We are using Ansible to deploy an image to Google Kubernetes Cluster (GKE).
We have setup Ubuntu 20.04 and Python 3.8.5.
playbook.main.yml:
---
- hosts: localhost
vars:
k8s_file_path: /home/pesinn/Documents/...
become: yes
become_method: sudo
roles:
- k8s
main.yml:
- name: First Deployment
k8s:
kubeconfig: /home/pesinn/.kube/config
src: "{{k8s_file_path}}/deployment.yml"
When trying to deploy the image defined in deployment.yml file, by running the playbook, we get this error:
kubernetes.config.config_exception.ConfigException: cmd-path: process returned 1
Cmd: /home/pesinn/y/google-cloud-sdk/bin/gcloud config config-helper --format=json
Stderr: WARNING: Could not open the configuration file: [/root/.config/gcloud/configurations/config_default].
ERROR: (gcloud.config.config-helper) You do not currently have an active account selected.
Please run:
$ gcloud auth login
What we've already done
Initialized the cloud: gcloud init
Logged in and chosen a project gcloud auth login
Run export GOOGLE_APPLICATION_CREDENTIALS="path_to_service_account_key.json"
Run gcloud container clusters get-credentials {gke_project} --region {region}
Run the playbook sudo ansible-playbook playbook.main.yml -vvv
Run gcloud config config-helper --format=json on the local machine without any problems
What is very strange here is that we're logged in for sure. We can access the GKE cluster through kubectl command on the local machine. However, Ansible complains about us not being logged in. Also, in the error logs, we see that it is trying to open /root/.config/gcloud/configurations/config_default. Our default config file is, on the other hand, located in the home folder.
This error occurs randomly. Sometimes Ansible can detect our login and deploys the image, but sometimes it gives us this error. Both scenarios can happens without any code changes being made.
For some reason, ansible does not use GCP's default environment variables for authentication.
You can set
GCP_AUTH_KIND
GCP_SERVICE_ACCOUNT_EMAIL
GCP_SERVICE_ACCOUNT_FILE
GCP_SCOPES
GCP_SERVICE_ACCOUNT_FILE is the equivalent of GOOGLE_APPLICATION_CREDENTIALS
Reference: https://docs.ansible.com/ansible/latest/scenario_guides/guide_gce.html#providing-credentials-as-environment-variables
Related
I have been learning Cloud-Init for several days to do an automatic deployment. To achieve this, and apply certain configurations, I am using Ansible playbooks. The problem that I have found is that I am not able to make the playbook run directly on the operating system that is being installed.
I leave you the user-data file that I am using.
#cloud-config
autoinstall:
version: 1
identity:
hostname: hostname
password: "$6$cOciYeIErEet80Rv$YX8qt6vizXgcUkgIPSKD1qNZNxe77tSWOY3k/0.i8D8EpApaGNuyucxJvONmZiRj4rVM3L6EE4sLKcnzYVcMj/ "
username: ubuntu
storage:
layout:
name: direct
locale: es_ES
timezone: "Europe/Madrid"
keyboard:
layout: es
packages:
- sshpass
- ansible
- git
late-commands:
- git clone https://github.com/MarcOrfilaCarreras/dotfiles /target/root/dotfiles
- ansible-playbook -i inventory-test /root/dotfiles/ansible/playbooks/docker.yml -u ubuntu -e "ansible_password=ubuntu" -e "ansible_become_pass=ubuntu"
PS: I am using Ubuntu Server 22.04, the Ansible command is temporary and only for testing and I know that I have to change the identity fields.
If you want to configure localhost, it's better to use local transport (which is -c local in command line).
Basically, change ansible call to:
ansible-playbook -i inventory-test /root/dotfiles/ansible/playbooks/docker.yml -c local
This will bypass all SSH things and run locally.
i've createa an eks cluster using terraform, and now i want to deploy this ansible role in the cluster
https://github.com/githubixx/ansible-role-traefik-kubernetes.
Unfortunatley, when running the command ansible-playbook i get this error message
status:Failure, message:forbidden: User system:anonymous cannot get path /apis
when the role tries to check if the namespace exists. If i run the same command using kubectl i correctly get an answer.
When using the kubectl cli in a Windows DOS prompt, I get a prompt to enter a username, that works fine but when I press enter after entering a username the prompt for a password appears and then immediately acts like I hit the enter key, no chance to enter the password, looks like this, from the screen print you can see that I am using kubectl version 1.15.
If I try this using Git Bash, it behaves the same but responds with the error shown below
Same deal where the password prompt is not waiting for input.
Anyone ever seen this or have any thoughts on how I can provide a username and password to kubectl without storing it a plain test in the config file?
Also, I am using a corporate Kubernates cluster, so no options to move to a more current version or do anything else that would require admin access.
Posting this answer as community wiki with general guidelines for issues similar to this:
TL;DR
The prompt for username and password is most probably caused by misconfigured .kube/config.
As for:
Anyone ever seen this or have any thoughts on how I can provide a username and password to kubectl without storing it a plain test in the config file?
There are a lot of possibilities for authentication in Kubernetes. All of them have some advantages and disadvantages. Please take a look on below links:
Medium.com: Comparing kubernetes authentication methods
Kubernetes.io: Authentication
The prompt for username and password can appear when .kube/config file is misconfigured. I included one possible reason below:
Starting with correctly configured .kube/config for a minikube instance.
apiVersion: v1
clusters:
- cluster:
certificate-authority: PATH_TO_SOMEWHERE/.minikube/ca.crt
server: https://172.17.0.3:8443
name: minikube
contexts:
- context:
cluster: minikube
user: minikube
name: minikube
current-context: minikube
kind: Config
preferences: {}
users:
- name: minikube
user:
client-certificate: PATH_TO_SOMEWHERE/client.crt
client-key: PATH_TO_SOMEWHERE/client.key
Issuing commands with above .kube/config should not prompt for user and password as below:
$ kubectl get pods
No resources found in default namespace.
