Sebastian Smith wrote: > The plugin layer in ReiserFS v4 [ snippage ] > Overall I think it's a cool idea... but I definately see why kernel > developers are complaining.
That sounds like a sweet system. But yeah, convincing a team as large as the kernel crew to take a big leap sounds like work. We'll see what comes of it. > Scott I'm not an expert on EXT3 (or anything really), but I think you > can tune journaling so that there is less risk of data loss. I'm not > sure if this would help in your case though. IIRC, there are 3 levels of journaling, ranging from less-journaling-but-fast to more-journaling-but slow. And IME, even with less journaling enabled, ext3 is a pretty slow candidate. I'm will to wager money the only reason it's caught on as much as it has it that it's the default setting on several widely used desktop distros, and it's backwards- and forwards-compatible with ext2. (Maybe that gets a "duh" from some people, but it's worth explaining how a sluggish journaling fs became so widely used.) I personally prefer using reiserfs on linux systems, but as my *BSD installbase has grown, I find I'm leaning more toward ext[23], as there is no reiserfs in OpenBSD nor FreeBSD.[*] I guess I'd really like to see one physical unix filesystems that can be used in a read/write capacity on a number of different unices and support unix-type filemodes. FAT* filesystems just aren't sufficient for a core unix filesystem; ufs is fine for some BSDs and Solaris, but I've had some issues with it in Linux. ext2 is almost guaranteed to work on any linux, ext2 support in the BSDs has been (IME) unreliable and unstable. And the developers don't seem apt to try too hard to stabilize a linux fs. :-) [*] Upcoming versions of FreeBSD are rumoured to support READ-ONLY reiserfs support. This is probably reiser3, of course. </stream> Cheers! Tim Hammerquist _______________________________________________ RLUG mailing list [email protected] http://lists.rlug.org/mailman/listinfo/rlug
