Dear CentOS Community,
As we enter the new year, I'm sure there's really only one thing on your mind, and so we'll start there.
As you are no doubt aware, the CentOS project has shifted focus from CentOS Linux - the RHEL rebuild - to CentOS Stream - the continuously delivered distribution that reflects what will be delivered in the next release of Red Hat Enterprise Linux (RHEL).
Many, many articles have been written about this, and I want to take an opportunity to call out some of the better ones, to help you understand what's happening, and where we go from here.
To those who claim that CentOS Stream will be somehow unstable, I would encourage you to read Brendan's article about how RHEL is made. Things that go into RHEL are not bleeding edge or continually shifting sands. They are small incremental changes which have been baked for a long time.
To those objecting to the term "rolling release", see Stef's article about continuous delivery, and how CentOS Linux and CentOS Stream related to RHEL.
And to those who are pre-judging CentOS Stream without the benefit of even trying it, you should read Jack's article about not knocking it until you try it. (Jack's an Ubuntu fan, but makes a lot of good points.)
Karsten has written an article about the various things that are kept in balance around the CentOS project, and some of the history that led to where we are.
Finally, Scott's article about ... well, all of it ... is definitely worth your time if you want to have a deeper understanding about why people are angry, and why they are right, and wrong, to be angry.
For those of you who are planning to move to Rocky, CloudLinux, or one of the other projects that has sprung up to take the place of CentOS Linux, we wish you - and these projects - all the best. But we caution you to understand that building an OS is a big project, and it's going to take a while for them to get where they're going. Please plan your migration accordingly.
There are other things happening in the CentOS community, but we understand that this one is pretty overshadowing right now.
Hyperscale SIG proposed
A group of developers has proposed a Hyperscale SIG, which will be voted on in Wednesday's board meeting. They propose to focus on solutions around large-scale infrastructures, such as those at organizations such as Facebook and Twitter.
If you are interested in this kind of SIG, and particularly if you are running a hyperscale infrastructure, we welcome your comments and participation.
CentOS Linux 8 (20-11) released
The fourth release of CentOS 8 is now available, as of December 7th. This release is labelled 8.2011 (ie, November 2020) and is based on the 8.3 release of RHEL.
Q1 CPE Priorities
In Q1, CPE will be working on the following priorities:
- CentOS Stream
- Noggin/AAA replacement
- Fedora-Messaging SchemasÂ
- Flatpak indexer code merge
- Debuginfo-d
- Datanomer & Datagrepper V.2
We'll be updating the centos-devel list as progress is made on these projects.
Happy New Year
We wish you a 2021 that is happy and productive, and hope to see you in person before the year is out. Thanks, as always, for being part of our community.
It's 2021, 🙂
Looks like the title corrected by the author.
After killing CentOS Linux with the worst possible timing and brigading the internet with Red Hatters that tell us how beautiful that murder was, could you at least spare as the FUD ("we caution you to understand that building an OS is a big project").
Thank you.
-- Chris
PS: My 2021 is indeed going to be very productive. Migrating systems off CentOS. You can be sure it won't be towards RHEL.
You hit on the things that I saw Chris.
1) Paid employees telling us how good this thing is. In all the uproar, there have been a few non-RH employees defending the move, but not very many.
2) Passive-aggressive comment about how hard it is to build an OS.
I read the links, and Scott's article had a lot of "But you're not paying for it!" from him in the comments. Truth will out.
Thanks for compiling a list of articles to be read but the damage has already been done and the trust has gone.
There is no point in staying with CentOS so the sooner Rocky get's into the ring the better.
Insisting in telling us how benefical the shift to CentOS Stream will be for CentOS Stream means that you aren't listening to the comunity at all.
We are all sure that this will be great to CentOS Stream, but we wanted and needed CentOS, not CentOS Stream.
Stop telling us we need CentOS Stream: it was already available before and I've never met anyone that was actually using it.
I use CentOS Stream but it is very unstable indeed.
I even tried to install new ISO versions of CentOS Stream and they simple don't find the package sources.
Let's see if there are something useful anyway.
After using CentOS for 10+ years you gave me the chance to look at other server OS. I never knew how well Debian performed compared to CentOS. Thanks RHEL.
CentOS Stream 9.0 release in 2021 Q2?
kernel version 5.11 up?
