Open ctalledo opened 2 years ago
Is this list sorted by "most important first"? :) If so, great! I came to this issue tracker to voice interest in Kube >1.21 support.
1.20 is already EOL and 1.21 will meet the same fate in about 2 months. (I'm particularly interested in 1.22 for the alpha apiserver tracing support).
I introduced Sysbox CE at my $dayjob to allow us to cohabitate some CI workloads alongside non-CI workloads in our baremetal clusters. It's working great, thanks! If a release candidate rolls around with Kube 1.22 support, I have access to a few 1.22+ baremetal clusters to test it on.
@samcday, thanks for the feedback!
I don't think the list is necessarily sorted based on priorities, but 1.22 support is certainly one of the most important issues for us to tackle.
Btw, we just addressed 2) and 6) in the last few hours and planning to jump into 1), 4) and 5) right after.
The 9 will help to streamline the installation, upgrade, and uninstallation of Sysbox on K8s. It will be nice! :)
Here is a good starting point, btw: https://github.com/helm/charts-repo-actions-demo
@rodnymolina @ctalledo hey - what's the current state about k8s 1.22.x support? I'm asking because we wanna use sysbox inside of AKS an MS deprecated 1.21.X. Thx in advance for your answer :)
Hi @immae1, support for v1.22 and v1.23 is there already.
Apologies, forgot to update this issue accordingly, will do so now.
Please see here for our official doc on this matter.
The following enhancements to sysbox-deploy-k8s are needed (as of 04/15/22):
1)
Add K8s 1.22 & 1.23 support.2) Add support for systemd-managed cgroups. 3) Add a way to configure sysbox-deploy-k8s (e.g., via K8s config map). 4) Handle scenarios where host uses mutliple kubelet systemd drop-in files (issue #525). 5) Handle installation behind HTTPS proxy (issue #522). 6) Have the sysbox-deploy-k8s daemonset monitor the kernel version periodically and if it detects changes, reinstall shiftfs with the appropriate version 7) Handle scenario where CRI-O is already installed in cluster. 8) Handle single-node clusters (i.e,. currently sysbox-deploy-k8s requires >= 2 nodes). 9) Create a Helm chart for sysbox-deploy-k8s. 10) Add support for more host distros (e.g., Debian, RedHat, Fedora, CentOS-Stream, etc.)