Hi,
I was following the instructions in ZFS Troubleshooting Guide on how to
replace a disk in the root pool on x86 system. I''m using OpenIndiana,
ZFS
pool v.28 with mirrored system rpool. The replacement disk is brand new.
root:~# zpool status
pool: rpool
state: DEGRADED
status: One or more devices has experienced an unrecoverable error. An
attempt was made to correct the error. Applications are unaffected.
action: Determine if the device needs to be replaced, and clear the errors
using ''zpool clear'' or replace the device with
''zpool replace''.
see: http://www.sun.com/msg/ZFS-8000-9P
scan: resilvered 17.6M in 0h0m with 0 errors on Wed Apr 11 17:45:16 2012
config:
NAME STATE READ WRITE CKSUM
rpool DEGRADED 0 0 0
mirror-0 DEGRADED 0 0 0
c2t5000CCA369C55DB8d0s0 OFFLINE 0 126 0
c2t5000CCA369D5231Cd0s0 ONLINE 0 0 0
errors: No known data errors
root:~#
I''m not very familiar with Solaris partitions and slices so somewhere
in
the format/partition commands I must to have made a mistake because when I
try to replace the disk I''m getting the following error:
root:~# zpool replace rpool c2t5000CCA369C55DB8d0s0 c2t5000CCA369C89636d0s0
invalid vdev specification
use ''-f'' to override the following errors:
/dev/dsk/c2t5000CCA369C89636d0s0 overlaps with
/dev/dsk/c2t5000CCA369C89636d0s2
root:~#
I used -f and it worked but I was wondering is there a way to completely
"reset" the new disk? Remove all partitions and start from scratch.
Thank you
Peter
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20120412/10a679aa/attachment.html>
Hi Peter, The root pool disk labeling/partitioning is not so easy. I don''t know which OpenIndiana release this is but in a previous Solaris release we had a bug that caused the error message below and the workaround is exactly what you did, use the -f option. We don''t yet have an easy way to clear a disk label, but sometimes I just create a new test pool on the problematic disk, destroy the pool, and start over with a more coherent label. This doesn''t work for scenarios. Some people use the dd command to wipe an existing label, but you must use it carefully. Thanks, Cindy On 04/12/12 11:35, Peter Wood wrote:> Hi, > > I was following the instructions in ZFS Troubleshooting Guide on how to > replace a disk in the root pool on x86 system. I''m using OpenIndiana, > ZFS pool v.28 with mirrored system rpool. The replacement disk is brand new. > > root:~# zpool status > pool: rpool > state: DEGRADED > status: One or more devices has experienced an unrecoverable error. An > attempt was made to correct the error. Applications are > unaffected. > action: Determine if the device needs to be replaced, and clear the errors > using ''zpool clear'' or replace the device with ''zpool replace''. > see: http://www.sun.com/msg/ZFS-8000-9P > scan: resilvered 17.6M in 0h0m with 0 errors on Wed Apr 11 17:45:16 2012 > config: > > NAME STATE READ WRITE CKSUM > rpool DEGRADED 0 0 0 > mirror-0 DEGRADED 0 0 0 > c2t5000CCA369C55DB8d0s0 OFFLINE 0 126 0 > c2t5000CCA369D5231Cd0s0 ONLINE 0 0 0 > > errors: No known data errors > root:~# > > I''m not very familiar with Solaris partitions and slices so somewhere in > the format/partition commands I must to have made a mistake because when > I try to replace the disk I''m getting the following error: > > root:~# zpool replace rpool c2t5000CCA369C55DB8d0s0 c2t5000CCA369C89636d0s0 > invalid vdev specification > use ''-f'' to override the following errors: > /dev/dsk/c2t5000CCA369C89636d0s0 overlaps with > /dev/dsk/c2t5000CCA369C89636d0s2 > root:~# > > I used -f and it worked but I was wondering is there a way to completely > "reset" the new disk? Remove all partitions and start from scratch. > > Thank you > Peter > > > _______________________________________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Actually, since I answer the root pool partitioning questions at least 52 times per year, I provided the steps for clearing the existing partitions in step 9, here: http://docs.oracle.com/cd/E23823_01/html/817-5093/disksxadd-2.html#disksxadd-40 How to Create a Disk Slice for a ZFS Root File System Step 9 uses the format-->disk-->partition-->modify option and sets the free hog space to slice 0. Then, you press return for each existing slice to zero them out. This creates one large slice 0. cs On 04/12/12 11:48, Cindy Swearingen wrote:> Hi Peter, > > The root pool disk labeling/partitioning is not so easy. > > I don''t know which OpenIndiana release this is but in a previous > Solaris release we had a bug that caused the error message below > and the workaround is exactly