Szabolcs Szakacsits wrote: > uses FUSE purposely to avoid kernel linking problems (Licensing). > > > However this is not true. Linux has a limited NTFS kernel module too. > Which is utterly useless, worse than FreeBSD's and Apple's piggybacked version. Most vendors just ship NTFS-3g and FUSE now, stripping out the in-kernel module, which I suspect will disappear soon. > We moved to a split user/kernel (hybrid) space implementation, thanks > to FUSE, because we can have better quality, more features, higher > performance, rapid mass deployment, being highly cross-platform at > a small fraction of the old developer resources. > Still, the main benefit is to avoid the rabid GPL problems in the Linux kernel. Performance is pretty good, though I'm still sure in-kernel and not through FUSE is faster or has less overhead. > NTFS-3G works on Linux, FreeBSD, NetBSD, OS X, Haiku, BeOS, and Solaris. > Shouldn't say "works" if there's a critical bug which risks stability on a particular platform and you know about it. > NTFS-3G is supported on the same level as any other platform. The > OpenSolaris FUSE kernel module has some issues but at seems Sun is > already working on those Indeed that's really what I meant. I can't trust it yet though unfortunately because of so many past problems specific to Solaris. I've always got the feeling that most your eggs are in the Linux basket, and the time it took to get it going with Sun is just unacceptable.
As a user and someone who'd rather not delve into strange code to fix something your team should had been aware of, it's just not a good situation to recommend NTFS-3g on Solaris still. If the problem was with the Solaris kernel, which I do not doubt, then when people have posted threads saying NTFS-3g hosed my volume or it is slower than mud, something more than "we're working on it" would be acceptable, or at least point to a bug specific to the ON consolidation in the bug database. James
