begin quoting Tracy R Reed as of Sat, Dec 15, 2007 at 07:16:29PM -0800: > SJS wrote: > >The promise of hardware RAID (for me) was transparency -- but this > >was never delivered, so far as I know. You had to have a RAID-aware > >OS to use hardware RAID, instead of having a device that could > >transparently give you RAID benefits on "legacy" and small systems. > > If by RAID-aware you mean having a RAID driver, this is true. Otherwise > the OS just sees a block device and had no idea that there is a RAID > behind it.
That's exactly what I mean. Requiring a driver renders the "otherwise" clause meaningless at best, disingenuous at worst. You can't yank an IDE/SCSI/FC drive out and drop a hardware RAID in. You just get hardware acceleration for what is basically a software RAID... > >Hm. I thought that it was that the Linux developers resisted the > >telescoping of their beloved layers... > > At the moment I think it is both. The layer issue alone would never stop > a few enterprising coders from integrating ZFS. But as it is GPLv3 and Um, no it isn't. > the rest of the kernel is GPLv2 there is a problem. In fact, ZFS has > already been implemented on Linux in user-space: > > http://zfs-on-fuse.blogspot.com/ Yes, we've discussed that here. But that makes ZFS a second-class filesystem in the linux world. > So there is definitely demand for it. It is unclear how this issue will > be resolved but I'm betting it will be someday. Well, yes, there's the degenerate solution for when (if!?) the copyrights expire on the linux kernel and the GPL no longer applies. -- Only another sixty years or so to go, eh? Stewart Stremler -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
