geom_stripe error
Alexandre Biancalana
biancalana at gmail.com
Tue Jul 18 16:52:38 UTC 2006
Here is the script used to create the stripe:
#!/bin/ksh
stripe_name=data
gconcat label -v concat1 /dev/ad1s1g /dev/ad3s1g
gconcat label -v concat0 /dev/ad0s2g /dev/ad1s2a /dev/ad2s1g /dev/da3s1d
disks="/dev/ad0s2a /dev/ad0s2b /dev/ad0s2d /dev/ad0s2e /dev/ad0s2f"
disks="$disks /dev/ad1s1a /dev/ad1s1b /dev/ad1s1d /dev/ad1s1e /dev/ad1s1f"
disks="$disks /dev/ad2s1a /dev/ad2s1b /dev/ad2s1d /dev/ad2s1e /dev/ad2s1f"
disks="$disks /dev/ad3s1a /dev/ad3s1b /dev/ad3s1d /dev/ad3s1e /dev/ad3s1f"
disks="$disks /dev/da0s1a /dev/da0s1b"
disks="$disks /dev/da1s1a /dev/da1s1b"
disks="$disks /dev/da2s1a /dev/da2s1b"
disks="$disks /dev/da3s1a /dev/da3s1b"
disks="$disks /dev/da4s1a /dev/da4s1b"
disks="$disks /dev/concat/concat0 /dev/concat/concat1"
echo "Clearing disks...."
gstripe destroy $stripe_name
gstripe clear $disks
echo "Creating stripe...."
gstripe label -v -s 131072 $stripe_name $disks
On 7/18/06, Eric Anderson <anderson at centtech.com> wrote:
>
> On 07/18/06 09:34, R. B. Riddick wrote:
> > --- Eric Anderson <anderson at centtech.com> wrote:
> >>> 157622337536/131072 appr. 1202563
> >>> 1202563 is congruent with 3 (modulo 32).
> >>> 158778490880/131072 appr. 1211383
> >>> 1211383 % 32 = 23
> >> So the offset is the number of records (16K based on stripe size in
> this
> >> case) offset, and not bytes? I was thinking bytes, which would make it
> >> drive number 5:
> >>
> > Hmm...
> >
> > I think the offsets in the log messages are in _bytes_. The stripe size
> is 128k
> > _bytes_. So the 157622337536-th byte is in the 1202563-rd stripe from
> the point
> > of view of the geom_stripe and the [1202563/32]-th stripe on the
> underlying
> > disk and the disk# should be [1202563%32].
> >
> > I dont see why/where ur 16k come into the computation...
>
> Oh - simple PEBKAC bug. :)
>
> Eric
>
>
>
> --
> ------------------------------------------------------------------------
> Eric Anderson Sr. Systems Administrator Centaur Technology
> Anything that works is better than anything that doesn't.
> ------------------------------------------------------------------------
> _______________________________________________
> freebsd-geom at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-geom
> To unsubscribe, send any mail to "freebsd-geom-unsubscribe at freebsd.org"
>
More information about the freebsd-geom
mailing list