.
I do love POSIX and often the first to defend some of Linux or BSDs idiosyncrasies but if there was ever anything that was worth someone saying "I know we've always done it this way but it's shit and it's about time we implemented a proper solution", it would be the dot prefix kludge. (I also know it's easier said than done and would take years for all the tooling to catch up)
Of course you do have system files, which are hidden from all standard folder listings, but aren't usually used by software, presumably because they're too damn difficult to find again (although I really hope google drive realizes at some point that desktop.ini is supposed to be a system file, not merely hidden).
(That said, I would still prefer that they not be.)
In fairness, I think that's the point, just like RHEL streams.
> There are times when you might want this, but it should be opt in not mandatory. I thinking specifically of lxd which is only shipped via snaps.
Yeah, decent idea, poor execution. Especially for system/infrastructure software like LXD. I want a stable kernel, a stable glibc, a stable LXD/Docker/k8s/whatever, and then on top, bleeding-edge applications that will not break the world if they autoupdate at random to a less-tested version.
The latest snap I had to get rid of was Visual Studio Code, because I was trying to work on an open source game with it, and I found out that if I launched the game from inside Visual Studio Code, my game wouldn't play sounds because it couldn't communicate with PulseAudio, and attempting to use ALSA just straight up gave me an error.
On the other hand, I've only had positive experiences with AppImage. Gives you an all-in-one image that you can directly execute if you like, and no sandboxing nonsense.
I expect you needed to do something like that for the VS Code snap.
https://firejail.wordpress.com/documentation-2/appimage-supp...
> Make each program do one thing well. To do a new job, build afresh rather than complicate old programs by adding new "features".
Both AppImage and Firejail do one thing and do it well, you can easily combine both and get what you want.
I realize I have to soon upgrade from Ubuntu 16.04, so a while ago I installed 19.10 to see what to roughly expect from 20.04, as there are quite a few changes. Trying things out, I installed VS Code from the Ubuntu Store; I was quite disappointed to see things going south right on the Welcome page, as the snap version of vscode couldn't even open a web browser (instead, Firefox promptly crashed every time).
Of course, being able to open a web browser from VS Code isn't a necessity, but for some reason for me it just seemed like the install was rather broken by default, and surely almost no one could have used it much not to notice a simple thing like that right away – and the thought that even if the problem was noticed and simply shrugged at didn't really make me want to see what else did not work.
I had run it via X11Forwarding (I had a contusion and used a puny laptop to connect to my desktop) and it was not a smooth ride, so it's not an entirely representative experience, but it shows that it's not all painless, yet.
The snap thing is a pain in the ass. I understand the need for something like snap or flatpak. I had software too new or too old that wouldn't work because some dependencies were not updated or were too new. Snaps can solve that by allowing the developers to provide everything you need (or everything you need that is not on your system). But why would I want a snap calculator or a snap system monitor? On 19.04 it would take a couple seconds to open the calculator...thank god they reverted those apps as normal packages.
Now I feel like I felt on Windows when I had to be extra careful installing software in case somethign weird came in the installer. What kind of package is this? Is it a snap? Can I install the normal package? Is the snap provided by someone trustworthy?
I also had to install Unity. Gnome lacks support for multiple monitors. Some stuff like the dash working only on the main monitor breaks completely my workflow. Almost for every action I want to make I have to change my focus to the main monitor.
Yes, applications remember which monitor and even location on the monitor they were last at.
I'll probably give Budgie a go.
On Unity it opens on the monitor you're focusing.
Sometimes it's a bit tricky. Let's say you're focusing on a Window on your main monitor but your mouse cursor is in the secondary monitor. If you hit the meta key, the menu will show on the secondary monitor. Now, if you open an application without using your mouse it'll open on your main monitor, where you had the focused window but if you click on the app it'll open on the secondary monitor because the click changes the focus.
Some stuff is not as polished as it should be but if we take in consideration that Unity has been abandoned for 4 years and before that it wasn't developed at all as all the effort was put into Unity8, I'd say it's a pretty damn good DE.
I hated it when it came out but now I can't live without it.
To be fair, my Thinkpad from work is the first laptop that has issues with when (dis)connecting a screen. When disconnecting I need to turn on the laptop screen before disconnecting; for connecting I can usually log in without screen and use Super+P to switch modes and that fixes it.
