indicate builds that can't push #3316

deads2k · 2015-06-18T19:17:48Z

This adds a target *ImageStreamNode into the ImagePipeline. It does this by creating an edge between an ImageStreamTagNode and an associated ImageStreamNode. When the graph is traversed to build the ImagePipeline, the edge is chased and resolved.

Output looks like this if the build won't be able to push:

In project foo

docker-nfs-server deploys docker-nfs-server:latest <- docker build of https://github.com/deads2k/docker-nfs-server 
  build 1 new for 14 minutes (can't push to image)
  #1 deployed 14 minutes ago - 1 pod

Warning: Some of your builds are pointing to image streams, but the administrator has not configured the integrated Docker registry (oadm registry).
To see more information about a Service or DeploymentConfig, use 'oc describe service <name>' or 'oc describe dc <name>'.
You can use 'oc get all' to see lists of each of the types described above.

@smarterclayton sidebar: why do we have veneers over the map traversals? "Math is hard"?

smarterclayton · 2015-06-18T19:23:26Z

pkg/api/graph/deployment.go

-	Image ImageTagLocation
-	Build *BuildConfigNode
+	Image       ImageTagLocation
+	ImageStream *ImageStreamNode


I'd really prefer this to be under ImageTagLocation, such that you can ask the imagetaglocation if it is resolved as being valid.

That obviates the need for tracking the image stream separately - it's just a check on whether an outbound edge exists when you create the pipeline.

So would I, but that means that you end up with a method on a shared interface that's only sometimes valid. That's your preference?

Also, why do we have the veneer on the graph. Straight graph traversal would make this point moot.

I'm thinking more behavior oriented - Resolved() true/false or Exists() true/false

----- Original Message -----

@@ -21,8 +21,9 @@ type DeploymentPipelineMap
map[*DeploymentConfigNode][]ImagePipeline
// ImagePipeline represents a build, its output, and any inputs. The input
// to a build may be another ImagePipeline.
type ImagePipeline struct {

Image ImageTagLocation

Build *BuildConfigNode

Image ImageTagLocation

ImageStream *ImageStreamNode

So would I, but that means that you end up with a method on a shared
interface that's only sometimes valid. That's your preference?

Also, why do we have the veneer on the graph. Straight graph traversal would
make this point moot.

Reply to this email directly or view it on GitHub:
https://github.com/openshift/origin/pull/3316/files#r32769193

deads2k · 2015-06-18T20:08:42Z

initial refactor done, looking at adding the overall check now.

deads2k · 2015-06-18T20:18:18Z

Updated text in description.

deads2k · 2015-06-19T19:37:48Z

I'd like to get #3334 and then rebase this. That will fix the imagepruning test.

deads2k · 2015-07-06T19:24:34Z

@ncdc builds on #3404 (in merge queue), so only second commit is interesting.

ncdc · 2015-07-07T13:47:40Z

Rebase pls?

deads2k · 2015-07-07T15:01:19Z

rebased

ncdc · 2015-07-07T15:54:13Z

pkg/cmd/cli/describe/projectstatus.go

+	imageStreamFailure := ""
+	// if we're using an image stream and that image stream is the internal registry and that registry doesn't exist
+	if (build.Parameters.Output.To != nil) && !pushTargetResolved {
+		imageStreamFailure = fmt.Sprintf(" (can't push to image)")


Do you plan to add any varargs or should this just be a plain string?

deads2k · 2015-07-07T17:15:18Z

comments addressed and one bug fixed.

ncdc · 2015-07-07T17:27:49Z

I don't think I have any additional comments. Ready to test in Jenkins?

deads2k · 2015-07-07T17:47:03Z

I don't think I have any additional comments. Ready to test in Jenkins?

Sure, though the merge task does the same.

deads2k · 2015-07-07T17:47:51Z

[test]

openshift-bot · 2015-07-07T17:50:41Z

continuous-integration/openshift-jenkins/test SUCCESS (https://ci.openshift.redhat.com/jenkins/job/test_pull_requests_origin/3562/)

deads2k · 2015-07-07T19:32:12Z

delete all flake re[test]

ncdc · 2015-07-07T20:03:33Z

Unhappy jenkins [test]

ncdc · 2015-07-07T21:03:25Z

Hub flake [test]

deads2k · 2015-07-08T11:50:20Z

green!

ncdc · 2015-07-08T13:07:54Z

pkg/image/graph/nodes/types.go

+	IsFound bool
+}
+
+func (n ImageStreamNode) Found() bool {


This isn't (yet) used anywhere, is it?

