Fix Deployment acceptance tests

[pdecat]

This PR fixes some failing deployment acceptance tests (tested on GKE 1.11.4-gke.8):

 make testacc TEST=./kubernetes TESTARGS='-run=TestAccKubernetesDeployment_.* -count=1'
==> Checking that code complies with gofmt requirements...
TF_ACC=1 go test ./kubernetes -v -run=TestAccKubernetesDeployment_.* -count=1 -timeout 120m
=== RUN   TestAccKubernetesDeployment_basic
--- PASS: TestAccKubernetesDeployment_basic (80.29s)
=== RUN   TestAccKubernetesDeployment_initContainer
--- PASS: TestAccKubernetesDeployment_initContainer (129.77s)
=== RUN   TestAccKubernetesDeployment_importBasic
--- PASS: TestAccKubernetesDeployment_importBasic (129.93s)
=== RUN   TestAccKubernetesDeployment_generatedName
--- PASS: TestAccKubernetesDeployment_generatedName (1.69s)
=== RUN   TestAccKubernetesDeployment_importGeneratedName
--- PASS: TestAccKubernetesDeployment_importGeneratedName (18.11s)
=== RUN   TestAccKubernetesDeployment_with_security_context
--- FAIL: TestAccKubernetesDeployment_with_security_context (600.93s)
        testing.go:518: Step 0 error: Error applying: 1 error(s) occurred:

                * kubernetes_deployment.test: 1 error(s) occurred:

                * kubernetes_deployment.test: Waiting for rollout to finish: 0 of 1 updated replicas are available...
=== RUN   TestAccKubernetesDeployment_with_container_liveness_probe_using_exec
--- PASS: TestAccKubernetesDeployment_with_container_liveness_probe_using_exec (3.57s)
=== RUN   TestAccKubernetesDeployment_with_container_liveness_probe_using_http_get
--- PASS: TestAccKubernetesDeployment_with_container_liveness_probe_using_http_get (3.54s)
=== RUN   TestAccKubernetesDeployment_with_container_liveness_probe_using_tcp
--- PASS: TestAccKubernetesDeployment_with_container_liveness_probe_using_tcp (3.54s)
=== RUN   TestAccKubernetesDeployment_with_container_lifecycle
--- FAIL: TestAccKubernetesDeployment_with_container_lifecycle (600.94s)
        testing.go:518: Step 0 error: Error applying: 1 error(s) occurred:

                * kubernetes_deployment.test: 1 error(s) occurred:

                * kubernetes_deployment.test: Waiting for rollout to finish: 0 of 1 updated replicas are available...
=== RUN   TestAccKubernetesDeployment_with_container_security_context
--- FAIL: TestAccKubernetesDeployment_with_container_security_context (600.97s)
        testing.go:518: Step 0 error: Error applying: 1 error(s) occurred:

                * kubernetes_deployment.test: 1 error(s) occurred:

                * kubernetes_deployment.test: Waiting for rollout to finish: 0 of 1 updated replicas are available...
=== RUN   TestAccKubernetesDeployment_with_volume_mount
--- PASS: TestAccKubernetesDeployment_with_volume_mount (4.70s)
=== RUN   TestAccKubernetesDeployment_with_resource_requirements
--- PASS: TestAccKubernetesDeployment_with_resource_requirements (3.51s)
=== RUN   TestAccKubernetesDeployment_with_empty_dir_volume
--- PASS: TestAccKubernetesDeployment_with_empty_dir_volume (3.51s)
FAIL
FAIL    github.com/terraform-providers/terraform-provider-kubernetes/kubernetes 2185.048s
make: *** [GNUmakefile:17: testacc] Error 1

The TestAccKubernetesDeployment_with_security_context and TestAccKubernetesDeployment_with_container_security_context test are timing out with containers in CrashLoopBackOff with the following container logs:

2018/12/07 19:00:24 [warn] 1#0: the "user" directive makes sense only if the master process runs with super-user privileges, ignored in /etc/nginx/nginx.conf:2
nginx: [warn] the "user" directive makes sense only if the master process runs with super-user privileges, ignored in /etc/nginx/nginx.conf:2
2018/12/07 19:00:24 [emerg] 1#0: mkdir() "/var/cache/nginx/client_temp" failed (13: Permission denied)
nginx: [emerg] mkdir() "/var/cache/nginx/client_temp" failed (13: Permission denied)

Using nginx images when security contexts define containers to run as non root user is known to cause issues:

• Don't use nginx image for the Pod example openshift/openshift-docs#10618
• https://medium.com/bitnami-perspectives/non-root-containers-to-show-openshift-some-love-3b32d7218ac6#c258

Work-around: use some other image such as redis.