But other than this first instance, I never heard of this being an issue and people here talk about it as if it's commonplace. I must have used or seen Linux used on at least 30 laptops with 2 (or more) screens without issue in the last ~10 years, counting the ones at my previous employer as one because they were all the same hardware even if people ran different distributions.
/var/snap is a subvolume. Purging snapd wants to remove the /var/snap directory, but it being mounted subvolume, it will fail. Purging snapd will therefore also fail.
Destroy (-r) the /var/snap subvolume before apt purge snapd.
Similarly, if using flatpak, create a new subvolume for /var/lib/flatpak before installing the first one. You don't need to snapshot your flatpaks together with the /.
https://github.com/cies/kubuntu-setup#remove-snap
This is needed because an every growing number op packages is "dependant" on it.
Here I show how to install Chromium as a DEB package from Debian (on a buntu):
I despise the bug, it causes so much trouble without a way to turn it off. Plus the switch to snap happened exactly during a critical vuln which meant people were running on critically vulnerable chromium for weeks.
I'm not against stuff changing, but tell us up front, explain what's changed and make sure that you're not breaking existing things or provide a way for people to keep their existing stuff work, a flag that says that following symbolic links is OK ... and put that UI in Chromium itself, not in some other box somewhere
A workound is posted at https://bugs.launchpad.net/snapd/+bug/1776873/comments/29 (simple patch & recompile). May be worth making a PPA for the fixed snapd if we can't get the Canonical dev team to fix the core issue.
It's extremely weird to see this kind of thing when it happens too, you usually trust Firefox to work!
I prefer the isolation of snaps so I’m willing to put a little work into it, but understand not everyone cares to do that.
It also causes data loss: https://bugs.launchpad.net/snapd/+bug/1616650
I have been using Xubuntu 20.04 for several weeks and it has worked well. Because that's a high end laptop I have not bothered to uninstall snap yet, although that has been on my mind.
I am surprised that the article writes there are snaps installed by default. Checking in Xubuntu 20.04 that's luckily not the case
$ snap list
No snaps are installed yet. Try 'snap install hello-world'.
$ apt show chromium-browser
Package: chromium-browser
Version: 80.0.3987.163-0ubuntu1
...
Pre-Depends: debconf, snapd
...
Description: Transitional package - chromium-browser -> chromium snap
This is a transitional dummy package. It can safely be removed.
.
chromium-browser is now replaced by the chromium snap.
Because hearing things like `sudo apt install chromium` actually aliases to using snap is disconcerting to say the least if true.
But I'd like to throw in a recommendation for Fedora. It includes Flatpak by default (which to me as always been the least objectionable sandboxing system) but nothing comes as a flatpak by default. You have to explicitly choose to use it.
Fedora's hardware support is second to none, and even supports in-OS BIOS updates.
I highly recommend it.
However, if you want to use chromium-browser you are out of luck. It is no longer available as a .deb package.
What is "performant" supposed to mean here?
Is Flatpak faster than Snap? more compact? simpler? more reliable/secure? easier to use? more efficient in terms of cpu/memory/communication/power/etc.? all of the above?
I really hate Snap/Flatpak conceptually. Use the package manager. That's what it's there for. FPM is a way to make package building easier.
Snap/Flat tools feel like they're about the same as Electron cancer; and I bet we'll see more closed source commercial stuff pushed to the Linux world via them as well.
By the way, what is it with Docker that makes it hard or impossible for it to be used for this exact purpose?
It doesn’t “suck,” and many of the ideas are compelling. But, the design and implementations could be better. I was interested in rkt until it was acqui-abandoned.
Snap suffers from the overhead of having to decompress a squashfs on program launch: https://forum.snapcraft.io/t/squashfs-is-a-terrible-storage-...
My experimetsn on the same laptop showed that..
Launching Skype installed from deb is almost instant.
From snap it takes around 10sec to get anything to show.
This is on quite fast nvme disk.
Yes, Flatpak is just bind mounts and namespaces. It doesn't have the overhead of the squashfs images.
It's also disingenuous to suggest a browser wrapped in a cryptocurrency promotional wallet thing as a comparable alternative.
