Live data from Hacker News

Ubuntu 21.04

ubuntu.com

231–240 of 322 posts

Re: Ubuntu 21.04

#231

Earlier quoted context omitted.

Disclaimer: I work for Red Hat >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 "majo…

Daniel, thanks for the comment and clarifications in (Googled your twitter account for your first name, hope that's OK). 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 benefi…

No worries.

Re: Ubuntu 21.04

#232

Earlier quoted context omitted.

From what I understand after briefly looking into this earlier this month the issue seems to be basically NVIDIA want to push their own buffer API (EGLStreams) rather than use the buffer API developers are actually using (GBM). 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.…

NVIDIA tends to be good (in an evil way) at proprietary lock-in, but they also tend to be good (in a genuine way) at making better thought-out and more advanced APIs. What's the balance of these two qualities in EGLStreams (as compared to GBM)?

I've looked into this myself before, because I was curious and had the same thought as you. I don't feel comfortable making any detailed claims about GBM vs EGLStreams and which is better, but I will say there are absolutely some tradeoffs between the two and neither is entirely without merit.

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.

Re: Ubuntu 21.04

#233
post #218

Earlier quoted context omitted.

It looked like bad sarcasm to me when I first read it. Perhaps other people thought the same.

I LOVE Ubuntu. 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…

That's clear! It's just that many of us have been conditioned by internet snark to read comments like the GP cynically, unless they contain disambiguating information (as your comment here does). You did nothing wrong, but I have a hunch that ambiguity may have been the source of some downvotes.

Re: Ubuntu 21.04

#234
post #226

Earlier quoted context omitted.

Disclaimer: I work for Red Hat >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 "majo…

> where did you hear that CentOS Stream didn't receive security patches? That is false... 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?

5 to 5.5 years - the same as RHEL "full support" phase.

Re: Ubuntu 21.04

#235

Earlier quoted context omitted.

I'm not real familiar with this stuff but I could've sworn that linux has been able to join Active Directory domains via samba for a long time.

I know you can, in a very basic sense, such that one can log in, after both setting up the OS, and then configuring the domain join after installing additional components. But there's two key things here: > 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 ar…

For a very long time the process on Linux has basically been making sure SSSD / SSSD-AD / realmd were installed, and then running,

    realm join mydomain.fqdn
Some distros defaulted to SMB (though which, I dont know), and occasionally the default configs of sssd.conf are not sufficient for your organization (such as if you want to use SSH keys via ldap, sudoers, dynamic dns, etc).

But generally you were seeing a bunch of guides because they were referring to truly ancient distros.

Re: Ubuntu 21.04

#236
post #13

Earlier quoted context omitted.

You always could! sssd’s ad integration has been really solid for years and if you have a huge number of machines you can stand up FreeIPA establish a trust between them. 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.

FWIW, GPO templates are what I consider the minimum to call something manageable, which is most of why I want things on Active Directory. This definitely moves this into "I might introduce this to our environment" level.

The problem is that GPOs are really a terrible way to do configuration, and people (like myself) who grew up on them simply were Stockholm'd into accepting them.

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.

Re: Ubuntu 21.04

#237
post #11

It's interesting to see Wayland enabled by default. 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 wo…

I've been using it with an AMD RX580 driving a 32" 4K monitor for a couple of years now and it just about works perfectly for my needs, including playing a few games through Steam. The only issue I have is that for some reason I occasionally cant alt-tab to the keepassxc window under gnome after I start it up. If you use Linux on the desktop and don't otherwise require an nvidia card (cuda, etc) it is worth while to…

There will be no new builds. There is no hardware :)

Re: Ubuntu 21.04

#238

Earlier quoted context omitted.

XWayland support on NVIDIA should be coming later this year: https://www.phoronix.com/scan.php?page=news_item&px=NVIDIA-G... 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.

So it sounds like folks with Nvidia cards should hold off on this version of Ubuntu?

Nope. Wayland is replaced with XOrg, if NVidia drivers are active. You don't even notice it.

I can switch from Intel only to Nvidia PRIME profile and after restart it just switches to XOrg.

Re: Ubuntu 21.04

#239
post #226

Earlier quoted context omitted.

> where did you hear that CentOS Stream didn't receive security patches? That is false... 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?

5 to 5.5 years - the same as RHEL "full support" phase.

5 years is half of the Ubuntu LTS and the previous CentOS Linux lifecycle. This is why many consider CentOS Stream to be a significant departure from CentOS Linux. Not saying it is a bad OS but it is no longer a free Linux operating system with long term support.

Re: Ubuntu 21.04

#240

Earlier quoted context omitted.

I don't have a firm answer, but considering how much Wayland pushes off on the compositor (seemingly everything, nothing is ever Wayland's fault or responsibility) I'd expect that the compositor (every desktop environment or WM also needs to be a compositor, now) crashing is identical to all of Xorg crashing, not just the DE or WM under Xorg crashing.

> (seemingly everything, nothing is ever Wayland's fault or responsibility) 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.

> 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.

Post reply on HN