OpenZFS - a perfect filesystem for containers


OpenZFS is a time proved filesystem, that has seen deployments in countless places, as storage array, single server local storage and even on desktop, as root filesystem. It has made filesystem administration simple and easy, compared to traditional, separate filesystem, volume manager and RAID. One, however, does not simple slap together disks and put them into production! To properly care for storage and data, the operator needs to understand core concepts behind OpenZFS and common scenarios involved.

To this end, typical pool layouts are going to be explained, along with features like snapshots, clones and checksumming. A problem of planning the pool and its growth is going to be explained. The presentation will also touch on what and how to monitor, giving examples of failures and how to fix them.

It is explained, why Ubuntu has made OpenZFS a first class filesystem in its 16.04 LTS release and why it has been made a default filesystem for LXD containers. LXD with an OpenZFS pool as a backend is going to be demonstrated, with explanation how OpenZFS makes it more efficient.

As a bonus a OpenZFS on / goodness will be showed off.

All the parts above will be acompanied by hands on presentations.

Ballroom B
Saturday, March 4, 2017 - 18:00 to 19:00