A little bit of background. I work at a company called MayaData (www.mayadata.io) who develops a very cool Open Source software called OpenEBS (CNCF Sandbox project) that simplifies the deployment of stateful applications on Kubernetes. You should check it out at www.openebs.io.
Kubernetes can be installed on any type of machine; be it a Virtual Machine, bare metal, or cloud machine. Kubernetes abstracts away most of the significant bits of a system, except storage. When it comes to storage, the main reason an abstraction will not work is that there is no uniqueness among the storage devices themselves. Every vendor and every virtualization platform implements it differently.
We hit this issue of unique virtual disks while deploying OpenEBS on Kubernetes backed by VMware VMS or other virtualization platforms because OpenEBS NDM is not able to uniquely identify the block devices themselves.
Changing the absolute configuration on the Virtual Machine can help you get around this issue.
Here are the steps to enable unique disk IDs in VMware via vSphere client:
This will assign WWN to each disk in the Virtual Machine.
That is it for today's tutorial. If you have any questions, feedback, or any topic that you feel I should cover next, feel free to comment on our blog or reach out to us on our Slack channel.