Date: Wed, 2 Jun 2004 20:51:06 +1000 (EST) From: Bruce Evans <bde@zeta.org.au> To: Brooks Davis <brooks@one-eyed-alien.net> Cc: Tuc at Beach House <tuc@ttsg.com> Subject: Re: 1TB issue Message-ID: <20040602203400.N24291@gamplex.bde.org> In-Reply-To: <20040602010339.GA29173@Odin.AC.HMC.Edu> References: <20040602004318.GA73188@panzer.kdm.org> <200406020052.i520qhXc038253@himinbjorg.tucs-beachin-obx-house.com> <20040602010339.GA29173@Odin.AC.HMC.Edu>
next in thread | previous in thread | raw e-mail | index | archive | help
> I'm not sure what the limit of a BSD label is, but it my be 1TB unless > you can up the block size (512K blocks plus signed ints yeilds 1TB > limits). You may want to use fdisk to make a slice for any large > partitions rather then trying to use disklabel. On the 2TB system I > have, I have one slice that is around 40GB that contains a disklabel and > the partions making up a normal OS install. A second slice that has a > 1.8TB UFS file system and no disklabel. The limit for a bsd label is 2^32-1 sectors (unless geom has broken it). With the normal sector size of 512 bytes, this gives a limit of 2TB less 512 bytes. With a larger sector size, the limit is larger. A sector size of 2048 is quite reasonable, since this is the default fragment size for ffs. This gives a limit of 8TB less 2048 bytes. The main 1TB limits are that RELENG_4 doesn't support disk addresses larger than 1TB, and ffs1 is broken for disk addresses larger than 1TB (disk addresses 31 bytes in units of fragments internally in ffs, so the limit with the default fragment size of 2048 is 4TB less 2048 bytes, but things break at 1TB due to bogus parentheses in the conversion from ffs addresses to disk addresses). Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040602203400.N24291>