[BUG] signal SIGSEGV: segmentation violation on node creation

[zewelor]

What did you do?
k3d create node test -c cluster

• How was the cluster created?

  • k3d create -x A -y B

• What did you do afterwards?

  • k3d commands?
  • docker commands?
  • OS operations (e.g. shutdown/reboot)?

What did you expect to happen?

Concise description of what you expected to happen after doing what you described above.

Screenshots or terminal output

k3d version v3.0.0-rc.2
k3s version v1.18.3-k3s1 (default)
~  k3d create node test -c cluster  ✔
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x8 pc=0xd5f9ae]

goroutine 1 [running]:
github.com/rancher/k3d/pkg/cluster.AddNodeToCluster(0x1155d00, 0xc000321e00, 0x1169f80, 0x1893b10, 0xc000422ea0, 0xc00044e500, 0xc000436c00, 0x0, 0x4, 0xc000239428)
/home/travis/gopath/src/github.com/rancher/k3d/pkg/cluster/node.go:43 +0x12ce
github.com/rancher/k3d/pkg/cluster.AddNodesToCluster(0x1155d40, 0xc000134010, 0x1169f80, 0x1893b10, 0xc000138b68, 0x1, 0x1, 0xc00044e500, 0xc000329600, 0x0, ...)
/home/travis/gopath/src/github.com/rancher/k3d/pkg/cluster/node.go:155 +0x1ea
github.com/rancher/k3d/cmd/create.NewCmdCreateNode.func1(0xc000439080, 0xc000437380, 0x1, 0x3)
/home/travis/gopath/src/github.com/rancher/k3d/cmd/create/createNode.go:50 +0xf6
github.com/spf13/cobra.(*Command).execute(0xc000439080, 0xc000437350, 0x3, 0x3, 0xc000439080, 0xc000437350)
/home/travis/gopath/src/github.com/rancher/k3d/vendor/github.com/spf13/cobra/command.go:846 +0x29d
github.com/spf13/cobra.(*Command).ExecuteC(0x185a5e0, 0x405abf, 0xc000114058, 0x0)
/home/travis/gopath/src/github.com/rancher/k3d/vendor/github.com/spf13/cobra/command.go:950 +0x349
github.com/spf13/cobra.(*Command).Execute(...)
/home/travis/gopath/src/github.com/rancher/k3d/vendor/github.com/spf13/cobra/command.go:887
github.com/rancher/k3d/cmd.Execute()
/home/travis/gopath/src/github.com/rancher/k3d/cmd/root.go:79 +0x2d
main.main()
/home/travis/gopath/src/github.com/rancher/k3d/main.go:27 +0x20

Which OS & Architecture?

• Linux amd64

Which version of k3d? v3.0.0-rc.2

• output of k3d --version

Which version of docker?

• output of docker version

[iwilltry42]

Hi @zewelor, good catch, thank you! :)
This is a bug in the exception handling for when the chosen cluster does not exist.
Fixed in the linked commit 2d0ec0a 👍

[zewelor]

Hmm I've tried to connect to existing cluster, but on remote computer. I've also set K3S_URL and K3S_TOKEN , prior to this command.

EDIT: I've tested in on rc.3 and can confirm now works, but cannot connect to remote cluster. How it should be done ?

[iwilltry42]

Oh okay, k3d v3 does not support this.
How did you set K3S_URL/K3S_TOKEN? 🤔
Currently, k3d v3 only supports connecting to other k3d clusters in the same docker evironment.

[zewelor]

I've set cluster via k3d on one computer, then get token via k3d get cluster --token and did export K3S_ env vars on other computer. So it's not possible to join worker, via k3d, to already existing cluster ?

[iwilltry42]

So there are several things to notice here:

• k3d is spawning containerized k3s nodes, so they don't honor the K3S_* environment variables that one specifies on host level
• k3d clusters (by default) have their own docker network and no ports exposed, so even traffic between different clusters/hosts may be limited
• You can add new nodes to existing clusters on the same host, that's not a problem.
• k3d v1.6 had an unsupported feature to add k3d nodes to existing (plain) k3s clusters, which seemed to work but was untested.