Vertica cluster migration using copycluster – Part 2

Posted on Posted in Data & Business Intelligence, DevOps & Networking

In the previous post I covered the preparations for the copycluster process.

This post will cover the actual copycluster process, the goal is to copy the Vertica database to another cluster.

First we have to stop the database on the target cluster (using the adminTools is the easiest way)

Then we have to prepare the configuration file on one of the source cluster nodes.

What needs to be set in the Database section:

  • dbName
  • dbUser
  • dbPassword

The mapping section is for matching the source cluster node names to the target cluster hosts, you can check the source node names using the following command:

The target hosts in the mapping section will be the hostnames or IP addresses of the target cluster, in our case its: vertica-01 – vertica-14

This is the final copycluster.ini configuration file:

The last step is to run the VBR command, as this is a long running process I prefer to run in the background and track the progress later (nohup will send the output to nohup.out in the current directory)

What actually happens in the background is a backup and restore process in this single command and you will start seeing rsync processes between the source and target nodes.

It took about 5 days to complete the copycluster process from the 14 physical cluster to to our small virtual target cluster, this was the output in case for the successful process:


Leave a Reply

Your email address will not be published. Required fields are marked *