and we don't log data, so our performance is reasonable. My team will try to generate some benchmarks while I am at Linux Kongress comparing the two approaches. We are on a recent 2.2 kernel also. Best, Hans "Stephen C. Tweedie" wrote: > Hi folks, > > OK, here it is --- have fun. > > ftp://ftp.uk.linux.org/pub/linux/sct/fs/jfs/release-0.0.1.tar.gz > > Read the README in the tarball FIRST!!! > > This is an initial, alpha release. I don't make any guarantees, but it > works for me. I can build a kernel on it, both on a 16MB 486 and on a > huge quad xeon with "make -j 16". I can pull the plug during the build, > or during a make clean, and everything comes back cleanly. Beyond that > I make no promises. :) > > See the README in the release tarball for known limitations and work > still to be done. In particular, this is NOT a high-performance release > since all data is journaled (and hence copied twice) as well as metadata > for now (this is entirely for testing purposes --- I need to get this > mode as debugged as possible before journaling metadata updates only). > > I'm off to the Darmstadt storage workshop and Linux Kongress tomorrow, > so I won't be particularly responsive to your emails for a week, but > even if you don't get a reply from me I _will_ be watching, honest. > > Sorry that this is against 2.2.2 for now: completing the 2.2.12 update > is at the top of my list to do when I get back, but it just didn't quite > make it to this release in time. > > OK, have fun breaking this code. :) > > --Stephen -- Get Linux (http://www.kernel.org) plus ReiserFS (http://devlinux.org/namesys). If you sell an OS or internet appliance, buy a port of ReiserFS! If you need customizations and industrial grade support, we sell them.
