Migrating to FreeBSD from Solaris 11: Part 1 - About

Where am I coming from?

I am currently in the process of switching from a Solaris 11 storage system to one running FreeBSD. I have used Solaris 11 as my primary storage appliance since 2010. In 2010 FreeNAS was still UFS based and ZFS support was considered experimental. Also IIRC, FreeBSD’s support of ZFS had just been released. My lack of experience/knowledge, aversion to risk of data, and the lure of native encryption influenced my decision to use Solaris.

Why am I switching?

In my not so humble opinion, I believe FreeBSD is the best platform for nearly all computing needs. The factual reasons include the stability, the community, the BSD license, the built-in type-2 hypervisor called bhyve, the solid ZFS implementation, the recent packages/ports, the whole system approach, and the methodical-forward-thinking development process make FreeBSD one, if not the best platform for computing.

I mentioned encryption.

My primary use for encryption is that it makes warranty returns on hard drives “No big deal” as Leslie Chow put it. You see, I buy cheap drives intended for light desktop use but then abuse them with 24/7 SAN & NAS use. I know I am a terrible person for the flagrant abuse of hard drives. That said, GELI will be used in place of the native encryption provided in the Solaris-fork.

What services am I replacing?

Currently files are shared over SMB and ZVols are served over Fiber Channel. However to make this project even more complex and challenging, this machine’s duties will expand to provide more services. The primary reason for adding services to this machine is to consolidate the always-on hardware to one server and one DAS.


What specific ‘core’ services will be hosted as virtual machines?

The above services and VMs seems like a lot to ask a single machine to do, however the host has 192GB of RAM and the data is stored on eight pairs of striped mirrors.

Anticipated pain points and what I plan to document in this series:

This series is being written to document the process to transition a large ZFS pool from one system to another, this includes the services which depend on that data, in the hopes that the series will inspire, inform, or help in some way.