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

fix: Add skaffold internal error and return that instead of user cancelled #6846

Merged
merged 10 commits into from
Nov 15, 2021
Merged

fix: Add skaffold internal error and return that instead of user cancelled #6846

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
fda2a58
fix: Add skaffold internal error and return that instead of user canc…
tejal29 Nov 11, 2021
3bf856f
change logic to detect skaffold deplpy status
tejal29 Nov 12, 2021
a991781
fix lint
tejal29 Nov 12, 2021
f6cdeed
rever intended change
tejal29 Nov 12, 2021
5e5a547
fix tests
tejal29 Nov 12, 2021
60127d3
fix race condition between deployment recorded as successful and pod …
tejal29 Nov 12, 2021
0b072af
fix lint
tejal29 Nov 12, 2021
1d2c886
fix struct lint
tejal29 Nov 12, 2021
221c085
code review comments
tejal29 Nov 12, 2021
05a8b52
fix lint size of struct
tejal29 Nov 13, 2021
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
42 changes: 28 additions & 14 deletions docs/content/en/api/skaffold.swagger.json
View file
Open in desktop

Large diffs are not rendered by default.

1 change: 1 addition & 0 deletions docs/content/en/docs/references/api/grpc.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -980,6 +980,7 @@ For Cancelled Error code, use range 800 to 850.<br>
| STATUSCHECK_UNKNOWN_UNSCHEDULABLE | 502 | Container is unschedulable due to unknown reasons |
| STATUSCHECK_CONTAINER_WAITING_UNKNOWN | 503 | Container is waiting due to unknown reason |
| STATUSCHECK_UNKNOWN_EVENT | 509 | Container event reason unknown |
| STATUSCHECK_INTERNAL_ERROR | 514 | Status Check internal error |
| DEPLOY_UNKNOWN | 504 | Deploy failed due to unknown reason |
| SYNC_UNKNOWN | 505 | SYNC failed due to known reason |
| BUILD_UNKNOWN | 506 | Build failed due to unknown reason |
Expand Down
58 changes: 38 additions & 20 deletions pkg/skaffold/kubernetes/status/status_check.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -67,9 +67,10 @@ const (
)

type counter struct {
total int
pending int32
failed int32
total int
pending int32
failed int32
cancelled int32
}

type Config interface {
Expand Down Expand Up @@ -220,17 +221,19 @@ func (s *monitor) statusCheck(ctx context.Context, out io.Writer) (proto.StatusC

ctx, cancel := context.WithCancel(ctx)
defer cancel()
var exitStatusCode proto.StatusCode

for _, d := range resources {
wg.Add(1)
go func(r *resource.Resource) {
defer wg.Done()
// keep updating the resource status until it fails/succeeds/times out
pollResourceStatus(ctx, s.cfg, r)
rcCopy := c.markProcessed(r.Status().Error())
rcCopy := c.markProcessed(ctx, r.StatusCode())
Copy link
Member

Choose a reason for hiding this comment

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

Can we move the logic here and in getSkaffoldDeployStatus() into counter? This scattering of logic is awkward:

  • markProcessed can check if ctx.Err() == context.Canceled to determine if the user cancelled
  • have markProcessed return (counter, failed bool) to replace the resourceFailed() check on line 235
  • move the exitStatusCode determination into counter
  • line 249 would return c.result()

Copy link
Contributor Author

Choose a reason for hiding this comment

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

We can't use markProcessed can check if ctx.Err() == context.Canceled to determine if the user cancelled since, we fail fast and cancel context if a deployment failed.
So at this point, its difficult to detect if a user hit cntrl C Vs, code cancelling all the status check for all resources.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

I had suggestions 2-4 implemented but then the unit tests were becoming more complex. Maybe the counter need to be renamed to deployState.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Ok, so i remember, why we cant have counter store the exitStatusCode. The counter gets updated in a go routine.
The exit code is stored in resources. It makes sense to declare another mutex to update the exit status code safely.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Ok,
I am addressed 1 and 2 from the suggestions.
For 3, 4:

  • I have declared a exitStatusCode in checkStatus and updated the code to only capture first failure.
  • kept the getDeployStatus as is.

s.printStatusCheckSummary(out, r, rcCopy)
// if one deployment fails, cancel status checks for all deployments.
if r.Status().Error() != nil && r.StatusCode() != proto.StatusCode_STATUSCHECK_USER_CANCELLED {
// if one resource fails, cancel status checks for all resources.
if resourceFailed(r.StatusCode()) {
exitStatusCode = r.StatusCode()
cancel()
}
}(d)
Expand All @@ -243,8 +246,7 @@ func (s *monitor) statusCheck(ctx context.Context, out io.Writer) (proto.StatusC

// Wait for all deployment statuses to be fetched
wg.Wait()
cancel()
Copy link
Contributor Author

Choose a reason for hiding this comment

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

There is a defer cancel at L223

return getSkaffoldDeployStatus(c, resources)
return getSkaffoldDeployStatus(c, exitStatusCode)
}

func getStandalonePods(ctx context.Context, client kubernetes.Interface, ns string, l *label.DefaultLabeller, deadlineDuration time.Duration) ([]*resource.Resource, error) {
Expand Down Expand Up @@ -382,18 +384,21 @@ func pollResourceStatus(ctx context.Context, cfg kubectl.Config, r *resource.Res
}
}

func getSkaffoldDeployStatus(c *counter, rs []*resource.Resource) (proto.StatusCode, error) {
func getSkaffoldDeployStatus(c *counter, sc proto.StatusCode) (proto.StatusCode, error) {
// return overall code cancelled if status check for all resources was cancelled
if int(c.cancelled) == c.total && c.total > 0 {
return proto.StatusCode_STATUSCHECK_USER_CANCELLED, fmt.Errorf("status check cancelled")
}
// return success if no failures find.
if c.failed == 0 {
return proto.StatusCode_STATUSCHECK_SUCCESS, nil
}
// construct an error message and return appropriate error code
err := fmt.Errorf("%d/%d deployment(s) failed", c.failed, c.total)
for _, r := range rs {
if r.StatusCode() != proto.StatusCode_STATUSCHECK_SUCCESS &&
r.StatusCode() != proto.StatusCode_STATUSCHECK_USER_CANCELLED {
return r.StatusCode(), err
}
if sc == proto.StatusCode_STATUSCHECK_SUCCESS || sc == 0 {
return proto.StatusCode_STATUSCHECK_INTERNAL_ERROR, err
}
return proto.StatusCode_STATUSCHECK_USER_CANCELLED, err
return sc, err
}

func getDeadline(d int) time.Duration {
Expand Down Expand Up @@ -483,8 +488,12 @@ func newCounter(i int) *counter {
}
}

func (c *counter) markProcessed(err error) counter {
if err != nil && err != context.Canceled {
func (c *counter) markProcessed(ctx context.Context, sc proto.StatusCode) counter {
if resourceCancelled(sc) {
log.Entry(ctx).Debug("marking resource status check cancelled", sc)
atomic.AddInt32(&c.cancelled, 1)
} else if resourceFailed(sc) {
log.Entry(ctx).Debugf("marking resource failed due to error code %s", sc)
atomic.AddInt32(&c.failed, 1)
}
atomic.AddInt32(&c.pending, -1)
Expand All @@ -493,12 +502,21 @@ func (c *counter) markProcessed(err error) counter {

func (c *counter) copy() counter {
return counter{
total: c.total,
pending: c.pending,
failed: c.failed,
total: c.total,
pending: c.pending,
failed: c.failed,
cancelled: c.cancelled,
}
}

func resourceFailed(sc proto.StatusCode) bool {
return sc != proto.StatusCode_STATUSCHECK_SUCCESS && sc != proto.StatusCode_STATUSCHECK_USER_CANCELLED
}

func resourceCancelled(sc proto.StatusCode) bool {
return sc == proto.StatusCode_STATUSCHECK_USER_CANCELLED
}