The TestAccKubernetesDeployment_with_container_lifecycle test is timing out with containers in CrashLoopBackOff because the gcr.io/google_containers/liveness image used in that test only contains the /server executable, but no shell, ls or date executables:

# kubectl --context=gke_cloudpublic-sandbox_us-east1_pdecat-terraform-kubernetes-testacc get pod tf-acc-test-nza00o9x1u-dcffd6bf6-jwmx8
NAME                                     READY   STATUS                                                   RESTARTS   AGE
tf-acc-test-nza00o9x1u-dcffd6bf6-jwmx8   0/1     PostStartHookError: command 'ls -al' exited with 126:    5          2m

# kubectl get event |grep "ls -al"
1m          1m           3       tf-acc-test-nza00o9x1u-dcffd6bf6-jwmx8.156e1e06346f7435          Pod          spec.containers{containername}   Warning   FailedPostStartHook      kubelet, gke-pdecat-terraform-kub-default-pool-318bb3a5-ns08   Exec lifecycle hook ([ls -al]) for Container "containername" in Pod "tf-acc-test-nza00o9x1u-dcffd6bf6-jwmx8_default(e5857cae-fa47-11e8-8ad1-42010a8e0070)" failed - error: command 'ls -al' exited with 126: , message: "rpc error: code = 2 desc = oci runtime error: exec failed: container_linux.go:247: starting container process caused \"exec: \\\"ls\\\": executable file not found in $PATH\"\n\r\n"

Work-around: use gcr.io/google_containers/busybox image.

[pdecat]

Acceptance tests results with these changes:

# make testacc TEST=./kubernetes TESTARGS='-run=TestAccKubernetesDeployment_.* -count=1'                                                                                               
==> Checking that code complies with gofmt requirements...
TF_ACC=1 go test ./kubernetes -v -run=TestAccKubernetesDeployment_.* -count=1 -timeout 120m
=== RUN   TestAccKubernetesDeployment_basic
--- PASS: TestAccKubernetesDeployment_basic (69.35s)
=== RUN   TestAccKubernetesDeployment_initContainer
--- PASS: TestAccKubernetesDeployment_initContainer (1.89s)
=== RUN   TestAccKubernetesDeployment_importBasic
--- PASS: TestAccKubernetesDeployment_importBasic (129.24s)
=== RUN   TestAccKubernetesDeployment_generatedName
--- PASS: TestAccKubernetesDeployment_generatedName (17.49s)
=== RUN   TestAccKubernetesDeployment_importGeneratedName
--- PASS: TestAccKubernetesDeployment_importGeneratedName (58.27s)
=== RUN   TestAccKubernetesDeployment_with_security_context
--- PASS: TestAccKubernetesDeployment_with_security_context (1.42s)
=== RUN   TestAccKubernetesDeployment_with_container_liveness_probe_using_exec
--- PASS: TestAccKubernetesDeployment_with_container_liveness_probe_using_exec (17.54s)
=== RUN   TestAccKubernetesDeployment_with_container_liveness_probe_using_http_get
--- PASS: TestAccKubernetesDeployment_with_container_liveness_probe_using_http_get (17.48s)
=== RUN   TestAccKubernetesDeployment_with_container_liveness_probe_using_tcp
--- PASS: TestAccKubernetesDeployment_with_container_liveness_probe_using_tcp (9.46s)
=== RUN   TestAccKubernetesDeployment_with_container_lifecycle
--- PASS: TestAccKubernetesDeployment_with_container_lifecycle (9.49s)
=== RUN   TestAccKubernetesDeployment_with_container_security_context
--- PASS: TestAccKubernetesDeployment_with_container_security_context (9.35s)
=== RUN   TestAccKubernetesDeployment_with_volume_mount
--- PASS: TestAccKubernetesDeployment_with_volume_mount (10.31s)
=== RUN   TestAccKubernetesDeployment_with_resource_requirements
--- PASS: TestAccKubernetesDeployment_with_resource_requirements (9.46s)
=== RUN   TestAccKubernetesDeployment_with_empty_dir_volume
--- PASS: TestAccKubernetesDeployment_with_empty_dir_volume (9.39s)
PASS
ok      github.com/terraform-providers/terraform-provider-kubernetes/kubernetes 370.187s

[pdecat]

I had to fix these before addressing #247 (shaving yaks here 😄 )

[alexsomesan]

Thanks for fixing these.
Running the internal CI for this PR. If all works well, we're good to go.

[alexsomesan]

CI is all green.
LGTM