gyptazy.ch is a Fediverse instance that uses the ActivityPub protocol. In other words, users at this host can communicate with people that use software like Mastodon, Pleroma, Friendica, etc. all around the world.
This server runs the snac software and there is no automatic sign-up process.
Coming to Red Hat Summit in #Denver? Join us for #Openshift Commons as well!
How to install/use Kubernetes Control Plane in FreeBSD with Linux worker nodes and Cilium - Article by Tony Norlin @tnorlin #Kubernetes #K8s https://medium.com/@norlin.t/kubernetes-on-freebsd-with-linux-worker-nodes-and-cilium-a87c50daef03
Interesting reading: Kubernetes and back - Why I don't run distributed systems
https://www.davd.io/posts/2024-03-20-kubernetes-and-back-why-i-dont-run-distributed-systems/
Being too long absent in this topic feels like starting from scratch again…
Being too long absent in this topic feels like starting from scratch again…
Why the hell is your Kubernetes API public?
https://leebriggs.co.uk/blog/2024/03/23/why-public-k8s-controlplane
Discussions: https://discu.eu/q/https://leebriggs.co.uk/blog/2024/03/23/why-public-k8s-controlplane
Typical modern software stack:
- They only provide a Dockerfile because it's so convoluted and patched together that it would be impossible to create a decent guide on how to make this stuff work without revealing just how much it's all held together by duct tape. Still, taking a peek at the Dockerfile will clearly reveal the chaos at play.
- Everything runs in separate containers, as it should, and they have automatic restarts, as they should. But then they're so tightly bound together with, you guessed it, duct tape, that when one of them fails, the container just restarts automatically, leaving the user wondering why everything failed, with only a cryptic error message.
- The main container is controlled by Supervisor - which is necessary because everything crashes very often, but the end user doesn't notice because Supervisor restarts the services. To outsiders, everything works fine (but occasionally throws some errors).
In short, the current trend is to sweep the dirt under the rug. A trendy rug, though!
#ModernSoftware #TechTrends #SoftwareStacks #Docker #Kubernetes #IT
Thinking through Yi Tay's post on managing infrastructure for compute heavy ML workloads. I'm utterly fascinated by the pre #kubernetes developments with Borg/Omega for batch processing. The infrastructure work behind borg/omega is mind-bogglingly complex. FWIW, so is Kubernetes.
[1] https://www.yitay.net/blog/training-great-llms-entirely-from-ground-zero-in-the-wilderness
[2] https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/44843.pdf
[3] https://storage.googleapis.com/gweb-research2023-media/pubtools/pdf/43438.pdf
But yeah, #BSD based systems like #FreeBSD or #NetBSD still work the same way like like 20 years before. For #Linux - I think #comtainers / #Kubernetes are/were the hot sh…. Things are more moving to #AI I guess… let’s see what the next big thing will be in a few month after the hype.
Software installation is to complicated. Lets use #docker.
Managing docker is to complicated. Lets use #kubernetes.
Managing kubernetes is to complicated. Lets use #portainer.
to be continued … 🖕
@lproven @theregister Plan9 is an #opensource OS with 38 system calls & few #linesofcode . Is not #UNIX . Need #Linux for VM only, to allow plan9 to run existing programs #mx #cluster #systemcalls #plan9 #filesystem #containers #docker #9front #kubernetes #snap #flatpak #git #unix #inferno #leansoftware #Oberon #a2 #multithreading #debian12 #Chrome #linesofcode #moreslaw #andalslaw #parallel_programming #softwaredevelopment #macOS #hypervisor #MicroVM #cubesos #linuxvm #pixie #zen #VMWare
Doug Rabson's (dfr) work building OCI container images of #FreeBSD for #podman. He unpacks what an OCI compatible container is. And what making light-weight container images of FreeBSD looks like. OCI containers are a pre-requisite for running #Kubernetes natively on FreeBSD. As is a container engine like podman.
@chris Apologies, but my German is relatively poor.
If I understand correctly, sending a task not simultaneously makes sense under maintenance.
I'm an #OpenStack and #Kubernetes expert. You want to update the master node one at a time, so that you don't lose your quorum while doing update. That's an example. On a large scale cluster, you don't want your repository to be flooded (ex: apt or docker registry) while performing the upgrade.
I am sorry if I mistaken the question.
@meka -- I was excited about this too. I think it's just the control plane for #kubernetes that is running on #FreeBSD. But it's also possible to bootstrap a node pool of #linux machines through #bhyve. With well-tunned Linux VMs, this could be a mighty on-prem compute cluster.