Skip to content

Lower threshold on backoff for sanity tests #1087

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Nov 23, 2022
Merged

Lower threshold on backoff for sanity tests #1087

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 6 additions & 2 deletions cmd/gce-pd-csi-driver/main.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -42,7 +42,9 @@ var (
httpEndpoint = flag.String("http-endpoint", "", "The TCP network address where the prometheus metrics endpoint will listen (example: `:8080`). The default is empty string, which means metrics endpoint is disabled.")
metricsPath = flag.String("metrics-path", "/metrics", "The HTTP path where prometheus metrics will be exposed. Default is `/metrics`.")

extraVolumeLabelsStr = flag.String("extra-labels", "", "Extra labels to attach to each PD created. It is a comma separated list of key value pairs like '<key1>=<value1>,<key2>=<value2>'. See https://cloud.google.com/compute/docs/labeling-resources for details")
errorBackoffInitialDurationMs = flag.Int("backoff-initial-duration-ms", 200, "The amount of ms for the initial duration of the backoff condition for controller publish/unpublish CSI operations. Default is 200.")
errorBackoffMaxDurationMs = flag.Int("backoff-max-duration-ms", 300000, "The amount of ms for the max duration of the backoff condition for controller publish/unpublish CSI operations. Default is 300000 (5m).")
extraVolumeLabelsStr = flag.String("extra-labels", "", "Extra labels to attach to each PD created. It is a comma separated list of key value pairs like '<key1>=<value1>,<key2>=<value2>'. See https://cloud.google.com/compute/docs/labeling-resources for details")

attachDiskBackoffDuration = flag.Duration("attach-disk-backoff-duration", 5*time.Second, "Duration for attachDisk backoff")
attachDiskBackoffFactor = flag.Float64("attach-disk-backoff-factor", 0.0, "Factor for attachDisk backoff")
Expand Down Expand Up @@ -122,7 +124,9 @@ func handle() {
if err != nil {
klog.Fatalf("Failed to get cloud provider: %v", err)
}
controllerServer = driver.NewControllerServer(gceDriver, cloudProvider)
initialBackoffDuration := time.Duration(*errorBackoffInitialDurationMs) * time.Millisecond
maxBackoffDuration := time.Duration(*errorBackoffMaxDurationMs) * time.Microsecond
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@judemars, looks like this should be time.Millisecond.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks! Sorry about that! Assigned #1098 to you.

controllerServer = driver.NewControllerServer(gceDriver, cloudProvider, initialBackoffDuration, maxBackoffDuration)
} else if *cloudConfigFilePath != "" {
klog.Warningf("controller service is disabled but cloud config given - it has no effect")
}
Expand Down
11 changes: 3 additions & 8 deletions pkg/gce-pd-csi-driver/controller.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,11 +37,6 @@ import (
gce "sigs.k8s.io/gcp-compute-persistent-disk-csi-driver/pkg/gce-cloud-provider/compute"
)

const (
errorBackoffInitialDuration = 200 * time.Millisecond
errorBackoffMaxDuration = 5 * time.Minute
)

type GCEControllerServer struct {
Driver *GCEDriver
CloudProvider gce.GCECompute
Expand Down Expand Up @@ -517,7 +512,7 @@ func (gceCS *GCEControllerServer) executeControllerPublishVolume(ctx context.Con

attached, err := diskIsAttachedAndCompatible(deviceName, instance, volumeCapability, readWrite)
if err != nil {
return nil, status.Error(codes.AlreadyExists, fmt.Sprintf("Disk %v already published to node %v but incompatbile: %v", volKey.Name, nodeID, err))
return nil, status.Error(codes.AlreadyExists, fmt.Sprintf("Disk %v already published to node %v but incompatible: %v", volKey.Name, nodeID, err))
}
if attached {
// Volume is attached to node. Success!
Expand Down Expand Up @@ -1596,8 +1591,8 @@ func pickRandAndConsecutive(slice []string, n int) ([]string, error) {
return ret, nil
}

func newCsiErrorBackoff() *csiErrorBackoff {
return &csiErrorBackoff{flowcontrol.NewBackOff(errorBackoffInitialDuration, errorBackoffMaxDuration)}
func newCsiErrorBackoff(initialDuration, errorBackoffMaxDuration time.Duration) *csiErrorBackoff {
return &csiErrorBackoff{flowcontrol.NewBackOff(initialDuration, errorBackoffMaxDuration)}
}

