Bug 1501792

Summary:	When imagechanges will trigger twice deploys.
Product:	OpenShift Container Platform	Reporter:	zhou ying <yinzhou>
Component:	openshift-controller-manager	Assignee:	Michal Fojtik <mfojtik>
Status:	CLOSED WONTFIX	QA Contact:	zhou ying <yinzhou>
Severity:	low	Docs Contact:
Priority:	low
Version:	3.7.0	CC:	aos-bugs, jminter, mfojtik, rromanni, yinzhou
Target Milestone:	---
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:	1466583	Environment:
Last Closed:	2019-08-23 12:58:26 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:	1466583
Bug Blocks:

Description zhou ying 2017-10-13 08:34:32 UTC

+++ This bug was initially created as a clone of Bug #1466583 +++

Description of problem:
Create app and set imagechange trigger with automatic, when image changes will trigger twice deploys.

Version-Release number of selected component (if applicable):
openshift v3.7.0-0.147.1
kubernetes v1.7.6+a08f5eeb62
etcd 3.2.1


How reproducible:
always

Steps to Reproduce:
1.Login in OpenShift and create project;
2.Create app :
`oc new-app centos/ruby-22-centos7~https://github.com/openshift/ruby-ex.git`
3.Trigger new build to make the image changes, check the DC.


Actual results:
3.When image changes will trigger the deployment twice.

[root@qe-yinzhou36-master-1 ~]# oc describe dc/ruby-ex
Name:        ruby-ex
Namespace:    testzy
Created:    4 minutes ago
Labels:        app=ruby-ex
Annotations:    openshift.io/generated-by=OpenShiftNewApp
Latest Version:    4
Selector:    app=ruby-ex,deploymentconfig=ruby-ex
Replicas:    1
Triggers:    Config, Image(ruby-ex@latest, auto=true)
Strategy:    Rolling
Template:
Pod Template:
  Labels:    app=ruby-ex
        deploymentconfig=ruby-ex
  Annotations:    openshift.io/generated-by=OpenShiftNewApp
  Containers:
   ruby-ex:
    Image:        docker-registry.default.svc:5000/testzy/ruby-ex@sha256:d07573ca263e7492b47b33789c69be975da55d077f08690decce52d68a1a7631
    Port:        8080/TCP
    Environment:    <none>
    Mounts:        <none>
  Volumes:        <none>

Deployment #4 (latest):
    Name:        ruby-ex-4
    Created:    29 seconds ago
    Status:        Complete
    Replicas:    1 current / 1 desired
    Selector:    app=ruby-ex,deployment=ruby-ex-4,deploymentconfig=ruby-ex
    Labels:        app=ruby-ex,openshift.io/deployment-config.name=ruby-ex
    Pods Status:    1 Running / 0 Waiting / 0 Succeeded / 0 Failed
Deployment #3:
    Created:    34 seconds ago
    Status:        Failed
    Replicas:    0 current / 0 desired
Deployment #2:
    Created:    2 minutes ago
    Status:        Complete
    Replicas:    0 current / 0 desired

Events:
  FirstSeen    LastSeen    Count    From                SubObjectPath    Type        Reason                Message
  ---------    --------    -----    ----                -------------    --------    ------                -------
  2m        2m        10    deploymentconfig-controller            Normal        DeploymentAwaitingCancellation    Deployment of version 2 awaiting cancellation of older running deployments
  2m        2m        1    deployer-controller                Normal        RolloutCancelled        ruby-ex-1: Rollout for "testzy/ruby-ex-1" cancelled
  2m        2m        1    deploymentconfig-controller            Normal        DeploymentCreated        Created new replication controller "ruby-ex-2" for version 2
  34s        34s        1    deploymentconfig-controller            Normal        DeploymentCreated        Created new replication controller "ruby-ex-3" for version 3
  33s        33s        1    deploymentconfig-controller            Normal        DeploymentCancelled        Cancelled deployment "ruby-ex-3" superceded by version 4
  33s        30s        10    deploymentconfig-controller            Normal        DeploymentAwaitingCancellation    Deployment of version 4 awaiting cancellation of older running deployments
  29s        29s        1    deploymentconfig-controller            Normal        DeploymentCreated        Created new replication controller "ruby-ex-4" for version 4
  29s        29s        1    deployer-controller                Normal        RolloutCancelled        ruby-ex-3: Rollout for "testzy/ruby-ex-3" cancelled

[root@qe-yinzhou36-master-1 ~]# oc get is ruby-ex -o yaml
apiVersion: v1
......
status:
  dockerImageRepository: docker-registry.default.svc:5000/testzy/ruby-ex
  tags:
  - items:
    - created: 2017-06-30T02:34:02Z
      dockerImageReference: docker-registry.default.svc:5000/testzy/ruby-ex@sha256:d07573ca263e7492b47b33789c69be975da55d077f08690decce52d68a1a7631
      generation: 1
      image: sha256:d07573ca263e7492b47b33789c69be975da55d077f08690decce52d68a1a7631
    - created: 2017-06-30T02:32:11Z
      dockerImageReference: docker-registry.default.svc:5000/testzy/ruby-ex@sha256:26beea1f9a4f4b4752ea2e07aaba0906b17f9f1d744ddcf813458b0f45c65c45
      generation: 1
      image: sha256:26beea1f9a4f4b4752ea2e07aaba0906b17f9f1d744ddcf813458b0f45c65c45
    tag: latest


Expected results:
3.Once image changes only trigger the deployment once.

Additional info:

--- Additional comment from Michal Fojtik on 2017-06-30 06:14:45 EDT ---

Steps 2 and 3 are confusing... The new-app should trigger a new build automatically. Are you triggering an extra build? That might result in the second deployment.

--- Additional comment from Michal Fojtik on 2017-06-30 06:16:53 EDT ---

If my comment is not valid and you don't trigger an extra build, can you please provide
the master log when this happens? We have similar issue in GitHub that is causing new-app test to flake hard. This looks very similar to that but we are not able to reproduce this. This only happens on GCE for us.

--- Additional comment from zhou ying on 2017-06-30 06:37 EDT ---



--- Additional comment from Michal Fojtik on 2017-06-30 11:21:43 EDT ---

The fix should be https://github.com/openshift/origin/pull/14882

We also need to update Ansible playbook and cut release to get this changes.

--- Additional comment from Michal Fojtik on 2017-07-03 04:34:10 EDT ---

This is basically a duplicate of #1463499

Comment 1 Michal Fojtik 2017-10-17 14:45:52 UTC

Why we have this clone?

Comment 2 Michal Fojtik 2017-10-18 08:49:45 UTC

The original bug was duplicate of https://bugzilla.redhat.com/show_bug.cgi?id=1463499 which was verified. Do you see this behavior in 3.7.0 (iow. we regressed?)

Comment 3 zhou ying 2017-10-18 09:21:56 UTC

We could reproduced with version v3.7.0-0.147.1, but for version openshift v3.7.0-0.155.0 can't reproduce again, so weird, will set it to low.

Comment 4 Ramón Román Nissen 2018-02-07 10:41:59 UTC

I have reproduced this on OCP 3.7.14 running in the CDK 3.3.0-1. The problem is that the first deployment gets cancelled automatically, and the second has problems pulling the image:

Failed to pull image "orders:latest": rpc error: code = 2 desc = unauthorized: authentication required

If I cancel the second deployment and trigger one manually, the deployment completes with no problem. This makes the use of the fabric8 maven plugin very difficult, as manual intervention is required.