Sure, but some people just want to have their problem fixed and they only have a sledgehammer in front of them. If it works, is it really that bad?
> a browser wrapped in a cryptocurrency promotional wallet thing
I think you have it the wrong way around. It's a browser containing an integration with a cryptocurrency wallet, not the other way around. You'd still be able to use it as a normal browser if you ignore the cryptocurrency stuff.
So shortly after the Mir fiasco, why does Canonical feel it necessary to come up with the next dumb idea?
They give the suggestion to remove via
sudo rm -rf /var/cache/snapd/ && sudo apt autoremove --purge snapd && rm -fr ~/snap
sudo bash -c "cat > /etc/apt/preferences.d/no-snapd.pref" << EOL
Package: snapd
Pin: origin ""
Pin-Priority: -1
EOLOther than that, I do not think other packages are affected.
My only con is that the defaults update all snaps like every day, and I really would like to have better control on that, because I'm always on mobile data.
No startup slowness, and you'll still get the vendor-provided updates as they're released.
tl;dr -> HashiCorp's various tools exist as snaps but none are published by HashiCorp. All are out of date. Some have incorrect metadata. Few provide any clue as to who or where the upstream is. There's usually not even a way to contact the snap author to submit patches or ask for an upstream link. eg https://snapcraft.io/nomad
There is a level of care here which I think is great. Some engineer somewhere made sure that the system would still work without snaps. This is a very Debian attitude which Ubuntu inherits from and which I would like to celebrate for a bit :-)
Only Canonical can run a store. Once you install something, Canonical decides, when packages on your computers will upgrade. Wrt control, it goes way beyond Apple store/Google Play/Microsoft Store.
No thanks. Flatpak can preserve the benefits of decentralized repositories and upgrade under user control, with all the ease of packaging and sandboxing of the new toys.
It seems like they also often end up going back to the community solutions/standards eventually, so what benefit do they see in doing this over and over again?
The sandboxing part was spun off to the bubblewrap project, which has been adopted by the community for other purposes, such as Gnome's Epiphany browser.
https://github.com/containers/bubblewrap
https://blogs.gnome.org/mcatanzaro/2020/03/11/epiphany-3-36-...
SnapCraft on the other hand is a project that requires copyright assignment to Canonical Ltd for contributions, which presumably has the intended effects of keeping non-Canonical contributions close to zero.
https://github.com/snapcore/snapd/blob/master/CONTRIBUTING.m...
https://blogs.gnome.org/alexl/2018/06/20/flatpak-a-history/
AppImage only solves the easy part of the generic-linux-app problem, namely running the packaged appimage; it does nothing to solve the hard parts: how do you build the application so that it will actually run on generic Linux distributions, and sandboxing.
> If you’re a Fedora user and you want to install Spotify, you’re told to go to https://snapcraft.io/spotify. Spotify doesn’t distribute RPM packages, appimage, Flatpak or anything useful to a Fedora user who wants to download it, or to a Fedora maintainer who wants to add it to a repository. Fedora users are told to go to what is essentially a commercial store operated by a RedHat competitor where stats tell them their distribution is only 7th best.
> We’re in luck, we can still download the .deb. If Spotify stops caring, what do we do? We move to snap because we have to? Will the snap store continue to let people download actual .snap files in the future or will that get locked down ? Will the snap store continue to operate without an Ubuntu One account or will we get vendor-locked ?
Canonical provides the tools and several companies are providing their software as snap packages.
Another example is JetBrains that provide all their development environments as snap packages.
It seems that canonical does best when they embrace the ecosystem and not try to replace it.
Canonical cannot be accused of NIH for Upstart, it was a widely adopted clean init system at the time and made many distro's boot times great again. Service configuration files were also very easy and clean.
Unity was a desktop environment that both was liked by many people and reused many components from Gnome, I don't really see a problem here.
A point could be made for Mir.
So it's wrong to say Canonical is guilty of NIH-syndrome with these two items, when their versions were out first.
