Add logs and describe for pgo (#96)

* Add logs and describe for pgo
* Release notes

Issues: [PGO-1175]

README.md

@@ -64,3 +64,10 @@ The `pgo` CLI supports all actively maintained versions of PGO v5+.
 
 For more about PGO, please see the
 [PGO Documentation](https://access.crunchydata.com/documentation/postgres-operator/).
+
+## Testing the Client
+
+The PGO client has several KUTTL tests that run through some common scenarios.
+In order to run these tests, we need a PGO operator or the equivalent running, i.e., `make deploy-dev`.
+
+Note: the `support export` test requires a PGO operator running in the `postgres-operator` namespace.

docs/config.toml

@@ -15,7 +15,7 @@ defaultContentLanguageInSubdir= false
 enableMissingTranslationPlaceholders = false
 
 [params]
-clientVersion = "0.4.1"
+clientVersion = "0.4.2"
 
 # crunchy-hugo-theme params
 editURL = "https://github.com/CrunchyData/postgres-operator/edit/master/docs/content/"

docs/content/reference/pgo_support_export.md

@@ -22,6 +22,7 @@ PostgresCluster.
     jobs.batch                                          [list]
     limitranges                                         [list]
     namespaces                                          [get]
+    networkpolicies.networking.k8s.io                   [list]
     nodes                                               [list]
     persistentvolumeclaims                              [list]
     poddisruptionbudgets.policy                         [list]
@@ -61,6 +62,11 @@ kubectl pgo support export daisy --output . --pg-logs-count 2
 # This is only required when monitoring is not deployed in the PostgresCluster's namespace.
 kubectl pgo support export daisy --monitoring-namespace another-namespace --output .
 
+# Operator namespace override
+# This is only required when the Operator is not deployed in the PostgresCluster's namespace.
+# This is used for getting the logs and specs for the operator pod(s).
+kubectl pgo support export daisy --operator-namespace another-namespace --output .
+
 ```
 ### Example output
 ```
@@ -91,11 +97,13 @@ Collecting services...
 Collecting endpoints...
 Collecting serviceaccounts...
 Collecting ingresses...
+Collecting networkpolicies...
 Collecting limitranges...
 Collecting events...
 Collecting Postgres logs...
 Collecting PostgresCluster pod logs...
 Collecting monitoring pod logs...
+Collecting operator pod logs...
 Collecting Patroni info...
 Collecting pgBackRest info...
 Collecting processes...
@@ -114,6 +122,7 @@ Collecting PGO CLI logs...
 ```
   -h, --help                          help for export
       --monitoring-namespace string   Monitoring namespace override
+      --operator-namespace string     Operator namespace override
   -o, --output string                 Path to save export tarball
   -l, --pg-logs-count int             Number of pg_log files to save (default 2)
 ```

