Ubuntu 21.04
ubuntu.com
ubuntu.com
I'm curious to know what state Wayland is in nowadays, especially regarding how well it works with Nvidia cards. The last time I tried it out was around 2017ish when Fedora enabled it by default. It was 100% not ready then as it wouldn't even boot into the desktop with my nvidia card.
Is it actually ready now? Does it work with nvidia cards? Does GPU acceleration work with nvidia cards? Are apps launching via xwayland gpu accelerated / working as expected? It sounds like OBS and some screenshot programs work now which is nice.
But fractional scaling with Nvidia did not work at all before 21.04 so I guess there is progress.
Additionally, NVIDIA seems to be putting a lot of effort into overall Wayland support (yay!): https://www.phoronix.com/scan.php?page=news_item&px=NVIDIA-4...
Hopefully, by next year, Wayland on NVIDIA should be "fully" supported. It's slow progress, but it is moving forward.
Wayland support is getting really good these days, I'd recommend giving it a try before going back to Xorg if needed.
This release is no different than any previous one for NVIDIA users
Depends on your needs and hardware. I have an Nvidia desktop and Intel Laptop both running Arch + GNOME 40. Intel is definitely less buggy, but I still find myself switching back to X11. For example, when I needed screen sharing for work video conferencing, it only works under X11(PipeWire support was buggy at the time). Wayland/GNOME also still have issues with mouse lag with high CPU.
It's getting better every day, but still rough around the edges.
I'm still on Xorg myself, but planning on switching to Wayland later this year once XWayland support is merged (also using NVIDIA card)
Edit: I've switched to PipeWire for Xorg with pipewire-pulse with zero issues so far.
Years ago I remember dreading firing up XFree86 on a machine with only 4MB, opting to use svgalib apps instead. But I've been using XFree86/Xorg for two decades now. I like new technology (I love my Toyota hybrid, for example, even though it lacks a clutch pedal). What tangible benefit does Wayland offer me?
- Reduced tearing
- Simpler architecture, better security
In theory, Wayland does provide few user-facing change (I think)
I can switch from Intel only to Nvidia PRIME profile and after restart it just switches to XOrg.
Before christmas I played around with Plasma on Wayland. It mostly works, but there were some weird quirky behavior. I didn't look much into it, since I'm so happy with Sway.
I hear people say Gnome is fine on Wayland.
As for Nvidia: can't help you with that, sorry.
With luck, this problem will soon be solved from the opposite direction:
> Looking forward
> Now that there is a pure GTK3 port of Emacs, GTK4 has naturally been released.
If you use one of the "big" compositors like GNOME or KDE I believe they have sort of added EGLStreams support so you can use it on NVIDIA but it's not perfect. If you use a more niche one, e.g sway, then it won't work on the NVIDIA proprietary drivers because they don't support GBM.
There is some news recently that NVIDIA may be coming around and are in the progress of implementing GBM support in their drivers, but it's not confirmed.
You can also use the nouveau nvidia drivers but from what I understand the performance is trash in 3d applications using modern cards, it seems more orientated to supporting older nvidia cards.
But not sure about the OpenGL (I think Nvidia do).
These apis let you export a FD for sharing. Pretty sure you can import (scanout compatible) objects into DRI/DRM/KMS. Which is what libgbm does for Mesa drivers.
Although most compositors (including wl_roots based ones) still use libgbm as they might also need to support embedded hardware. (eg: Tizen).
(In the original release of Vulkan intel has a weird "unofficial" extension that let you export a FD. and khr had a VK_cube demo that could directly use KMS with Vulkan using this "extension". This was removed as it was only there to test Vulkan before any window manager integration was implemented)
The thing that is worth noting is how this situation came to be. That is, there was at one point a conference where the community invited AMD, Intel, and NVidia to discuss the future of this API and Nvidia was a no show for some reason. Those who showed ended up producing GBM, Nvidia pulled EGLStreams out of a hat, and the rest is history.
Who knows, maybe it was something as simple as scheduling and the people at Nvidia that needed to be there couldn't make it? Maybe in another timeline this mess could have been avoided entirely.
Money and proximity to hardware development would be the competing factors, and normally I'd like to pin the attribution on them, but there's just one small problem: I've been on a bunch of committees and they have pushed my estimation of typical committee dysfunction to extraordinary levels.
