Skip to content
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

Jump to bottom

Issue 71614: Protect log message maps #71617

Merged
merged 1 commit into from
Dec 3, 2018
Merged

Issue 71614: Protect log message maps #71617

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.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
6 changes: 6 additions & 0 deletions pkg/kubelet/kuberuntime/kuberuntime_manager.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,7 @@ import (
"errors"
"fmt"
"os"
"sync"
"time"

cadvisorapi "github.com/google/cadvisor/info/v1"
Expand Down Expand Up @@ -131,6 +132,7 @@ type kubeGenericRuntimeManager struct {

// Time last per-container error message was printed
errorPrinted map[string]time.Time
errorMapLock sync.Mutex
}

// KubeGenericRuntime is a interface contains interfaces for container runtime and command.
Expand Down Expand Up @@ -830,6 +832,8 @@ func (m *kubeGenericRuntimeManager) killPodWithSyncResult(pod *v1.Pod, runningPo
}

func (m *kubeGenericRuntimeManager) cleanupErrorTimeouts() {
m.errorMapLock.Lock()
defer m.errorMapLock.Unlock()
for name, timeout := range m.errorPrinted {
if time.Now().Sub(timeout) >= identicalErrorDelay {
delete(m.errorPrinted, name)
Expand Down Expand Up @@ -886,6 +890,8 @@ func (m *kubeGenericRuntimeManager) GetPodStatus(uid kubetypes.UID, name, namesp

// Get statuses of all containers visible in the pod.
containerStatuses, err := m.getPodContainerStatuses(uid, name, namespace)
m.errorMapLock.Lock()
defer m.errorMapLock.Unlock()
if err != nil {
lastMsg, ok := m.lastError[podFullName]
if !ok || err.Error() != lastMsg || time.Now().Sub(m.errorPrinted[podFullName]) >= identicalErrorDelay {
Expand Down
12 changes: 12 additions & 0 deletions pkg/kubelet/remote/remote_runtime.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,6 +21,7 @@ import (
"errors"
"fmt"
"strings"
"sync"
"time"

"google.golang.org/grpc"
Expand All @@ -40,6 +41,7 @@ type RemoteRuntimeService struct {
lastError map[string]string
// Time last per-container error message was printed
errorPrinted map[string]time.Time
errorMapLock sync.Mutex
}

const (
Expand Down Expand Up @@ -236,8 +238,10 @@ func (r *RemoteRuntimeService) StopContainer(containerID string, timeout int64)
ctx, cancel := getContextWithTimeout(t)
defer cancel()

r.errorMapLock.Lock()
delete(r.lastError, containerID)
delete(r.errorPrinted, containerID)
r.errorMapLock.Unlock()
_, err := r.runtimeClient.StopContainer(ctx, &runtimeapi.StopContainerRequest{
ContainerId: containerID,
Timeout: timeout,
Expand All @@ -256,8 +260,10 @@ func (r *RemoteRuntimeService) RemoveContainer(containerID string) error {
ctx, cancel := getContextWithTimeout(r.timeout)
defer cancel()

r.errorMapLock.Lock()
delete(r.lastError, containerID)
delete(r.errorPrinted, containerID)
r.errorMapLock.Unlock()
_, err := r.runtimeClient.RemoveContainer(ctx, &runtimeapi.RemoveContainerRequest{
ContainerId: containerID,
})
Expand Down Expand Up @@ -287,6 +293,8 @@ func (r *RemoteRuntimeService) ListContainers(filter *runtimeapi.ContainerFilter

// Clean up any expired last-error timers
func (r *RemoteRuntimeService) cleanupErrorTimeouts() {
r.errorMapLock.Lock()
defer r.errorMapLock.Unlock()
for ID, timeout := range r.errorPrinted {
if time.Now().Sub(timeout) >= identicalErrorDelay {
delete(r.lastError, ID)
Expand All @@ -304,6 +312,8 @@ func (r *RemoteRuntimeService) ContainerStatus(containerID string) (*runtimeapi.
ContainerId: containerID,
})
r.cleanupErrorTimeouts()
r.errorMapLock.Lock()
defer r.errorMapLock.Unlock()
if err != nil {
// Don't spam the log with endless messages about the same failure.
lastMsg, ok := r.lastError[containerID]
Expand Down Expand Up @@ -491,6 +501,8 @@ func (r *RemoteRuntimeService) ContainerStats(containerID string) (*runtimeapi.C
ContainerId: containerID,
})
r.cleanupErrorTimeouts()
r.errorMapLock.Lock()
defer r.errorMapLock.Unlock()
if err != nil {
lastMsg, ok := r.lastError[containerID]
if !ok || err.Error() != lastMsg || time.Now().Sub(r.errorPrinted[containerID]) >= identicalErrorDelay {
Expand Down