forked from rook/rook
Travis Nielsen
291bf9e061
The osd prepare job detects the failure domain for the OSD by querying the node topology labels. This topology is then assigned to the OSD daemon in the CRUSH map. Previously, the affinity was required to be set in the cluster CR, but it was very difficult to get right. Now the operator will enforce the correct topology label on the OSD daemon nodeAffinity by setting the label of the lowest topology in the hierarcy. For example, if there are region, zone, and rack labels, the rack label would be used to set the node affinity for the OSD daemon. If an OSD prepare job is run in rack1, the corresponding OSD daemon will have node affinity to rack1 to ensure the same topology. Previously, the OSD could have ended up in rack2 unless the storageClassDeviceSet placement was very carefully crafted. Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
424 lines
12 KiB
Go
424 lines
12 KiB
Go
/*
|
|
Copyright 2018 The Rook Authors. All rights reserved.
|
|
|
|
Licensed under the Apache License, Version 2.0 (the "License");
|
|
you may not use this file except in compliance with the License.
|
|
You may obtain a copy of the License at
|
|
|
|
http://www.apache.org/licenses/LICENSE-2.0
|
|
|
|
Unless required by applicable law or agreed to in writing, software
|
|
distributed under the License is distributed on an "AS IS" BASIS,
|
|
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
See the License for the specific language governing permissions and
|
|
limitations under the License.
|
|
*/
|
|
|
|
// Package k8sutil for Kubernetes helpers.
|
|
package k8sutil
|
|
|
|
import (
|
|
"context"
|
|
"reflect"
|
|
"testing"
|
|
|
|
rookv1 "github.com/rook/rook/pkg/apis/rook.io/v1"
|
|
optest "github.com/rook/rook/pkg/operator/test"
|
|
"github.com/stretchr/testify/assert"
|
|
v1 "k8s.io/api/core/v1"
|
|
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
|
|
"k8s.io/client-go/kubernetes/fake"
|
|
)
|
|
|
|
func createNode(nodeName string, condition v1.NodeConditionType, clientset *fake.Clientset) error {
|
|
ctx := context.TODO()
|
|
node := &v1.Node{
|
|
ObjectMeta: metav1.ObjectMeta{
|
|
Name: nodeName,
|
|
},
|
|
Status: v1.NodeStatus{
|
|
Conditions: []v1.NodeCondition{
|
|
{
|
|
Type: condition, Status: v1.ConditionTrue,
|
|
},
|
|
},
|
|
},
|
|
}
|
|
_, err := clientset.CoreV1().Nodes().Create(ctx, node, metav1.CreateOptions{})
|
|
return err
|
|
}
|
|
|
|
func TestValidNode(t *testing.T) {
|
|
nodeA := "nodeA"
|
|
nodeB := "nodeB"
|
|
|
|
storage := rookv1.StorageScopeSpec{
|
|
Nodes: []rookv1.Node{
|
|
{
|
|
Name: nodeA,
|
|
},
|
|
{
|
|
Name: nodeB,
|
|
},
|
|
},
|
|
}
|
|
var placement rookv1.Placement
|
|
// set up a fake k8s client set and watcher to generate events that the operator will listen to
|
|
clientset := fake.NewSimpleClientset()
|
|
|
|
nodeErr := createNode(nodeA, v1.NodeReady, clientset)
|
|
assert.Nil(t, nodeErr)
|
|
nodeErr = createNode(nodeB, v1.NodeNetworkUnavailable, clientset)
|
|
assert.Nil(t, nodeErr)
|
|
validNodes := GetValidNodes(storage, clientset, placement)
|
|
assert.Equal(t, len(validNodes), 1)
|
|
}
|
|
|
|
func testNode(taints []v1.Taint) v1.Node {
|
|
n := v1.Node{}
|
|
n.Spec.Taints = append(n.Spec.Taints, taints...)
