Date: Tue, 09 Dec 2014 13:41:15 +0100 From: Mark Martinec <Mark.Martinec+freebsd@ijs.si> To: freebsd-fs@freebsd.org Subject: Re: ZFS scrub on new disks gives cksum errors Message-ID: <5216a88b2411cb197af1170e5a81bab0@mailbox.ijs.si> In-Reply-To: <20141209021651.4304976.79605.2556@denninger.net> References: <20141208155230.GA574@iozz.us> <5485CB35.5010608@multiplay.co.uk> <CACfj5vLXnAZgkbGnw-B0GTAsj7cHh097zzig54rcAUt=7PFHoA@mail.gmail.com> <20141209021042.GA649@iozz.us> <20141209021651.4304976.79605.2556@denninger.net>
next in thread | previous in thread | raw e-mail | index | archive | help
The key here is 'true on-line' (double-conversion). The waveform itself does not matter much, can be sine or pseudo-sine. Line-interactive and off-line UPS devices with line filters are not good candidates when quality of grid is poor. They handle clean true power outages just fine, but dealing with brownouts and electrical noise calls for a better UPS. Mark 2014-12-09 03:16, Karl Denninger wrote > Stepped-wave ups power is perfectly fine to feed a switching power > supply -- in fact most run COOLER on it than on sine wave power! >=20 > --=C2=A0Karl > From: Brian N > Sent: Monday, December 8, 2014 20:11 > To: Sean Chittenden > Cc: freebsd-fs@freebsd.org > Subject: Re: ZFS scrub on new disks gives cksum errors >=20 > On Mon, Dec 08, 2014 at 09:09:51AM -0800, Sean Chittenden wrote: >> I had this exact problem at home on my FreeNAS host when a box is=20 >> plugged >> in to residential power that is "not clean." If you plug this host in=20 >> to a >> UPS it should clear things up instantly (find a sine wave UPS, not a >> digital stepping supply). There's something about the fluctuating=20 >> power or >> brownouts that ZFS is well suited to detecting (and became terrifying=20 >> to me >> because wtf happened in the past pre-ZFS??). >>=20 >> -sc >=20 > Interesting. I have a UPS (APC Back-UPS ES 550 (stepped)) on the same=20 > circuit > (residential power) that routinely gives me power outage messages where= =20 > an > outage lasts 1-2 seconds. Other electrical devices (lights, etc) are=20 > not > affected and I just thought that the UPS was very sensitive. Thanks for= =20 > the > reply.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5216a88b2411cb197af1170e5a81bab0>