type NoopMonitor struct {
status.NoopMonitor
}
Expand Down
119 changes: 56 additions & 63 deletions pkg/skaffold/kubernetes/status/status_check_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -230,97 +230,82 @@ func TestGetDeployments(t *testing.T) {
}
}

func TestGetDeployStatus(t *testing.T) {
func TestExitErrorMessage(t *testing.T) {
tests := []struct {
description string
counter *counter
deployments []*resource.Resource
sc proto.StatusCode
expected string
expectedCode proto.StatusCode
shouldErr bool
}{
{
description: "one error",
counter: &counter{total: 2, failed: 1},
deployments: []*resource.Resource{
resource.NewResource("foo", resource.ResourceTypes.Deployment, "test", time.Second).
WithPodStatuses([]proto.StatusCode{proto.StatusCode_STATUSCHECK_NODE_DISK_PRESSURE}),
},
description: "one error",
counter: &counter{total: 2, failed: 1},
expected: "1/2 deployment(s) failed",
expectedCode: proto.StatusCode_STATUSCHECK_NODE_DISK_PRESSURE,
sc: proto.StatusCode_STATUSCHECK_POD_INITIALIZING,
expectedCode: proto.StatusCode_STATUSCHECK_POD_INITIALIZING,
shouldErr: true,
},
{
description: "no error",
counter: &counter{total: 2},
deployments: []*resource.Resource{
withStatus(
resource.NewResource("r1", resource.ResourceTypes.Deployment, "test", 1),
&proto.ActionableErr{ErrCode: proto.StatusCode_STATUSCHECK_SUCCESS},
),
withStatus(
resource.NewResource("r2", resource.ResourceTypes.Deployment, "test", 1),
&proto.ActionableErr{ErrCode: proto.StatusCode_STATUSCHECK_SUCCESS},
),
},
description: "no error",
sc: proto.StatusCode_STATUSCHECK_SUCCESS,
expectedCode: proto.StatusCode_STATUSCHECK_SUCCESS,
counter: &counter{total: 2},
},
{
description: "multiple errors",
counter: &counter{total: 3, failed: 2},
expected: "2/3 deployment(s) failed",
deployments: []*resource.Resource{
resource.NewResource("foo", resource.ResourceTypes.Deployment, "test", time.Second).
WithPodStatuses([]proto.StatusCode{proto.StatusCode_STATUSCHECK_NODE_DISK_PRESSURE}),
},
expectedCode: proto.StatusCode_STATUSCHECK_NODE_DISK_PRESSURE,
description: "multiple errors",
counter: &counter{total: 3, failed: 2},
expected: "2/3 deployment(s) failed",
sc: proto.StatusCode_STATUSCHECK_CONFIG_CONNECTOR_FAILED,
expectedCode: proto.StatusCode_STATUSCHECK_CONFIG_CONNECTOR_FAILED,
shouldErr: true,
},
{
description: "0 deployments",
counter: &counter{},
description: "0 deployments",
counter: &counter{total: 0},
expectedCode: proto.StatusCode_STATUSCHECK_SUCCESS,
},
{
description: "unable to retrieve pods for deployment",
counter: &counter{total: 1, failed: 1},
deployments: []*resource.Resource{
withStatus(
resource.NewResource("deployment", resource.ResourceTypes.Deployment, "test", 1),
&proto.ActionableErr{ErrCode: proto.StatusCode_STATUSCHECK_DEPLOYMENT_FETCH_ERR},
),
},
shouldErr: true,
description: "unable to retrieve pods for deployment",
counter: &counter{total: 1, failed: 1},
sc: proto.StatusCode_STATUSCHECK_DEPLOYMENT_FETCH_ERR,
expectedCode: proto.StatusCode_STATUSCHECK_DEPLOYMENT_FETCH_ERR,
shouldErr: true,
},
{
description: "one deployment failed and others cancelled and or succeeded",
counter: &counter{total: 3, failed: 2},
deployments: []*resource.Resource{
withStatus(
resource.NewResource("deployment-cancelled", resource.ResourceTypes.Deployment, "test", 1),
&proto.ActionableErr{ErrCode: proto.StatusCode_STATUSCHECK_USER_CANCELLED},
),
withStatus(
resource.NewResource("deployment-success", resource.ResourceTypes.Deployment, "test", 1),
&proto.ActionableErr{ErrCode: proto.StatusCode_STATUSCHECK_SUCCESS},
),
withStatus(
resource.NewResource("deployment", resource.ResourceTypes.Deployment, "test", 1),
&proto.ActionableErr{ErrCode: proto.StatusCode_STATUSCHECK_DEPLOYMENT_FETCH_ERR},
),
},
description: "one deployment failed and others cancelled and or succeeded",
counter: &counter{total: 3, failed: 2},
sc: proto.StatusCode_STATUSCHECK_NODE_DISK_PRESSURE,
expectedCode: proto.StatusCode_STATUSCHECK_NODE_DISK_PRESSURE,
expected: "2/3 deployment(s) failed",
shouldErr: true,
},
{
description: "deployments did not stabilize within deadline returns the pod error",
counter: &counter{total: 1, failed: 1},
sc: proto.StatusCode_STATUSCHECK_UNHEALTHY,
expected: "1/1 deployment(s) failed",
expectedCode: proto.StatusCode_STATUSCHECK_UNHEALTHY,
shouldErr: true,
},
{
description: "user cancelled session",
counter: &counter{total: 2, failed: 0, cancelled: 2},
sc: proto.StatusCode_STATUSCHECK_USER_CANCELLED,
expected: "status check cancelled",
expectedCode: proto.StatusCode_STATUSCHECK_USER_CANCELLED,
shouldErr: true,
expectedCode: proto.StatusCode_STATUSCHECK_DEPLOYMENT_FETCH_ERR,
},
}