Whether the tech is any good is another matter.
https://blogs.gnome.org/alexl/2018/06/20/flatpak-a-history/
According to https://en.wikipedia.org/wiki/Snappy_(package_manager), Snap was publicly revealed on 9 December 2014, whereas the design of xdg-app (FlatPak's old name) was already discussed publicly at GUADEC 2013.
https://www.superlectures.com/guadec2013/sandboxed-applicati...
While I don't like the tech, the real reason Snap pushes me away from Ubuntu is the fear of this future where the app store dream turned out to be a failure again. What happens to the distro then?
Along with the native package "bundles" (e.g. dev-tools, ruby-basic, containers-virt, etc.) that include built-in dependencies, Flatpak/FlatHub apps seem to work better than Snap/Snapcraft in my experience.
I wonder what is about to happen to the desktop experience for Clear Linux.
People like me who would like a system that works their way, even if it's not the same way as the distribution maintainers - nope. Believe it or not, I don't want my desktop to be full of animations, I want updates to happen on my schedule, and I care about having the active title bar a different colour to the rest.
People who believe in Free As In Freedom - it's still technically free, but for all practical purposes it might as well not be if you have to recompile a core package for things like using a non-standard home directory path. This is one of many things where I sigh and go Freedom wasnt meant to be like this.
The average non-techy user - uses windows anyway, and has access to a much larger software ecosystem as a result.
Not fun to think that could have happened on a production env if people didn't disable snap.
That does not forbid others to package LXD independently. Debian has been close to packaging LXD at https://wiki.debian.org/LXD It is a matter of picking up interest to complete the work.
It's not much of a problem for me since I believe that keeping diversity of Web browser is necessary for the healthy Web ecosystem and I force myself to use Firefox. And I really hate Chromium force me to waste CPU time on some kind of test payload occasionaly, it use 100% of a core and there is no way to disable it.
I'm fine with restricting myself to use Firefox since I'm not a Web frontend developer so I don't need to test Chromium behaviours myself. But others may not.
2. Is the whole packaging a pain? Just do the work once and create the snapcraft.yaml packaging file that describes the whole process. You can even use the snap build server to rebuild fresh snap packages (in six architectures) as soon as you make a commit to your repository.
3. Are you a software company that wants to distribute your software? You can do all the snap building in-house.
4. The same snap works in many many distributions.
Flatpak exists and works. The application I'm working on is distributed on Flathub, and I don't have the time nor the interest to learn yet another packaging system.
Ubuntu would do well trying to not just lead (which they do in some aspects, I'm sure) but also try to follow when it's more appropriate. The push towards Unity and now Snap just causes more fragmentation in the Linux community.
http://manpages.ubuntu.com/manpages/trusty/man8/update-alter...
https://www.reddit.com/r/linuxquestions/comments/g7cv6k/debs...
The deb tries to install the Chromium snap package ;)
Inspired by this post I checked what is actually offered as a snap.
This one surprised me: https://snapcraft.io/aws-kernel
Who can explain me how I would use a kernel running as a snap?
Well, last updated 2017, so I don't think I want to try that one.
Just tried to install 20.04 on LVM. It gets in to a completely broken state if you misconfigure the network due to an error in an exception handler in an Python error handling view in the UI. Doesn't exactly inspire confidence.
a) is it possible to use Chromium without snaps?
b) Can anybody please describe the general experience with Centos for a media production machine - needs to run NVIDIA GPU support, Ardour, low latency jackd and Davinci Resolve on XFCE.
I think most users will be fine. Those of us who need more fine grained control, can have it.
Annoying but harmless: it's start time has multiplied.
Workflow-breaking: Chromium now can't see my NFS mounts anymore.
It's not just about having a pretty home directory.
It's distributed as AppImage for Linux. Very much recommended.
In this month only I had to give up using firefox because uglyBar and ubuntu because of snaps.
There won't be anywhere to run soon.
sudo apt install gnome-software-plugin-flatpakufw < firewall-config
===============
I like & prefer Ubuntu. But where else is Fedora/CentOS/other distros ahead of Ubuntu?
I strongly recommend anyone similarly frustrated to check out debian, which is a fantastic distro. Thanks to Kevin for posting this, but if you're using Ubuntu and disabling snap, you're fighting against the current and I have to imagine it's going to be increasingly difficult with subsequent releases.
That’s a bit rich: are they not the #1 consumer distro, which hardly implies they are failing to execute. A successful product has missteps, so what.
