End 2 End tests speedup (zalando#1180)

* Improving end 2 end tests, especially speed of execution and error, by implementing proper eventual asserts and timeouts.
* Add documentation for running individual tests
* Fixed String encoding in Patorni state check and error case
* Printing config as multi log line entity, makes it readable and grepable on startup
* Cosmetic changes to logs. Removed quotes from diff. Move all object diffs to text diff. Enabled padding for log level.
* Mount script with tools for easy logaccess and watching objects.
* Set proper update strategy for Postgres operator deployment.
* Move long running test to end. Move pooler test to new functions.
* Remove quote from valid K8s identifiers.

Author: Jan-M

Makefile

@@ -24,12 +24,16 @@ PKG := `go list ./... | grep -v /vendor/`
 
 ifeq ($(DEBUG),1)
 	DOCKERFILE = DebugDockerfile
-	DEBUG_POSTFIX := -debug
+	DEBUG_POSTFIX := -debug-$(shell date hhmmss)
 	BUILD_FLAGS += -gcflags "-N -l"
 else
 	DOCKERFILE = Dockerfile
 endif
 
+ifeq ($(FRESH),1)
+  DEBUG_FRESH=$(shell date +"%H-%M-%S")
+endif
+
 ifdef CDP_PULL_REQUEST_NUMBER
 	CDP_TAG := -${CDP_BUILD_VERSION}
 endif
@@ -66,7 +70,7 @@ docker: ${DOCKERDIR}/${DOCKERFILE} docker-context
 	echo "Version ${VERSION}"
 	echo "CDP tag ${CDP_TAG}"
 	echo "git describe $(shell git describe --tags --always --dirty)"
-	cd "${DOCKERDIR}" && docker build --rm -t "$(IMAGE):$(TAG)$(CDP_TAG)$(DEBUG_POSTFIX)" -f "${DOCKERFILE}" .
+	cd "${DOCKERDIR}" && docker build --rm -t "$(IMAGE):$(TAG)$(CDP_TAG)$(DEBUG_FRESH)$(DEBUG_POSTFIX)" -f "${DOCKERFILE}" .
 
 indocker-race:
 	docker run --rm -v "${GOPATH}":"${GOPATH}" -e GOPATH="${GOPATH}" -e RACE=1 -w ${PWD} golang:1.8.1 bash -c "make linux"
@@ -97,4 +101,4 @@ test:
 	GO111MODULE=on go test ./...
 
 e2e: docker # build operator image to be tested
-	cd e2e; make e2etest
+	cd e2e; make e2etest

e2e/Dockerfile

@@ -14,11 +14,15 @@ RUN apt-get update \
            python3-setuptools \
            python3-pip \
            curl \
+           vim \
     && pip3 install --no-cache-dir -r requirements.txt \
     && curl -LO https://storage.googleapis.com/kubernetes-release/release/v1.18.0/bin/linux/amd64/kubectl \
     && chmod +x ./kubectl \
     && mv ./kubectl /usr/local/bin/kubectl \
     && apt-get clean \
     && rm -rf /var/lib/apt/lists/*
 
-ENTRYPOINT ["python3", "-m", "unittest", "discover", "--start-directory", ".", "-v"]
+# working line
+# python3 -m unittest discover -v --failfast -k test_e2e.EndToEndTestCase.test_lazy_spilo_upgrade --start-directory tests
+ENTRYPOINT ["python3", "-m", "unittest"]
+CMD ["discover","-v","--failfast","--start-directory","/tests"]

e2e/README.md

@@ -12,6 +12,10 @@ Docker.
 Docker
 Go
 
+# Notice
+
+The `manifest` folder in e2e tests folder is not commited to git, it comes from `/manifests`
+
 ## Build test runner
 
 In the directory of the cloned Postgres Operator repository change to the e2e
@@ -35,6 +39,46 @@ In the e2e folder you can invoke tests either with `make test` or with:
 To run both the build and test step you can invoke `make e2e` from the parent
 directory.
 
