CERN's Quiet Exit from Red Hat Exposes a Decade-Old Mistake
Abandoning Scientific Linux Was a Mistake
CERN is moving over 2,200 accelerator control systems to Debian 13 by the end of 2026, after RHEL's CPU baseline changes threatened costly hardware replacements. The author argues that abandoning Scientific Linux a decade ago was a mistake: it wasn't just duplicated engineering, but infrastructure that gave the scientific community a credible exit from vendor lock-in. That option value vanished when CentOS changed, and rebuilding it is far harder than keeping it alive.
Scientific Linux gave the scientific community its own implementation of a RHEL-compatible computing environment. More importantly, it kept alive the people, processes, infrastructure, and institutional knowledge needed to maintain one. That capability had option value.
- darkwi11ow
Scientific Linux was great! Sad to see the blog in its memory having so much llm-speak. I hope one day writing without llm will become a trend.
- forgotpwd16
Article quite a mess and, honestly, largely devoid of content. For starters, complains that SL should've continued because other community options might eventually be discontinued/change (as happened with CentOS). Like, duh. By this reasoning every effort should be duplicated to guard against the possibility that one option might disappear. In the end though all those RHEL rebuilds depend on Red Hat maintaining RHEL. The article never notes this and indirectly equalizes the effort and existence of a a genuinely independent community-driven distro (Debian, that CERN decided to move their accelerator-control computers) and RHEL rebuild (e.g., AlmaLinux, that CERN will keep the rest of infrastructure on, e.g. data centers, compute farms, etc). And here's the misguided part. If SL as still active it'll have been necessary to recompile numerous upstream RHEL source packages to remove the microarchitecture restrictions. This will've turned SL to entire new, independently maintained distro. The thing they wanted to avoid.
- _joel
Well they could have used Debian from the start. I get why, but it's such a strange argument, the way it's posited.
- intoXbox
The benefits of SL described are pretty handwaivy to me. Having used a fork of scientific Linux, this was a huge pain because of outdated openSSL packages. Now you could argue this deserves being maintained separately, but I think it makes total sense to use the security- benefits of long term distros and only modify the parts required for reproducible scientific work.
I was really glad when the vendor started unofficially supporting OpenSuse so I could finally use git on these distros.
- tempay
I think the premise of this article is wrong. Scientific Linux was important once upon a time but these days with userspace containers there is no need for the same level of homogeneity. In practice everyone is using Apptainer on machines that don’t have a singular purpose.