examine nil pointer before applying backend protocol configurations #678

freehan · 2019-03-12T00:27:36Z

Mitigates #675

rramkumar1 · 2019-03-12T00:42:25Z

pkg/healthchecks/healthchecks.go

@@ -62,6 +62,10 @@ const (
 	// backends, the port or named port specified in the Backend Service is
 	// used for health checking.
 	UseServingPortSpecification = "USE_SERVING_PORT"
+
+	newHealthCheckErrorMessageTemplate = "the %v health check configuration on the exiting health check %v is nil. " +
+		"This is usually caused by application protocol changed. " +


what do you mean by exiting?

rramkumar1 · 2019-03-12T00:42:42Z

pkg/healthchecks/healthchecks.go

@@ -62,6 +62,10 @@ const (
 	// backends, the port or named port specified in the Backend Service is
 	// used for health checking.
 	UseServingPortSpecification = "USE_SERVING_PORT"
+
+	newHealthCheckErrorMessageTemplate = "the %v health check configuration on the exiting health check %v is nil. " +
+		"This is usually caused by application protocol changed. " +


"..caused by an application..."

rramkumar1 · 2019-03-12T00:42:56Z

pkg/healthchecks/healthchecks.go

+
+	newHealthCheckErrorMessageTemplate = "the %v health check configuration on the exiting health check %v is nil. " +
+		"This is usually caused by application protocol changed. " +
+		"Please revert the change on application protocol to avoid this error message."


"...revert the change on the application..."

Suggested change

"Please revert the change on application protocol to avoid this error message."

"Please revert the change on the application's protocol to avoid this error message."

I don't have a lot of insight into how the ingress-gce works internally, but I think a TODO on handling this type of change would be helpful. I think progressing from HTTP-1.1 to HTTP-2 (soon HTTP-3) or introducing changes on the protocol to adopt a service mesh will be a common occurrence.

yeah. Agreed. In the long run we want to support this. But currently, due to GCE api changes, it was broken. So we need to put this short term fix in place so that it can be cherry picked.

Can we add the TODO just so we remember to do it?

cmluciano · 2019-03-12T13:53:42Z

pkg/healthchecks/healthchecks.go

@@ -62,6 +62,10 @@ const (
 	// backends, the port or named port specified in the Backend Service is
 	// used for health checking.
 	UseServingPortSpecification = "USE_SERVING_PORT"
+
+	newHealthCheckErrorMessageTemplate = "the %v health check configuration on the exiting health check %v is nil. " +
+		"This is usually caused by application protocol changed. " +


Suggested change

"This is usually caused by application protocol changed. " +

"This is usually caused by a change of the application's protocol." +

cmluciano · 2019-03-12T13:58:56Z

pkg/healthchecks/healthchecks.go

+
+	newHealthCheckErrorMessageTemplate = "the %v health check configuration on the exiting health check %v is nil. " +
+		"This is usually caused by application protocol changed. " +
+		"Please revert the change on application protocol to avoid this error message."


Suggested change

"Please revert the change on application protocol to avoid this error message."

"Please revert the change on the application's protocol to avoid this error message."

I don't have a lot of insight into how the ingress-gce works internally, but I think a TODO on handling this type of change would be helpful. I think progressing from HTTP-1.1 to HTTP-2 (soon HTTP-3) or introducing changes on the protocol to adopt a service mesh will be a common occurrence.

freehan · 2019-03-12T17:32:13Z

fixed.

mariantalla · 2019-03-13T14:30:50Z

Hi, is there anything missing for this to be /lgtmed?
cc @MrHohn @thockin

mariantalla · 2019-03-14T12:26:27Z

Assigning for lgtm (please reassign if I've done this wrong):
/assign @MrHohn

MrHohn · 2019-03-14T17:45:12Z

This LGTM, but will defer to @rramkumar1 to make the final call.
/assign @rramkumar1

bowei · 2019-03-14T19:15:12Z

pkg/healthchecks/healthchecks.go

@@ -62,6 +62,10 @@ const (
 	// backends, the port or named port specified in the Backend Service is
 	// used for health checking.
 	UseServingPortSpecification = "USE_SERVING_PORT"
+
+	newHealthCheckErrorMessageTemplate = "the %v health check configuration on the existing health check %v is nil. " +
+		"This is usually caused by an application protocol changed. " +


can we give the name of the actual field involved, this error is a bit vague sounding. are we talking about the GCP object, the protocol used by the app itself, etc

HealthCheck object contains HTTPHealthCheck configurations. It has bunch of fields.