docs/content/reference/pgo_version.md

@@ -31,7 +31,7 @@ pgo version
 ```
 ### Example output
 ```
-Client Version: v0.4.1
+Client Version: v0.4.2
 Operator Version: v5.5.0
 ```
 

docs/content/releases/0.4.1.md

@@ -1,7 +1,7 @@
 ---
 title: "0.4.1"
 draft: false
-weight: 996
+weight: 995
 ---
 
 [Crunchy Postgres for Kubernetes]: https://www.crunchydata.com/products/crunchy-postgresql-for-kubernetes
@@ -16,4 +16,5 @@ For more information about using the CLI and the various commands available, ple
 Additionally, please see the [CPK documentation](https://access.crunchydata.com/documentation/postgres-operator/latest) for information about [getting started](https://access.crunchydata.com/documentation/postgres-operator/latest/quickstart/) with Crunchy Postgres for Kubernetes.
 
 ## Bug 
+
 - The `support export` command would fail fatally if the user didn't have permissions to list postgresclusters at the cluster-scope. This release fixes that so our postgrescluster list fails gracefully (like our other atttempts to gather Kubernetes resources).

docs/content/releases/0.4.2.md

@@ -0,0 +1,20 @@
+---
+title: "0.4.2"
+draft: false
+weight: 994
+---
+
+[Crunchy Postgres for Kubernetes]: https://www.crunchydata.com/products/crunchy-postgresql-for-kubernetes
+[`pgo` CLI documentation]: https://access.crunchydata.com/documentation/postgres-operator-client/latest
+
+Crunchy Data announces the release of `pgo`, Postgres Operator Client from Crunchy Data 0.4.2.
+
+Built as a `kubectl` plugin, the `pgo` CLI facilitates the creation and management of PostgreSQL clusters created using [Crunchy Postgres for Kubernetes][].
+
+For more information about using the CLI and the various commands available, please see the [`pgo` CLI documentation][].
+
+Additionally, please see the [CPK documentation](https://access.crunchydata.com/documentation/postgres-operator/latest) for information about [getting started](https://access.crunchydata.com/documentation/postgres-operator/latest/quickstart/) with Crunchy Postgres for Kubernetes.
+
+## Changes
+
+- The `support export` command now captures logs from and descriptions for the PGO pod.

internal/cmd/client_version.go

@@ -15,4 +15,4 @@
 package cmd
 
 // store the current PGO CLI version
-const clientVersion = "v0.4.1"
+const clientVersion = "v0.4.2"

internal/cmd/export.go

@@ -38,7 +38,9 @@ import (
 	apierrors "k8s.io/apimachinery/pkg/api/errors"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/apis/meta/v1/unstructured"
+	"k8s.io/apimachinery/pkg/labels"
 	"k8s.io/apimachinery/pkg/runtime/schema"
+	"k8s.io/apimachinery/pkg/selection"
 	"k8s.io/apimachinery/pkg/util/duration"
 	"k8s.io/cli-runtime/pkg/printers"
 	"k8s.io/client-go/discovery"
@@ -126,6 +128,22 @@ var clusterNamespacedResources = []schema.GroupVersionResource{{
 	Resource: "serviceaccounts",
 }}
 
+// Resources specifically for the operator;
+// currently only pods, but leaving as is to allow expansion as requested.
+var operatorNamespacedResources = []schema.GroupVersionResource{{
+	Group:    appsv1.SchemeGroupVersion.Group,
+	Version:  appsv1.SchemeGroupVersion.Version,
+	Resource: "deployments",
+}, {
+	Group:    appsv1.SchemeGroupVersion.Group,
+	Version:  appsv1.SchemeGroupVersion.Version,
+	Resource: "replicasets",
+}, {
+	Group:    corev1.SchemeGroupVersion.Group,
+	Version:  corev1.SchemeGroupVersion.Version,
+	Resource: "pods",
+}}
+
 // These "removed" GVRs are for making our CLI backwards compatible with older PGO versions.
 var removedNamespacedResources = []schema.GroupVersionResource{{
 	Group:    batchv1beta1.SchemeGroupVersion.Group,
@@ -142,6 +160,10 @@ var otherNamespacedResources = []schema.GroupVersionResource{{
 	Group:    networkingv1.SchemeGroupVersion.Group,
 	Version:  networkingv1.SchemeGroupVersion.Version,
 	Resource: "ingresses",
+}, {
+	Group:    networkingv1.SchemeGroupVersion.Group,
+	Version:  networkingv1.SchemeGroupVersion.Version,
+	Resource: "networkpolicies",
 }, {
 	Group:    corev1.SchemeGroupVersion.Group,
 	Version:  corev1.SchemeGroupVersion.Version,
@@ -168,6 +190,7 @@ PostgresCluster.
     jobs.batch                                          [list]
     limitranges                                         [list]
     namespaces                                          [get]
+    networkpolicies.networking.k8s.io                   [list]
     nodes                                               [list]
     persistentvolumeclaims                              [list]
     poddisruptionbudgets.policy                         [list]
@@ -215,6 +238,9 @@ PostgresCluster.
 	var monitoringNamespace string
 	cmd.Flags().StringVarP(&monitoringNamespace, "monitoring-namespace", "", "", "Monitoring namespace override")
 
+	var operatorNamespace string
+	cmd.Flags().StringVarP(&operatorNamespace, "operator-namespace", "", "", "Operator namespace override")
+
 	cmd.Args = cobra.ExactArgs(1)
 
 	cmd.Example = internal.FormatExample(`# Short Flags
@@ -227,6 +253,11 @@ kubectl pgo support export daisy --output . --pg-logs-count 2
