I'm curious where ZFS development is going. I've been reading through the lists, and watching Oracle, Nexenta, Illumos, and OpenIndiana for signs of life.
The feeling I get is that while there is plenty of userland work being done, there is next to nothing on ZFS development outside of the Oracle camp. So, I decided to see if I could help set something in motion. Agreeing with some of the opinions expressed here, (http://opensolaris.org/jive/thread.jspa?messageID=508798񼍾 ) I contacted Erik Trimble and we had a very quick/brief discussion that we want to bring to the list so a more public and wide-scope discussion can happen. I have my ideas of where I'd like to see ZFS go, Erik doesn't fully agree, and has other ideas as well. We both know that the rest of the community will have further ideas on what should be happening, and that's why we're discussing it here. However, I think it's imperative that we don't fracture ZFS into 4 different OpenSource versions that are all incompatible with each-other. I'd like to lay out some groundwork for this thread to keep it manageable; 1) This thread is about ZFS, not generic Solaris development, userland or non-ZFS development in _x_Solaris. In my mind, I see these issues as pressing and needing addressing; 1) Action. We need an official "we're moving on" type of agreement, so we stop waiting for Oracle to do something. I don't think we're going to see v31 anytime soon, let's stop waiting for it. What version do we branch from? 2) Home? ZFS requires a home. What is the home? Is it official? Will it always be able to live here? This home would also be in charge of version/change management. A new version system will need to be created to diff from the Oracle efforts going forward. 3) Generic? Can this home be generic (i.e. lessen or remove the bias to Solaris)? Is it practical? FreeBSD has a very robust ZFS copy, running v28 in the beta builds. Pjd has done a lot of work porting this, and I know other people will do a lot of work porting ZFS in the future. Can we minimize the efforts of porting ZFS so the work can go into development and features, not constantly adapting changes to the OS in use? More than just Solaris developers want to contribute to ZFS. 4) Legal? Is there anything in ZFS that needs to be removed to ensure it has a long vibrant life in the Open Source community? Do we need a "Lite" version much like 4.4 BSD Lite to escape AT&T? This may need to spawn into a separate thread, as I've already seen a few epic threads about that topic alone. I think that's plenty to get this started.
_______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
