Date: Tue, 16 Nov 2010 18:32:42 +0100 From: Ivan Voras <ivoras@freebsd.org> To: freebsd-questions@freebsd.org Subject: Re: ZFS and 4k sector drives Message-ID: <ibuf7q$kuu$1@dough.gmane.org> In-Reply-To: <alpine.BSF.2.00.1011152022140.78383@tensor.gdynia.pl> References: <4CE145C5.8040602@midsummerdream.org> <alpine.BSF.2.00.1011152022140.78383@tensor.gdynia.pl>
next in thread | previous in thread | raw e-mail | index | archive | help
On 11/15/10 20:24, Wojciech Puchar wrote: >> work with ZFS raidz2 pools. It seems that most of the 4k sector drives >> are using emulation, and reporting 512 byte sectors to the OS instead >> of their native 4k size. I know someone who had an issue trying to >> insert one of these drives into a running ZFS pool with other 512 byte >> sector drives with bad results. > > ZFS use 4k blocks. Actually, no, it's much worse: ZFS uses variable block size upto 128 KiB, meaning not only does it not write or align data on a sensible boundary like 4 KiB, both its metadata and data IO can write on any offset and length which is a multiple of 512 bytes. References: http://www.solarismen.de/archives/5-Solaris-and-the-new-4K-Sector-Disks-e.g.-WDxxEARS-Part-2.html http://permalink.gmane.org/gmane.os.freebsd.devel.geom/4318
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?ibuf7q$kuu$1>