func (_ *csiErrorBackoff) backoffId(nodeId, volumeId string) csiErrorBackoffId {
Expand Down
2 changes: 2 additions & 0 deletions pkg/gce-pd-csi-driver/controller_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2149,6 +2149,8 @@ type backoffTesterConfig struct {
}

func newFakeCsiErrorBackoff(tc *clock.FakeClock) *csiErrorBackoff {
errorBackoffInitialDuration := 200 * time.Millisecond
errorBackoffMaxDuration := 5 * time.Minute
return &csiErrorBackoff{flowcontrol.NewFakeBackOff(errorBackoffInitialDuration, errorBackoffMaxDuration, tc)}
}

Expand Down
5 changes: 3 additions & 2 deletions pkg/gce-pd-csi-driver/gce-pd-driver.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,6 +16,7 @@ package gceGCEDriver

import (
"fmt"
"time"

csi "github.com/container-storage-interface/spec/lib/go/csi"
"google.golang.org/grpc/codes"
Expand Down Expand Up @@ -148,13 +149,13 @@ func NewNodeServer(gceDriver *GCEDriver, mounter *mount.SafeFormatAndMount, devi
}
}

func NewControllerServer(gceDriver *GCEDriver, cloudProvider gce.GCECompute) *GCEControllerServer {
func NewControllerServer(gceDriver *GCEDriver, cloudProvider gce.GCECompute, errorBackoffInitialDuration, errorBackoffMaxDuration time.Duration) *GCEControllerServer {
return &GCEControllerServer{
Driver: gceDriver,
CloudProvider: cloudProvider,
seen: map[string]int{},
volumeLocks: common.NewVolumeLocks(),
errorBackoff: newCsiErrorBackoff(),
errorBackoff: newCsiErrorBackoff(errorBackoffInitialDuration, errorBackoffMaxDuration),
}
}

Expand Down
6 changes: 5 additions & 1 deletion pkg/gce-pd-csi-driver/gce-pd-driver_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,6 +16,7 @@ package gceGCEDriver

import (
"testing"
"time"

gce "sigs.k8s.io/gcp-compute-persistent-disk-csi-driver/pkg/gce-cloud-provider/compute"
)
Expand Down Expand Up @@ -43,7 +44,10 @@ func initBlockingGCEDriver(t *testing.T, cloudDisks []*gce.CloudDisk, readyToExe
func initGCEDriverWithCloudProvider(t *testing.T, cloudProvider gce.GCECompute) *GCEDriver {
vendorVersion := "test-vendor"
gceDriver := GetGCEDriver()
controllerServer := NewControllerServer(gceDriver, cloudProvider)
errorBackoffInitialDuration := 200 * time.Millisecond
errorBackoffMaxDuration := 5 * time.Minute

controllerServer := NewControllerServer(gceDriver, cloudProvider, errorBackoffInitialDuration, errorBackoffMaxDuration)
err := gceDriver.SetupGCEDriver(driver, vendorVersion, nil, nil, controllerServer, nil)
if err != nil {
t.Fatalf("Failed to setup GCE Driver: %v", err)
Expand Down
3 changes: 2 additions & 1 deletion test/sanity/sanity_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,6 +21,7 @@ import (
"path"
"strings"
"testing"
"time"

"github.com/google/uuid"
"google.golang.org/grpc"
Expand Down Expand Up @@ -64,7 +65,7 @@ func TestSanity(t *testing.T) {

//Initialize GCE Driver
identityServer := driver.NewIdentityServer(gceDriver)
controllerServer := driver.NewControllerServer(gceDriver, cloudProvider)
controllerServer := driver.NewControllerServer(gceDriver, cloudProvider, 1*time.Millisecond, 5*time.Minute)
nodeServer := driver.NewNodeServer(gceDriver, mounter, deviceUtils, metadataservice.NewFakeService(), mountmanager.NewFakeStatter(mounter))
err = gceDriver.SetupGCEDriver(driverName, vendorVersion, extraLabels, identityServer, controllerServer, nodeServer)
if err != nil {
Expand Down