In my neck of the woods (Central Europe, so not a place with cheap and abundant hardware like North America or Asia) you can find plenty of decently priced hardware in stock if you're an educated consumer and willing to look around.
Sure, you won't find any discrete GPUs anywhere at MSRP thanks to mining and scalpers, BUT, you can find golden nuggets like the Zen2 based 6-core 4650 Pro Ryzen APUs for 220 Euros which is a cracking CPU with good integrated graphics, or you can buy a gaming laptop with the latest gen 8 core Ryzen 5800H and an RTX 3060, 16Gb RAM and 512Gb SSD for about 1200 Euros or even less which IMHO is insane value for this level of performance considering the current apocaliptic market.
This was 2 weeks ago on kubuntu 21.04 beta.
But I’m on manjaro now and for some reason I have less issues on my laptop than I had with ubuntu 20.04.
Considering I probably have to fool with X every six months or so, a lot will depend on what you mean by "ready."
Personally, I'm not ready for Wayland. It doesn't solve any problem I have. And might create problems I don't have. If I was in the market for new problems I'd probably try Arch again ;)
Some way of signing into Ubuntu with a custom SAML provider would be very cool. Windows supports it via their "credential provider"[0] framework and Google uses that to provide sign-in for Windows with your GSuite credentials.
[0]: https://docs.microsoft.com/en-us/windows/win32/secauthn/cred...
A bit frustrating, as you can otherwise join a windows 10 desktop (or vm) to just azure ad for sso etc.
[1] https://docs.microsoft.com/en-us/azure/active-directory-doma...
well that depends, the higher subscriptions include it for free.
sadly not for desktops, btw. you can create your own pam provider (grant_type=password) and put it into sssd (for local cache), but that does not work with 2fa.
I see this as a strict improvement for adoption of Linux workstations in the corporate world
MSIs still dont work on Ubuntu. GPOs are still very limited-- mostly just HBAC type stuff.
The stuff that works is the important stuff, but if people want to roll out software to Linux they aren't doing it with AD. They're doing it with a CM tool like puppet or ansible.
Also, it's kind of buggy. I have some Ubuntu 20 machines set to do AD logins here and more often than not it just stalls until it times out on login with no useful diagnostic messages emitted. Very frustrating.
Some environments mandate this stuff, and there's frankly no excuse for NOT supporting Kerberos / LDAP / centralized auth. The choice wasnt whether to support "stuff like this", but whether the support would be first class or second rate.
There is a large link to "Ubuntu Desktop" at the top left. I'm not sure why this was so hard for the GP. Maybe it could be slightly easier if you didn't have to click on the big Download link in the first place, but it seems to require very low effort as far as I can tell. Maybe he's running on a textmode browser with no Javascript support?
I can totally imagine this costing several would-be-users
That's the ec2 image over at amazon.
(Mind you, there are plenty of “retail” installs of Ubuntu — but individuals will often just upgrade Ubuntu to a new release using built-in tooling, rather than downloading a fresh release-image. They only need to visit the site once—the first time they install.)
I think, if you’re visiting ubuntu.com nowadays, you’re most likely to be looking for an Ubuntu Server image.
If Ubuntu depended on random users pointing their browsers to ubuntu.com and then clicking a download link to download and burn an ISO, they'd be long, long gone. There hasn't been any real $$$ in that since, well, forever?
Is this really aligned with Canonical's (and Ubuntu's) philosophy and goals?
The snarky answer and the sincere answer are the same for once: Yes. If you want that, go to Debian.
Completely unrelated: I just discovered Rofi (extensible Spotlight-ish launcher for Linux) and it’s lovely, especially when bound to Super+Space.
In the mean time, there is an unofficial PPA https://launchpad.net/~shemgp/+archive/ubuntu/gnome-40
https://www.phoronix.com/scan.php?page=news_item&px=Ubuntu-2...
Anybody know what happened to that idea? It's been a long time desire of mine to have a secure Linux laptop without needing to enter my password twice during startup and I was hoping to get it with this release, but when I tried the beta the installer did not allow it.