+To run the end 2 end test and keep the kind state execute:
+```bash
+NOCLEANUP=True ./run.sh
+```
+
+## Run indidual test
+
+After having executed a normal E2E run with `NOCLEANUP=True` Kind still continues to run, allowing you subsequent test runs.
+
+To run an individual test, run the following command in the `e2e` directory
+
+```bash
+NOCLEANUP=True ./run.sh main tests.test_e2e.EndToEndTestCase.test_lazy_spilo_upgrade
+```
+
+## Inspecting Kind
+
+If you want to inspect Kind/Kubernetes cluster, use the following script to exec into the K8s setup and then use `kubectl`
+
+```bash
+./exec_into_env.sh
+
+# use kube ctl
+kubectl get pods
+
+# watch relevant objects
+./scripts/watch_objects.sh
+
+# get operator logs
+./scripts/get_logs.sh
+```
+
+## Cleaning up Kind
+
+To cleanup kind and start fresh
+
+```bash
+e2e/run.sh cleanup
+```
+
 ## Covered use cases
 
 The current tests are all bundled in [`test_e2e.py`](tests/test_e2e.py):

e2e/exec_into_env.sh

@@ -0,0 +1,14 @@
+#!/bin/bash
+
+export cluster_name="postgres-operator-e2e-tests"
+export kubeconfig_path="/tmp/kind-config-${cluster_name}"
+export operator_image="registry.opensource.zalan.do/acid/postgres-operator:latest"
+export e2e_test_runner_image="registry.opensource.zalan.do/acid/postgres-operator-e2e-tests-runner:0.3"
+
+docker run -it --entrypoint /bin/bash --network=host -e "TERM=xterm-256color" \
+    --mount type=bind,source="$(readlink -f ${kubeconfig_path})",target=/root/.kube/config \
+    --mount type=bind,source="$(readlink -f manifests)",target=/manifests \
+    --mount type=bind,source="$(readlink -f tests)",target=/tests \
+    --mount type=bind,source="$(readlink -f exec.sh)",target=/exec.sh \
+    --mount type=bind,source="$(readlink -f scripts)",target=/scripts \
+    -e OPERATOR_IMAGE="${operator_image}" "${e2e_test_runner_image}"

e2e/run.sh

@@ -9,6 +9,10 @@ IFS=$'\n\t'
 readonly cluster_name="postgres-operator-e2e-tests"
 readonly kubeconfig_path="/tmp/kind-config-${cluster_name}"
 readonly spilo_image="registry.opensource.zalan.do/acid/spilo-12:1.6-p5"
+readonly e2e_test_runner_image="registry.opensource.zalan.do/acid/postgres-operator-e2e-tests-runner:0.3"
+
+export GOPATH=${GOPATH-~/go}
+export PATH=${GOPATH}/bin:$PATH
 
 echo "Clustername: ${cluster_name}"
 echo "Kubeconfig path: ${kubeconfig_path}"
@@ -19,12 +23,7 @@ function pull_images(){
   then
     docker pull registry.opensource.zalan.do/acid/postgres-operator:latest
   fi
-
   operator_image=$(docker images --filter=reference="registry.opensource.zalan.do/acid/postgres-operator" --format "{{.Repository}}:{{.Tag}}" | head -1)
-
-  # this image does not contain the tests; a container mounts them from a local "./tests" dir at start time
-  e2e_test_runner_image="registry.opensource.zalan.do/acid/postgres-operator-e2e-tests-runner:latest"
-  docker pull ${e2e_test_runner_image}
 }
 
 function start_kind(){
@@ -36,12 +35,17 @@ function start_kind(){
   fi
 
   export KUBECONFIG="${kubeconfig_path}"
-  kind create cluster --name ${cluster_name} --config kind-cluster-postgres-operator-e2e-tests.yaml
-  kind load docker-image "${operator_image}" --name ${cluster_name}
+  kind create cluster --name ${cluster_name} --config kind-cluster-postgres-operator-e2e-tests.yaml  
   docker pull "${spilo_image}"
   kind load docker-image "${spilo_image}" --name ${cluster_name}
 }
 
