Hide Forgot
Description of problem: After a node is marked as unschedulable and pods have been evacuated, scaling (up or down) of a replicationController results in pods getting schedules on the node that is unschedulable. Version-Release number of selected component (if applicable): atomic-openshift-3.1.0.4-1.git.15.5e061c3.el7aos.x86_64 How reproducible: evacuate pods from a node, scale a replicationController Steps to Reproduce: [root@test-master-0 ~]# oc -n testing-2 get pods -o wide NAME READY STATUS RESTARTS AGE NODE mongodb-1-0zxcg 1/1 Running 0 1m test-node-primary-0.example.com mongodb-1-60t65 1/1 Running 0 46s test-node-primary-2.example.com mongodb-1-ru4ru 1/1 Running 0 1m test-node-primary-2.example.com mongodb-1-wk558 1/1 Running 0 17h test-node-primary-1.example.com nodejs-mongodb-example-6-6mmze 1/1 Running 2 20h test-node-primary-1.example.com nodejs-mongodb-example-6-e6j0w 1/1 Running 0 3m test-node-primary-0.example.com [root@test-master-0 ~]# oadm manage-node test-node-primary-2.example.com --schedulable=false NAME LABELS STATUS AGE test-node-primary-2.example.com kubernetes.io/hostname=test-node-primary-2.example.com,region=primary,zone=default Ready,SchedulingDisabled 16d [root@test-master-0 ~]# oadm manage-node test-node-primary-2.example.com --evacuate Migrating these pods on node: test-node-primary-2.example.com NAME READY STATUS RESTARTS AGE mongodb-1-60t65 1/1 Running 0 1m mongodb-1-ru4ru 1/1 Running 0 2m [root@test-master-0 ~]# oc -n testing-2 get pods -o wide NAME READY STATUS RESTARTS AGE NODE mongodb-1-0zxcg 1/1 Running 0 2m test-node-primary-0.example.com mongodb-1-g5o4r 1/1 Running 0 22s test-node-primary-1.example.com mongodb-1-rxrdb 1/1 Running 0 22s test-node-primary-0.example.com mongodb-1-wk558 1/1 Running 0 17h test-node-primary-1.example.com nodejs-mongodb-example-6-6mmze 1/1 Running 2 20h test-node-primary-1.example.com nodejs-mongodb-example-6-e6j0w 1/1 Running 0 4m test-node-primary-0.example.com [root@test-master-0 ~]# oc -n testing-2 scale --replicas=3 rc mongodb-1 replicationcontroller "mongodb-1" scaled [root@test-master-0 ~]# oc -n testing-2 get pods -o wide NAME READY STATUS RESTARTS AGE NODE mongodb-1-rxrdb 1/1 Running 0 1m test-node-primary-0.example.com mongodb-1-v4l7b 1/1 Running 0 26s test-node-primary-2.example.com mongodb-1-wk558 1/1 Running 0 17h test-node-primary-1.example.com nodejs-mongodb-example-6-6mmze 1/1 Running 2 20h test-node-primary-1.example.com nodejs-mongodb-example-6-e6j0w 1/1 Running 0 4m test-node-primary-0.example.com [root@test-master-0 ~]# oc get nodes NAME LABELS STATUS AGE test-master-0.example.com kubernetes.io/hostname=test-master-0.example.com,region=master,zone=default Ready,SchedulingDisabled 16d test-node-infra-0.example.com kubernetes.io/hostname=test-node-infra-0.example.com,region=infra,zone=default Ready 16d test-node-infra-1.example.com kubernetes.io/hostname=test-node-infra-1.example.com,region=infra,zone=default Ready 16d test-node-primary-0.example.com kubernetes.io/hostname=test-node-primary-0.example.com,region=primary,zone=default Ready 16d test-node-primary-1.example.com kubernetes.io/hostname=test-node-primary-1.example.com,region=primary,zone=default Ready 16d test-node-primary-2.example.com kubernetes.io/hostname=test-node-primary-2.example.com,region=primary,zone=default Ready,SchedulingDisabled 16d Actual results: pods get schedule on test-node-primary-2.example.com Expected results: no pod scheduled on test-node-primary-2.example.com Additional info: [root@test-master-0 ~]# oc -n testing-2 describe pod mongodb-1-v4l7b Name: mongodb-1-v4l7b Namespace: testing-2 Image(s): registry.access.redhat.com/rhscl/mongodb-26-rhel7:latest Node: test-node-primary-2.example.com/10.19.0.249 Start Time: Thu, 07 Apr 2016 07:56:06 +0000 Labels: deployment=mongodb-1,deploymentconfig=mongodb,name=mongodb Status: Running Reason: Message: IP: 10.1.5.70 Replication Controllers: mongodb-1 (3/3 replicas created) Containers: mongodb: Container ID: docker://65299d2d4e6362dc7c8ff09c6b9fa02a7b95809f798bd76b4e8bd92061bd20f8 Image: registry.access.redhat.com/rhscl/mongodb-26-rhel7:latest Image ID: docker://19c92ed464ccfaa085af8ed8cca18edfa242c337c9fcab6c9c7dd8b5cb2b9c3c QoS Tier: cpu: BestEffort memory: Guaranteed Limits: memory: 512Mi Requests: memory: 512Mi State: Running Started: Thu, 07 Apr 2016 07:56:09 +0000 Ready: True Restart Count: 0 Environment Variables: MONGODB_USER: userLUV MONGODB_PASSWORD: SgNJjwXpFoaO2Cm2 MONGODB_DATABASE: sampledb MONGODB_ADMIN_PASSWORD: YgfBUE2gH4CpNKNc Conditions: Type Status Ready True Volumes: default-token-uu427: Type: Secret (a secret that should populate this volume) SecretName: default-token-uu427 Events: FirstSeen LastSeen Count From SubobjectPath Reason Message ───────── ──────── ───── ──── ───────────── ────── ─────── 10m 10m 1 {scheduler } Scheduled Successfully assigned mongodb-1-v4l7b to test-node-primary-2.example.com 7m 7m 1 {scheduler } FailedScheduling Failed for reason Region and possibly others 7m 7m 1 {kubelet test-node-primary-2.example.com} implicitly required container POD Pulled Container image "openshift3/ose-pod:v3.1.0.4" already present on machine 7m 7m 1 {kubelet test-node-primary-2.example.com} implicitly required container POD Started Started with docker id 1ab9cda791de 7m 7m 1 {kubelet test-node-primary-2.example.com} implicitly required container POD Created Created with docker id 1ab9cda791de 6m 6m 1 {kubelet test-node-primary-2.example.com} spec.containers{mongodb} Pulled Container image "registry.access.redhat.com/rhscl/mongodb-26-rhel7:latest" already present on machine 6m 6m 1 {kubelet test-node-primary-2.example.com} spec.containers{mongodb} Created Created with docker id 65299d2d4e63 6m 6m 1 {kubelet test-node-primary-2.example.com} spec.containers{mongodb} Started Started with docker id 65299d2d4e63
I haven't been able to reproduce this.
As the environment has died, I cant replicate. Will reopen if I got it replicated on a new environment. Thanks for the help!
Trying to reproduce the issue on origin v1.1.6 and ose v3.1.1.6. With the same result not being able to reproduce it.