Easy LDAP/AD support and TPM encryption were the two showstoppers whenever I asked IT departments at previous workplaces to provide me with a Linux laptop, as their engineers did not know how to set them up but they had compliance check-boxes they needed to fill in regardless of utility (antivirus being a third one but it turns out these days AV companies actually provide Linux builds). Now that the LDAP/AD stuff is sorted out, fingers crossed for TPM encryption.
Only with regards to that sentence (as I don't know much about TPM): You can simply opt for full disk encryption during install, and then you can also select "automatically log me in on boot" (not exact wording) when creating your user. So you enter the encryption password and that's it. Not sure why you would _need_ to enter password twice with encryption?
That's how I handle it. Screen lock, or returning from suspend still prompts for user password as usual.
There is a GPO template. I think I'm in love.
The thing that’s changing with this release is the ability to support some GPOs with adsys. sssd has/had some basic gpo support as it relates to auth but not really for system config.
This definitely moves this into "I might introduce this to our environment" level.
We have linux servers on our domain, specifically becaeuse there are certain things they do better than windows, and im a tool for a trade kinda guy.
But authentication and management of creds is a hard requirement.
The rest i can work with. We setup Ansible to handle much of the day to day config things. GPO"s might be nice.
At one and for years i was using PBIS, and whatever it was called before it was bought out, forked and named PBIS - https://github.com/BeyondTrust/pbis-open
I liked it because it kept GID/UID's consistent across the domain. But it was flaky for samba shares specifically ad group enumeration. There are paid add-ons that leverage GPO but never went that route.
More recently we are just leveraging SSSD/Chrony/Smb/Nmb to do the same, there are a few tweaks, but its solid. Its actually more stable now.
Some GPO's would be nice I supposed.
The moment you learn about things like Ansible, and then ask "wait, what if I want to programatically make GPOs"-- and try to do so-- you realize just how horrendously bad they are.
Their permissions inheritance is terrible, so trying to do least-privelege / granular rights for admins is awful.
Their supercedence structure is confusing, overly complex, and naturally inclines people towards abusing "enforce" and "block" in the worst possible ways.
Un-screwing GPO inheritance messes is essentially impossible, so in production the answer is generally "create a new GPO" rather than "clean up the old mess".
And the settings structure itself is a terrible organizational mess; things "sort of" have a logical structure, except for the thousand odd exceptions. You can't set preferences on items that may not exist (such as disabling insecure services from old OSes) without generating a bunch of logged errors. You can't really target your GPO without some rather expensive client-side queries.
Logging what is actually happening (e.g. why did that printer mapping fail) is awful, which I have generally seen solved by writing one's own drive and printer mapping scripts.
The whole thing is a relic of the 90s that never quite got finished, polished, or maintained and mostly exists because it was "good enough" and there were no real alternatives. There still aren't, really, unless you want a third-party solution which is inevitably almost as big of a maintenance burden as GPO-- and GPO being free and mostly reliable makes it easy to just accept its awfulness.
I've yet to see something as comprehensive, navigable, and flexible. Sure, it can be horrendously complicated, especially if you chose not to clean up the old mess, but it's still... basically better than everything else.
Setting up SCCM is an exercise in hating your life but once you have the servers up it’s stupidly powerful.
In the Linux world — Ansible if you want push, Puppet if you want pull.
> Ubuntu machines can join an Active Directory (AD) domain at installation for central configuration.
This means it's now a ready-to-go functionality of the OS, and can be done simply without going to read a blog article on how to do it, post-setup.
> Ubuntu 21.04 adds the ability to configure system settings from an AD domain controller. Using a Group Policy Client, system administrators can specify security policies on all connected clients
And this means one can deploy Ubuntu machines on an Active Directory network at scale. Rather than hand-crafting one Ubuntu server that connects to your AD, ability to set policies with GPOs is the baseline to be able to say, deploy Ubuntu desktops to an entire department. Group Policy is the "pets vs cattle" difference in a Windows-based network.
Fedora did this few releases ago, as a part of the new-at-the time OOB wizard (you could use AD or FreeIPA, autodetected). Good to see that Ubuntu does similarly.
> can be done simply without going to read a blog article on how to do it, post-setup.
