tldr: their infra that interfaces directly with accelerator uses really old computers and they are unsupported in newer versions of rhel. Only those machines are affected by these good news.
Personally, this looks weird, they are planning to run like 20 years old machines for another 10 to save money on replacing them while at the same time they are planning (or already in the process) migration to k8s?
How much RAM do you think they have? I think a top-notch scientific machine from literally 20 years ago might have like 4GB RAM, maybe a bit more. I'm not sure about non-control-plane kubelet runtime overhead, I see online sources showing numbers as little as 30MB, and as big as 1GB. But even with 1GB you will still have 3GB for work.
For regular k8s a few years ago (in my own tests) it hovered around 500-600 megabytes, and some cpu as well (can't say how much would it use on machines this old) so the thing that does nothing but controls what will run on a machine consumes like 1/8th of ram. Also their infra sounds as static as it gets and machines themselves are not pets - they are family members at this point, what would the benefits of k8s be?
Compared to what? Manually running software on machines? Automatic distribution of work, automatic networking, linking/transfer of resources, etc. I think their scenarios are often oneshot tasks that need to run somewhere and produce results that get used in other tasks or aggregated somewhere, etc.
what a weird title. the actual article is fact base but the title is just weird. why is it good (or bad) news? why not just say "CERN moves over 2,200 control system to drbian 13 to support the older hardware"
Almalinux is used for desktop computers (along with RHEL and windows).
This is about the control system computers, which have dedicated hardware for various tasks.
It was planned for them to also use almalinux, but they dropped support for a bunch of old hardware architectures. This made Alma unsuitable as some of these custom computers just would not be supported. The cost/risk of redesigning just would not make sense.
Hence: plan B, which is Debian and its many supported hardware architectures.
16 Comments
esc@piefed.social · 17 pts · 23d
tldr: their infra that interfaces directly with accelerator uses really old computers and they are unsupported in newer versions of rhel. Only those machines are affected by these good news.
Personally, this looks weird, they are planning to run like 20 years old machines for another 10 to save money on replacing them while at the same time they are planning (or already in the process) migration to k8s?
lil_baka@ani.social · 5 pts · 22d
What's the problem with running those 20 year old machines as kubelets?
esc@piefed.social · 4 pts · 22d
It actually requires not an insignificant amount of resources for 20 y.o. machines?
lil_baka@ani.social · 4 pts · 22d
How much RAM do you think they have? I think a top-notch scientific machine from literally 20 years ago might have like 4GB RAM, maybe a bit more. I'm not sure about non-control-plane kubelet runtime overhead, I see online sources showing numbers as little as 30MB, and as big as 1GB. But even with 1GB you will still have 3GB for work.
esc@piefed.social · 4 pts · 22d
For regular k8s a few years ago (in my own tests) it hovered around 500-600 megabytes, and some cpu as well (can't say how much would it use on machines this old) so the thing that does nothing but controls what will run on a machine consumes like 1/8th of ram. Also their infra sounds as static as it gets and machines themselves are not pets - they are family members at this point, what would the benefits of k8s be?
lil_baka@ani.social · 5 pts · 22d
Compared to what? Manually running software on machines? Automatic distribution of work, automatic networking, linking/transfer of resources, etc. I think their scenarios are often oneshot tasks that need to run somewhere and produce results that get used in other tasks or aggregated somewhere, etc.
SlurpingPus@lemmy.world · 2 pts · 22d
There's decades of effort already done on this kind of software in science. Idk why Kubernetes would suddenly be needed for that.
user_123@lemmy.world · 1 pts · 22d
exdor@programming.dev · 3 pts · 22d
Lol that is pretty important context XD
terabyterex@lemmy.world · 11 pts · 22d
what a weird title. the actual article is fact base but the title is just weird. why is it good (or bad) news? why not just say "CERN moves over 2,200 control system to drbian 13 to support the older hardware"
illusionist@lemmy.zip · 9 pts · 23d
Supporting red hat costs 5.5 Mio for the next ~8 years, and the switch to debian? What would replacing the old hardware (~20 years old) cost?
It's bad for red hat, good for debian and suse wasn't even considered?
Tabula_stercore@lemmy.world · 9 pts · 22d
Weird. I thought CERN used alma linux; https://linux.web.cern.ch/almalinux/
Scafir@discuss.tchncs.de · 15 pts · 22d
Almalinux is used for desktop computers (along with RHEL and windows). This is about the control system computers, which have dedicated hardware for various tasks. It was planned for them to also use almalinux, but they dropped support for a bunch of old hardware architectures. This made Alma unsuitable as some of these custom computers just would not be supported. The cost/risk of redesigning just would not make sense. Hence: plan B, which is Debian and its many supported hardware architectures.
kiri@ani.social · 9 pts · 23d
CERN's Black Holes now will be based on debian 13)
gole@lemmy.zip · 4 pts · 22d
Aside: I recently fired up a server with Debian 13 and the size was only 900MB compared to 2G of Debian 11, 12 and Ubuntu. Anyone knows what changed?
Zubgub@lemmy.blahaj.zone · 3 pts · 22d
13 dropped i386 support, but I'm not sure if that would account for that much.
somegeek@programming.dev · 1 pts · 22d
Debian my beloved
ultimate_worrier@lemmy.dbzer0.com · -2 pts · 22d