I'm immensely pleased with the approach they've chosen, because it offers the possibility of improving your stripe size, without requiring commitment to. New bl

I'm immensely pleased with the approach they've chosen, because it offers the possibility of improving your stripe size, without requiring commitment to. New blocks are written with the new stripe size, so all you have to do to improve your stripe width is rewrite all of your files (and deal with the snapshot churn that will cause). It also offers the tempting possibility of the reverse, where you can create new files with stripes and then rewrite all your data to reduce capacity. I used to work at Isilon, and the design of this feature looks very familiar to me: That's a good thing. I'm excited for the possibilities for ZFS going forward. I hope they'll come up with a restriper tool that will allow you to rewrite all of your files with the new block layout without breaking snapshots, etc, but either way, this is a huge advancement for my own personal use of ZFS.