> I don't want to run that garbage on my desktop.
So don’t. Why complain that others do? I use Ubuntu because it works and I can mostly find information about how to do what I want. There are major aspects of Ubuntu I don’t like (Gnome, Snap) but selecting a distro is all about choosing your compromises. I have tried Debian and other distros, but I tend to go back to Ubuntu because it works best for me.
All the hard work to make it a viable OS is done by Debian. Canonical just adds some polish and then wrecks it all with poor design decisions over and over again.
Just like windows Shuttleworth admires and seeks emulate so very much.
It's called a network externality. Microsoft achieved this with piracy based market penetration of dos and it worked great for them and they've built on it ever since.
Shuttleworth understood the importance of getting established as being popular when he launched and would press install disks, as many as you asked for, and ship them to you at his own cost. As one example of nakedly going after market share and spending resource to do so.
Bug #1 in the ubu bug tracker is literally "windows is the most popular os."
Separate to the marketing, which is worth discussing on this site because some of us actually care about what works and why so wish to discuss it, let's talk engineering decisions.
Between 2 choices that are technically about equal. Choose the one that is more popular. Many feet trample more bugs. Better support. More likely to be around after $time_period. If it needs to work with something else, the managers of the something else project will likely suppor the more popular first etc. Obviously popularity is not the only concern but it really does count for something, ignore it as a dimension in your decision process at your peril. I use linux, and ubuntu as it happens on this laptop. I'm aware windows and osx are more popular and that popularity makes certain things easier. For /my/ purposes and to /my/ taste linux and ubuntu are worth paying that cost to have installed here and I'm very comfortable with that decision.
Just quietly, perhaps people who don't care for business decisions and engineering decisions are on the wrong website? There's plenty of places "boosters" can go to do that.
There was zero documentation on the change and the error you received attempting to install an extension was basically "operation failed". I discovered the cause only because there was an open bug about it on the Debian bug tracker where the maintainer refused to acknowledge the problem. Eventually, sane minds prevailed and that stupid patch was reverted.
So, unfortunately - you'll end up having to deal with people that refuse to look at things from the user's perspective no matter what distro you use.
And yes - I'm still bitter. :-/
That said, I think Debian's occasional messups are far less egregious and damaging than Ubuntu's though.
LTS on Ubuntu was always better than Debian, I mean saying that Ubuntu is just a repackage of Debian is very short sighted, especially on the security side, Canonical security team is top notch.
They need to, for a long time, Ubuntu has been shipped with EOL Kernel versions https://ubuntu.com/kernel/lifecycle
But what benefits Red Hat in the enterprise world is to their detriment in the consumer world. There's a reason the Debian/Ubuntu package ecosystem is richer and more featureful than RPM, and this is why Ubuntu dominates in the container space--because almost any piece of software that one could expect to have been packaged has been packaged as a .deb and already exists in the default package archives. I can't count the number of times I couldn't find an RPM--certainly not in the default repositories (RHEL, CentOS, or even Fedora), but not even in the third-party community repositories. And those that do exist are of lesser quality than the comparable .deb, for various reasons. (That is, the long-tail of packages is of higher quality for Debian.)
By pushing Snap, Canonical is definitely going astray. Ubuntu's competitive advantage is the Debian package ecosystem. Both Canonical and Red Hat seem to underestimate the role and importance of their respective packaging ecosystems. How many projects to revolutionize or replace RPM/Yum/whatever at Red Hat have crashed and burned? Many, though it's hard to count because half-way through they often realize what they're trying to do is functionally or even technically impossible (as with their aborted 2017 plans for RPM package streams), and scale things back to iterative improvements.
Containers are a security nightmare, and pretty much the only reason to pay Canonical and Red Hat licensing fees is for security and bug fix maintenance of their package archives. On our large Kubernetes clusters at work there are thousands of open CVEs for the containers that are being run, and we'll have to boil the oceans to get them all updated, let alone keep them updated. But updating packages is as simple as an apt-get/yum upgrade[1], and rarely do you have to worry about anything breaking, especially relative to the pain that updating containers regularly brings.