We were on CentOS until recently, to which we had been migrated from CAOS Linux, to which we had been migrated from Whitebox Linux, to which we moved from Redhat Linux when they discontinued that in favour of RHEL and its subscription model. So, we were a long time RH / RH variant user.
However, for us the writing was on the wall when RH took over the CentOS project as a corporate product. The introduction of systemd with centos-8 in itself had little to do with our later decisions but it was the proximate cause of re-evaluating RH as an unwanted business partner. Which RH had become by virtue of taking over CentOS; and let us be clear, that is exactly what happened as is evidenced by these events.
So, this current development comes as no surprise to us. The 'stream' approach to binding your customers in an intractable web that requires constant near real-time monitoring by the software distributor, with the unavoidable privacy and security implications appertaining thereto, is evident in Microsoft 10 (and every app downloaded to a smartphone). RH stream will follow much the same path, as will RHEL or whatever RH chooses to call it. At some point, the insatiable desire for 'rents' may see RH move entirely into the "cloud" and metal hosted user systems deprecated.
Recall that a commercial corporation has but one objective: to get the most money, from the most people, for the longest time, at the least cost to itself. Every action taken by a corporation has to be viewed in the light of how that act furthers the corporation's goal. Every other motive publicly claimed is no more than propaganda when it is not outright deceit.
The existence of CentOS in its present form is undoubtedly viewed as a parasitic loss of revenue by RH owners/management. That they themselves depend upon the vast body of intellectual material provided at no cost by countless others only registers as an irrelevance, if it registers at all. This outcome was inevitable from the moment CentOS was surrendered by us to RH.
Brilliantly worded.
The big difference I found between 8 and 8 stream is that some services and applications seems to use debug logging by default. If I am to migrate to steam I need to better understand what differences there will be between stream and rhel so I can config as needed. I recall some debug logs creating issues for me in the past.
This WILL result in short-term gains for RedHat sales, as some will be forced to purchase RHEL licenses. However, RedHat botched the messaging and DID NOT have plans in place to help those using CentOS for cost/sensitive applications migrate to RHEL.
CentOS has SIGNIFICANT market share, far greater than RHEL. Many businesses rely on it and many supercomputers run it. Migration to anything else, including RHEL, will incur significant cost which is difficult for non-commercial entities especially in these Covid-challenged times.
Destroying CentOS destroyed goodwill. This will be VERY HARD, if not impossible, to replace and will cause long-term damage to RedHat and IBM. Some customers will spurn RHEL even if offered inexpensive licenses.
Many had planned for a long life cycle for CentOS 8. For example my products run for over 10 and I need a stable, very-long-term O/S. CentOS was our solution due to cost. These changes ARE CAUSING PROBLEMS with my customers and our teams.
We are NOT cloud-based because of hardware constraints, so we run on bare metal or in simple VMs. We feel leftout.
Issues:
1. horrible messaging
2. lack of RHEL migration options
3. not listening to CentOS community
4. significant loss of goodwill
5. loss of trust in RedHat
6. loss of options for long-term
I expect there will be many business school papers describing what went wrong with this move and how it had a significant negative impact on RedHat long-term.
I hope I am wrong, but time will tell.
I am curious why people don't seem to be willing to pay Red Hat for RHEL when they are running a for-profit business or well funded education or government program.
Usually the same people pay for hardware, and other systems software or application software, and often consulting. When Red Hat creates a product that these users find valuable, and does the work to build a compatible ecosystem of hardware and software, and provide updates, and security fixes.
Yet, it seems to me many people think that they should be able to use something for free instead of paying, what usually amounts to 1-5% of the total cost of the hardware and software running on systems, for the capabilities of RHEL.
It seems to me if people did not value the work Red Hat did to create RHEL, and a primary objective was to "not pay for Linux", they would have already been using a Debian based version of Linux. But I am assuming they chose CentOS because it was RHEL-like.
They have always had, and still have, the option to pay a small amount of money for RHEL.
For true non-profit, developers of all types, and small business type users I am hopeful Red Hat will offer something free or very very low cost. It would be smart to do so.
But all the for-profit business users, I think the free-ride of Red Hat value should be over. My advice is -- Get on with it, and don't sweat the small stuff (the 1-5% of your cost) ! Go use Linux to solve much bigger challenges and problems.
