hi,
your output doesn't show it, but format(1m)'s VERIFY tries to read
the Primary Label (PL) and Backup Labels (BL)... and VERIFY prefixes
the output you show w/ a string saying where the info was read from.
VTOC-style disk labels on x86/solaris show a size that is 1 sector less than
if the label had been written on SPARC/Solaris.
were all the disks formatted under the same {architecture}/solaris config?
lastly, from the zfs-basics i know, zfs creates both a "header" and
"trailer" on a Volume (using round-up/down techniques) so that issues
like what you're seeing become masked. eg, a disk's size can be "off by 1",
yet zfs loses this "rounding err" due to its headers/trailers.
/andrew
On 11/05/09 15:06, Ron Mexico wrote:
> I have 24 identical Western Digital drives connected to a Dell SAS 5/E HBA.
>
> Three of the drives list the following disk information when using the verify
> utility under the format command:
>
> Volume name = < >
> ascii name = <ATA-WDC WD2002FYPS-0-5G04-1.82TB>
> bytes/sector = 512
> sectors = 3907029166
> accessible sectors = 3907029133
> Part Tag Flag First Sector Size Last Sector
> 0 usr wm 34 1.82TB 3907012749
> 1 unassigned wm 0 0 0
> 2 unassigned wm 0 0 0
> 3 unassigned wm 0 0 0
> 4 unassigned wm 0 0 0
> 5 unassigned wm 0 0 0
> 6 unassigned wm 0 0 0
> 8 reserved wm 3907012750 8.00MB 3907029133
>
>
> The remaining 21 drives show this disk info:
>
> Volume name = < >
> ascii name = <ATA-WDC WD2002FYPS-0-5G04-1.82TB>
> bytes/sector = 512
> sectors = 3907029166
> accessible sectors = 3907029134
> Part Tag Flag First Sector Size Last Sector
> 0 usr wm 256 1.82TB 3907012750
> 1 unassigned wm 0 0 0
> 2 unassigned wm 0 0 0
> 3 unassigned wm 0 0 0
> 4 unassigned wm 0 0 0
> 5 unassigned wm 0 0 0
> 6 unassigned wm 0 0 0
> 8 reserved wm 3907012751 8.00MB 3907029134
>
>
> I know this isn't going to be an issue when creating a raidz pool, but if I
> have to replace a failed disk with one that has one less accessible sector,
> won't that cause problems?
>
> According the ZFS best practices guide: "The size of the replacements vdev,
> measured by usable sectors, must be the same or greater than the vdev being
> replaced. This can be confusing when whole disks are used because different
> models of disks may provide a different number of usable sectors."
>
> Can anyone shed some light on this?
--
Andrew Rutz andrew.rutz at sun.com
Solaris RPE Ph: (x64089) 512-401-1089
Austin, TX 78727 Fax: 512-401-1452
