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

docker: kill signal API should include timeout context #9502

Merged
merged 1 commit into from
Dec 2, 2020
Merged

docker: kill signal API should include timeout context #9502

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
1 change: 1 addition & 0 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -72,6 +72,7 @@ BUG FIXES:
* csi: Fixed a bug where `nomad volume detach` would not accept prefixes for the node ID parameter. [[GH-9041](https://github.com/hashicorp/nomad/issues/9041)]
* csi: Fixed a bug where `nomad alloc status -verbose` would display an error when querying volumes. [[GH-9354](https://github.com/hashicorp/nomad/issues/9354)]
* csi: Fixed a bug where queries for CSI plugins could be interleaved, resulting in inconsistent counts of plugins. [[GH-9438](https://github.com/hashicorp/nomad/issues/9438)]
* driver/docker: Fixed a bug where the Docker daemon could block longer than the `kill_timeout`. [[GH-9502](https://github.com/hashicorp/nomad/issues/9502)
* driver/docker: Fixed a bug where the default `image_delay` configuration was ignored if the `gc` configuration was not set. [[GH-9101](https://github.com/hashicorp/nomad/issues/9101)]
* driver/raw_exec: Fixed a bug where raw_exec attempts to create a freezer cgroups for the tasks even when `no_cgroups` is set. [[GH-9328](https://github.com/hashicorp/nomad/issues/9328)]
* ui: Fixed a bug in the volume status page where read allocations and write allocations were not displayed. [[GH-9377](https://github.com/hashicorp/nomad/issues/9377)]
Expand Down
5 changes: 4 additions & 1 deletion drivers/docker/driver.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1461,7 +1461,10 @@ func (d *Driver) SignalTask(taskID string, signal string) error {
return fmt.Errorf("failed to parse signal: %v", err)
}

return h.Signal(sig)
// TODO: review whether we can timeout in this and other Docker API
// calls without breaking the expected client behavior.
// see https://github.com/hashicorp/nomad/issues/9503
return h.Signal(context.Background(), sig)
tgross marked this conversation as resolved.
Show resolved Hide resolved
}

func (d *Driver) ExecTask(taskID string, cmd []string, timeout time.Duration) (*drivers.ExecTaskResult, error) {
Expand Down
14 changes: 9 additions & 5 deletions drivers/docker/handle.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -103,7 +103,7 @@ func (h *taskHandle) Exec(ctx context.Context, cmd string, args []string) (*driv
return execResult, nil
}

func (h *taskHandle) Signal(s os.Signal) error {
func (h *taskHandle) Signal(ctx context.Context, s os.Signal) error {
// Convert types
sysSig, ok := s.(syscall.Signal)
if !ok {
Expand All @@ -116,8 +116,9 @@ func (h *taskHandle) Signal(s os.Signal) error {

dockerSignal := docker.Signal(sysSig)
opts := docker.KillContainerOptions{
ID: h.containerID,
Signal: dockerSignal,
ID: h.containerID,
Signal: dockerSignal,
Context: ctx,
}
return h.client.KillContainer(opts)

Expand All @@ -127,7 +128,10 @@ func (h *taskHandle) Signal(s os.Signal) error {
func (h *taskHandle) Kill(killTimeout time.Duration, signal os.Signal) error {
// Only send signal if killTimeout is set, otherwise stop container
if killTimeout > 0 {
if err := h.Signal(signal); err != nil {
ctx, cancel := context.WithTimeout(context.Background(), killTimeout)
defer cancel()

if err := h.Signal(ctx, signal); err != nil {
// Container has already been removed.
if strings.Contains(err.Error(), NoSuchContainerError) {
h.logger.Debug("attempted to signal nonexistent container")
Expand All @@ -146,7 +150,7 @@ func (h *taskHandle) Kill(killTimeout time.Duration, signal os.Signal) error {
select {
case <-h.waitCh:
return nil
case <-time.After(killTimeout):
case <-ctx.Done():
}
}

Expand Down