This tutorial, provides detailed instructions on how to setup a Kubernetes and OpenEBS Cluster using the Vagrantfile provided in this folder. The following types of VMs can be launched:
- Kubernetes Master
- Kubernetes Minion
- OpenEBS Maya Master
- OpenEBS Storage Host
Depending on your need, you can also use this Vagrantfile, to only setup a Kubernetes Cluster or just OpenEBS Cluster or both. The number of nodes under each category is also configurable.
The Kubernetes cluster is setup, in this Vagrantfile using "kubeadm".
The following instructions have been verified on :
- Laptop installed with Ubuntu 16.04
- Laptop installed Windows 10 and Ubuntu 16.04 VM (created using VMWare Player 12)
Verify that you have the required software installed, with atleast minimum specified version.
1.Vagrant (>=1.9.1)
2.VirtualBox 5.1
3.Git
ubuntu-host:~/$ vagrant version
Installed Version: 1.9.1
Latest Version: 1.9.1
You're running an up-to-date version of Vagrant!
ubuntu-host:~/$ vboxmanage --version
5.1.14r112924
ubuntu-hos:~/$ git version
git version 2.7.4
ubuntu-host:~/$
Setup your local directory, where the demo code will be downloaded. Let us call this as $demo-folder
mkdir $demo-folder
cd $demo-folder
git clone https://github.com/openebs/openebs.git
cd openebs/k8s/demo
ls $demo-folder/openebs/k8s/demo/Vagrantfile
ubuntu-host:~/$ cd $demo-folder/openebs/k8s/demo/
ubuntu-host:~/demo-folder/openebs/k8s/demo$ vagrant status
Current machine states:
kubemaster-01 not created (virtualbox)
kubeminion-01 not created (virtualbox)
omm-01 not created (virtualbox)
osh-01 not created (virtualbox)
osh-02 not created (virtualbox)
This environment represents multiple VMs. The VMs are all listed
above with their current state. For more information about a specific
VM, run `vagrant status NAME`.
By default, 5 VMs will be created as shown above. To create these VMs, you will need around 10GB RAM and 6 CPU. Depending on your system configuration, you can edit the Vagrantfile to create lesser or more nodes. Please check the customization instructions here
If you are installing for the first time on the host, the vagrant will need to download the virtualbox images. If you are running on a low speed internet, the download can take considerable time. We recommend that you perform the following two steps, prior to issuing the "vagrant up" command below. Both the images put together will require around 1500MB.
cd $demo-folder/openebs/k8s/demo
vagrant box add openebs/k8s-1.5.5
vagrant box add openebs/openebs-0.2
ubuntu-host:~/demo-folder/openebs/k8s/demo$ vagrant box list | grep openebs
openebs/k8s-1.5.5 (virtualbox, 2017033101)
openebs/openebs-0.2 (virtualbox, 2017033102)
ubuntu-host:~/demo-folder/openebs/k8s/demo$
It is possible that, you may have a later versions downloaded. Thats ok!
cd $demo-folder/openebs/k8s/demo
vagrant up
This step, can take few minutes depending on your network speed. The required vagrant plugins are installed (for caching the vagrant boxes), the ubuntu-xenial box is download and the required software packages for each of the node are downloaded and installed.
ubuntu-host:~/demo-folder/openebs/k8s/demo$ vagrant status
Current machine states:
kubemaster-01 running (virtualbox)
kubeminion-01 running (virtualbox)
omm-01 running (virtualbox)
osh-01 running (virtualbox)
osh-02 running (virtualbox)
This environment represents multiple VMs. The VMs are all listed
above with their current state. For more information about a specific
VM, run `vagrant status NAME`.
ubuntu-host:~/demo-folder/openebs/k8s/demo$
Note: The network slowness can abort the process of installation, since a default timeout of 300 seconds is specified for the VMs to be launched. This can be modified by editing the Vagrantfile ( vmCfg.vm.boot_timeout ) value. If the installation aborts in between, make sure to clean-up before starting with the installation again.
cd $demo-folder/openebs/k8s/demo
vagrant destroy
Use the kubectl from within the Kubernetes master to get the current status.
ubuntu-host:~/demo-folder/openebs/k8s/demo$ vagrant ssh kubemaster-01
... snipped ...
ubuntu@kubemaster-01:~$ kubectl get nodes
NAME STATUS AGE
kubemaster-01 Ready,master 1h
kubeminion-01 Ready 57m
ubuntu@kubemaster-01:~$
Check the status of the system pods.
ubuntu@kubemaster-01:~$ kubectl get pods --all-namespaces
The above command will ouput the status of the pods created:
NAMESPACE NAME READY STATUS RESTARTS AGE
kube-system dummy-2088944543-1t7hc 1/1 Running 0 22m
kube-system etcd-kubemaster-01 1/1 Running 0 21m
kube-system kube-apiserver-kubemaster-01 1/1 Running 0 22m
kube-system kube-controller-manager-kubemaster-01 1/1 Running 0 21m
kube-system kube-discovery-1769846148-nzbd1 1/1 Running 0 22m
kube-system kube-dns-2924299975-jt7wz 4/4 Running 0 20m
kube-system kube-proxy-5wrcp 1/1 Running 0 15m
kube-system kube-proxy-p2n9v 1/1 Running 0 20m
kube-system kube-scheduler-kubemaster-01 1/1 Running 0 21m
kube-system weave-net-6t9pz 2/2 Running 0 15m
kube-system weave-net-frt55 2/2 Running 0 20m
For more usage options of kubectl refer to:
https://kubernetes.io/docs/user-guide/kubectl-cheatsheet/
Use the mayactl from within the OpenEBS Maya Master to get the current status.
ubuntu-host:~/demo-folder/openebs/k8s/demo$ vagrant ssh omm-01
... snipped ...
ubuntu@omm-01:~$
ubuntu@omm-01:~$ maya omm-status
Name Address Port Status Leader Protocol Build Datacenter Region
omm-01.global 172.28.128.4 4648 alive true 2 0.5.5 dc1 global
m-apiserver listening at http://172.28.128.4:5656
Use the below command to check the openebs storage host status
ubuntu@omm-01:~$
ubuntu@omm-01:~$ maya osh-status
ID DC Name Class Drain Status
e57020e9 dc1 osh-02 <none> false ready
b6789013 dc1 osh-01 <none> false ready
ubuntu@omm-01:~$
Check out the issues section for known issues and possible workarounds for some of the commom glitches while setting up kubernetes clusters on Ubunut/VirtualBox.
https://github.com/openebs/openebs/issues?utf8=%E2%9C%93&q=label%3Ademo%2Fk8s%20
Some of the issues observed:
- Unable to upgrade to VirtualBox 5.1
- weave pod, is stuck in CrashLoop