skaffold dev causes a panic

[paulgrav]

Expected behavior

Build succeeds and deploy succeeds. No panic.

Actual behavior

Build succeeds, but deploy causes skaffold panic

Information

• Skaffold version: v2.6.1
• Operating system: macOS 13.4.1 (22F82)
• Installed via: Homebrew
• Contents of skaffold.yaml:

apiVersion: skaffold/v4beta6
kind: Config
metadata:
  name: REDACTED
build:
  artifacts:
    - image: REDACTED
      context: .
      buildpacks:
        builder: paketobuildpacks/builder:tiny
manifests:
  kustomize:
    paths:
      - skaffold/k8s/overlay/dev
deploy:
  docker:
    images:
      - REDACTED
test:
  - image: REDACTED
    custom:
      - command: |-
          skaffold/install-trivy.sh
          trivy image -q --severity HIGH --ignore-unfixed $IMAGE
      - command: |-
          skaffold/install-trivy.sh
          trivy image -q --severity CRITICAL --ignore-unfixed --exit-code=1 $IMAGE

profiles:
  - name: nonprod
    manifests:
      kustomize:
       paths:
         - skaffold/k8s/overlay/nonprod
    deploy:
      kubectl:
        defaultNamespace: akamai-datastream-telemetry
        hooks: &hooks
          before:
            - host:
                command:
                  - "sh"
                  - -c
                  - |
                    chmod +x skaffold/create-gcp-serviceaccount.sh
                    GCP_PROJECT=REDACTED skaffold/create-gcp-serviceaccount.sh
                os: [darwin, linux]
          after: &listkservice
            - host:
                command:
                  - sh
                  - -c
                  - |
                    for i in $SKAFFOLD_NAMESPACES; do kubectl get kservice/REDACTED -n $i; done

  - name: prod
    manifests:
      kustomize:
       paths:
         - skaffold/k8s/overlay/prod
    deploy:
      kubectl:
        hooks:
          before:
            - host:
                command:
                  - "sh"
                  - -c
                  - |
                    chmod +x skaffold/create-gcp-serviceaccount.sh
                    GCP_PROJECT= REDACTED skaffold/create-gcp-serviceaccount.sh
                os: [darwin, linux]
          after: *listkservice

  - name: test
    activation:
      - command: deploy
    manifests:
      kustomize:
       paths:
         - skaffold/k8s/overlay/test
    deploy:
      kubectl:
        defaultNamespace: XXX
        hooks:
          <<: *hooks

Steps to reproduce the behavior

1. skaffold dev

They are tagged and referenced by a unique, local only, tag instead.
See https://skaffold.dev/docs/pipeline-stages/taggers/#how-tagging-works  subtask=-1 task=Deploy
Starting deploy...
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x2 addr=0x0 pc=0x104ccab3c]

goroutine 1 [running]:
github.com/GoogleContainerTools/skaffold/v2/pkg/skaffold/docker/logger.(*Logger).Stop(0x1400014e400)
        /private/tmp/skaffold-20230705-4219-jq79ds/pkg/skaffold/docker/logger/log.go:176 +0x2c
github.com/GoogleContainerTools/skaffold/v2/pkg/skaffold/log.LoggerMux.Stop({0x140004dfd00, 0x1, 0x1})
        /private/tmp/skaffold-20230705-4219-jq79ds/pkg/skaffold/log/logger_mux.go:40 +0x80
github.com/GoogleContainerTools/skaffold/v2/pkg/skaffold/runner.(*SkaffoldRunner).Dev(0x14000abcc40, {0x106602270, 0x140003c5340}, {0x1065dda60, 0x14000a13450}, {0x14000a2e040, 0x1, 0x1})
        /private/tmp/skaffold-20230705-4219-jq79ds/pkg/skaffold/runner/dev.go:307 +0x1070
github.com/GoogleContainerTools/skaffold/v2/cmd/skaffold/app/cmd.runDev.func5({0x106610420, 0x14000abcc40}, {0x14000c583f0, 0x1, 0x1})
        /private/tmp/skaffold-20230705-4219-jq79ds/cmd/skaffold/app/cmd/dev.go:71 +0x224
github.com/GoogleContainerTools/skaffold/v2/cmd/skaffold/app/cmd.withRunner({0x106602270, 0x140003c5340}, {0x1065dda60, 0x14000a13450}, 0x14000d9d3d0)
        /private/tmp/skaffold-20230705-4219-jq79ds/cmd/skaffold/app/cmd/runner.go:54 +0x180
github.com/GoogleContainerTools/skaffold/v2/cmd/skaffold/app/cmd.runDev({0x106602270, 0x140003c5340}, {0x1065dda60, 0x14000a13450})
        /private/tmp/skaffold-20230705-4219-jq79ds/cmd/skaffold/app/cmd/dev.go:66 +0x25c
github.com/GoogleContainerTools/skaffold/v2/cmd/skaffold/app/cmd.(*builder).NoArgs.func1(0x0, {0x0, 0x0, 0x0})
        /private/tmp/skaffold-20230705-4219-jq79ds/cmd/skaffold/app/cmd/commands.go:138 +0x90
github.com/GoogleContainerTools/skaffold/v2/cmd/skaffold/app/cmd.applyPostRunHooks.func1(0x14000243200, {0x14000a7e000, 0x0, 0x1})
        /private/tmp/skaffold-20230705-4219-jq79ds/cmd/skaffold/app/cmd/commands.go:185 +0x58
github.com/spf13/cobra.(*Command).execute(0x14000243200, {0x14000a69fe0, 0x1, 0x1})
        /private/tmp/skaffold-20230705-4219-jq79ds/vendor/github.com/spf13/cobra/command.go:940 +0x90c
github.com/spf13/cobra.(*Command).ExecuteC(0x14000242900)
        /private/tmp/skaffold-20230705-4219-jq79ds/vendor/github.com/spf13/cobra/command.go:1068 +0x758
github.com/spf13/cobra.(*Command).ExecuteContextC(0x14000242900, {0x106602270, 0x140003c5340})
        /private/tmp/skaffold-20230705-4219-jq79ds/vendor/github.com/spf13/cobra/command.go:1001 +0x78
github.com/GoogleContainerTools/skaffold/v2/cmd/skaffold/app.Run({0x1065dc840, 0x140001ae008}, {0x1065dc840, 0x140001ae010})
        /private/tmp/skaffold-20230705-4219-jq79ds/cmd/skaffold/app/skaffold.go:51 +0x498
main.main()
        /private/tmp/skaffold-20230705-4219-jq79ds/cmd/skaffold/skaffold.go:49 +0x244

Note that I’m using OrbStack and skaffold doesn’t seem to be aware of my docker context, so I have to manually set DOCKER_HOST via export DOCKER_HOST=$(docker context ls --format json | grep 'Current":true' | jq -r '.DockerEndpoint'). My DOCKER_HOST is set to unix:///Users/gravepa/.orbstack/run/docker.sock

[mboulton-fathom]

I've submitted a PR to fix the panic, but for me the root problem was that I named my images like

build:
  artifacts:
    - image: example.com/repo/imagename

which gave the error

creating container in local docker: Error response from daemon: Invalid container name (example.com/repo/imagename), only [a-zA-Z0-9][a-zA-Z0-9_.-] are allowed

This builds fine but fails when you do skaffold run/etc

[morsok]

I also have the same segmentation fault when trying to use scaffold dev to deploy to docker with docker-compose (I use docker-compose to try to bind volumes).

Even running with verbosity to trace I have no idea what's happening, there's no log not even a trace log from docker, it segfaults directly.