On Sun, Jun 29, 2008 at 10:05 AM, Steven Hartland <killing_at_multiplay.co.uk> wrote: > > ----- Original Message ----- From: "Søren Schmidt" <sos_at_FreeBSD.ORG> >>> >>> Although it is more visible, personally I would prefer it to just fail >>> instead of proceeding. RAID5 is not RAID5 without parity so why even >>> allow it to continue and hence risk such an unrecoverable situation? >> >> Well, this has been rehashed many times before, it has been disabled, put >> a warning in the boot log, warning in the docs, all 3 was the favorite at >> the time it was done. >> >> I'm all ears for what the decision might be this time, just get consensus >> and I'll flip the right switch. > > As a small expansion to this, a good compromise might be to disable it > unless a sysctl flag is explicitly set. Similar to how the sysctl > kern.geom.debugflags=16 works? Then for those that know the implications > they can still have the functionality but prevents someone stumbling > into a nasty situation without being aware of the issues. > > Regards > Steve That wouldn't have helped me in this situation because I didn't set anything up via geom, because I was blissfully unaware and didn't ask how anyone to setup a RAID config properly before now. That idea probably wouldn't be a bad one though... -GarrettReceived on Sun Jun 29 2008 - 17:32:23 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:32 UTC