[1] If the container uses Ubuntu, Red Hat, etc you can sometimes just rebuild the container to get the newer packages. But that assumes you control the container image. Most containers come from third-party, decentralized sources (that's the point!). But Docker Hub doesn't cajole and coordinate container owners to update their crappy images. It's no substitute for the orchestration of people that are traditional package repositories.
So, LTS/stable.
I'd expect most people tech-savvy enough to install Ubuntu would also have a decent enough internet to download a ~700mb file.
This was also around the time I would often brick my primary (only) workstation for whatever reason. Having a properly mastered Live CD was super useful.
I would order at least two with every release cycle for a few years at least.
Thankfully, I saw the light early with Ubuntu-server, and stayed with Debian. Ubuntu-desktop makes for a good enough live / recovery / troubleshooting environment, but not sure I’d use it for anything more.
we installed ubuntu with friends in freakin junior high school. it's not like it's rocket science...
OTOH the village where I grew up as a child only got DSL > 512k around 2008 iirc.
So yeah, lot of people got those.
You could equally say that all the hard work to make a viable OS is done by Linux, so screw Debian? Last I hurd, the GNU developed OS is unviable (no 64 bit, no SMP).
Or equally say that all the hard work to make the majority of end-user programs (you know, the raison d’être for an OS) is done by other open source projects, not Debian, so screw Debian...
These are open source projects, with cross-pollination everywhere, each with their own opinions on licensing. Ubuntu mostly helps the ecosystem, and certainly isn’t a parasitic player (although like all, they are not perfect).
Why bag on Ubuntu just because it happens to be popular? Should we also cancel all the other Debian based distros?
PS: complaining about upstart shows you are just being biased (or perhaps misinformed). Canonical were developing upstart before systemd was developed - and systemd was developed by RedHat. The main con given against upstart was not technical, but due to licensing. “In terms of overall feature[s] there is really rather little to distinguish upstart from systemd” https://wiki.debian.org/Debate/initsystem/upstart
No, you couldn't say that. Without toolchains, userlands, and packaging, a kernel is pretty worthless. The barest bones you can go is still gcc, linux, uclibc, and busybox. There is more code that goes into a computer running linux, then there is in the linux kernel. By a wide margin.
FSF does fabulous work, which we are all appreciative of, but some decisions are peeing in the open source pool.
I think RMS creates unnecessary division against Linux and Linus for what I feel are poor reasons. I went to a lecture by him where he spent half his time being negative towards Linux and Linus (that felt like he was just pissed off because Linux was popular) and a bit because Linus had used the GPL2 (not trivial to change, and you don’t get change by attack). Being negative towards the people who are on your own side is wrong IMHO. It could equally be argued that Debian should be called Debian/Linux. Edit: I just found a quote from Linus about RMS that summarises what I wished to say here: “It's not passion for something, it becomes passion against something else.“ - http://torvalds-family.blogspot.com/2008/11/black-and-white....
PS: I totally admire RMS and his relentless idealism. He has given so much to the world, and the faults I see in him are interwoven with the strengths I see: I’m not sure the faults could be mitigated without badly weakening the virtues.
GNOME history: https://unix.stackexchange.com/questions/141114/what-is-the-...
To return to topic: “How to install Flatpak apps on Ubuntu 20.04 LTS”: https://jatan.blog/2020/04/25/how-to-install-flatpak-apps-on...
FWIW the whole GNU/Linux pedantry bothers me too.
Android went from non GNU absolutism to GNU LGPL for its standard java library a few years ago:
https://arstechnica.com/tech-policy/2016/01/android-n-switch...
> That’s a bit rich: are they not the #1 consumer distro, which hardly implies they are failing to execute.
No, that's Debian doing almost all the work.
90% of the packages in Ubuntu are simply taken from Debian without significant modification.
Snaps may be a pain sometimes, but Gnome seems to be working like a charm...
I wouldn't use any other DE, at this point.KDE has always been cluttered and XFCE is buggy and not particularly intuitive.
Nothing as big as a DE is perfect, thankfully it's broadly moving in the right direction.
Like a Juju one? :)
People complain like this because they have no real control of their own lives. It makes them feel smart, if only they were in control, then things would be better. It would be so easy, the people in charge must be stupid. It comes from a lack of experience and the inability to understand the challenges in those positions.