I never had a problem with setting up AD membership after-the-fact, except for three things that bothered me:
1) you still needed a local (root or sudoer) account in Ubuntu (but not in Fedora),
2) the setup didn't set up domain admins (or other group) neither as sudoers, nor as PolicyKit admins (these GUI prompts for admin password that you get from Gnome Software or Disks, or other privileged apps). These had to be done manually,
3) Files doesn't use AD Kerberos ticket when accessing file shares. It will just asks the user again; kind of defeats the point of Kerberos.
I'm really interested whether these are fixed.
Log in, get group and netgroup membership, apply GPO based access control, access ssh keys, apply sudo rules.
> after both setting up the OS, and then configuring the domain join
realmd join ad.example.com
Add-Computer -Domain ad.example.com
Huh?!? Are you manually joining every machine to the domain and manually configuring every machine by hand? Even if you don't want to bother with config management like Chef/Puppet/Ansible for post-install config you at very least could have leveraged preseed files to automate the install, domain join, and applying your configs.
Like it's great that Ubuntu is supporting a subset of GPO but just like in the Windows world you can't configure everything with it and still need SCCM and Linux has solutions abound for that functionality.
Genuine question, wouldn't that require a kerberos ticket to work? Or at least some high privilege password that you might not want to keep around your preseed files? Bootstrapping trust is a tricky problem. Most that I've seen tend to go with some custom solution.
We don't broadcast the join password and rotate it every so often but we also don't really care if it gets out.
Having an OU that computers are provisioned in, and then moved out of later on would have help me a lot.
realm join mydomain.fqdn
But generally you were seeing a bunch of guides because they were referring to truly ancient distros.
Does anyone know why I can find the admx and get going or is it reading the standard windows policies? Can't seem to find any articles/documentation.
Release notes might be more useful for this crowd.
> Comments should get more thoughtful and substantive, not less, as a topic gets more divisive.
The parent comment is neither thoughtful nor substantive, merely agreeing with the grandparent -- a verbose equivalent of a "This!" comment. A better comment might include the commenter's reasoning, their experiences with other platforms and their pros and cons, or how they have used Ubuntu personally or professionally, for example.
> Eschew flamebait. Avoid unrelated controversies and generic tangents.
The parent's implicit dismissal of Windows or Mac desktop platforms as un-enjoyable could be interpreted as flamebait, since choice of operating system can be a common point of contention.
Hopefully this explanation helps. Since the comment was low-quality but not flagrantly bad, I assume it will pick up one or two downvotes but not a flag, and will sink naturally down the discussion as higher-quality posts supersede it.
Subjective interpretation is now a basis for downvoting and being labelled flamebait.
Meanwhile the most upvoted comment is about the graphic looking like hairy testicles(which it does).
I don't think people realize how important a well supported DESKTOP experience is for open source adoption by the mainstream users.
The mainstream doesn't care about zshell or fish shell or vim.
All they see is the graphical desktop experience, for their email or browser.
If that market can be brought into the open source world away from the giant companies that would be a huge win for open source.
And Ubuntu is currently the only real game in town, and having used it myself since v11 or something it's come a long long way, to the point where my grandmother can use it.
I kind of get the downvotes.
“It’s a curious thing about our industry: not only do we not learn from our mistakes, we also don’t learn from our successes.” – Keith Braithwaite
CentOS rolling release will still be so slow that itll lag a lot behind Fedora.
Its just that CentOS will now be afew months ahead of RHEL instead of afew months behind RHEL.
The chart looks more like this:
Fedora ---------------------------------------------------------------------> CentOS Stream --> RHEL
It's definitely not "exactly midstream".
There's a new Linux distribution called CentOS Stream, but it's not the same distribution as CentOS, and it doesn't have the same goals of CentOS. It would be correct to say that CentOS is being replaced by CentOS Stream, but that means that CentOS is no more.
You may argue that this is pedantic though, and that the phrase "the death of CentOS" is intended to communicate something which isn't accurate. And, yeah, that might be the case.
---
Disclaimer: We use CentOS on almost all of our servers at production level for 10+ years.
The main problem with CentOS is not moving into a rolling release schedule, but change of its place in the ecosystem.
Before, CentOS was the last tier. Fedora was testing ideas, RedHat was implementing them, and CentOS was following the trail by porting them later. There was an unwritten agreement that RedHat didn't prevent CentOS' development, and CentOS didn't port everything at day 1, so they were in a mutualistic state. Moreover, CentOS enjoyed a ~10 year support on every release, so it was the soul-successor of the original RedHat from the olden times.
