thr3ads.net - zfs discuss - [zfs-discuss] can''t import zpool after upgrade to solaris 10u6 [Jan 2009]

If this information is useful, please help other people find it:
Share via:

Steve Goldthorpe

2009-Jan-08 17:34 UTC

[zfs-discuss] can''t import zpool after upgrade to solaris 10u6

Here''s what I did:
* had a t1000 with a zpool under /dev/dsk/c0t0d0s7 on solaris 10u4
* re-installed with solaris 10u6 (disk layout unchanged)
* imported zpool with zpool import -f (I''m forever forgetting to export
them first) - this was ok
* re-installed with solaris 10u6 and more up-to-date patches (again forgetting
to export it)

When I do zpool import i get the following:
# zpool import         
  pool: zpool
    id: 17419375665629462002
 state: FAULTED
status: The pool metadata is corrupted.
action: The pool cannot be imported due to damaged devices or data.
        The pool may be active on another system, but can be imported using
        the ''-f'' flag.
   see: http://www.sun.com/msg/ZFS-8000-72
config:

        zpool       FAULTED  corrupted data
          c0t0d0s7  ONLINE

So I thought I''d done something wrong, however checked the partition
layout and it''s not changed.  However after doing a bit of poking
about, I''ve found some weird stuff - what zdb -l is showing and
what''s actually on the disk doesn''t seem to tally - I
can''t find that transaction ID from zdb and there seems to be a mixture
of version 4 and version 10 uberblocks on disk (and they all have bigger
transaction IDs than zdb is showing).

Am I missing something?

-Steve

# zdb -l /dev/dsk/c0t0d0s7
--------------------------------------------
LABEL 0
--------------------------------------------
    version=4
    name=''zpool''
    state=0
    txg=1809157
    pool_guid=17419375665629462002
    top_guid=12174008987990077602
    guid=12174008987990077602
    vdev_tree
        type=''disk''
        id=0
        guid=12174008987990077602
        path=''/dev/dsk/c0t0d0s7''
        devid=''id1,sd at n5000cca321ca2647/h''
        whole_disk=0
        metaslab_array=14
        metaslab_shift=30
        ashift=9
        asize=129904410624
        DTL=24
--------------------------------------------
LABEL 1
--------------------------------------------
    version=4
    name=''zpool''
    state=0
    txg=1809157
    pool_guid=17419375665629462002
    top_guid=12174008987990077602
    guid=12174008987990077602
    vdev_tree
        type=''disk''
        id=0
        guid=12174008987990077602
        path=''/dev/dsk/c0t0d0s7''
        devid=''id1,sd at n5000cca321ca2647/h''
        whole_disk=0
        metaslab_array=14
        metaslab_shift=30
        ashift=9
        asize=129904410624
        DTL=24
--------------------------------------------
LABEL 2
--------------------------------------------
--------------------------------------------
LABEL 3
--------------------------------------------

-- (sample output from a little script i knocked up)

Uberblock Offset: 0020000 (131072)
Uber version: 4
Transaction group: 1831936
Timestamp: 2008-11-20:11:14:49
GUID_SUM: 9ab0d28ccc7d2e94

Uberblock Offset: 0020400 (132096)
Uber version: 4
Transaction group: 1831937
Timestamp: 2008-11-20:11:14:54
GUID_SUM: 9ab0d28ccc7d2e94
...
Uber version: 10
Transaction group: 114560
Timestamp: 2009-01-07:09:59:11
GUID_SUM: 9f8d9ef301489223

Uberblock Offset: 0e18400 (14779392)
Uber version: 10
Transaction group: 114561
Timestamp: 2009-01-07:09:59:41
GUID_SUM: 9f8d9ef301489223
...
-- 
This message posted from opensolaris.org

Steve Goldthorpe

2009-Jan-09 16:06 UTC

head link

[zfs-discuss] can''t import zpool after upgrade to solaris 10u6

After having a think I''ve come up with the following hypothesis:

1) When I was on Solaris 10u4 things were working fine.
2) When I re-installed with Solaris 10u6 and imported the zpool (with zpool
import -f), it created a zpool.cache file and didn''t update the on disk
data structures for some reason.
3) When I re-installed Solaris 10u6, I lost the zpool.cache file and now zfs
looks at the data structures on the disk and they are inconsistent.