The shortcomings of snap are well documented. I assumed anyone reading the comments would have been aware of them. But apparently some people will jump on any chance to virtue-signal.
I'm seriously considering not taking the 20.04 LTS release and either using 18.04 until it's untenable or switching to something else.
I think this is ok though. If ubuntu were better thought out, the linux ecosystem might be less vibrant than it is. I would like to think other distros are learning from these failures. Personally ive had excellent luck with scientific linux (CentOs based).
But have you been following last year's improvements to GNOME's performance and responsiveness? A lot of it is Canonical's devs bringing their experience from Unity.
20.04 should be better still (I've yet to upgrade).
And Void Linux/ i3-gaps before that.
And Linux BBQ / OpenBox before that.
there is no need for latency-hiding animations and subsequently trying to make them run smoothly on the gpu if there's no perceptible latency.
I hear a lot of praise for Unity and I'm the kind of person who enjoy trying out new stuff and Linux Desktops is no exception.
For me, Unity was broken because of alt-tab (behavior and lack of configurability).
It might work for everyone else but when I want to switch back to the last or second last thing I worked with I want that done now.
I don't want to look at the tab switcher to ponder what to do next, just alt-tab, done.
This has worked consistently in every Windows since at least 3.1 (the first my family owned), and in every Linux desktop environment I've used except Unity and Gnome 3. And in Gnome 3 it was at least configurable.
This might seem trivial to a lot of you but to keep focus I keep one application maximized most of the time. I don't use them side by side. Then when I need to reference something (Jira, vendor documentation etc) I alt-tab. Same goes for slack.
It is true that the Debian people are doing a great job.
> [...] if you're using Ubuntu and disabling snap, you're fighting against the current and I have to imagine it's going to be increasingly difficult with subsequent releases.
Actually, snap was harder to remove in the previous release: you had to rebuild certain packages (actually, just pulseaudio, so it only matters for desktops) to get rid of the dependency, but it seems now that it's just a couple of apt commands, so you have to give Canonical credits for making it easier.
That being said, it's not as bright as I thought it was - see my other comment.
python3-ubuntu-image
xubuntu-desktop
xubuntu-core
vanilla-gnome-desktop
ubuntustudio-desktop-core
ubuntustudio-desktop
ubuntukylin-desktop
ubuntu-unity-desktop
ubuntu-snappy-cli
ubuntu-snappy
ubuntu-mate-desktop
ubuntu-mate-core
ubuntu-core-launcher
ubuntu-budgie-desktop
snapd-xdg-open
snapcraft
snap-confine
qml-module-snapd
plasma-discover-backend-snap
lxd
lubuntu-desktop
libsnapd-qt1
kubuntu-desktop
ember
cyphesis-cpp
chromium-browser
ubuntu-server
ubuntu-desktop-minimal
ubuntu-desktop
ubuntu-core-snapd-units
livecd-rootfs
maas
apparmor
libsnapd-glib1
gnome-software-plugin-snap
command-not-found
sigh...
Edit: Please note that many of these are "leaf" packages, by which I mean that no other packages depend on them.
Not at my computer now, but that cannot be correct. Could recommended packages be included in that list?
I did apt-purge snapd on my xubuntu 20.04 and it did not pull out anything unrelated.
When you install something later an it tries to pull in something later you can always try --no-install-recommends.
In the worst case there is eqivs https://wiki.debian.org/Packaging/HackingDependencies but it has been 3+ years that I needed to use that the last time.
I tossed Gentoo and KDE (this is not a Gentoo endorsement, it was just a "hey I wonder what Gentoo's been up to in the last dozen years since I last used it") on a spare laptop. It turns out that KDE is amazing now. It's seriously the best DE I've ever used, and I'm a Mac user! (Half of the utilities I install out of the box on a fresh macOS are built in, and the annoying stuff that used to be editing arcane files is now easy preference settings. It's actually great.)
What the hell are Ubuntu doing shipping Gnome (with the ugliest custom theme known to man, to boot)? Admittedly it was my own ignorance, for which they are not responsible, but their mindshare and bad choice tainted my whole view of the state of the art for a long time.
