Skip to content

Collect boot logs even when bootstrap fails #1383

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
May 21, 2021
Merged

Collect boot logs even when bootstrap fails #1383

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
35 changes: 20 additions & 15 deletions test/e2e/azure_logcollector.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -45,11 +45,16 @@ type AzureLogCollector struct{}
func (k AzureLogCollector) CollectMachineLog(ctx context.Context, managementClusterClient client.Client, m *clusterv1.Machine, outputPath string) error {
var errors []error

if err := collectLogsFromNode(ctx, managementClusterClient, m, outputPath); err != nil {
am, err := getAzureMachine(ctx, managementClusterClient, m)
if err != nil {
return err
}

if err := collectLogsFromNode(ctx, managementClusterClient, m, am, outputPath); err != nil {
errors = append(errors, err)
}

if err := collectBootLog(ctx, m, outputPath); err != nil {
if err := collectBootLog(ctx, am, outputPath); err != nil {
errors = append(errors, err)
}

Expand All @@ -58,17 +63,14 @@ func (k AzureLogCollector) CollectMachineLog(ctx context.Context, managementClus
}

// collectLogsFromNode collects logs from various sources by ssh'ing into the node
func collectLogsFromNode(ctx context.Context, managementClusterClient client.Client, m *clusterv1.Machine, outputPath string) error {
func collectLogsFromNode(ctx context.Context, managementClusterClient client.Client, m *clusterv1.Machine, am *v1alpha4.AzureMachine, outputPath string) error {
cluster, err := util.GetClusterFromMetadata(ctx, managementClusterClient, m.ObjectMeta)
if err != nil {
return err
}

Logf("INFO: Collecting logs for machine %s in cluster %s in namespace %s\n", m.GetName(), cluster.Name, cluster.Namespace)
isWindows, err := isNodeWindows(ctx, managementClusterClient, m)
if err != nil {
return err
}
isWindows := isNodeWindows(am)

controlPlaneEndpoint := cluster.Spec.ControlPlaneEndpoint.Host
hostname := m.Spec.InfrastructureRef.Name
Expand Down Expand Up @@ -108,18 +110,19 @@ func collectLogsFromNode(ctx context.Context, managementClusterClient client.Cli
return kinderrors.AggregateConcurrent(linuxLogs(execToPathFn))
}

func isNodeWindows(ctx context.Context, managementClusterClient client.Client, m *clusterv1.Machine) (bool, error) {
func isNodeWindows(am *v1alpha4.AzureMachine) bool {
return am.Spec.OSDisk.OSType == azure.WindowsOS
}

func getAzureMachine(ctx context.Context, managementClusterClient client.Client, m *clusterv1.Machine) (*v1alpha4.AzureMachine, error) {
key := client.ObjectKey{
Namespace: m.Spec.InfrastructureRef.Namespace,
Name: m.Spec.InfrastructureRef.Name,
}

azMachine := &v1alpha4.AzureMachine{}
if err := managementClusterClient.Get(ctx, key, azMachine); err != nil {
return false, err
}

return azMachine.Spec.OSDisk.OSType == azure.WindowsOS, nil
err := managementClusterClient.Get(ctx, key, azMachine)
return azMachine, err
}

func linuxLogs(execToPathFn func(outputFileName string, command string, args ...string) func() error) []func() error {
Expand Down Expand Up @@ -251,8 +254,10 @@ func windowsNetworkLogs(execToPathFn func(outputFileName string, command string,
}

// collectBootLog collects boot logs of the vm by using azure boot diagnostics
func collectBootLog(ctx context.Context, m *clusterv1.Machine, outputPath string) error {
resourceId := strings.TrimPrefix(*m.Spec.ProviderID, azure.ProviderIDPrefix)
func collectBootLog(ctx context.Context, am *v1alpha4.AzureMachine, outputPath string) error {
Logf("INFO: Collecting boot logs for AzureMachine %s\n", am.GetName())

resourceId := strings.TrimPrefix(*am.Spec.ProviderID, azure.ProviderIDPrefix)
resource, err := autorest.ParseResourceID(resourceId)
if err != nil {
return errors.Wrap(err, "failed to parse resource id")
Expand Down
5 changes: 5 additions & 0 deletions test/e2e/azure_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,6 +31,7 @@ import (
. "github.com/onsi/gomega"
corev1 "k8s.io/api/core/v1"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/types"
"k8s.io/utils/pointer"
capi_e2e "sigs.k8s.io/cluster-api/test/e2e"
"sigs.k8s.io/cluster-api/test/framework/clusterctl"
Expand Down Expand Up @@ -75,6 +76,10 @@ var _ = Describe("Workload cluster creation", func() {
})

AfterEach(func() {
if result.Cluster == nil {
// this means the cluster failed to come up. We make an attempt to find the cluster to be able to fetch logs for the failed bootstrapping.
_ = bootstrapClusterProxy.GetClient().Get(ctx, types.NamespacedName{Name: clusterName, Namespace: namespace.Name}, result.Cluster)
}
dumpSpecResourcesAndCleanup(ctx, specName, bootstrapClusterProxy, artifactFolder, namespace, cancelWatches, result.Cluster, e2eConfig.GetIntervals, skipCleanup)
Expect(os.Unsetenv(AzureResourceGroup)).NotTo(HaveOccurred())
Expect(os.Unsetenv(AzureVNetName)).NotTo(HaveOccurred())
Expand Down