Could the above have actually happened?  It would explain what I''m
seeing.

-Steve
-- 
This message posted from opensolaris.org

Steve Goldthorpe

2009-Jan-10 23:59 UTC

head link

[zfs-discuss] can''t import zpool after upgrade to solaris 10u6

There''s definately something strange going on as these are the only
uberblocks I can find by scanning /dev/dsk/c0t0d0s7 - nothing to conflict with
my theory so far:

TXG: 106052 TIME: 2009-01-04:11:06:12 BLK: 0e29000 (14848000) VER: 10 GUID_SUM:
9f8d9ef301489223 (11497020190282519075)
TXG: 106052 TIME: 2009-01-04:11:06:12 BLK: 0e69000 (15110144) VER: 10 GUID_SUM:
9f8d9ef301489223 (11497020190282519075)
TXG: 106053 TIME: 2009-01-04:11:06:42 BLK: 0e29400 (14849024) VER: 10 GUID_SUM:
9f8d9ef301489223 (11497020190282519075)
TXG: 106053 TIME: 2009-01-04:11:06:42 BLK: 0e69400 (15111168) VER: 10 GUID_SUM:
9f8d9ef301489223 (11497020190282519075)
skipped 248 blocks...
TXG: 114710 TIME: 2009-01-07:11:14:10 BLK: 0e1d800 (14800896) VER: 10 GUID_SUM:
9f8d9ef301489223 (11497020190282519075)
TXG: 114710 TIME: 2009-01-07:11:14:10 BLK: 0e5d800 (15063040) VER: 10 GUID_SUM:
9f8d9ef301489223 (11497020190282519075)
TXG: 114715 TIME: 2009-01-07:11:15:41 BLK: 0e1ec00 (14806016) VER: 10 GUID_SUM:
9f8d9ef301489223 (11497020190282519075)
TXG: 114715 TIME: 2009-01-07:11:15:41 BLK: 0e5ec00 (15068160) VER: 10 GUID_SUM:
9f8d9ef301489223 (11497020190282519075)

TXG: 1830158 TIME: 2008-11-20:08:46:41 BLK: 0023800 (145408) VER: 4 GUID_SUM:
9ab0d28ccc7d2e94 (11146640579909987988)
TXG: 1830158 TIME: 2008-11-20:08:46:41 BLK: 0063800 (407552) VER: 4 GUID_SUM:
9ab0d28ccc7d2e94 (11146640579909987988)
TXG: 1830382 TIME: 2008-11-20:09:05:20 BLK: 003b800 (243712) VER: 4 GUID_SUM:
9ab0d28ccc7d2e94 (11146640579909987988)
TXG: 1830382 TIME: 2008-11-20:09:05:20 BLK: 007b800 (505856) VER: 4 GUID_SUM:
9ab0d28ccc7d2e94 (11146640579909987988)
skipped 248 blocks...
TXG: 1832026 TIME: 2008-11-20:11:22:18 BLK: 0036800 (223232) VER: 4 GUID_SUM:
9ab0d28ccc7d2e94 (11146640579909987988)
TXG: 1832026 TIME: 2008-11-20:11:22:18 BLK: 0076800 (485376) VER: 4 GUID_SUM:
9ab0d28ccc7d2e94 (11146640579909987988)
TXG: 1832027 TIME: 2008-11-20:11:22:19 BLK: 0036c00 (224256) VER: 4 GUID_SUM:
9ab0d28ccc7d2e94 (11146640579909987988)
TXG: 1832027 TIME: 2008-11-20:11:22:19 BLK: 0076c00 (486400) VER: 4 GUID_SUM:
9ab0d28ccc7d2e94 (11146640579909987988)