what you did, use the -f option. > > We don''t yet have an easy way to clear a disk label, but > sometimes I just create a new test pool on the problematic disk, > destroy the pool, and start over with a more coherent label. > This doesn''t work for scenarios. Some people use the dd command > to wipe an existing label, but you must use it carefully. > > Thanks, > > Cindy > > On 04/12/12 11:35, Peter Wood wrote: >> Hi, >> >> I was following the instructions in ZFS Troubleshooting Guide on how to >> replace a disk in the root pool on x86 system. I''m using OpenIndiana, >> ZFS pool v.28 with mirrored system rpool. The replacement disk is >> brand new. >> >> root:~# zpool status >> pool: rpool >> state: DEGRADED >> status: One or more devices has experienced an unrecoverable error. An >> attempt was made to correct the error. Applications are >> unaffected. >> action: Determine if the device needs to be replaced, and clear the >> errors >> using ''zpool clear'' or replace the device with ''zpool replace''. >> see: http://www.sun.com/msg/ZFS-8000-9P >> scan: resilvered 17.6M in 0h0m with 0 errors on Wed Apr 11 17:45:16 2012 >> config: >> >> NAME STATE READ WRITE CKSUM >> rpool DEGRADED 0 0 0 >> mirror-0 DEGRADED 0 0 0 >> c2t5000CCA369C55DB8d0s0 OFFLINE 0 126 0 >> c2t5000CCA369D5231Cd0s0 ONLINE 0 0 0 >> >> errors: No known data errors >> root:~# >> >> I''m not very familiar with Solaris partitions and slices so somewhere in >> the format/partition commands I must to have made a mistake because when >> I try to replace the disk I''m getting the following error: >> >> root:~# zpool replace rpool c2t5000CCA369C55DB8d0s0 >> c2t5000CCA369C89636d0s0 >> invalid vdev specification >> use ''-f'' to override the following errors: >> /dev/dsk/c2t5000CCA369C89636d0s0 overlaps with >> /dev/dsk/c2t5000CCA369C89636d0s2 >> root:~# >> >> I used -f and it worked but I was wondering is there a way to completely >> "reset" the new disk? Remove all partitions and start from scratch. >> >> Thank you >> Peter >> >> >> _______________________________________________ >> zfs-discuss mailing list >> zfs-discuss at opensolaris.org >> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss > _______________________________________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
"Hung-Sheng Tsao (Lao Tsao 老曹) Ph.D."
2012-Apr-12 18:59 UTC
[zfs-discuss] Replacing root pool disk
hi by default the disk partition s2 cover the whole disk this is fine for ufs for LONG time. Now zfs does not like this overlap so you just need to run format then delete s2 or use s2 and delete all other partitions (by default when you run format/fdisk it create s2 whole disk and s7 for boot, so there is also overlap between s2 and s7:-( ) ZFS root need to fix this problem that require partition and not the whole disk (without s?) my 2c regards On 4/12/2012 1:35 PM, Peter Wood wrote:> Hi, > > I was following the instructions in ZFS Troubleshooting Guide on how > to replace a disk in the root pool on x86 system. I''m using > OpenIndiana, ZFS pool v.28 with mirrored system rpool. The replacement > disk is brand new. > > root:~# zpool status > pool: rpool > state: DEGRADED > status: One or more devices has experienced an unrecoverable error. An > attempt was made to correct the error. Applications are > unaffected. > action: Determine if the device needs to be replaced, and clear the errors > using ''zpool clear'' or replace the device with ''zpool replace''. > see: http://www.sun.com/msg/ZFS-8000-9P > scan: resilvered 17.6M in 0h0m with 0 errors on Wed Apr 11 17:45:16 2012 > config: > > NAME STATE READ WRITE CKSUM > rpool DEGRADED 0 0 0 > mirror-0 DEGRADED 0 0 0 > c2t5000CCA369C55DB8d0s0 OFFLINE 0 126 0 > c2t5000CCA369D5231Cd0s0 ONLINE 0 0 0 > > errors: No known data errors > root:~# > > I''m not very familiar with Solaris partitions and slices so somewhere > in the format/partition commands I must to have made a mistake because > when I try to replace the disk I''m getting the following error: > > root:~# zpool replace rpool c2t5000CCA369C55DB8d0s0 > c2t5000CCA369C89636d0s0 > invalid vdev specification > use ''-f'' to override the following errors: > /dev/dsk/c2t5000CCA369C89636d0s0 overlaps with > /dev/dsk/c2t5000CCA369C89636d0s2 > root:~# > > I used -f and it worked but I was wondering is there a way to > completely "reset" the new disk? Remove all partitions and start from > scratch. > > Thank you > Peter > > > _______________________________________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20120412/f2d2045a/attachment.html> -------------- next part -------------- A non-text attachment was scrubbed... Name: laotsao.vcf Type: text/x-vcard Size: 483 bytes Desc: not available URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20120412/f2d2045a/attachment.vcf>
