Node Installation Guide

Navigation:  Installation >

Node Installation Guide

Previous pageReturn to chapter overviewNext page

Note that SELinux and SSH PermitRootLogin are expected to be enabled on the node.

 

Note that if you install KuberDock master with CEPH then node installation script will setup CEPH on this node automatically.

 

Log in to master server console as root and perform the following steps:

 

1. Copy nginx users’ SSH key to the node running the following command:

 

ssh-copy-id -i /var/lib/nginx/.ssh/id_rsa.pub [email protected]_node

 

where your_node is node IP address.

 

Note. Nginx user requires root access to the node. The key is generated automatically during master server installation.

 

When key installation succeeded you will see the following message: Number of key(s) added: 1

 

2. Log in to KuberDock as Admin.

 

Note. To restore Admin password, perform the following steps:

 

Log in to master server console and run the command:

 

python /var/opt/kuberdock/manage.py reset-password

 

Enter new password;

Retype new password.

 

New password will be saved.

 

5. Click Add node.

 

add_node

 

6. Specify “uname -n” of the node in Node hostname field. Note that “uname -n” should have records in direct and reverse DNS zones or, if you don't have an access to DNS,  in /etc/hosts of all servers in cluster:

 

enter_node_name

 

When deploying in non-AWS environment with ZFS as a Local Storage backend, an existing free block device (like /dev/sdc or /dev/sdd) should be specified. You can add multiple devices to the node:

 

block_device

 

The devices are needed to create zpool. In AWS, all that is required for zpool provisioning will be done automatically.

 

Choose Kube Type for this node from drop-down menu. You can find information on how to add new Kube Types in Configure Kube Types section.

 

Note that only the pods with the Kube Type chosen for the node, can be located on that node.

 

choose_kube_type

7. Click Add to start node installation.

 

Note that deploy script will execute clean up of a node and clear all data to perform clean installation. That is why the node that was previously added will be cleared.

 

add_cancel

 

8. Wait until the node is installed - its Status will change from pending for running.

 

Note that during installation process (which can take up to an hour) the node will be rebooted. During the reboot node status may be displayed incorrectly (if it takes more than one minute), it will be changed automatically right after reboot completed.

 

node_status_running

 

You can go to the Node page while its status is still pending to monitor installation process log.

 

If node installation failed, you can find the reason in the log as well.

 

If you can't identify the error, contact our support team at https://helpdesk.cloudlinux.com. Choose department “KuberDock” for your ticket and attach installation log as part of the ticket.

 

Note that memory swap is disabled on the nodes by default, which is required to ensure strict limitation of RAM allocated to users' containers. It is strongly recommended not to enable memory swap on the nodes.