Allow specifying containerd snapshotter (and auto-detect the best snapshotter) #924
Labels
kind/enhancement
An improvement to existing functionality
Milestone
Comments
|
Thanks for the updates @AkihiroSuda, with those changes appears to be working well: vagrant@k3s:~$ cat /etc/lsb-release
DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=19.04
DISTRIB_CODENAME=disco
DISTRIB_DESCRIPTION="Ubuntu 19.04"
vagrant@k3s:~$ sudo apt-get install uidmap
Reading package lists... Done
Building dependency tree
Reading state information... Done
uidmap is already the newest version (1:4.5-1.1ubuntu2).
0 upgraded, 0 newly installed, 0 to remove and 89 not upgraded.
vagrant@k3s:~$ cat /etc/subuid
vagrant:100000:65536
vagrant@k3s:~$ curl -sfL https://github.com/rancher/k3s/releases/download/v0.10.1-rc1/k3s -o k3s
vagrant@k3s:~$ chmod a+x k3s
vagrant@k3s:~$ rm -rf .rancher/
vagrant@k3s:~$ ./k3s server --rootless >k3s.log 2>&1 &
[1] 2183
vagrant@k3s:~$ export KUBECONFIG=~/.kube/k3s.yaml
vagrant@k3s:~$ ./k3s kubectl get all -A
NAMESPACE NAME READY STATUS RESTARTS AGE
kube-system pod/local-path-provisioner-58fb86bdfd-sl9cp 1/1 Running 0 5m58s
kube-system pod/coredns-57d8bbb86-kq266 1/1 Running 0 5m58s
kube-system pod/helm-install-traefik-nlz7b 0/1 Completed 0 5m58s
kube-system pod/svclb-traefik-m6mcl 3/3 Running 0 5m40s
kube-system pod/traefik-65bccdc4bd-ksl4n 1/1 Running 0 5m40s
NAMESPACE NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
default service/kubernetes ClusterIP 10.43.0.1 <none> 443/TCP 6m15s
kube-system service/kube-dns ClusterIP 10.43.0.10 <none> 53/UDP,53/TCP,9153/TCP 6m14s
kube-system service/traefik LoadBalancer 10.43.177.137 127.0.0.1 80:32698/TCP,443:30443/TCP,8080:32326/TCP 5m40s
NAMESPACE NAME DESIRED CURRENT READY UP-TO-DATE AVAILABLE NODE SELECTOR AGE
kube-system daemonset.apps/svclb-traefik 1 1 1 1 1 <none> 5m40s
NAMESPACE NAME READY UP-TO-DATE AVAILABLE AGE
kube-system deployment.apps/local-path-provisioner 1/1 1 1 6m14s
kube-system deployment.apps/coredns 1/1 1 1 6m14s
kube-system deployment.apps/traefik 1/1 1 1 5m40s
NAMESPACE NAME DESIRED CURRENT READY AGE
kube-system replicaset.apps/local-path-provisioner-58fb86bdfd 1 1 1 5m58s
kube-system replicaset.apps/coredns-57d8bbb86 1 1 1 5m58s
kube-system replicaset.apps/traefik-65bccdc4bd 1 1 1 5m40s
NAMESPACE NAME COMPLETIONS DURATION AGE
kube-system job.batch/helm-install-traefik 1/1 18s 6m14s^ cc @ShylajaDevadiga |
|
With --rootless mode v0.10.0, enable ip fprwarding for the servicelb pods. |
|
@ShylajaDevadiga I don't think that relates to this issue. Could you reopen? |
|
Sorry about that, I had asked Shylaja to test & close if working (fix for helm pod errors is in v0.10.1). |
|
yes, also, manually specifying snapshotter should be also supported. |
|
I'm using nanoPC T4 and when i run also when i use ctr i import image.tar, see that when I use what can I do to specifying snapshotter now or maybe we should use |
|
i have such problem too |
Closed
|
The current master branch supports |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Is your feature request related to a problem? Please describe.
Ubuntu and Debian (since 10.0) supports mounting overlayfs in user namespaces.
This allows k3s Rootless-mode to use containerd
overlaysnapshotter.However, containerd v1.3.0 started to call
mknod 0 0to create overlay whiteout files which is even not supported by Ubuntu and Debian: containerd/containerd#3762So k3s rootless-mode cannot run some containers including helm which is required for deploying traefik:
Describe the solution you'd like
Although the issue is likely to be fixed in containerd v1.3.1 or v1.3.2 (containerd/containerd#3763), it would be good to allow k3s user to specify containerd
nativesnapshotter as a workaround.Also,
k3sshould detect the best snapshotter automatically.The text was updated successfully, but these errors were encountered: