"Waiting for SSH to be available" message

Rancher 2.x
1

Rancher version: v.2.6.3
Installation option:

  • Helm version: v3.8.0
  • k3s version: v1.22.6+k3s1
    Openstack:
  • Microstack Installation - version ussuri, Revision 245

I want to integrate OpenStack Cloud Provider in the Rancher interface. To do that in the Drivers tab (on Rancher Interface) I enabled the option " OpenStack ".
Then, I created a " Node Template " with the following options:

authURL: https://XX:5000/v3
availabilityZone: nova
domainId: default
flavorId: 1
floatingipPool: XX
imageName: cirros
ipVersion: 4
keypairName: <keyPair to SSH to the instance>
netId: <network "test" ID>
password: XX
privateKeyFile: <content of XX.pem>
region: XX
secGroups: default
sshPort: 22
sshUser: <user to OpenStack machine>
tenantId: XX
userDomainId: default
username: admin

Then, I created a Cluster with “ RKE1 ” and OpenStack options. I selected Node Template that I have created and options “etcd”, “Control Plane” and “Worker”. In " Cluster Options ", I selected " External " in the Cloud Provider field and edit the YAML to introduce the following configuration:

cloud_provider:
 name: openstack
 openstackCloudProvider:
    global:
      username: admin
      password: XX
      auth-url: https://XX:5000/v3
      tenant-id:  XX
      domain-id: default
    load_balancer:
      subnet-id: XX
    block_storage:
      ignore-volume-az: true
    route:
      router-id: XX
bastion_host:
  port: '22'
  ssh_agent_auth: true
  user: <User of instance>

When I clicked on "Create " and verified on " Provisioning Log " in the Cluster, this is the output:

image

In Rancher Interface, the " Provisioning Log " shows the message " Waiting for SSH to be available… ".

1 Like
2

Hello !

I also try to deploy a Kubernetes cluster in an OpenStack Project, with a private network, throught a bastion host.
Without success.
Like I posted in Rancher 2.3.5, Openstack and Bastion/Jump hosts

When I copy/paste the bastion section in the cluster config in the Rancher Web UI, the ssh_key element is removed…

I don’t konw if it the only reason why it doesn’t work, but that’s all I can see.

Does someone knows if there are logs somwhere ? The Provisoning Logs tab is empty in the Web UI.

3

Hi.

At the moment this problem is already advanced (my problem was that my Rancher server didn’t have the IP route to the network inside OpenStack). But this integration between Rancher and OpenStack is not clear and in my case I still have problems and not a successful integration.
But you could see this link: Deploying Rancher On OpenStack - vkhitrin.com

In cluster config is normal that password rancher and ssh key content will don’t appear, since they are sensitive information.

You should check the routing between networks and also if your cluster config is correct.

The logs you can see:
Logs openstack: Find nova.log file

Logs Rancher
Ref.: Rancher Docs: Logging
#kubectl -n cattle-system logs -l app=rancher -c rancher
#kubectl logs -n cattle-system

4

Strange, the cluster I tried to create is not there anymore, I have deleted it.
But I see logs from rancher trying to connect to a cluster :

2022/03/04 10:46:12 [ERROR] Failed to handling tunnel request from 10.42.1.99:49818: response 401: failed authentication
2022/03/04 10:46:13 [ERROR] Failed to handling tunnel request from 10.42.1.99:52124: response 400: cluster not found
2022/03/04 10:46:15 [ERROR] Failed to dial steve aggregation server: websocket: bad handshake
2022/03/04 10:45:57 [ERROR] Failed to dial steve aggregation server: websocket: bad handshake