# zdb -l /dev/dsk/c0t0d0s7
--------------------------------------------
LABEL 0
--------------------------------------------
version=4
name=''zpool''
state=0
txg=1809157
pool_guid=17419375665629462002
top_guid=12174008987990077602
guid=12174008987990077602
vdev_tree
type=''disk''
id=0
guid=12174008987990077602
path=''/dev/dsk/c0t0d0s7''
devid=''id1,sd at n5000cca321ca2647/h''
whole_disk=0
metaslab_array=14
metaslab_shift=30
ashift=9
asize=129904410624
DTL=24
--------------------------------------------
LABEL 1
--------------------------------------------
version=4
name=''zpool''
state=0
txg=1809157
pool_guid=17419375665629462002
top_guid=12174008987990077602
guid=12174008987990077602
vdev_tree
type=''disk''
id=0
guid=12174008987990077602
path=''/dev/dsk/c0t0d0s7''
devid=''id1,sd at n5000cca321ca2647/h''
whole_disk=0
metaslab_array=14
metaslab_shift=30
ashift=9
asize=129904410624
DTL=24
--------------------------------------------
LABEL 2
--------------------------------------------
--------------------------------------------
LABEL 3
--------------------------------------------

-Steve
> After having a think I''ve come up with the following
> hypothesis:
> 
> 1) When I was on Solaris 10u4 things were working
> fine.
> 2) When I re-installed with Solaris 10u6 and imported
> the zpool (with zpool import -f), it created a
> zpool.cache file and didn''t update the on disk data
> structures for some reason.
> 3) When I re-installed Solaris 10u6, I lost the
> zpool.cache file and now zfs looks at the data
> structures on the disk and they are inconsistent.
> 
> Could the above have actually happened?  It would
> explain what I''m seeing.
> 
> -Steve-- 
This message posted from opensolaris.org

2009-Jan-11 00:18 UTC

head link

[zfs-discuss] can''t import zpool after upgrade to solaris 10u6

Hi Gold,
9987988 sounds factual to me...

IMHO,
z

----- Original Message ----- 
From: "Steve Goldthorpe" <sun at waistcoat.org.uk>
To: <zfs-discuss at opensolaris.org>
Sent: Saturday, January 10, 2009 6:59 PM
Subject: Re: [zfs-discuss] can''t import zpool after upgrade to solaris
10u6

