CentOS-8 1911 Released As Rebuild Off Red Hat Enterprise Linux 8.1

[anarki2]

In practice, it was released on Monday. At least that's what /etc/centos-release told me that day. Needless to say, I was kinda confused, as there was no official word anywhere yet at the time.

[Britoid]

Canonical gets it worse than Red Hat, I do see plenty of companies who are using Ubuntu, with zero intention of either contributing to Ubuntu or ever paying Canonical a dime.

If you want to contribute to CentOS, contribute upstream.

Red Hat's ethos has always been that if you wanted to improve RHEL.

[anarki2]

You assume wrong. I smell a PEBKAC error.

I don't understand your attitude, CentOS 8.0 factually has at least 2 critical bugs: the first one is the one mentioned by the dude you responded to. Try installing from netinstall, it'll never work without the workaround in that bug ticket. You "smell" that everyone in that ticket is an idiot? How about you try it for yourself, genius?

Second, the kernel never updates. Just never. Unless you stumble upon this article:

https://wiki.centos.org/Manuals/Rele...s/CentOSStream

You're left with an unpatched kernel. Sweet, huh?

Just for sh*ts and giggles, CentOS 8 also breaks in UEFI mode in Hyper-V. It only boots like every 20th time, until then, it's a reboot loop. Too bad Hyper-V stops the VM after like 5 failed attempts, so often times, your VM will fail to reboot or start up, and you have to manually try starting it up until it finally succeeds. You have to do with to every C8 VM, after every reboot.

Fourth, the installer explicitly asks you if you want to enable AppStream, yet it enables it every time, even if you say NO. Which wouldn't bother me this f*cking much if AppStream wouldn't break various 3rd party repos, like PostgreSQL.

https://noobient.com/2019/11/26/post...-8-and-rhel-8/

These four issues all came from my VERY limited testing so far. I can't imagine how many bugs are out there I didn't have to deal with yet.

TLDR C7 was extremely reliable, but C8 is definitely a heck of a mess at the moment.

[kgonzales]

I don't understand your attitude, CentOS 8.0 factually has at least 2 critical bugs: the first one is the one mentioned by the dude you responded to. Try installing from netinstall, it'll never work without the workaround in that bug ticket. You "smell" that everyone in that ticket is an idiot? How about you try it for yourself, genius?

Second, the kernel never updates. Just never. Unless you stumble upon this article:

https://wiki.centos.org/Manuals/Rele...s/CentOSStream

You're left with an unpatched kernel. Sweet, huh?

Just for sh*ts and giggles, CentOS 8 also breaks in UEFI mode in Hyper-V. It only boots like every 20th time, until then, it's a reboot loop. Too bad Hyper-V stops the VM after like 5 failed attempts, so often times, your VM will fail to reboot or start up, and you have to manually try starting it up until it finally succeeds. You have to do with to every C8 VM, after every reboot.

Fourth, the installer explicitly asks you if you want to enable AppStream, yet it enables it every time, even if you say NO. Which wouldn't bother me this f*cking much if AppStream wouldn't break various 3rd party repos, like PostgreSQL.

https://noobient.com/2019/11/26/post...-8-and-rhel-8/

These four issues all came from my VERY limited testing so far. I can't imagine how many bugs are out there I didn't have to deal with yet.

TLDR C7 was extremely reliable, but C8 is definitely a heck of a mess at the moment.

Sounds like everyone using CentOS got what they paid for.

[anarki2]

Sounds like everyone using CentOS got what they paid for.

Or... you could just admit you're a cynical clueless idiot adding nothing to the conversation except low effort trolling.

[Zyklon]

So set aside all that stupid bullshit gaga above:
Does anyone know why just some parts of the Gnome System have been updated to Gnome 3.32 and why other programs still are 3.30 or 3.28?

[CommunityMember]

Does anyone know why just some parts of the Gnome System have been updated to Gnome 3.32 and why other programs still are 3.30 or 3.28?

It is not an entirely satisfactory reason, but that is what upstream has (CentOS follows RHEL). If you want to know why RH chose those versions you are probably going to have to open a case under your RH support contract to get a definitive answer (rather then just guesses).

