ZFS was the star of the Silicon Valley OpenSolaris User Group meeting last week. A great presentation was made that provided the 'talk' around The Last Word in File Systems technical article posted on the ZFS community site.
ZFS addresses the problems of data corruption, management, and read/write speed. The technology uses transactional all-or-nothing commits that are numbered, object-based, and grouped for efficiency and accuracy.
The most numerous questions were around the following topics:
- RAID-Z: how does it work? how is it different from the metadata concepts?
- Pools: how do pools relate to blocks? How do they replace volumes?
- Performance: how do I move disks, split writes, rotate and replace?
- Administration: why do I still have to script it?
The discussion was lively and energetic. It went on for over three hours and chitchat continued well past midnight. For my part, I met the engineers and described to them the request for downloadable HTML of the ZFS Administration Guide that came across the docs-discuss-AT-opensolaris.org alias the previous morning. Watch the tagged PDF thread for the latest status on that project.