> There''s definately something strange going on as these are the
only
> uberblocks I can find by scanning /dev/dsk/c0t0d0s7 - nothing to conflict 
> with my theory so far:
>
> TXG: 106052 TIME: 2009-01-04:11:06:12 BLK: 0e29000 (14848000) VER: 10 
> GUID_SUM: 9f8d9ef301489223 (11497020190282519075)
> TXG: 106052 TIME: 2009-01-04:11:06:12 BLK: 0e69000 (15110144) VER: 10 
> GUID_SUM: 9f8d9ef301489223 (11497020190282519075)
> TXG: 106053 TIME: 2009-01-04:11:06:42 BLK: 0e29400 (14849024) VER: 10 
> GUID_SUM: 9f8d9ef301489223 (11497020190282519075)
> TXG: 106053 TIME: 2009-01-04:11:06:42 BLK: 0e69400 (15111168) VER: 10 
> GUID_SUM: 9f8d9ef301489223 (11497020190282519075)
> skipped 248 blocks...
> TXG: 114710 TIME: 2009-01-07:11:14:10 BLK: 0e1d800 (14800896) VER: 10 
> GUID_SUM: 9f8d9ef301489223 (11497020190282519075)
> TXG: 114710 TIME: 2009-01-07:11:14:10 BLK: 0e5d800 (15063040) VER: 10 
> GUID_SUM: 9f8d9ef301489223 (11497020190282519075)
> TXG: 114715 TIME: 2009-01-07:11:15:41 BLK: 0e1ec00 (14806016) VER: 10 
> GUID_SUM: 9f8d9ef301489223 (11497020190282519075)
> TXG: 114715 TIME: 2009-01-07:11:15:41 BLK: 0e5ec00 (15068160) VER: 10 
> GUID_SUM: 9f8d9ef301489223 (11497020190282519075)
>
> TXG: 1830158 TIME: 2008-11-20:08:46:41 BLK: 0023800 (145408) VER: 4 
> GUID_SUM: 9ab0d28ccc7d2e94 (11146640579909987988)
> TXG: 1830158 TIME: 2008-11-20:08:46:41 BLK: 0063800 (407552) VER: 4 
> GUID_SUM: 9ab0d28ccc7d2e94 (11146640579909987988)
> TXG: 1830382 TIME: 2008-11-20:09:05:20 BLK: 003b800 (243712) VER: 4 
> GUID_SUM: 9ab0d28ccc7d2e94 (11146640579909987988)
> TXG: 1830382 TIME: 2008-11-20:09:05:20 BLK: 007b800 (505856) VER: 4 
> GUID_SUM: 9ab0d28ccc7d2e94 (11146640579909987988)
> skipped 248 blocks...
> TXG: 1832026 TIME: 2008-11-20:11:22:18 BLK: 0036800 (223232) VER: 4 
> GUID_SUM: 9ab0d28ccc7d2e94 (11146640579909987988)
> TXG: 1832026 TIME: 2008-11-20:11:22:18 BLK: 0076800 (485376) VER: 4 
> GUID_SUM: 9ab0d28ccc7d2e94 (11146640579909987988)
> TXG: 1832027 TIME: 2008-11-20:11:22:19 BLK: 0036c00 (224256) VER: 4 
> GUID_SUM: 9ab0d28ccc7d2e94 (11146640579909987988)
> TXG: 1832027 TIME: 2008-11-20:11:22:19 BLK: 0076c00 (486400) VER: 4 
> GUID_SUM: 9ab0d28ccc7d2e94 (11146640579909987988)
>
> # zdb -l /dev/dsk/c0t0d0s7
> --------------------------------------------
> LABEL 0
> --------------------------------------------
> version=4
> name=''zpool''
> state=0
> txg=1809157
> pool_guid=17419375665629462002
> top_guid=12174008987990077602
> guid=12174008987990077602
> vdev_tree
> type=''disk''
> id=0
> guid=12174008987990077602
> path=''/dev/dsk/c0t0d0s7''
> devid=''id1,sd at n5000cca321ca2647/h''
> whole_disk=0
> metaslab_array=14
> metaslab_shift=30
> ashift=9
> asize=129904410624
> DTL=24
> --------------------------------------------
> LABEL 1
> --------------------------------------------
> version=4
> name=''zpool''
> state=0
> txg=1809157
> pool_guid=17419375665629462002
> top_guid=12174008987990077602
> guid=12174008987990077602
> vdev_tree
> type=''disk''
> id=0
> guid=12174008987990077602
> path=''/dev/dsk/c0t0d0s7''
> devid=''id1,sd at n5000cca321ca2647/h''
> whole_disk=0
> metaslab_array=14
> metaslab_shift=30
> ashift=9
> asize=129904410624
> DTL=24
> --------------------------------------------
> LABEL 2
> --------------------------------------------
> --------------------------------------------
> LABEL 3
> --------------------------------------------
>
> -Steve
>
>> After having a think I''ve come up with the following
>> hypothesis:
>>
>> 1) When I was on Solaris 10u4 things were working
>> fine.
>> 2) When I re-installed with Solaris 10u6 and imported
>> the zpool (with zpool import -f), it created a
>> zpool.cache file and didn''t update the on disk data
>> structures for some reason.
>> 3) When I re-installed Solaris 10u6, I lost the
>> zpool.cache file and now zfs looks at the data
>> structures on the disk and they are inconsistent.
>>
>> Could the above have actually happened?  It would
>> explain what I''m seeing.
>>
>> -Steve
> -- 
> This message posted from opensolaris.org
> _______________________________________________
> zfs-discuss mailing list
> zfs-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss

zfs discuss - Jan 2009 - can''t import zpool after upgrade to solaris 10u6

[zfs-discuss] can''t import zpool after upgrade to solaris 10u6

[zfs-discuss] can''t import zpool after upgrade to solaris 10u6

[zfs-discuss] can''t import zpool after upgrade to solaris 10u6

[zfs-discuss] can''t import zpool after upgrade to solaris 10u6