|
|
return n
|
|
}
|
|
|
|
func taintReservedForRook() v1.Taint {
|
|
return v1.Taint{Key: "reservedForRook", Effect: v1.TaintEffectNoSchedule}
|
|
}
|
|
|
|
func taintReservedForOther() v1.Taint {
|
|
return v1.Taint{Key: "reservedForNOTRook", Effect: v1.TaintEffectNoSchedule}
|
|
}
|
|
|
|
func taintAllWellKnown() []v1.Taint {
|
|
taints := []v1.Taint{}
|
|
for _, t := range WellKnownTaints {
|
|
taints = append(taints, v1.Taint{
|
|
// assume the "worst" with NoExecute
|
|
Key: t, Effect: v1.TaintEffectNoExecute,
|
|
})
|
|
}
|
|
return taints
|
|
}
|
|
|
|
func taints(taints ...v1.Taint) []v1.Taint {
|
|
list := []v1.Taint{}
|
|
for _, t := range taints {
|
|
list = append(taints, t)
|
|
}
|
|
return list
|
|
}
|
|
|
|
func tolerateRook() []v1.Toleration {
|
|
return []v1.Toleration{{Key: "reservedForRook"}}
|
|
}
|
|
|
|
func TestNodeIsTolerable(t *testing.T) {
|
|
type args struct {
|
|
node v1.Node
|
|
tolerations []v1.Toleration
|
|
ignoreWellKnownTaints bool
|
|
}
|
|
tests := []struct {
|
|
name string
|
|
args args
|
|
want bool
|
|
}{
|
|
{name: "tolerate node w/o taints", args: args{
|
|
node: v1.Node{},
|
|
tolerations: tolerateRook(),
|
|
ignoreWellKnownTaints: false,
|
|
}, want: true},
|
|
{name: "tolerate node w/ rook taint", args: args{
|
|
node: testNode(taints(taintReservedForRook())),
|
|
tolerations: tolerateRook(),
|
|
ignoreWellKnownTaints: false,
|
|
}, want: true},
|
|
{name: "do not tolerate rook taint", args: args{
|
|
node: testNode(taints(taintReservedForRook())),
|
|
tolerations: nil,
|
|
ignoreWellKnownTaints: false,
|
|
}, want: false},
|
|
{name: "do not tolerate other taint", args: args{
|
|
node: testNode(taints(taintReservedForRook(), taintReservedForOther())),
|
|
tolerations: tolerateRook(),
|
|
ignoreWellKnownTaints: false,
|
|
}, want: false},
|
|
{name: "do not tolerate node w/ known taints", args: args{
|
|
node: testNode(taintAllWellKnown()),
|
|
tolerations: nil,
|
|
ignoreWellKnownTaints: false,
|
|
}, want: false},
|
|
{name: "do not tolerate node w/ known taints 2", args: args{
|
|
node: testNode(taintAllWellKnown()),
|
|
tolerations: tolerateRook(),
|
|
ignoreWellKnownTaints: false,
|
|
}, want: false},
|
|
{name: "tolerate node w/ known taints and rook taint", args: args{
|
|
node: testNode(taintAllWellKnown()),
|
|
tolerations: tolerateRook(),
|
|
ignoreWellKnownTaints: true,
|
|
}, want: true},
|
|
{name: "do not tolerate node w/ known taints and rook taint", args: args{
|
|
node: testNode(append(taintAllWellKnown(), taintReservedForRook())),
|
|
tolerations: nil,
|
|
ignoreWellKnownTaints: true,
|
|
}, want: false},
|
|
{name: "tolerate node w/ known taints and rook taint", args: args{
|
|
node: testNode(append(taintAllWellKnown(), taintReservedForRook())),
|
|
tolerations: tolerateRook(),
|
|
ignoreWellKnownTaints: true,
|
|
}, want: true},
|
|
{name: "do not tolerate node w/ known and other taints", args: args{
|
|
node: testNode(append(taintAllWellKnown(), taintReservedForRook(), taintReservedForOther())),
|
|
tolerations: tolerateRook(),
|
|
ignoreWellKnownTaints: true,
|
|
}, want: false},
|
|
}
|
|
for _, tt := range tests {
|
|
t.Run(tt.name, func(t *testing.T) {
|
|
if got := NodeIsTolerable(tt.args.node, tt.args.tolerations, tt.args.ignoreWellKnownTaints); got != tt.want {
|
|
t.Errorf("NodeIsTolerable() = %v, want %v", got, tt.want)
|
|
}
|
|
})
|
|
}
|
|
}
|
|
|
|
func TestNodeIsReady(t *testing.T) {
|
|
assert.True(t, NodeIsReady(v1.Node{Status: v1.NodeStatus{Conditions: []v1.NodeCondition{
|
|
{Type: v1.NodeReady, Status: v1.ConditionTrue},
|
|
}}}))
|
|
