PWX-38297

cmd/operator/operator.go

@@ -32,6 +32,7 @@ import (
 	"github.com/libopenstorage/operator/drivers/storage"
 	_ "github.com/libopenstorage/operator/drivers/storage/portworx"
 	"github.com/libopenstorage/operator/pkg/apis"
+	"github.com/libopenstorage/operator/pkg/controller/portworxdiag"
 	"github.com/libopenstorage/operator/pkg/controller/storagecluster"
 	"github.com/libopenstorage/operator/pkg/controller/storagenode"
 	_ "github.com/libopenstorage/operator/pkg/log"
@@ -52,6 +53,7 @@ const (
 	flagEnableProfiling          = "pprof"
 	flagDisableCacheFor          = "disable-cache-for"
 	defaultLockObjectName        = "openstorage-operator"
+	flagEnableDiagController     = "diag-controller"
 	defaultResyncPeriod          = 30 * time.Second
 	defaultMetricsPort           = 8999
 	defaultPprofPort             = 6060
@@ -110,6 +112,10 @@ func main() {
 			Name:  flagEnableProfiling,
 			Usage: "Enable Portworx Operator profiling using pprof (default: false)",
 		},
+		cli.BoolFlag{
+			Name:  flagEnableDiagController,
+			Usage: "Enable Portworx Diag Controller (default: false)",
+		},
 		cli.StringFlag{
 			Name:  flagDisableCacheFor,
 			Usage: "Comma separated object types to disable from cache to reduce memory usage, for example \"Pod,ConfigMap,Deployment,PersistentVolume\"",
@@ -146,6 +152,8 @@ func run(c *cli.Context) {
 		}()
 	}
 
+	diagControllerEnabled := c.Bool(flagEnableDiagController)
+
 	config, err := rest.InClusterConfig()
 	if err != nil {
 		log.Fatalf("Error getting cluster config: %v", err)
@@ -183,6 +191,15 @@ func run(c *cli.Context) {
 		log.Fatalf("Error registering CRD's for StorageNode controller: %v", err)
 	}
 
+	var diagController portworxdiag.Controller
+	if diagControllerEnabled {
+		diagController = portworxdiag.Controller{Driver: d}
+		err = diagController.RegisterCRD()
+		if err != nil {
+			log.Fatalf("Error registering CRDs for PortworxDiag controller: %v", err)
+		}
+	}
+
 	// TODO: Don't move createManager above register CRD section. This part will be refactored because of a bug,
 	// similar to https://github.com/kubernetes-sigs/controller-runtime/issues/321
 	mgr, err := createManager(c, config)
@@ -256,6 +273,12 @@ func run(c *cli.Context) {
 		log.Fatalf("Error initializing storage node controller: %v", err)
 	}
 
+	if diagControllerEnabled {
+		if err := diagController.Init(mgr); err != nil {
+			log.Fatalf("Error initializing portworx diag controller: %v", err)
+		}
+	}
+
 	if err := storageClusterController.StartWatch(); err != nil {
 		log.Fatalf("Error start watch on storage cluster controller: %v", err)
 	}
@@ -264,6 +287,12 @@ func run(c *cli.Context) {
 		log.Fatalf("Error starting watch on storage node controller: %v", err)
 	}
 
+	if diagControllerEnabled {
+		if err := diagController.StartWatch(); err != nil {
+			log.Fatalf("Error starting watch on portworx diag controller: %v", err)
+		}
+	}
+
 	if c.BoolT(flagMigration) {
 		log.Info("Migration is enabled")
 		migrationHandler := migration.New(&storageClusterController)

deploy/crds/core_v1_storagecluster_crd.yaml

@@ -154,6 +154,16 @@ spec:
                           their place. Once the new pods are available, it then proceeds onto other
                           StorageCluster pods, thus ensuring that at least 70% of original number of
                           StorageCluster pods are available at all times during the update.
+                      disruption:
+                        type: object
+                        description: >-
+                          The default behavior is non-disruptive upgrades. This setting disables the default
+                          non-disruptive upgrades and reverts to the previous behavior of upgrading nodes in
+                          parallel without worrying about disruption.
+                        properties:
+                          allow:
+                            type: boolean
+                            description: Flag indicates whether updates are non-disruptive or disruptive.
               deleteStrategy:
                 type: object
                 description: Delete strategy to uninstall and wipe the storage cluster.

