Attention to anyone using
#kubernetes on
@alpinelinux - kubernetes 1.34 has some breaking changes in the kubelet configuration. We have updated the ungleich kubernetes handbook on
https://redmine.ungleich.ch/projects/open-infrastructure/wiki/The_ungleich_kubernetes_infrastructure#How-to-upgrade-a-kubernetes-cluster with the instructions to upgrade.
The ungleich kubernetes infrastructure - Open Infrastructure - ungleich redmine
Redmine
Cluster usage of the
#p10 #kubernetes cluster as of today: 819GB RAM actively and 30 CPUs actively used. Thus about 27GB of RAM per active CPU are used. Over are the times that the 1 CPU / 2 GB RAM rule was appropriate.
Axpo verified - they made a mistake on synchronising a machine on the 220kV line and thus causing an emergency switch to turn off power.
This affected the whole back valley (place10, place6, place7)
Axpo switched over to emergency - we got power back.
Storage cluster for VMs is back in write mode, I/O performance currently degraded.
We got an update that the problem comes from Axpo affecting the grid to which we are also connected.
Affected: place10 (VM, kubernetes based services)
Not affected: place5 (colocation)
The whole region of Glarus is having a power outage.
All systems are still up, running on battery power at the moment. We are in touch with the grid and local power plant to get insight on what is going on.
So
#openwrt does not (yet) work with
#cdist. A short report can be found on
https://ungleich.ch/u/blog/2024-08-02-openwrt-no-cdist-yet/ - let's see what the future brings.
cdist does not (yet) fully support OpenWrt | ungleich.ch
Affordable VM hosting in Glarus, Switzerland, operated by 100% hydropower. Devuan, debian, freebsd, ubuntu, centos. Our tech stack is fully open source.
Nico -telmich- Schottelius