K8s controller failed to bootstrap with 2.8/edge
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Canonical Juju |
Fix Released
|
Undecided
|
Unassigned |
Bug Description
Error pulling the image when bootstrapping a controller from 2.8/edge
$ microk8s.kubectl describe pods controller-0 -n controller-
Name: controller-0
Namespace: controller-
Priority: 0
Node: canonical/
Start Time: Wed, 17 Jun 2020 18:34:45 +0200
Labels: controller-
Annotations: juju.io/controller: b68567ec-
Status: Pending
IP: 10.1.73.30
IPs:
IP: 10.1.73.30
Controlled By: StatefulSet/
Containers:
mongodb:
Container ID: containerd:
Image: jujusolutions/
Image ID: docker.
Port: 37017/TCP
Host Port: 0/TCP
Command:
mongod
Args:
-
-
-
-
--port=37017
--journal
-
--quiet
-
--ipv6
--auth
-
-
--bind_ip_all
State: Waiting
Reason: CrashLoopBackOff
Last State: Terminated
Reason: Error
Exit Code: 1
Started: Wed, 17 Jun 2020 18:35:13 +0200
Finished: Wed, 17 Jun 2020 18:35:13 +0200
Ready: False
Restart Count: 2
Limits:
memory: 1536Mi
Requests:
memory: 1536Mi
Liveness: exec [mongo --port=37017 --ssl --sslAllowInval
Readiness: exec [mongo --port=37017 --ssl --sslAllowInval
Environment: <none>
Mounts:
/var/lib/juju from storage (rw)
/
/
/
/
api-server:
Container ID:
Image: jujusolutions/
Image ID:
Port: <none>
Host Port: <none>
Command:
/bin/sh
Args:
-c
export JUJU_DATA_
export JUJU_TOOLS_
mkdir -p $JUJU_TOOLS_DIR
cp /opt/jujud $JUJU_TOOLS_
echo Installing Dashboard...
export gui='/var/
mkdir -p $gui
curl -sSf -o $gui/gui.tar.bz2 --retry 10 --noproxy 127.0.0.
[ -f $gui/gui.tar.bz2 ] && sha256sum $gui/gui.tar.bz2 > $gui/jujugui.sha256
[ -f $gui/jujugui.sha256 ] && (grep 'e3215baf556a8b
test -e $JUJU_DATA_
$
State: Waiting
Reason: ImagePullBackOff
Ready: False
Restart Count: 0
Limits:
memory: 1536Mi
Requests:
memory: 1536Mi
Environment: <none>
Mounts:
/var/lib/juju from storage (rw)
/
/
/
/
/
Conditions:
Type Status
Initialized True
Ready False
ContainersReady False
PodScheduled True
Volumes:
storage:
Type: PersistentVolum
ClaimName: storage-
ReadOnly: false
controller-
Type: Secret (a volume populated by a Secret)
SecretName: controller-secret
Optional: false
controller-
Type: Secret (a volume populated by a Secret)
SecretName: controller-secret
Optional: false
controller-
Type: ConfigMap (a volume populated by a ConfigMap)
Name: controller-
Optional: false
controller-
Type: ConfigMap (a volume populated by a ConfigMap)
Name: controller-
Optional: false
default-
Type: Secret (a volume populated by a Secret)
SecretName: default-token-f5snz
Optional: false
QoS Class: Burstable
Node-Selectors: <none>
Tolerations: node.kubernetes
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning FailedScheduling <unknown> default-scheduler running "VolumeBinding" filter plugin for pod "controller-0": pod has unbound immediate PersistentVolum
Warning FailedScheduling <unknown> default-scheduler running "VolumeBinding" filter plugin for pod "controller-0": pod has unbound immediate PersistentVolum
Normal Scheduled <unknown> default-scheduler Successfully assigned controller-
Warning BackOff 30s (x2 over 39s) kubelet, canonical Back-off restarting failed container
Normal Pulling 30s (x2 over 42s) kubelet, canonical Pulling image "jujusolutions/
Warning Failed 29s (x2 over 40s) kubelet, canonical Failed to pull image "jujusolutions/
Warning Failed 29s (x2 over 40s) kubelet, canonical Error: ErrImagePull
Normal Pulled 17s (x3 over 43s) kubelet, canonical Container image "jujusolutions/
Normal Created 16s (x3 over 43s) kubelet, canonical Created container mongodb
Normal BackOff 16s (x4 over 39s) kubelet, canonical Back-off pulling image "jujusolutions/
Warning Failed 16s (x4 over 39s) kubelet, canonical Error: ImagePullBackOff
Normal Started 16s (x3 over 42s) kubelet, canonical Started container mongodb
Changed in juju: | |
status: | Fix Committed → Fix Released |
The problem here is probably that the 2.8.1 docker image (docker. io/jujusolution s/jujud- operator: 2.8.1.3802) isn't uploaded to docker.io.
I'm not sure whether this is an issue w/ the CI not having a job to upload an image for edge, or an issue w/ the image getting tagged the wrong way for the edge snap. Either way, we should probably fix it so that folks can do testing on k8s w/ the edge snap!