This might help, too: http://constantin.glez.de/blog/2011/03/how-set-zfs-root-pool-mirror-oracle-solaris-11-express On Apr 12, 2012, at 11:04 AM, Cindy Swearingen wrote:> Actually, since I answer the root pool partitioning questions > at least 52 times per year, I provided the steps for clearing > the existing partitions in step 9, here: > > http://docs.oracle.com/cd/E23823_01/html/817-5093/disksxadd-2.html#disksxadd-40 > > How to Create a Disk Slice for a ZFS Root File System > > Step 9 uses the format-->disk-->partition-->modify option and > sets the free hog space to slice 0. Then, you press return for > each existing slice to zero them out. This creates one large > slice 0. > > cs > > On 04/12/12 11:48, Cindy Swearingen wrote: >> Hi Peter, >> >> The root pool disk labeling/partitioning is not so easy. >> >> I don''t know which OpenIndiana release this is but in a previous >> Solaris release we had a bug that caused the error message below >> and the workaround is exactly what you did, use the -f option. >> >> We don''t yet have an easy way to clear a disk label, but >> sometimes I just create a new test pool on the problematic disk, >> destroy the pool, and start over with a more coherent label. >> This doesn''t work for scenarios. Some people use the dd command >> to wipe an existing label, but you must use it carefully. >> >> Thanks, >> >> Cindy >> >> On 04/12/12 11:35, Peter Wood wrote: >>> Hi, >>> >>> I was following the instructions in ZFS Troubleshooting Guide on how to >>> replace a disk in the root pool on x86 system. I''m using OpenIndiana, >>> ZFS pool v.28 with mirrored system rpool. The replacement disk is >>> brand new. >>> >>> root:~# zpool status >>> pool: rpool >>> state: DEGRADED >>> status: One or more devices has experienced an unrecoverable error. An >>> attempt was made to correct the error. Applications are >>> unaffected. >>> action: Determine if the device needs to be replaced, and clear the >>> errors >>> using ''zpool clear'' or replace the device with ''zpool replace''. >>> see: http://www.sun.com/msg/ZFS-8000-9P >>> scan: resilvered 17.6M in 0h0m with 0 errors on Wed Apr 11 17:45:16 2012 >>> config: >>> >>> NAME STATE READ WRITE CKSUM >>> rpool DEGRADED 0 0 0 >>> mirror-0 DEGRADED 0 0 0 >>> c2t5000CCA369C55DB8d0s0 OFFLINE 0 126 0 >>> c2t5000CCA369D5231Cd0s0 ONLINE 0 0 0 >>> >>> errors: No known data errors >>> root:~# >>> >>> I''m not very familiar with Solaris partitions and slices so somewhere in >>> the format/partition commands I must to have made a mistake because when >>> I try to replace the disk I''m getting the following error: >>> >>> root:~# zpool replace rpool c2t5000CCA369C55DB8d0s0 >>> c2t5000CCA369C89636d0s0 >>> invalid vdev specification >>> use ''-f'' to override the following errors: >>> /dev/dsk/c2t5000CCA369C89636d0s0 overlaps with >>> /dev/dsk/c2t5000CCA369C89636d0s2 >>> root:~# >>> >>> I used -f and it worked but I was wondering is there a way to completely >>> "reset" the new disk? Remove all partitions and start from scratch. >>> >>> Thank you >>> Peter >>> >>> >>> _______________________________________________ >>> zfs-discuss mailing list >>> zfs-discuss at opensolaris.org >>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >> _______________________________________________ >> zfs-discuss mailing list >> zfs-discuss at opensolaris.org >> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss > _______________________________________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20120412/a07dfe0d/attachment-0001.html>
Cindy Swearingen wrote:> > We don''t yet have an easy way to clear a disk label, ...dd if=/dev/zero of=... on the 1st and last 10% (roughly) of the disk has worked fine for me. -- Roberto Waltman
Thank you all for the replies. I''ll try the suggested solutions. -- Peter On Thu, Apr 12, 2012 at 2:06 PM, Roberto Waltman <lists at rwaltman.com> wrote:> Cindy Swearingen wrote: > >> >> We don''t yet have an easy way to clear a disk label, ... >> > > dd if=/dev/zero of=... on the 1st and last 10% (roughly) of the disk has > worked fine for me. > > -- > Roberto Waltman > > ______________________________**_________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/**mailman/listinfo/zfs-discuss<http://mail.opensolaris.org/mailman/listinfo/zfs-discuss> >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20120412/0f4d65dc/attachment.html>