Red Hat's Karsten Wade, a Senior Community Architect and member of the CentOS board, has defended the decision to kill off CentOS Linux in favour of CentOS Stream, saying the two projects were "antithetical" and Stream is a satisfactory replacement in most cases.
CentOS Linux is downstream of Red Hat Enterprise Linux (RHEL), whereas CentOS Stream, introduced in September 2019, is upstream, a late development build of what will shortly go into RHEL (unless problems are discovered).
All CentOS variants are free, and CentOS Linux is understandably popular, combining the stability of RHEL with free availability. For example, according to statistics from W3Techs CentOS has an 18.5 per cent share of websites, compared to Red Hat's 1.5 per cent share. Earlier this month Red Hat declared that CentOS Linux would be phased out in favour of Stream.
Wade explained the necessity for CentOS Stream as a way of making it easier for the community to contribute to RHEL. He also said that "as a project, trying to do two antithetical things at once would mean doing both poorly," implying this was the reason for the abandonment of CentOS Linux.
He confirmed the decision was driven by Red Hat, which "approached the CentOS Project with its plan" but said "the CentOS Board signed on to it."
Acknowledging that the lack of CentOS Linux creates an "availability gap," Wade nevertheless said he is confident that Stream can cover "95 per cent (or so) of current user workloads" and referred to a post by Stef Walter, Director of Linux Engineering, which described Stream as RHEL with a continuous delivery model, stating: "The whole point of continuous delivery is to make each release as stable as the one before."
Wade also said Red Hat will make additional solutions available – presumably meaning more affordable licensing for RHEL in some scenarios.
Is concern about doing two things badly really the reason for scrapping CentOS Linux, or is it an attempt to sell more RHEL licences? The community is not convinced by Wade's argument that the two are antithetical.
There is particular upset that CentOS 8 support has been curtailed. "People are complaining because you are suddenly killing CentOS 8 which has been released last year with the promise of binary compatibility to RHEL 8 and security updates until 2029," said a comment on Wade's post.
Maintaining an open source project such as RHEL involves a complex balance of commercial and community considerations. Red Hat's success has hinged on its ability to manage this. Red Hat builds on work freely given by others; equally those who build free distributions from the work of Red Hat engineers are in a sense riding on that commercially-supported input. When we asked Hayden Young, part of the team for the would-be CentOS replacement Rocky Linux, how if at all Red Hat would benefit from the fork, he said that "In some respects, we don't give much back to them at all … but I think we're creating a project that's going to make people say, I like this, the paid one is probably going to give me so much more."
The difficulty for Red Hat is that while from a commercial perspective it may hurt to support a project that creates a free alternative to its main commercial product, the risk is that without CentOS Linux, users will shift to alternatives to RHEL as well.
"I have over 300,000 Centos nodes that require Long term support as it's impossible to turn them over rapidly. I also have 154,000 RHEL nodes. I now have to migrate 454,000 nodes over to Ubuntu because Redhat just made the dumbest decision short of letting IBM acquire them I've seen … nothing like millions in lost revenue from a single customer," said another commenter on the post.
Making major changes to an open source ecosystem that works is not without peril. ®