Now, CentOS moved to pre-RH position. So Fedora experiments, CentOS makes the Beta & RC testing and RedHat gets more thoroughly tested patches and, that's it. CentOS is moving to a Debian Testing meets Arch Linux position. It's neither stable as Debian Testing, Nor supported like Arch and lacks any official support and possibly no security patch support.
This is problematic for many places since CentOS was the RPM Equivalent of Debian Stable. Now, there's no RedHat based free and community-driven and community-supported distro. People who can't use CentOS in its future state will either migrate to RedHat or to Ubuntu or Debian Stable.
For us, and for other data centers which do the same thing as us, current situation is a very big let's wait and see game.
For the health of the ecosystem, we need another fully free (as in beer & as in speech) and fully supported distribution. Hope Rocky can fill that void.
I'll continue to use Debian on my personal systems, for foreseeable future.
Software development and verification is huge in scientific computing. It's not just "Meh, let's port it in a weekend and be done with it".
maybe time to get on with it.
Nevertheless, I'll try to call around to see who can start shortly. :)
not to rub salt into the wound but relying on a single vendor could be considered technical debt.
>CentOS makes the Beta & RC testing and RedHat gets more thoroughly tested patches and, that's it. CentOS is moving to a Debian Testing meets Arch Linux position. It's neither stable as Debian Testing, Nor supported like Arch and lacks any official support and possibly no security patch support.
This isn't correct.
Debian Testing is a true rolling release distribution for the next "major" version of Debian. If you install Debian testing, what you're getting is a hybrid between Debian N and Debian N+1, with package versions that at any point in time may or may not be similar to those in _either_ Debian N or Debian N+1, since they get continually updated up until the stabilization phase.
That is not what CentOS Stream is.
CentOS Stream is a rolling release for the next minor (_not_ major) release of RHEL, and follows the same development process, including the exact same CI and testing scrutiny that was required to update a package in RHEL internally. It's basically taking the development process which used to be internal, and opening it up to everyone else.
Unlike Debian Testing, CentOS Stream is _not_ a hybrid between major releases of RHEL (say, RHEL 8 and RHEL 9). It's frozen to a major release. So CentOS Stream 8 will track the development of RHEL 8.3, 8.4, 8.5 and so on, and CentOS Stream 9 will track the development of RHEL 9.1, 9.2, 9.3 and so on. And like both RHEL and current CentOS, that means that the updates will only fall into the categories of backported bugfixes, security fixes, support for new hardware, and on very rare occasions individual backported features.
This is more significantly stable than Debian Testing - it is less "Debian Testing meets Arch" but rather "old CentOS meets Debian Testing".
Where did you hear that CentOS Stream didn't receive security patches? That is not true...
Actually, the initial communication of this issue was so vague from our perspective, so this is what I and my colleagues understood.
Again, thanks for clarifying, because I personally don't want to bash CentOS, but want to understand what's happening and continue to use it. Maybe it would be beneficial to disseminate this in a more visible and more understandable way.
> And further - where did you hear that CentOS Stream didn't receive security patches? That false...
I didn't hear, but as I said, CentOS Stream was presented as a proving-grounds distribution and, I understood that it'll receive security updates in a best-effort basis.
The news came in a crashing way and the initial roadmap didn't communicated well to the outside world in the beginning. To be frank, a lot of people felt betrayed by IBM/RH. When a company announces a big paradigm shift and cuts the support for the latest release at the end of the year without further explanation besides marketing speak, thinking otherwise is pretty hard.
Hope you understand the frustration.
Cheers
If he wanted me to remove it, I would have happily done so.
Also, I just pasted his nick to Google and it came on top. So I presume he didn’t try to hide his name. If I have sensed the contrary, I would not dig one step further.
It's not false under the context of long term support which is why I highlighted so in the OP. How long will each CentOS Stream release be supported? How long with each CentOS Stream release receive security patches?
No it isn't. Ubuntu LTS is supported for 5 years.
https://ubuntu.com/blog/what-is-an-ubuntu-lts-release
>An Ubuntu LTS is a commitment from Canonical to support and maintain a version of Ubuntu for five years.
---
>Not saying it is a bad OS but it is no longer a free Linux operating system with long term support.
