No, a container image for 18.04 is not enough. It means that it will use more disk space unnecessarily, have a performance hit from the container, have issues with Mesa <=> drm version mismatch, not have the latest OGL / Vulkan features, have issues with compiling with older unsupported version of GCC / LLVM for shaders etc. In short not workable.
64 bit only is fine for some open source software and generally all server based software, but for certain segments of the desktop market it is a showstopper.
In computer science, it is common to compartmentalize software. The benefits of compartmentalization are so important that outweigh any issues in increased disk space (minimal in LXD for multiple containers), overhead (again minimal). Memory might be a small issue (can fit about a dozen containers in 8GB RAM).
If personally you do not see the benefits of compartmentalization, then these are probably not suitable for you.
Consider though all the efforts in virtual machines which could have been avoided if people installed all software on the same baremetal server, in the single OS installation.
VMs take lots of resources (dedicated 2GB RAM and 20GB space). But with containers it is about 150MB RAM and 250MB space. And additional containers do not use any additional space due to copy-on-write.
I think the trainwreck is bigger than this thread. Google: Ubuntu steam news. I feel a bit sorry for Ubuntu marketing and evangelists. They are having a bad week.
The process is simple for a Linux literate person. But the target audience for the ubuntu desktop is not these users. Show this process you are linking to the Linux noobie that just wants to run his games and he will be back on windows in no time. Containers work yeah. But the average user could only use them if the set up is done auto-magically with everything completely transparent to them. The tech is not there yet.
200% agreed, I’m a subscriber. It’s the Economist of linux news. They are not much into distribution news though and have zero coverage of this so far. EDIT: wrong: https://lwn.net/Articles/791509/
And all those echo-chamber headlines are a symptom of a terrible process by Ubuntu devs, who appear a bit clueless (focusing on the 1% of users who use 32 bit hardware instead of the real problem, another dev needing to actually try to run games on a pure 64 bit system before he could see the problem …). It looks very unimpressive and very out of control.
Your lxd posts look interesting, but it’s kind of pre-alpha stage; you have a technical proof of concept, and I wonder if this informed the decision, or if you are heroically trying to rescue the situation.
And PS Steam and Wine dropping Ubuntu is genuine news, IMO.
My guess is that they will piggyback on the effort done on Debian libraries and make full QA tests for their hardware only which should be much more tractable? Otherwise it will just be them rebuilding the Debian libraries. System76 are a hardware vendor after all.
How will pop support multiarch, when it is too much work for Ubuntu?
If it is not a rhetorical question, you can find their response here
We at System76 are still planning what to do about this. We will certainly have our future releases continue to support Steam, and all games therein, with the same performance you would expect with current releases.
They also stated elsewhere that they will also keep the 32-bit toolchain support.
There is an echo chamber of the Ubuntu haters.
Here is a recent case.
Someone discovered the /etc/motd in Ubuntu and was crying out loud about the “addition of ads to the motd of Ubuntu server”. Twitter was on fire.
And here is the coverage by LWN, https://lwn.net/Articles/726902/
I will wait to see if lwn has any substantive comment on dropping multi-arch. The lwn comments are usually highly informed, and at this point they are also pointing out the Valve decision. I think everyone is surprised and confused by the decision, including lwn readers. There are definitely ubuntu haters, but the reddit and discourse comments are more than just haters. I’m not a hater, I love ubuntu. I use it everywhere, except on Raspberry pi. But I am also confused, surprised and mostly sad. I don’t have anything more to add, except that unless something big changes, I won’t recommend Ubuntu anymore, and I will move away from it because this decision and the way it has been made has shaken me. I see a half-baked thought-bubble reached by people who didn’t think enough and who failed to reach out to developers and projects which could have provided some different perspectives. The discovery process that is now happening is not very impressive. What really are the chances of discovering every printer which is going to break by asking passers-by on Discourse? Even if the decision was judged to be correct after that process, some convincing alternatives should have been ready at the time of the announcement. You can mention echo chamber, but what I see is group-think. I see a distribution thoughtful enough to include nvidia binaries in the 19.10 ISO and yet which will break Steam and Wine (in the opinion of Steam and Wine). There would have to be an awesome justification for this to make sense.