assert.False(t, NodeIsReady(v1.Node{Status: v1.NodeStatus{Conditions: []v1.NodeCondition{
|
|
{Type: v1.NodeReady, Status: v1.ConditionFalse},
|
|
}}}))
|
|
assert.False(t, NodeIsReady(v1.Node{Status: v1.NodeStatus{Conditions: []v1.NodeCondition{
|
|
{Type: v1.NodeReady, Status: v1.ConditionUnknown},
|
|
}}}))
|
|
// if `Ready` condition does not exist, must assume that node is not ready
|
|
assert.False(t, NodeIsReady(v1.Node{Status: v1.NodeStatus{Conditions: []v1.NodeCondition{
|
|
{Type: v1.NodeDiskPressure, Status: v1.ConditionTrue},
|
|
}}}))
|
|
// if `Ready` condition is not accompanied by a status, must assume that node is not ready
|
|
assert.False(t, NodeIsReady(v1.Node{Status: v1.NodeStatus{Conditions: []v1.NodeCondition{
|
|
{Type: v1.NodeDiskPressure},
|
|
}}}))
|
|
}
|
|
|
|
func TestGetRookNodesMatchingKubernetesNodes(t *testing.T) {
|
|
ctx := context.TODO()
|
|
clientset := optest.New(t, 3) // create nodes 0, 1, and 2
|
|
rookNodes := []rookv1.Node{}
|
|
|
|
getNode := func(name string) v1.Node {
|
|
n, err := clientset.CoreV1().Nodes().Get(ctx, name, metav1.GetOptions{})
|
|
assert.NoError(t, err)
|
|
return *n
|
|
}
|
|
|
|
// no rook nodes specified
|
|
nodes, err := GetKubernetesNodesMatchingRookNodes(rookNodes, clientset)
|
|
assert.NoError(t, err)
|
|
assert.Empty(t, nodes)
|
|
|
|
// more rook nodes specified than nodes exist
|
|
rookNodes = []rookv1.Node{
|
|
{Name: "node0"},
|
|
{Name: "node2"},
|
|
{Name: "node5"}}
|
|
nodes, err = GetKubernetesNodesMatchingRookNodes(rookNodes, clientset)
|
|
assert.NoError(t, err)
|
|
assert.Len(t, nodes, 2)
|
|
assert.Contains(t, nodes, getNode("node0"))
|
|
assert.Contains(t, nodes, getNode("node2"))
|
|
|
|
// rook nodes match k8s nodes
|
|
rookNodes = []rookv1.Node{
|
|
{Name: "node0"},
|
|
{Name: "node1"},
|
|
{Name: "node2"}}
|
|
nodes, err = GetKubernetesNodesMatchingRookNodes(rookNodes, clientset)
|
|
assert.NoError(t, err)
|
|
assert.Len(t, nodes, 3)
|
|
assert.Contains(t, nodes, getNode("node0"))
|
|
assert.Contains(t, nodes, getNode("node1"))
|
|
assert.Contains(t, nodes, getNode("node2"))
|
|
|
|
// no k8s nodes exist
|
|
clientset = optest.New(t, 0)
|
|
nodes, err = GetKubernetesNodesMatchingRookNodes(rookNodes, clientset)
|
|
assert.NoError(t, err)
|
|
assert.Len(t, nodes, 0)
|
|
}
|
|
|
|
func TestRookNodesMatchingKubernetesNodes(t *testing.T) {
|
|
ctx := context.TODO()
|
|
clientset := optest.New(t, 3) // create nodes 0, 1, and 2
|
|
|
|
getNode := func(name string) v1.Node {
|
|
n, err := clientset.CoreV1().Nodes().Get(ctx, name, metav1.GetOptions{})
|
|
assert.NoError(t, err)
|
|
return *n
|
|
}
|
|
n0 := getNode("node0")
|
|
n0.Labels = map[string]string{v1.LabelHostname: "node0-hostname"}
|
|
n1 := getNode("node1")
|
|
n2 := getNode("node2")
|
|
n2.Labels = map[string]string{v1.LabelHostname: "node2"}
|
|
k8sNodes := []v1.Node{n0, n1, n2}
|
|
|
|
// no rook nodes specified for input
|
|
rookStorage := rookv1.StorageScopeSpec{
|
|
Nodes: []rookv1.Node{},
|
|
}
|
|
retNodes := RookNodesMatchingKubernetesNodes(rookStorage, k8sNodes)
|
|
assert.Len(t, retNodes, 0)
|
|
|
|
// all rook nodes specified
|
|
rookStorage.Nodes = []rookv1.Node{
|
|
{Name: "node0"},
|
|
{Name: "node1"},
|
|
{Name: "node2"}}
|
|
retNodes = RookNodesMatchingKubernetesNodes(rookStorage, k8sNodes)
|
|
assert.Len(t, retNodes, 3)
|
|
// this should return nodes named by hostname if that is available
|
|
assert.Contains(t, retNodes, rookv1.Node{Name: "node0-hostname"})
|
|
assert.Contains(t, retNodes, rookv1.Node{Name: "node1"})
|
|
assert.Contains(t, retNodes, rookv1.Node{Name: "node2"})
|
|
|
|
// more rook nodes specified than exist
|
|
rookStorage.Nodes = []rookv1.Node{