Ubuntu LTS is suppored for 5 years, Debian Stable is supported for 5 years, and OpenSUSE Leap is supported 5 years (as far as I can tell - the only documentation I found said "up to" 60 months).
CentOS Stream absolutely provides "long term" support.
>ESM is available through an Ubuntu Advantage for Infrastructure subscription for physical servers, virtual machines, containers and desktops, and is free for personal use.
https://ubuntu.com/security/esm
Note that if you click through "personal use" means "up to 3 machines" and obviously doesn't apply to infrastructure. RHEL has free "personal use" subscriptions too, except they apply to up to 16 machines.
And also:
> Initially, free subscription is available for Ubuntu 14.04 LTS only.
Thanks, I recognized it first! It should be more clearly advertised.
I think you just made a new malapropism by getting the "sole" in "sole successor" mixed up with the homophone "soul" and then relating it to "spiritual successor"
Both "sole successor" and "spiritual successor" are accurate descriptions in this case, so it works beautifully.
Being tired and having a different mother tongue has its perks, it seems. :)
The fact that it's a rolling release is not really the problem per-se, it's more that you can no longer expect that CentOS n == RHEL n. It's not a drop-in replacement. You can't expect that something that works in RHEL will work in CentOS and vice-versa.
For people who ran CentOS because they liked it over the competition it may not be a deal breaker. For people like me who only used it because it was "free RHEL" the new rolling version is effectively useless. And I'm 100% certain that it's exactly what RedHat/IBM counted on: no more free candy, just buy a license.
Do you mean that it's not 100% Open Source anymore? That you can't build your RHEL anymore?
One thing I don't understand is why they vehemently deny it was the reason.
Whatever your personal preferences are, it's obvious that this change doesn't fit well with that crowd.
Red Hat has every right do do that change, but we shouldn't pretend that this is a good fit for the traditional centos crowd.
A distant friend of mine was a Sailor of a big cargo ship around 2007 (i think), he had just a crappy Satellite connection. But he wanted an easy Linux next to his XP, so he ordered a FREE! ubuntu-cd to the next port, it was there, he is now a Dev specialized in shipping-load-unload-software.
https://lwn.net/Articles/849639/
Ed: or audio: yes?
> Pipewire support is now enabled which restore working screen recording and allow better audio handling for sandboxed applications
From release notes: https://discourse.ubuntu.com/t/hirsute-hippo-release-notes/1... posted by @spivak in another comment.
https://res.cloudinary.com/canonical/image/fetch/f_auto,q_au...
Definitely changing that after install, I don't want to be greeted by a giant parcel of pubes when I log in.
I know this is a really dumb thing to comment on when there's so much actual substantive change in this release, but this image is literally the first impression new users will get for the next six months, and will be the thumbnail for tons of tutorials and such during the next year. It probably could've done with another revising pass.
Unlike most other semiaquatic animals, hippos have very little hair. The skin is 6 cm (2 in) thick, providing it great protection against conspecifics and predators.
How did this pass any UX testing?
Ubuntu/Ubuntu Studio is a decent enough distro for audio production, so it would be sad if this was broken.
This is why you can get PCs from enterprise PC vendors like Dell and HP shipped with Ubuntu.
I found the downloads page[1] but it still says Beta, unsure what's going on.
Question: Is this a GNOME architecture issue? I hope KDE could avoid this issues, atm on Xorg I can restart the window manager and the desktop shell without losing any data.
If Linux would have decided to use Mir would have at least shared same "server" code and probably also get around this issues but is hard to promote quality when Gnome and RedHat has a different idea.
That's because there is no such thing as "wayland" actually running on your machine -- it's simply a protocol for compositors and applications to talk to each other.
"Wayland" is also an implementation. The protocol and implementation share a name. By the look of things, Gnome does indeed require it (that is, various "libwayland" packages) on Ubuntu.
Wikipedia:
> Wayland is a communication protocol that specifies the communication between a display server and its clients, as well as a C library implementation of that protocol.
That's the C library also named Wayland and made by the people guiding the Wayland spec on which Gnome depends, no? Or are there three things going by the name Wayland, one of them not actually associated with the Wayland project?
AFAIUI, e.g. KDE doesn't actually use any of that. (I could be wrong, I honestly don't care that much about the minutiae.)
