Wow. What a week. I’m almost not sure to start but let’s give this a go.
Red Hat’s had a pretty hard week convincing Centos users that their announcement Tues (15th Dec) deprecating Centos 8 (and Centos downstream in general), is A Good Thing(tm). How did this come about?
Centos is one of the most used enterprise-level Linux distribution without being commercial. It’s used by anyone from tinkerers to large enterprises, and everything in between. Many have staked their companies and products on Centos, a dependency which has now landed them in a bit of hot water.
Centos started out purely as a community-based binary-compatible release of Red Hat’s RHEL in the 2000’s. But after a few community issues, an offer from Red Hat to’ partner’ in the production of the distribution seemed like a very good idea. And it has been so far, as Centos devs have had employment, and Centos has gained structure and quality control under the stewardship of Red Hat.
But Red Hat has now decided that the downstream version Centos 8 will be deprecated as of Dec 2021 (the EoS date for Centos 8 was originally 2029) in favour of Centos Stream, a rolling-release version of Centos that was introduced in Sep 2019. Red Hat CTO Chris Wright indicated that Stream will not be a replacement for CentOS Linux.
CentOS Stream isn’t a replacement for CentOS Linux; rather, it’s a natural, inevitable next step intended to fulfill the project’s goal of furthering enterprise Linux innovation. Stream shortens the feedback loop between developers on all sides of the RHEL landscape, making it easier for all voices, be they large partners or individual contributors, to be heard as we craft future versions of RHEL.
Essentially Centos Stream is upstream of RHEL and will be a beta version for the commercial solution. Reactions?
The community (as well as many small and large commercial operations using Centos) are not happy. There’s been a very vocal outcry regarding the announcement. And Red Hat are lucky that they don’t have a comments section on the corporate announcements page. I think the most important factor here is not that Red Hat has made this change, but they’ve done it in such a way that essentially pulls the rug out from under the feet of users, with very little notice. While Centos 7 support continues until 2024, the fairly new Centos 8 only has 1 year left. And initially at least, Red Hat’s indication of alternate solutions is quite non existent at the moment.
On the other hand, and in mitigation, Red Hat has every right to make this change. They own the Centos project and can do with it what they like. Some might say that they’ve been used by the community with little return, seeing as there’s little to no community input to Centos. And this may be a reason for the change. Stream should provide more of an opportunity for community input.
Chef cofounder Adam Jacob just released a statement saying he believes this model is the correct one:
“If I do an open source strategy for a company ever again, I will own the upstream, it will be fully open source, and I’ll happily collaborate with anyone downstream.” But not just an open upstream–it’s also important to, “Produce a commercial distribution collaborate on downstream non-commercial ones, in the open,”
I happen to think this is the correct way as well. But as mentioned previously, the way Red Hat have gone about the change is not great. Is this an example of their new master IBM’s style of operation?
So how do existing Centos 8 users proceed?
In some cases, Centos 8 could be swapped out to Centos Stream with little change or impact. But for most, the beta-ish nature of stream won’t be good enough for their requirements.
RHEL’s source remains (and always should remain) open, so there’s nothing stopping other’s from just putting up another Centos … and that’s exactly what the community has gone and done. In short order too.
Starting with Rocky Linux, Centos’ original co-founder Gregory Kurtzer has initialised this project in honour of the other (now-deceased) co-founder Rock McGough. And the project has won a major amount of support in a very short period time. This distribution is likely to be a natural extension and conversion of Centos 8 for most of the community looking to stay on the RHEL platform.
Project Lenix has been created by the Cloud Linux team, noted for their Centos-optimised distribution for cloud environments. This may be another natural route for existing users.
The next option is Oracle Linux which has been available since 2006. It is 100% application binary compatible with Red Hat Enterprise Linux (RHEL) and it provides an equivalent to each RHEL release. And no, you don’t need to sign any agreement with Oracle for using Oracle Linux. The only decision here is how comfortable you are with Oracle’s stewardship of this product?
There are numerous other non-RHEL options out there including Ubuntu, Debian and Suse, however the migration path for these will not be seamless. Ubuntu though is likely to be the most popular option for those not choosing the RHEL-based route.
But we’ll have to see how this plays out over the next few months. Will Red Hat come to the party with options? Or will migrations to RHEL-based forks be the main option?