 # This is only required when monitoring is not deployed in the PostgresCluster's namespace.
 kubectl pgo support export daisy --monitoring-namespace another-namespace --output .
 
+# Operator namespace override
+# This is only required when the Operator is not deployed in the PostgresCluster's namespace.
+# This is used for getting the logs and specs for the operator pod(s).
+kubectl pgo support export daisy --operator-namespace another-namespace --output .
+
 ### Example output
 ┌────────────────────────────────────────────────────────────────
 | PGO CLI Support Export Tool
@@ -255,11 +286,13 @@ Collecting services...
 Collecting endpoints...
 Collecting serviceaccounts...
 Collecting ingresses...
+Collecting networkpolicies...
 Collecting limitranges...
 Collecting events...
 Collecting Postgres logs...
 Collecting PostgresCluster pod logs...
 Collecting monitoring pod logs...
+Collecting operator pod logs...
 Collecting Patroni info...
 Collecting pgBackRest info...
 Collecting processes...
@@ -287,6 +320,7 @@ Collecting PGO CLI logs...
 		writeDebug(cmd, fmt.Sprintf("Flag - Output Directory: %s\n", outputDir))
 		writeDebug(cmd, fmt.Sprintf("Flag - Num Logs: %d\n", numLogs))
 		writeDebug(cmd, fmt.Sprintf("Flag - Monitoring Namespace: %s\n", monitoringNamespace))
+		writeDebug(cmd, fmt.Sprintf("Flag - Operator Namespace: %s\n", operatorNamespace))
 
 		namespace, err := config.Namespace()
 		if err != nil {
@@ -433,6 +467,29 @@ Collecting PGO CLI logs...
 			err = gatherPodLogs(ctx, clientset, monitoringNamespace, util.LabelMonitoring, "monitoring", tw, cmd)
 		}
 
+		// get operator Pod logs and descriptions
+		if operatorNamespace == "" {
+			operatorNamespace = namespace
+		}
+		// Operator and Operator upgrade pods should have
+		// "postgres-operator.crunchydata.com/control-plane" label
+		// but with different values
+		if err == nil {
+			req, _ := labels.NewRequirement(util.LabelOperator,
+				selection.Exists, []string{},
+			)
+			nsListOpts := metav1.ListOptions{
+				LabelSelector: req.String(),
+			}
+			err = gatherNamespacedAPIResources(ctx, dynamicClient,
+				operatorNamespace, "operator", operatorNamespacedResources,
+				nsListOpts, tw, cmd)
+		}
+		if err == nil {
+			writeInfo(cmd, "Collecting operator pod logs...")
+			err = gatherPodLogs(ctx, clientset, operatorNamespace, util.LabelOperator, "operator", tw, cmd)
+		}
+
 		// Exec resources
 		if err == nil {
 			err = gatherPatroniInfo(ctx, clientset, restConfig, namespace, clusterName, tw, cmd)

internal/util/naming.go

@@ -31,6 +31,9 @@ const (
 
 	// LabelMonitoring is used to identify monitoring Pods
 	LabelMonitoring = "app.kubernetes.io/name=postgres-operator-monitoring"
+
+	// LabelOperator is used to identify operator Pods
+	LabelOperator = "postgres-operator.crunchydata.com/control-plane"
 )
 
 const (

testing/kuttl/e2e/support-export/01--support_export.yaml

@@ -2,12 +2,12 @@
 apiVersion: kuttl.dev/v1beta1
 kind: TestStep
 commands:
-- script: kubectl-pgo --namespace $NAMESPACE support export kuttl-support-cluster -o .
+- script: kubectl-pgo --namespace $NAMESPACE --operator-namespace postgres-operator support export kuttl-support-cluster -o .
 - script: tar -xzf ./crunchy_k8s_support_export_*.tar.gz
 - script: |
     CLEANUP="rm -r ./kuttl-support-cluster ./crunchy_k8s_support_export_*.tar.gz"
     check_file() {
-      if [ ! -s ./kuttl-support-cluster/"${1}" ]
+      if [ ! -s ./"${1}" ]
       then
         echo "Expected ${1} file to not be empty"
         eval "$CLEANUP"
@@ -29,23 +29,29 @@ commands:
     }
 
     # check that the cluster-names file exists and is not empty
-    check_file "cluster-names"
+    check_file "kuttl-support-cluster/cluster-names"
 
     # check that the system-time file exists and is not empty
-    check_file "system-time"
+    check_file "kuttl-support-cluster/system-time"
 
     # check that the context file exists and is not empty
-    check_file "current-context"
+    check_file "kuttl-support-cluster/current-context"
 
     # check that the patroni info file exists and is not empty
-    check_file "patroni-info"
+    check_file "kuttl-support-cluster/patroni-info"
 
     # check that the pgbackrest info file exists and is not empty
-    check_file "pgbackrest-info"
+    check_file "kuttl-support-cluster/pgbackrest-info"
 
     # check that the plugin list file exists and is not empty
     # the file will at least include kubectl-pgo
-    check_file "plugin-list"
+    check_file "kuttl-support-cluster/plugin-list"
+    
+    # check that the operator file exists and is not empty
+    # the list file will not be empty for the requested Kubernetes types
+    check_file "operator/deployments/list"
+    check_file "operator/replicasets/list"
+    check_file "operator/pods/list"
 
     # check for expected gzip compression level
     FILE_INFO=$(file ./crunchy_k8s_support_export_*.tar.gz)
@@ -85,7 +91,7 @@ commands:
     fi
 
     # check that the events file exist and is not empty
-    check_file "events"
+    check_file "kuttl-support-cluster/events"
 
     EVENTS="./kuttl-support-cluster/events"
     # check that the events file contains the expected string