>[...] To show our commitment to this dual licensing model, the KDE Free Qt Foundation was founded in 1998. It is a non-profit foundation ensuring that Qt will always be available under open source licensing terms. The agreement that governs this foundation has stayed mainly unchanged over the last 17 years. As a lot of things have changed during these years, we have been working with KDE over the last year to create a new and updated agreement that takes todays realities better into account.[...]
From: https://www.qt.io/blog/2016/01/13/new-agreement-with-the-kde...
The updated contract mentioned in that blog post is available here: https://kde.org/community/whatiskde/Software_License_Agreeme...
That's been the case with KDE for 15-20 years now. KDE 3.5 was a great environment (and Trinity (TDE) is a modernized fork of it).
Note that, this year, KDE added telemetry to their Plasma desktop environment. Of course, it's opt-in, so it must be acceptable, right? Well, of course, users who objected to the telemetry found bugs that caused data to be recorded even when disabled.
KDE's response was to ban said users from reddit.com/r/kde and call them "paranoid schizos." (The mods there are KDE members wearing "KDE developer" flair, not random Redditors.)
So, despite using and recommending KDE for almost 2 decades, it's hard for me to do so any longer. I wholeheartedly recommend checking out TDE instead.
It sounds a bit hyperbolic, but I can't recall a time when telemetry was added to a system and then later dialed back in any way. It always increases.
If I dabble with debian or kubuntu I will make sure to mitigate it, thank you for making it known, keep up this kind of good work!
It’s a real shame that they found it necessary to even build a telemetry client. I switched to free software after years on macOS (I remember upgrading to System 7) because of all the phone-home that Catalina STILL does even with iCloud, Siri, analytics/crashes, Screen Time, iMessage/FaceTime, Location Services, ntp, App Store, and software update all disabled.
iOS apps in the App Store are all allowed to phone-home like mad, too, and Apple permits this on the basis of you “opting in” to it in the App Store TOS, as if we have any sort of choice on iPhones. It’s totally endemic, and entirely undermines the credibility of Apple’s claims to caring about user privacy.
Seems like spying on users is getting heavily normalized these days. :(
Do you have examples of what you mean?
1. KDE's telemetry is opt-in (as you mentioned)
2. All data is anonymized
3. You can select which information you're comfortable to share
4. If you have it turned off, the data is recorded on your hard disk.
5. That data was always recorded in order to enable other features like "Recently used documents".
Here's some explanation:
https://www.reddit.com/r/kde/comments/fmgyy9/kuserfeedback/
Honestly, KDE has been one of the most open, privacy aware and idealistic communities out there. The criticism is quite unfair.
(I used to be a KDE contributor)
> users who objected to the telemetry found bugs that caused data to be recorded even when disabled.
> KDE's response was to ban said users from reddit.com/r/kde and call them "paranoid schizos."
I agree. The data should not even be generated if the analytics are not opted-in-to.
I do give them credit for making the system opt-in. They deserve that.
It was like a different gnome - quickly reaching the desktop and lots of nice differences (like the privacy menu wasn't crafted by marketing and legal)
And that is doable now, considering Thunderbolt. Hell, Oneplus should try to push OxygenOS to be tablet-like and this would set them apart from everyone.
Upstart was started alongside or even before systemd, if I recall correctly.
But still, emptying a directory is not how I expect to disable what is a system service. It should be systemctl disable netplan...
I think emptying would be better than deleting the directory since it'll probably just get recreated on an update.
Seriously, a whole bar for a WATCH? How come good old Gnome2 did it better 15 years ago, and had a terrific hierarchical menu, to add?
I think this model is underrated, for all that it can sometimes be annoying. Consistency is valuable. Constant change is not good, even when the changes themselves are positive. But it does mean you'll sometimes be left with out-of-date software.
Edit: Oh, I should mention that you can also use Debian testing to get frequent updates. Primary issue here there is Debian Testing actually gets security updates later than Debian Stable.
upstart was pretty good. It just lost the popularity contest with systemd. I'm not sure if there's really anything serious to complain here about.
Please read the HN guidelines: "Please don't post shallow dismissals, especially of other people's work". Link at the bottom-left.
Do you know Red Hat and google were using upstart? This companies must be hyper incompetent if they run garbage.