This isn't (yet) used anywhere, is it?

Not yet. The missing resource status check requires it.

ncdc · 2015-07-08T13:23:13Z

[merge]

openshift-bot · 2015-07-08T13:27:20Z

continuous-integration/openshift-jenkins/merge SUCCESS (https://ci.openshift.redhat.com/jenkins/job/merge_pull_requests_origin/2687/) (Image: devenv-fedora_1954)

openshift-bot · 2015-07-08T13:27:21Z

Evaluated for origin up to 0f7089d

Merged by openshift-bot

smarterclayton reviewed Jun 18, 2015
View reviewed changes

deads2k force-pushed the status-for-missing-registry branch from 9318df5 to 839290d Compare June 18, 2015 20:07

deads2k force-pushed the status-for-missing-registry branch from 839290d to 07b9272 Compare June 18, 2015 20:17

deads2k force-pushed the status-for-missing-registry branch from 07b9272 to 7e432c2 Compare July 6, 2015 19:21

deads2k force-pushed the status-for-missing-registry branch from 7e432c2 to c3c4721 Compare July 7, 2015 15:01

ncdc reviewed Jul 7, 2015
View reviewed changes

deads2k force-pushed the status-for-missing-registry branch from 3cb55aa to 207d2d2 Compare July 7, 2015 17:47

indicate builds that can't push

0f7089d

deads2k force-pushed the status-for-missing-registry branch from 207d2d2 to 0f7089d Compare July 8, 2015 12:26

ncdc reviewed Jul 8, 2015
View reviewed changes

openshift-bot pushed a commit that referenced this pull request Jul 8, 2015

Merge pull request #3316 from deads2k/status-for-missing-registry

e66e8f2

Merged by openshift-bot

openshift-bot merged commit e66e8f2 into openshift:master Jul 8, 2015

deads2k deleted the status-for-missing-registry branch July 31, 2015 18:31

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

indicate builds that can't push #3316

indicate builds that can't push #3316

deads2k commented Jun 18, 2015

smarterclayton Jun 18, 2015

smarterclayton Jun 18, 2015

deads2k Jun 18, 2015

smarterclayton Jun 18, 2015

deads2k commented Jun 18, 2015

deads2k commented Jun 18, 2015

deads2k commented Jun 19, 2015

deads2k commented Jul 6, 2015

ncdc commented Jul 7, 2015

deads2k commented Jul 7, 2015

ncdc Jul 7, 2015

deads2k commented Jul 7, 2015

ncdc commented Jul 7, 2015

deads2k commented Jul 7, 2015

deads2k commented Jul 7, 2015

openshift-bot commented Jul 7, 2015

deads2k commented Jul 7, 2015

ncdc commented Jul 7, 2015

ncdc commented Jul 7, 2015

deads2k commented Jul 8, 2015

ncdc Jul 8, 2015

deads2k Jul 8, 2015

ncdc commented Jul 8, 2015

openshift-bot commented Jul 8, 2015

openshift-bot commented Jul 8, 2015

indicate builds that can't push #3316

indicate builds that can't push #3316

Conversation

deads2k commented Jun 18, 2015

smarterclayton Jun 18, 2015

Choose a reason for hiding this comment

smarterclayton Jun 18, 2015

Choose a reason for hiding this comment

deads2k Jun 18, 2015

Choose a reason for hiding this comment

smarterclayton Jun 18, 2015

Choose a reason for hiding this comment

deads2k commented Jun 18, 2015

deads2k commented Jun 18, 2015

deads2k commented Jun 19, 2015

deads2k commented Jul 6, 2015

ncdc commented Jul 7, 2015

deads2k commented Jul 7, 2015

ncdc Jul 7, 2015

Choose a reason for hiding this comment

deads2k commented Jul 7, 2015

ncdc commented Jul 7, 2015

deads2k commented Jul 7, 2015

deads2k commented Jul 7, 2015

openshift-bot commented Jul 7, 2015

deads2k commented Jul 7, 2015

ncdc commented Jul 7, 2015

ncdc commented Jul 7, 2015

deads2k commented Jul 8, 2015

ncdc Jul 8, 2015

Choose a reason for hiding this comment

deads2k Jul 8, 2015

Choose a reason for hiding this comment

ncdc commented Jul 8, 2015

openshift-bot commented Jul 8, 2015

openshift-bot commented Jul 8, 2015