It's generally accepted that two things can break a ZFS pool: the use of non-ECC RAM; and storage devices which do not respect cache flushes. You're aiming for a system which is likely to have both. The latter by virtue of the fact that most USB <=> SATA controllers often don't respect sync requests. Both of these can lead to broken metadata/uberblocks and hence a broken pool.
Not that ZFS will be any *less* reliable than any given file system, it's just more likely to be able to tell you about the problems. Cheers, On 9 February 2011 00:44, Jacob Ritorto <jacob.rito...@gmail.com> wrote: > Hi, > I was thinking of building a minimal low performance/experimental > zfs filer on a beagleboard ( http://beagleboard.org ) or something similar > with a 2tb mirror disk set attached via usb and a serial console. Are there > any projects or plans to compile/port OpenIndiana to arm? If not, I see > that Ubuntu already works on beagleboard, so that, plus the Livermore > zfs-on-linux port ( http://zfsonlinux.org ) would perhaps do the trick for > now, I guess.. > > thoughts? > > thx > jake > > _______________________________________________ > OpenIndiana-discuss mailing list > OpenIndiana-discuss@openindiana.org > http://openindiana.org/mailman/listinfo/openindiana-discuss > _______________________________________________ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