Yes. And always has been since like forever ago. I'm all for choice and diversity, but I've personally never understood all the love and excitement and attention that GNOME has gotten ever since Canonical got behind it and dropped Unity.
I'm certainly not it's target audience - running XFCE, LXDE/LXQT, and i3/Sway over the years - but it really gobsmacks me to hear the cheerleading over what seem like really basic fixes, like "now the single-threaded shell crashes to login less often! Horray!"
Cheap basic mice seemed to work well.
I think it's great that Firefox is bundled with Ubuntu, but I have to manually update Firefox each time a new release comes out because the version that's installed is a canonical-controlled version that stays 'stuck' on a old version and doesn't have an auto-update mechanism.
Other than that, I love the OS. I use it as my daily driver since it has a good community around it, and if you're having trouble, there is a good community-powered Q&A site for it.
Doesn't Snap solve this?
Source: https://github.com/mozilla/gecko-dev/commit/e6b8e53803e10244...
https://blog.mozilla.org/futurereleases/2016/04/21/firefox-d...
How so? I use the default Firefox on Ubuntu 20.04 and it's updated through apt like almost all other applications/libs I've installed and it stays pretty up-to-date. E.g. Firefox started offering version 88.0 on the general release channel three days and there ISN'T YET a corresponding update to the package repo, however I expect it'll most likely pop up there within the week.
This is par for the course on some distros. Fedora, for example, has its own repository for Firefox that is sometimes behind the mainline release of Firefox (although in my experience it's usually only a few days).
The advantage of the process is that the Fedora Firefox repo actually has its own QA testing, so they occasionally catch some things that might impact Fedora users specifically, that otherwise might go unnoticed if I were to switch to the Firefox flatpak.
What? It is updated just like any other software by using the package manager
https://packages.ubuntu.com/focal/firefox
I actively don't want random app-specific auto-updaters; I want one package manager.
Apropos of which, does anyone know why trackpad interfaces are an area of active change? It really surprised me when the trackpad got disabled by closing and opening the lid on 20.10. I naively think of trackpad hardware as mostly a solved problem.
But last time I checked a year ago, some poor lone soul was trying to gofundme his effort to provide better drivers/firmware.
Trackpad and sleep-mode are the two things I miss from windows/OSX. It's just more polished over there, specially OSX.
Many applications built with Flutter? Such as?
It appears to me that you've clipped the quote to change the meaning.
> Firefox, OBS Studio and many applications built with Electron and Flutter take advantage of Wayland automatically, for smoother graphics and better fractional scaling.
It implies that there are already many application built with Electron and Flutter.
My question is simply: what Flutter applications?
https://github.com/canonical/ubuntu-desktop-installer
Not sure if that's already in this version.
Once the next LTS is out, I usually wait a few more months. I'm not a big fan of having my workstation not being able to do basic things. ( no hard feelings, I trust the process )
(Using Xfce, but also have KDE and Gnome installed too.)
The fractional scaling is garbage in this setup; and not just because you put the numbers in backwards from what you'd expect(2x is 2x smaller, and 0.5 seems to be 150% larger). It seems like it just scales the rendered output; everything is blurry! Scaling the DPI up to 144 however makes almost everything nice. The term, Chrome, VSCode, Intellij and etc all seem to adjust properly based on the DPI setting. The fractional scaling has a multiplier effect on that though, and so then reducing the DPI makes it smaller but still blurry. It also causes a lot of lag.
So, fractional scaling in XUbuntu seems to be scaling the video output and not just the UI like you would expect. Adjusting this in VirtualBox itself leads to better performance.
Distros should not include filesystems that aren't in the kernel in their official installer. Fedora made the right decisions with btrfs. If users want zfs they don't need the official installer for it.
btrfs and zfs are.. not comparable.
It's supposedly meant to "just work" but out of the box it did not. Which is a very common story with ubuntu sadly.
Ubuntu ships ZFS module in kernel package and it works without having to be compiled nor it needs header package.
# dpkg -L "linux-modules-$(uname -r)" | grep zfs.ko
/lib/modules/5.4.0-70-generic/kernel/zfs/zfs.koBut more importantly, missing ZFS support should not brick the filesystem. The new kernel just can't mount it, which can be overcome by rebooting to an older kernel. And then addressing whatever the issue is.