deploy/crds/portworx.io_portworxdiags.yaml

@@ -0,0 +1,144 @@
+---
+apiVersion: apiextensions.k8s.io/v1
+kind: CustomResourceDefinition
+metadata:
+  annotations:
+    controller-gen.kubebuilder.io/version: v0.11.3
+  creationTimestamp: null
+  name: portworxdiags.portworx.io
+spec:
+  group: portworx.io
+  names:
+    kind: PortworxDiag
+    listKind: PortworxDiagList
+    plural: portworxdiags
+    shortNames:
+    - pxdiag
+    singular: portworxdiag
+  scope: Namespaced
+  versions:
+  - additionalPrinterColumns:
+    - description: Status of the Portworx diag collection.
+      jsonPath: .status.phase
+      name: Status
+      type: string
+    - description: Age of the diag resource.
+      jsonPath: .metadata.creationTimestamp
+      name: Age
+      type: date
+    name: v1
+    schema:
+      openAPIV3Schema:
+        description: PortworxDiag represents a portworx diag
+        properties:
+          apiVersion:
+            description: 'APIVersion defines the versioned schema of this representation
+              of an object. Servers should convert recognized schemas to the latest
+              internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources'
+            type: string
+          kind:
+            description: 'Kind is a string value representing the REST resource this
+              object represents. Servers may infer this from the endpoint the client
+              submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds'
+            type: string
+          metadata:
+            type: object
+          spec:
+            description: PortworxDiagSpec is the spec used to define a portworx diag.
+            properties:
+              portworx:
+                description: Configuration for diags collection of the main Portworx
+                  component.
+                properties:
+                  generateCore:
+                    description: Generates the core dump as well when collecting the
+                      diags. Could be useful to analyze the current state of the system.
+                    type: boolean
+                  nodes:
+                    description: Nodes for which the diags need to be collected. If
+                      a volume selector is also specified, then both the selectors
+                      will be honored and the selected nodes will be a union of both
+                      selectors.
+                    properties:
+                      all:
+                        description: Select all nodes in the Portworx cluster. If
+                          set to true, other selectors are ignored.
+                        type: boolean
+                      ids:
+                        description: Ids of the nodes to be selected.
+                        items:
+                          type: string
+                        type: array
+                      labels:
+                        additionalProperties:
+                          type: string
+                        description: Labels of the volumes to be selected.
+                        type: object
+                    type: object
+                  volumes:
+                    description: Volumes for which the diags need to be collected.
+                    properties:
+                      ids:
+                        description: Ids of the volumes to be selected.
+                        items:
+                          type: string
+                        type: array
+                      labels:
+                        additionalProperties:
+                          type: string
+                        description: Labels of the volumes to be selected.
+                        type: object
+                    type: object
+                  collectPodLogs:
+                    description: Flag indicating whether we want to collect pod logs too or not.
+                    type: boolean  
+                type: object
+            type: object
+          status:
+            description: PortworxDiagStatus is the status of a portworx diag.
+            properties:
+              clusterUuid:
+                description: UUID of the Portworx cluster. This is useful to find
+                  the uploaded diags.
+                type: string
+              message:
+                description: Optional message used to give the reason for any failure.
+                type: string
+              nodes:
+                description: Status of the diags collection from all the selected
+                  nodes.
+                items:
+                  description: Status of the diags collection from a single node.
+                  properties:
+                    message:
+                      description: Optional message used to give the reason for any
+                        failure.
+                      type: string
+                    nodeId:
+                      description: ID of the node for which the diag status is reported.
+                      type: string
+                    status:
+                      description: One word status of the diags collection on the
+                        node.
+                      type: string
+                  type: object
+                type: array
+              collectPodLogs:
+                description: Status of the diags collection from all the pods with label 'portworx'.
+                properties:
+                  message:
+                    description: Optional message used to give the reason for any failure.
+                    type: string
+                  status:
+                    description: One word status of the diags collection on the pod.
+                    type: string
+                type: object  
+              phase:
+                description: One word status of the entire diags collection job.
+                type: string
+            type: object
+        type: object
+    served: true
+    storage: true
+    subresources:
+      status: {}