Editing .kube/config and changing:
user: minikube
to:
user: not-minikube
Will lead to:
$ kubectl get pods
Please enter Username: minikube
Please enter Password:
Correctly configuring .kube/config is heavily dependent on a solution used (like minikube, kubeadm provisioned cluster, and a managed cluster like GKE). Please refer to official documentation of solution used.
I found this post because I had the same problem, and I knew the cause after seeing people's answers. Comment to add.
I fixed the kube config with the command
aws eks update-kubeconfig --region region-code --name my-cluster
Create a kubeconfig for Amazon EKS
I am unable to run ping commands from a ansible host (using localhost, see below).
I built a simple playbook to run ping using the command module:
---
#
- name: GET INFO
hosts: localhost
tasks:
- name: return motd to registered var
command: "/usr/bin/ping 10.39.120.129"
register: mymotd
- name: debug output
debug: var=mymotd
However, I this error: "ping: socket: Operation not permitted"
Seems like there is a permissions issue. However, looking at the /usr/bin directory, it looks like ping would be executable to me:
"-rwxr-xr-x. 1 root root 66176 Aug 4 2017 ping",
I cannot become or use sudo, it seems like tower is locked down for that and I don't have the authority to change it either.
Anyone have any suggestions? What brought me to this, is that I am trying to run ping in a custom module and getting a similar issue.
Thanks
ping binary needs to have the SETUID bit set to be fully runable as a normal user, which is not the case on your server.
You need to run as root:
chmod u+s $(which ping)
If you don't have root access and cannot have this done by an admin, I'm affraid you're stuck... unless the server you are trying to ping is a machine you can manage with ansible.
In this later case, there is a ping module you can use. It is not ICMP ping as said in the doc. See if this can be used in your situation.
One of the numerous ref I could find about ping permissions: https://ubuntuforums.org/showthread.php?t=927709
I used pip to install Ansible in MacOS. But I cannot find the /etc/ansible folder. Neither the inventory file.
I want to run my playbook in minikube environment. But the playbook returns,
[WARNING]: No inventory was parsed, only implicit localhost is available
[WARNING]: provided hosts list is empty, only localhost is available. Note that the implicit localhost does not match 'all'
[WARNING]: Could not match supplied host pattern, ignoring: 192.168.99.105
How to solve this issue?
I looked into this matter and using Ansible for managing minikube is not an easy topic. Let me elaborate on that:
The main issue is cited below:
Most Ansible modules that execute under a POSIX environment require a Python interpreter on the target host. Unless configured otherwise, Ansible will attempt to discover a suitable Python interpreter on each target host the first time a Python module is executed for that host.
-- Ansible Docs
What that means is that most of the modules will be unusable. Even ping
Steps to reproduce:
Install Ansible
Install Virtualbox
Install minikube
Start minikube
SSH into minikube
Configure Ansible
Test
Install Ansible
As the original poster said it can be installed through pip.
For example:
$ pip3 install ansible
Install VirtualBox
Please download and install appropriate version for your system.
Install minikube
Please follow this site: Kubernetes.io
Start minikube
You can start minikube by invoking command:
$ minikube start --vm-driver=virtualbox
Parameter --vm-driver=virtualbox is important because it will be useful later for connecting to the minikube.
Please wait for minikube to successfully deploy on the Virtualbox.
SSH into minikube
It is necessary to know the IP address of minikube inside the Virtualbox.
One way of getting this IP is:
Open Virtualbox
Click on the minikube virtual machine for it to show
Enter root for account name. It should not ask for password
Execute command: $ ip a | less and find the address of network interface. It should be in format of 192.168.99.XX
From terminal that was used to start minikube please run below command:
$ minikube ssh
Command above will ssh to newly created minikube environment and it will store a private key in location:
HOME_DIRECTORY .minikube/machines/minikube/id_rsa
id_rsa will be needed to connect to the minikube
Try to login to minikube by invoking command:
ssh -i PATH_TO/id_rsa docker#IP_ADDRESS
If login has happened correctly there should be no issues with Ansible
Configure Ansible
For using ansible-playbook 2 files will be needed:
Hosts file with information about hosts
Playbook file with statements what you require from Ansible to do
Example hosts file:
[minikube_env]
minikube ansible_host=IP_ADDRESS ansible_ssh_private_key_file=./id_rsa
[minikube_env:vars]
ansible_user=docker
ansible_port=22
The ansible_ssh_private_key_file=./id_rsa will tell Ansible to use ssh key from file with correct key to this minikube instance.
Note that this declaration will need to have id_rsa file in the same location as rest of the files.
Example playbook:
- name: Playbook for checking connection between hosts
hosts: all
gather_facts: no
tasks:
- name: Task to check the connection
ping:
You can test the connection by invoking command:
$ ansible-playbook -i hosts_file ping.yaml
Above command should fail because there is no Python interpreter installed.
fatal: [minikube]: FAILED! => {"ansible_facts": {"discovered_interpreter_python": "/usr/bin/python"}, "changed": false, "module_stderr": "Shared connection to 192.168.99.101 closed.\r\n", "module_stdout": "/bin/sh: /usr/bin/python: No such file or directory\r\n", "msg": "The module failed to execute correctly, you probably need to set the interpreter.\nSee stdout/stderr for the exact error", "rc": 127}
There is a successful connection between Ansible and minikube but there is no Python interpreter to back it up.
There is a way to use Ansible without Python interpreter.
This Ansible documentation is explaining the use of raw module.