On Fri, 17 Feb 2006, Vincent Meanie wrote:

> I have reviewed many faq's and searched the mailing list archives but found
> nothing.
> 
> I am currently in the process of a major upgrade from an existing system that
> has been in collocation for three years, openbsd 3.5 with Pentium 3 hardware
> and 400gb total storage. The system hardware replacing it will have a hardware
> sata raid controller, and the original plan was to present the Array as one
> 2.1tb drive to Openbsd 3.8. The initial headache was obtaining all the
> hardware.
> 
> Reviewing documentation revealed this gem:
> 
> > 14.7 - What are the issues regarding large drives with OpenBSD?
> 
> > OpenBSD supports an individual file system of up to 231-1, or 2,147,483,647
> > sectors, and as each sector is 512 bytes, that's a tiny > amount less than
> > 1T.
> > 
> > There is also a 1T limit on the size of the physical disk, although under
> > *some* circumstances, that may not cause you problems up to 2T, although
> > this
> > is not guaranteed.
> 
> Is this hard limit because of issues with the filesystem? Would it be
> possible, limitations with the controller aside, to present the array as three
> 700mb slices and combine them with CCD? Or would I be faced with the same
> limitation because of the large disk size.

Yes. The 1TB disk limit is a limitation caused by the fact that a 32
bit signed number is used to address disk sectors. 2^31 * 512 = 1TB.

> The array is running raid 5 which is the reason for the large disk size.

Just divide the array up into logical volumes, but do not combine them
using ccd(4), but by mounting.

Lets's say you create 3 logical volumes.  You create filesystems on
them, and mount them. Since your setup probably has some top level
dirs anyway, you can use those as mount points. It requires some
planning, but should not be that hard. 

Not that creating very large filesystems also has some drawbacks:
mostly very long fsck times and high memory consumption. Also, after a
crash, fsck'ing the logical volumes can take VERY long, since by
default the system thinks the logical volmes are separate disks and
fsck them in parallel, so all disk i/o will go to the same disks. Use
the -l parameter to fsck to avoid that. 

        -Otto

> 
> I am really stuck on this problem as Openbsd is my primary OS choice, but this
> is a deal breaker and I would like to explore all options before having to
> move to another OS.
> 
> Thank you for your time.

Reply via email to