+function load_operator_image() {
+  echo "Loading operator image"
+  export KUBECONFIG="${kubeconfig_path}"
+  kind load docker-image "${operator_image}" --name ${cluster_name}
+}
+
 function set_kind_api_server_ip(){
   echo "Setting up kind API server ip"
   # use the actual kubeconfig to connect to the 'kind' API server
@@ -52,35 +56,36 @@ function set_kind_api_server_ip(){
 }
 
 function run_tests(){
-  echo "Running tests..."
-
+  echo "Running tests... image: ${e2e_test_runner_image}"
   # tests modify files in ./manifests, so we mount a copy of this directory done by the e2e Makefile
 
   docker run --rm --network=host -e "TERM=xterm-256color" \
   --mount type=bind,source="$(readlink -f ${kubeconfig_path})",target=/root/.kube/config \
   --mount type=bind,source="$(readlink -f manifests)",target=/manifests \
   --mount type=bind,source="$(readlink -f tests)",target=/tests \
   --mount type=bind,source="$(readlink -f exec.sh)",target=/exec.sh \
-  -e OPERATOR_IMAGE="${operator_image}" "${e2e_test_runner_image}"
-  
+  --mount type=bind,source="$(readlink -f scripts)",target=/scripts \
+  -e OPERATOR_IMAGE="${operator_image}" "${e2e_test_runner_image}" ${E2E_TEST_CASE-} $@
 }
 
-function clean_up(){
+function cleanup(){
   echo "Executing cleanup"
   unset KUBECONFIG
   kind delete cluster --name ${cluster_name}
   rm -rf ${kubeconfig_path}
 }
 
 function main(){
-
-  trap "clean_up" QUIT TERM EXIT
-
-  time pull_images
-  time start_kind
-  time set_kind_api_server_ip
-  run_tests
+  echo "Entering main function..."
+  [[ -z ${NOCLEANUP-} ]] && trap "cleanup" QUIT TERM EXIT
+  pull_images
+  [[ ! -f ${kubeconfig_path} ]] && start_kind
+  load_operator_image
+  set_kind_api_server_ip
+
+  shift
+  run_tests $@
   exit 0
 }
 
-"$@"
+"$1" $@

e2e/scripts/cleanup.sh

@@ -0,0 +1,7 @@
+#!/bin/bash
+kubectl delete postgresql acid-minimal-cluster
+kubectl delete deployments -l application=db-connection-pooler,cluster-name=acid-minimal-cluster
+kubectl delete statefulsets -l application=spilo,cluster-name=acid-minimal-cluster
+kubectl delete services -l application=spilo,cluster-name=acid-minimal-cluster
+kubectl delete configmap postgres-operator
+kubectl delete deployment postgres-operator

e2e/scripts/get_logs.sh

@@ -0,0 +1,2 @@
+#!/bin/bash
+kubectl logs $(kubectl get pods -l name=postgres-operator --field-selector status.phase=Running -o jsonpath='{.items..metadata.name}')

e2e/scripts/watch_objects.sh

@@ -0,0 +1,19 @@
+#!/bin/bash
+
+watch -c "
+kubectl get postgresql
+echo
+echo -n 'Rolling upgrade pending: '
+kubectl get statefulset -o jsonpath='{.items..metadata.annotations.zalando-postgres-operator-rolling-update-required}'
+echo
+echo
+kubectl get pods -o wide
+echo
+kubectl get statefulsets
+echo
+kubectl get deployments
+echo
+kubectl get pods -l name=postgres-operator -o jsonpath='{.items..metadata.annotations.step}'
+echo
+kubectl get pods -l application=spilo -o jsonpath='{.items..spec.containers..image}'
+"