In this case, the controller thinks that HTTP2HealthCheck configuration is nil because it was not vendored.

bowei · 2019-03-14T19:17:09Z

pkg/healthchecks/healthchecks.go

 		v.HTTPHealthCheck = computealpha.HTTPHealthCheck(*hc.HttpsHealthCheck)
 	case annotations.ProtocolHTTP2:
+		if hc.Http2HealthCheck == nil {
+			err = fmt.Errorf(newHealthCheckErrorMessageTemplate, annotations.ProtocolHTTP2, hc.Name)


can we not return immediately (same in all case above)

freehan · 2019-03-14T23:02:04Z

Ready for next round

rramkumar1 · 2019-03-14T23:11:55Z

pkg/healthchecks/healthchecks.go

@@ -62,6 +62,10 @@ const (
 	// backends, the port or named port specified in the Backend Service is
 	// used for health checking.
 	UseServingPortSpecification = "USE_SERVING_PORT"
+
+	newHealthCheckErrorMessageTemplate = "the %v health check configurations on the existing health check %v is nil. " +


"the %v health check configuration..."

rramkumar1 · 2019-03-14T23:12:39Z

pkg/healthchecks/healthchecks.go

+
+	newHealthCheckErrorMessageTemplate = "the %v health check configuration on the exiting health check %v is nil. " +
+		"This is usually caused by application protocol changed. " +
+		"Please revert the change on application protocol to avoid this error message."


Can we add the TODO just so we remember to do it?

freehan · 2019-03-15T00:17:15Z

Done

rramkumar1 · 2019-03-15T17:12:40Z

/lgtm
/approve

k8s-ci-robot · 2019-03-15T17:13:53Z

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: freehan, rramkumar1

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

~~OWNERS~~ [freehan,rramkumar1]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

Cherrypick #678 into release-1.5

k8s-ci-robot added cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. size/S Denotes a PR that changes 10-29 lines, ignoring generated files. approved Indicates a PR has been approved by an approver from all required OWNERS files. labels Mar 12, 2019

k8s-ci-robot requested review from MrHohn and thockin March 12, 2019 00:27

rramkumar1 reviewed Mar 12, 2019

View reviewed changes

nikopen mentioned this pull request Mar 12, 2019

[Failing test] gce-cos-<branch>-ingress are timing out kubernetes/kubernetes#75186

Closed

cmluciano reviewed Mar 12, 2019

View reviewed changes

freehan force-pushed the nil-pointer branch from 2b17999 to e55e043 Compare March 12, 2019 17:32

freehan force-pushed the nil-pointer branch from e55e043 to c518487 Compare March 12, 2019 17:37

k8s-ci-robot assigned MrHohn Mar 14, 2019

k8s-ci-robot assigned rramkumar1 Mar 14, 2019

bowei reviewed Mar 14, 2019

View reviewed changes

freehan force-pushed the nil-pointer branch from c518487 to d297145 Compare March 14, 2019 22:25

rramkumar1 reviewed Mar 14, 2019

View reviewed changes

rramkumar1 changed the title ~~examine nil pointer before applying configurations~~ examine nil pointer before applying backend protocol configurations Mar 14, 2019

examine nil pointer before applying configurations

f929fd3

freehan force-pushed the nil-pointer branch from d297145 to f929fd3 Compare March 14, 2019 23:18

k8s-ci-robot added size/M Denotes a PR that changes 30-99 lines, ignoring generated files. and removed size/S Denotes a PR that changes 10-29 lines, ignoring generated files. labels Mar 14, 2019

k8s-ci-robot added the lgtm "Looks good to me", indicates that a PR is ready to be merged. label Mar 15, 2019

k8s-ci-robot merged commit d2150fc into kubernetes:master Mar 15, 2019

freehan mentioned this pull request Mar 15, 2019

Cherrypick #678 into release-1.5 #691

Merged

rramkumar1 added a commit that referenced this pull request Mar 15, 2019

Merge pull request #691 from freehan/cherrypick

cff9722

Cherrypick #678 into release-1.5

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

examine nil pointer before applying backend protocol configurations #678

examine nil pointer before applying backend protocol configurations #678

freehan commented Mar 12, 2019

rramkumar1 Mar 12, 2019

rramkumar1 Mar 12, 2019

rramkumar1 Mar 12, 2019

cmluciano Mar 12, 2019

freehan Mar 12, 2019

rramkumar1 Mar 14, 2019

cmluciano Mar 12, 2019

cmluciano Mar 12, 2019

freehan commented Mar 12, 2019

mariantalla commented Mar 13, 2019

mariantalla commented Mar 14, 2019

MrHohn commented Mar 14, 2019

bowei Mar 14, 2019

freehan Mar 14, 2019

bowei Mar 14, 2019

freehan commented Mar 14, 2019

rramkumar1 Mar 14, 2019

rramkumar1 Mar 14, 2019

freehan commented Mar 15, 2019

rramkumar1 commented Mar 15, 2019

k8s-ci-robot commented Mar 15, 2019

	"Please revert the change on application protocol to avoid this error message."
	"Please revert the change on the application's protocol to avoid this error message."

	"This is usually caused by application protocol changed. " +
	"This is usually caused by a change of the application's protocol." +

examine nil pointer before applying backend protocol configurations #678

examine nil pointer before applying backend protocol configurations #678

Conversation

freehan commented Mar 12, 2019

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

freehan commented Mar 12, 2019

mariantalla commented Mar 13, 2019

mariantalla commented Mar 14, 2019

MrHohn commented Mar 14, 2019

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

freehan commented Mar 14, 2019

Choose a reason for hiding this comment

Choose a reason for hiding this comment

freehan commented Mar 15, 2019

rramkumar1 commented Mar 15, 2019

k8s-ci-robot commented Mar 15, 2019