On Sun, Dec 27, 2009 at 2:15 AM, Michael Disserman <[email protected]> wrote:
> thanks for the info. I can't resize a slices until I resize a partition. But 
> I don't know how to resize it. If I just delete and recreate it with fdisk, 
> it changes a number of blocks per cylinder and the old layout becomes 
> completely inaccessible.

Yeah, I just tried it on Linux dom0, resizing the LVM from 1G to 10G,
and got what you describe :D
There's a workardound though. Use parted with sectors as unit. Here's
what I did (the pool name is test, on opensolaris snv_125)

#========================================================
#parted /dev/dsk/c7t5d0p0
(parted) unit s
unit s
(parted) print
print
Model: Generic Ide (ide)
Disk /dev/dsk/c7t5d0p0: 20971520s
Sector size (logical/physical): 512B/512B
Partition Table: msdos

Number  Start  End       Size      Type     File system  Flags
 1      2048s  2097151s  2095104s  primary               boot

(parted) rm 1
rm 1
(parted) mkpart primary solaris 2048s 100%
mkpart primary solaris 2048s 100%
(parted) toggle 1 boot
toggle 1 boot
(parted) print
print
Model: Generic Ide (ide)
Disk /dev/rdsk/c7t5d0p0: 20971520s
Sector size (logical/physical): 512B/512B
Partition Table: msdos

Number  Start  End        Size       Type     File system  Flags
 1      2048s  20964824s  20962777s  primary               boot

(parted) quit
quit
#========================================================


then I adjust the slices to be like the original one with "format", plus
#zpool export test
#zpool import test
#zpool set autoexpand=on test

By specifing sectors as unit to parted, modifying partition table and
slices to start on the same place as the original, and setting
autoexpand property to on (as Erik mentioned), the pool was extended
correctly. I haven't tested this on rpool yet (there's no "safe"
snapshot - extend feature on Linux LVM, and I like to keep my rpool
size just the way it is), but it should work.

-- 
Fajar
_______________________________________________
xen-discuss mailing list
[email protected]

Reply via email to