Skip to content
This repository has been archived by the owner on Sep 5, 2024. It is now read-only.

rio --system-namespace is stuck during installation [v0.6.0-rc1] #765

Closed
izaac opened this issue Nov 13, 2019 · 5 comments
Closed

rio --system-namespace is stuck during installation [v0.6.0-rc1] #765

izaac opened this issue Nov 13, 2019 · 5 comments
Assignees
Labels
bug Something isn't working
Milestone

Comments

@izaac
Copy link

izaac commented Nov 13, 2019

Describe the bug
use the --system-namespace global option for rio install. rio takes a long time and it get stuck here:

rio --system-namespace riosystemtest install
Deploying Rio control plane....
| Waiting for system components. Not ready deployments: [rio-system/autoscaler rio-system/buildkitd rio-system/cert-manager ...]

To Reproduce
Steps to reproduce the behavior:

  1. rio --system-namespace riosystemtest install
  2. wait until the installation progress
  3. Cancel with Ctrl-C the install seems to be finished but deploying services is not working.

Expected behavior

Get rio installed in a different namespace and functional.

Kubernetes version & type (GKE, on-prem): kubectl version

Client Version: version.Info{Major:"1", Minor:"16", GitVersion:"v1.16.2", GitCommit:"c97fe5036ef3df2967d086711e6c0c405941e14b", GitTreeState:"clean", BuildDate:"2019-10-15T19:18:23Z", GoVersion:"go1.12.10", Compiler:"gc", Platform:"linux/amd64"}
Server Version: version.Info{Major:"1", Minor:"14+", GitVersion:"v1.14.8-gke.12", GitCommit:"188432a69210ca32cafded81b4dd1c063720cac0", GitTreeState:"clean", BuildDate:"2019-11-07T19:27:01Z", GoVersion:"go1.12.11b4", Compiler:"gc", Platform:"linux/amd64"}

Type:
Rio version: rio info

Rio Version: v0.6.0-rc1 (1dec45ce)
Rio CLI Version: v0.6.0-rc1 (1dec45ce)
Cluster Domain: xxxxxx.on-rio.io
Cluster Domain IPs: x.x.x.x
System Namespace: riosystemtest
Wildcard certificates: true

Additional context
rio systemlogs
output:

rio-controller | time="2019-11-13T16:32:54Z" level=info msg="Starting rio-controller, version: v0.6.0-rc1, git commit: 1dec45ce"
rio-controller | time="2019-11-13T16:32:54Z" level=info msg="Updating CRD services.rio.cattle.io"
rio-controller | time="2019-11-13T16:32:54Z" level=info msg="Updating CRD stacks.rio.cattle.io"
rio-controller | I1113 16:32:54.463657       1 leaderelection.go:241] attempting to acquire leader lease  riosystemtest/rio...
rio-controller | time="2019-11-13T16:32:54Z" level=info msg="listening at :443"
rio-controller | I1113 16:32:54.473414       1 leaderelection.go:251] successfully acquired lease riosystemtest/rio
rio-controller | time="2019-11-13T16:32:56Z" level=info msg="Starting /v1, Kind=ConfigMap controller"
rio-controller | E1113 16:33:00.835267       1 reflector.go:123] github.com/rancher/wrangler-api/pkg/generated/informers/externalversions/factory.go:119: Failed to list *v1alpha1.TrafficSplit: the server could not find the requested resource (get trafficsplits.split.smi-spec.io)
rio-controller | E1113 16:33:01.972029       1 reflector.go:123] github.com/rancher/wrangler-api/pkg/generated/informers/externalversions/factory.go:119: Failed to list *v1alpha1.TrafficSplit: the server could not find the requested resource (get trafficsplits.split.smi-spec.io)
rio-controller | E1113 16:33:03.025807       1 reflector.go:123] github.com/rancher/wrangler-api/pkg/generated/informers/externalversions/factory.go:119: Failed to list *v1alpha1.TrafficSplit: the server could not find the requested resource (get trafficsplits.split.smi-spec.io)
rio-controller | E1113 16:33:04.028511       1 reflector.go:123] github.com/rancher/wrangler-api/pkg/generated/informers/externalversions/factory.go:119: Failed to list *v1alpha1.TrafficSplit: the server could not find the requested resource (get trafficsplits.split.smi-spec.io)
@izaac izaac added the bug Something isn't working label Nov 13, 2019
@izaac izaac added this to the v0.6.0 milestone Nov 13, 2019
@davidnuzik davidnuzik modified the milestones: v0.6.0, v0.6.x Nov 13, 2019
@izaac
Copy link
Author

izaac commented Nov 14, 2019

I'm testing this with a cluster with more resources as I'm seeing gloo failed to deploy due to not enough CPU.

@izaac
Copy link
Author

izaac commented Nov 14, 2019

@StrongMonkey on a cluster with enough resources I still get stuck as I originally reported but If i cancel and start deploying services that work.

In a cluster of 3 worker nodes with 1 CPU each I see gloo failed to deploy. So that may be the reason I ended not being able to deploy services in it.

default    Warning    FailedScheduling    gloo-6657cbbb49-9lch2    0/3 nodes are available: 3 Insufficient cpu.

@StrongMonkey
Copy link
Contributor

@rancher-max available to test on master.

@rancher-max
Copy link
Contributor

This no longer hangs.

However, rio -s ps -w does not show workloads from the system namespace when I install this way. Now I have to specify the namespace: rio -n riosystemtest ps -w.
rio info properly shows the system namespace:

Rio Version: fad34c11-dirty (fad34c11)
Rio CLI Version: dev (HEAD)
Cluster Domain: xxxxxx.on-rio.io
Cluster Domain IPs: x.x.x.x
System Namespace: riosystemtest
Wildcard certificates: true

@rancher-max
Copy link
Contributor

Working in master

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
bug Something isn't working
Projects
None yet
Development

No branches or pull requests

4 participants