for _, test := range tests {
testutil.Run(t, test.description, func(t *testutil.T) {
testEvent.InitializeState([]latestV1.Pipeline{{}})
errCode, err := getSkaffoldDeployStatus(test.counter, test.deployments)
actual, err := getSkaffoldDeployStatus(test.counter, test.sc)
t.CheckError(test.shouldErr, err)
t.CheckDeepEqual(test.expectedCode, actual)
if test.shouldErr {
t.CheckErrorContains(test.expected, err)
t.CheckDeepEqual(test.expectedCode, errCode)
}
})
}
Expand Down Expand Up @@ -529,29 +514,37 @@ func TestResourceMarkProcessed(t *testing.T) {
tests := []struct {
description string
c *counter
err error
sc proto.StatusCode
expected counter
}{
{
description: "when deployment failed, counter is updated",
c: newCounter(10),
err: errors.New("some ae"),
sc: proto.StatusCode_STATUSCHECK_DEADLINE_EXCEEDED,
expected: counter{total: 10, failed: 1, pending: 9},
},
{
description: "when deployment is cancelled, failed is not updated",
c: newCounter(10),
sc: proto.StatusCode_STATUSCHECK_USER_CANCELLED,
expected: counter{total: 10, failed: 0, pending: 9},
},
{
description: "when deployment is successful, counter is updated",
c: newCounter(10),
sc: proto.StatusCode_STATUSCHECK_SUCCESS,
expected: counter{total: 10, failed: 0, pending: 9},
},
{
description: "counter when 1 deployment is updated correctly",
c: newCounter(1),
sc: proto.StatusCode_STATUSCHECK_SUCCESS,
expected: counter{total: 1, failed: 0, pending: 0},
},
}
for _, test := range tests {
testutil.Run(t, test.description, func(t *testutil.T) {
t.CheckDeepEqual(test.expected, test.c.markProcessed(test.err), cmp.AllowUnexported(counter{}))
t.CheckDeepEqual(test.expected, test.c.markProcessed(context.Background(), test.sc), cmp.AllowUnexported(counter{}))
})
}
}
Expand Down
Loading