Just to throw in another data point:
RHEL on AWS is 626% for a t3.micro or 156% for m5.large compared to a free Linux baseline.
I need to get to a m5.metal to get down to a +2.5%.
(on demand in eu-west-1)
-- Chris
You are comparing the cost of RHEL to the cost of the VM only I think in that analyis.
Which in some cases may be the entire the cost for a company.
But I think most companies have costs for other software, or costs for developers to write software, costs for hardware (or VM's if using cloud), and possibly costs for consultants to implement the software and applications.
The the point I was trying to make, is for people to look at the cost of RHEL, as part of the cost of the overall solution, and see if the % is worth the value that one gets from it. If not, then using Debian based alternatives would seem appealing. But if someone wants the widespread compatibility of RHEL, I still think they should be ok paying for it, just like the pay AWS for VM, or AWS for a database service. Maybe all the people that think like me are already paying Red Hat and are not involved in the CentOS dynamics.
Because the decision to use CentOS was often made under some constrains (budget, for example) and under some general rules and expectations. CentOS was fit for the purpose, RHEL was not.
Then the rules changed. People who promoted CentOS inside their organizations and for their customers, were burned. Expecting any love for Redhat now from them is delusional. Now they won't be giving Redhat any money by principle.
Congrats, Redhat, burning bridges en-masse is not a small feat.
I continually see reference to the financial impact of CentOS, but it should always be reiterated that Red Hat explicitly has kept CentOS from accepting direct financial donations. The financial issues from earlier in CentOS' life were no longer relevant.
If RH wanted revenue neutrality for the costs of running the CentOS Project, it should have reached out to the EL userbase years ago.
This only leaves CentOS Linux's negative impact on RHEL sales. Plenty of discussions have occurred on the centos and centos-devel lists about the problems that exist with the RHEL subscription model -- CentOS Linux's mere *existence* doesn't change that, since it can be duplicated with a crowd-funded effort elsewhere. OTOH, the negative effects of this move impact goodwill toward RH, and by extension reduce others' willingness to be more open with their pocketbooks than they need to be.
I find it very fascinating that out of all the uproar, it's paid Red Hat folks that are defending Stream with very, very little people on the outside even defending it at all. The amount of disdain and distrust you've created is actually impressive. Imagine all the people walking away from the EL ecosystem as a whole because of this move.
You might as well rename CentOS to something else, since the "C" for "Community" isn't actually a reality with this project, your team, and with Red Hat.
I have to comment on that. Although it is not what I wanted. I live in Brazil. And with that decision you were able to overcome this current president who is governing. Because the president of my country is stupid. I have always believed in CentOS. I indicated. But are you still believing that these changes will attract users? I have already migrated a machine to Debian. Although it was not what I wanted. There are other good distributions, Ubuntu, Debian ... I simply believe that you lost CentOS and RedHat. I was not the one who made the decision to change one of the best Linux distributions. They bear the consequences in the future.
We are switching to Oracle Linux which is a good alternative to Centos Linux.
One thing I caution you on is that if you don't buy other things from Oracle - like their databases, applications or hardware, don't expect to be treated with any priority for anything. Also, I would be fairly leery of them continuing to support something that makes them no money. All I am saying is be aware.
I can't help but facepalm when we are being treated like casual Windows users. We are all aware of what CentOS stream is! The only one confused here is RH. And for good reason (they aren't paying attention!)
Both CloudLinux and RockyLinux have plenty of experience in this regard. In the case of CloudLinux, most of their work is already done cause they do it for cloudlinux already.
On that topic, CloudLinux has given an official name to their OS(Lenix was a temp internal name), It is called Alma Linux
But its good to know that RH is worried about us planning our migration. Wish they were even 1/1000th as worried when they decided to end support for CentOS 8 in only 1 year, before even CentOs 7! And right after people migrated from EOL CentOS 6!
This issue they still refuse to address and pretend it doesn't exist.
given the regression my proposal for a new name :
BentOS
for Beta Enterprise Operating System
USD$350 per server per annum is simply too expensive for a struggling small business.
The good news is that there's many other viable options, both existing and emerging.
The $349 option is only applicable for instances deployed on a physical system. RH also state it's not intended for production environments.
For any virtual / cloud use (which will be the vast majority), it's a minimum of $799 per year, per instance.