|
|
{Name: "node0-hostname"},
|
|
{Name: "node2"},
|
|
{Name: "node5"}}
|
|
retNodes = RookNodesMatchingKubernetesNodes(rookStorage, k8sNodes)
|
|
assert.Len(t, retNodes, 2)
|
|
assert.Contains(t, retNodes, rookv1.Node{Name: "node0-hostname"})
|
|
assert.Contains(t, retNodes, rookv1.Node{Name: "node2"})
|
|
|
|
// no k8s nodes specified
|
|
retNodes = RookNodesMatchingKubernetesNodes(rookStorage, []v1.Node{})
|
|
assert.Len(t, retNodes, 0)
|
|
}
|
|
|
|
func TestGenerateNodeAffinity(t *testing.T) {
|
|
type args struct {
|
|
nodeAffinity string
|
|
}
|
|
tests := []struct {
|
|
name string
|
|
args args
|
|
want *v1.NodeAffinity
|
|
wantErr bool
|
|
}{
|
|
{
|
|
name: "GenerateNodeAffinity",
|
|
args: args{
|
|
nodeAffinity: "rook.io/ceph=true",
|
|
},
|
|
want: &v1.NodeAffinity{
|
|
RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
|
|
NodeSelectorTerms: []v1.NodeSelectorTerm{
|
|
{
|
|
MatchExpressions: []v1.NodeSelectorRequirement{
|
|
{
|
|
Key: "rook.io/ceph",
|
|
Operator: v1.NodeSelectorOpIn,
|
|
Values: []string{"true"},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
wantErr: false,
|
|
},
|
|
{
|
|
name: "FailGenerateNodeAffinity",
|
|
args: args{
|
|
nodeAffinity: "rook.io/ceph,cassandra=true",
|
|
},
|
|
want: nil,
|
|
wantErr: true,
|
|
},
|
|
{
|
|
name: "GenerateNodeAffinityWithKeyOnly",
|
|
args: args{
|
|
nodeAffinity: "rook.io/ceph",
|
|
},
|
|
want: &v1.NodeAffinity{
|
|
RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
|
|
NodeSelectorTerms: []v1.NodeSelectorTerm{
|
|
{
|
|
MatchExpressions: []v1.NodeSelectorRequirement{
|
|
{
|
|
Key: "rook.io/ceph",
|
|
Operator: v1.NodeSelectorOpExists,
|
|
},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
wantErr: false,
|
|
},
|
|
}
|
|
for _, tt := range tests {
|
|
t.Run(tt.name, func(t *testing.T) {
|
|
got, err := GenerateNodeAffinity(tt.args.nodeAffinity)
|
|
if (err != nil) != tt.wantErr {
|
|
t.Errorf("GenerateNodeAffinity() error = %v, wantErr %v", err, tt.wantErr)
|
|
return
|
|
}
|
|
if !reflect.DeepEqual(got, tt.want) {
|
|
t.Errorf("GenerateNodeAffinity() = %v, want %v", got, tt.want)
|
|
}
|
|
})
|
|
}
|
|
}
|
|
|
|
func TestGetNotReadyKubernetesNodes(t *testing.T) {
|
|
ctx := context.TODO()
|
|
clientset := optest.New(t, 0)
|
|
|
|
//when there is no node
|
|
nodes, err := GetNotReadyKubernetesNodes(clientset)
|
|
assert.NoError(t, err)
|
|
assert.Equal(t, 0, len(nodes))
|
|
|
|
//when all the nodes are in ready state
|
|
clientset = optest.New(t, 2)
|
|
nodes, err = GetNotReadyKubernetesNodes(clientset)
|
|
assert.NoError(t, err)
|
|
assert.Equal(t, 0, len(nodes))
|
|
|
|
//when there is a not ready node
|
|
node := &v1.Node{
|
|
ObjectMeta: metav1.ObjectMeta{
|
|
Name: "failed",
|
|
},
|
|
Status: v1.NodeStatus{
|
|
Conditions: []v1.NodeCondition{
|
|
{
|
|
Type: v1.NodeReady, Status: v1.ConditionFalse,
|
|
},
|
|
},
|
|
},
|
|
}
|
|
_, err = clientset.CoreV1().Nodes().Create(ctx, node, metav1.CreateOptions{})
|
|
assert.NoError(t, err)
|
|
nodes, err = GetNotReadyKubernetesNodes(clientset)
|
|
assert.NoError(t, err)
|
|
assert.Equal(t, 1, len(nodes))
|
|
|
|
// when all the nodes are not ready
|
|
allNodes, err := clientset.CoreV1().Nodes().List(ctx, metav1.ListOptions{})
|
|
assert.NoError(t, err)
|
|
for _, n := range allNodes.Items {
|
|
n.Status.Conditions[0].Status = v1.ConditionFalse
|
|
updateNode := n
|
|
_, err := clientset.CoreV1().Nodes().Update(ctx, &updateNode, metav1.UpdateOptions{})
|
|
assert.NoError(t, err)
|
|
}
|
|
nodes, err = GetNotReadyKubernetesNodes(clientset)
|
|
assert.NoError(t, err)
|
|
assert.Equal(t, 3, len(nodes))
|
|
}
|