[mroche]

Second, the kernel never updates. Just never. Unless you stumble upon this article

That article is specific to the Stream ISOs, the standard ISOs work just fine, and adding the Stream repos doesn't affect anything. Have you tried to see if the CentOS Stream 191219 iso's rectify that issue?

Try installing from netinstall, it'll never work without the workaround in that bug ticket.

This is solved with the 8.1.1911 boot image. Source: I just did while only doing the two following things: enabled network device and set the package source to "Closest mirror". Chose Minimal as the installation type and kicked it off just fine. It installed the latest packages without issue.

Which wouldn't bother me this f*cking much if AppStream wouldn't break various 3rd party repos, like PostgreSQL.

The PostgreSQL team tells you to do disable the conflicting module on the current download page. Whether or not that was added after Nov 26 (time of that post), I can't say for certain. The Wayback Machine's last archival of that page was from before PG12 was released (Jul 22), but it didn't have a module step in their instructions. If a 3rd party repo conflicts with AppStream, it's up to those parties to alert users to that.

https://www.postgresql.org/download/linux/redhat/

However, a contributor noted this on Jul 25th on Planet PostgreSQL. I'm assuming based on the contents of the post, that a decision was made before the release of PG12 and that the download instructions reflected that (as they currently do).

https://people.planetpostgresql.org/...on-RHEL-8.html

And as linked in that post, a corresponding Bugzilla report from mid-June:

https://bugzilla.redhat.com/show_bug.cgi?id=1718201

Just for sh*ts and giggles, CentOS 8 also breaks in UEFI mode in Hyper-V. It only boots like every 20th time, until then, it's a reboot loop. Too bad Hyper-V stops the VM after like 5 failed attempts, so often times, your VM will fail to reboot or start up, and you have to manually try starting it up until it finally succeeds. You have to do with to every C8 VM, after every reboot.

Just for personal curiousity, as I don't use HyperV, have you tried this with the 8.1 release? If it's still occuring and you have time, I would try this with RHEL 8.1 (using dev subscription) to validate that it's just a CentOS issue. Have you cross-checked your process with Red Hat's instructions from 8.0?

https://developers.redhat.com/rhel8/...-rhel8-hyperv/

Cheers,
Mike

[Paradigm Shifter]

Have to say that release 8 has been a complete mess in my experience as well with CentOS. From having to manually search out the details of repos just to get it to install with the netinstaller, to having to go to a second console and restart systemd-logind or otherwise get a black screen after login... CentOS8 has been a wholly unsatisfying experience.

I'll give 8.1 a go, but it isn't high on my priority list simply because of the hot mess the previous release was (also goes for RHEL8 as well...) also, I will never forget all the fun I had when RHEL would always kernel panic during boot (even to install) on the SuperMicro socket G34 boards I was running. That pushed me away from RedHat for years, and once you've found something else that works, it is hard to justify the shift back.

I really need to pick up a junk laptop or something that I can leave "broken" for days when something goes wrong, because potentially losing a production system is infuriating for a supposed Enterprise OS.

[ZeroPointEnergy]

What you don't see and the "sales suits" do see are the very large companies who are literally telling Red Hat that they will use CentOS, they will not contribute to open source (not even to CentOS itself), and they don't see a reason to change while CentOS exists. And the cost for THAT is immense. It's in the tens of millions of dollars per year sucked out of the open source ecosystem.

That is some A grade linear thinking there, if you think getting rid of CentOS would translate into more sales. Are you one of those sale suits?

First, if there is no money or no need for support there will simply be no sale. If CentOS doesn't exist they would use Debian or something else. BUT the second that changes and they need support, and that happens all the time, if they use CentOS they will change to RHEL. If they don't use CentOS or it doesn't exist or is not compatible anymore they will most likely not.

The availability of free distros is not "money sucked out if the open source ecosystem". It is a prerequisite for that market to even exist. If the was no free distro we could as well just still be stuck with AIX, Solaris and the like and then no money would ever make its way into companies that translate that into work on open source software.