You’re really committed to ubuntu and it seems you are confident there is a workable technical solution. I really like the way you engage here, thanks. I won’t take any more of your time, I think the communication effort needs to be with Valve and Wine.
Now, I’ve been using Ubuntu for close to fifteen years now. I don’t participate in community discussions, ever, nor do I vote in polls if you ever run those, I don’t know. I registered for this post just to express my annoyance.
But I have maintained hundreds of ubuntu servers along the years and currently run ubuntu on my desktops and laptops, as well as maintain around 20 servers currently. I found it flabbergasting how out of touch ubuntu has always been with it’s users who aren’t the vocal minority. Mir, wayland, gnome shell, unity, amazon, you name it. Some might even describe these changes as hostility against it’s users, like dogfooding the end users - which shouldn’t happen.
I’m so close to being done with ubuntu and its complex committees and politburo-like policing. If this doesn’t get reversed, I’ll likely move all my personal devices over to some other distro as well as servers I maintain, as I like to use the same distros on every device.
I need to add here that the planning to phase out i386 in 19.10 was drafted and planned more than a year ago. As a community it looks bad to get loud voices at the last stages while before it was all calm. I strongly despise the “stop power” attitude, that is, the negativity I see in copy-pasted posts. It’s not a trend, it is bad quality click-resharing.
Ok, You guys decided something 1 year ago, so we (I think a lot off pp have the same opinion) now realise that:
we can´t run properlly some apps like wine or steam
The solution:
We start using others distros
What the people are traing to say to you is this decision going to break our workflow.
I accept that we don’t pay for this (I can’t by a subscrition for my desktops; can only donate money), soo you guys need to make pratical decisions and cut costs, but this decisions point us to one thing: Ubuntu is moving form a All in one system to a IoT/Server/Cloud oriented distro, that don’t suits the needs of the majority of your user base.
Sorry man, but from my point of viewe this is a bad decision (like kill Unity). I’m going to move on to another distro, because of this.
Is only my opinion.
Need money to support the costs?! Tell the community to make donations like append with Ubuntu edge phone.
From the first second of existence, Ubuntu has achieved an undisputed worldwide fame for excellence because of the user experience. Ubuntu won the desktop war against other distributions, without any doubt. And now, really… Are they beginning to throw away any of the great work that has been done? I’m a Ubuntu 18.04 LTS user but I’ll move to another distro on 20.04 if they keep this wrong direction. In fact, I’ll move to the distro Valve/Steam will recommend.
Since much of the discussion in this thread is based upon the 18 June announcement,
and since that announcement has been superseded by the 24 June update and newer discussion thread,
I’m going to end this thread here to avoid confusion.
There are several interesting side-discussions going on - those who wish to continue to develop those conversations, PM me and I’ll happily migrate those posts to a new thread.
I will wait to see if lwn has any substantive comment on dropping multi-arch. The lwn comments are usually highly informed, and at this point they are also pointing out the Valve decision. I think everyone is surprised and confused by the decision, including lwn readers. There are definitely ubuntu haters, but the reddit and discourse comments are more than just haters. I’m not a hater, I love ubuntu. I use it everywhere, except on Raspberry pi. But I am also confused, surprised and mostly sad. I don’t have anything more to add, except that unless something big changes, I won’t recommend Ubuntu anymore, and I will move away from it because this decision and the way it has been made has shaken me. I see a half-baked thought-bubble reached by people who didn’t think enough and who failed to reach out to developers and projects which could have provided some different perspectives. The discovery process that is now happening is not very impressive. What really are the chances of discovering every printer which is going to break by asking passers-by on Discourse? Even if the decision was judged to be correct after that process, some convincing alternatives should have been ready at the time of the announcement. You can mention echo chamber, but what I see is group-think. I see a distribution thoughtful enough to include nvidia binaries in the 19.10 ISO and yet which will break Steam and Wine (in the opinion of Steam and Wine). There would have to be an awesome justification for this to make sense.