/var/lib/rancher/k3s/ is 9GB on master, 12GB on worker nodes after 60 Days

[admantium-sg]

Environmental Info:
K3s Version:
k3s version v1.17.2+k3s1 (cdab19b)

Node(s) CPU architecture, OS, and Version:
Linux k3s-server 4.19.0-8-amd64 #1 SMP Debian 4.19.98-1 (2020-01-26) x86_64 GNU/Linux

Cluster Configuration:
1 master, 2 nodes

Describe the bug:
On the master node, /var/lib/rancher/k3s/ consumes 9GB after 60 days of operations.
On the worker nodes, /var/lib/rancher/k3s/ consumes 12GB after 60 days of operations.

How to clean up old files? E.g. there seems to be a lot of snapshot data in /var/lib/rancher/k3s/agent/containerd/io.containerd.snapshotter.v1.overlay

Steps To Reproduce:

• Install K3S
• Run for 60 days

Expected behavior:

• Less space required

Actual behavior:

• 9GB required

Additional context / logs:

• n.a.

[brandond]

See the discussion on image garbage collection here: #1900 (comment)

[admantium-sg]

Thank you very much!

I executed k3s crictl rmi --prune on each machine.

• master: 9GB => 1.4GB
• worker: 12GB => 2.6GB

[admantium-sg]

Please close this ticket.