If you use DKMS, then you might have issues, but there's no need with Ubuntu (and other distros like Manjaro) anyway.
Btrfs is probably never the right decision - use the FS that's right for you and your workload.
And you're posting anti-btrfs FUD.
https://lwn.net/Articles/824855/
https://facebookmicrosites.github.io/btrfs/docs/btrfs-facebo...
Even if it's the only deployment in the world, it's more than "never". They provide strong technical reasons for using it.
did they fix all the issues with Wayland on desktop yet? Like HiDPI being broken, Electron-based apps being blurry, broken color management, broken 60+fps, etc etc
Ubuntu 21.04 will be supported for 9 months until January 2022. If you need Long Term Support, it is recommended you use Ubuntu 20.04 LTS instead.
Edit: Ninja'd by jpace121, should have refreshed before commenting.
Why so hostile?
"22 April 2021: Today, Canonical released Ubuntu 21.04 with native Microsoft Active Directory integration, Wayland graphics by default, and a Flutter application development SDK. Separately, Canonical and Microsoft announced performance optimization and joint support for Microsoft SQL Server on Ubuntu."
So, the major changes done to the distribution is to suck up Microsoft and Google at the same time, and integrate whatever shitty stack that they want to push.
For example, Microsoft SQL, if you are on Linux, you have the full power of Postgres, Mariadb, ..., who in his right mind would willing-fully use Ms SQL on Linux?
And so, going one step further and spend resources to work on that as priority and market it?
As a very long term Linux and Open Source proponent, I have the feeling that it is a hold up.
As someone who works in a .NET shop, it's been possible for a while to run MS SQL Server on Linux, anyway. This leads to a smaller overall footprint for those of us using dockerfiles to spin up test environments and such.
It's not an operating system's job to tell people what to run on it, it's job is to provide a platform.
If you want to go around yelling at people for using MS SQL then that's your choice right? Don't expect others to do the same.
> who in his right mind would willing-fully use Ms SQL on Linux?
The anti-anything-Microsoft rhetoric got old a long time ago, and TBH, comes across as petulent and a bit childish. Having actually used it for several years, I can say that MS SQL is a really good database - it has features and integrations that Postgres doesn't, and vice-versa. And if nothing else, having SQL Server on Linux is great for running on Docker when testing.
FWIW, I'm a huge fan of Postgres, and it's my database of choice - nobody is forcing you to use SQL Server just because it's better supported.
It seems to me that changes like this only serve to bolster the popularity of Linux.
What do you think of the actual release-notes ? https://discourse.ubuntu.com/t/hirsute-hippo-release-notes/1...
"Microsoft SQL Server on Ubuntu delivers an optimised, cost-effective database platform"
Of course not, it's not free so it's not cost effective compared to the best database, which is PostgreSQL.
Integration with Active Directory group policy seems also a really bad idea since it allows insecure proprietary software to control your Linux machines.
And plenty of business apps straight up require Microsoft SQL Server. You should be thrilled those now no longer need a Windows license, just the SQL Server license.
This is a smart move by Canonical to court enterprise customers, especially with the Red Hat change of ownership chaos and churn.
Active Directory has its own share of problems, but the benefit of having a central location for policies which enforce them on the connected machines is a security plus despite the addition of attack surface by itself.
While I first thought that AD integration regards group policy looks like a hard architecture mismatch, it is interestingly not. Take the Windows Registry with millions of keys vs. the thousands of etc files and their thousand sub settings.
I still miss some of the tool ecosystem (like schema/data diffing tools from RedGate) from my SQL Server days.
Also proprietary software is not insecure just be being proprietary.
And if that's the extent of your understanding of what the software does, then you aren't really qualified to give an opinion on its security.
"Richard M. Stallman, frequently known as RMS, has been a dangerous force in the free software community for a long time. He has shown himself to be misogynist, ableist, and transphobic, among other serious accusations of impropriety." - https://rms-open-letter.github.io/
Because of his history of abuse (per Fedora) Fedora will not attend any FSF events, will not attend any event Stallman attends and will not attend any events by any Organization Stallman is connected with.
Has ubuntu taken a similar stance? I couldn't find a similar Ubuntu open letter on this.