Hi Folks,
Begin forwarded message:
> From: Edward Irvine <eirvine at tpg.com.au>
> Date: 12 December 2007 8:44:57 AM
> To: ictim.unix at aec.gov.au
> Subject: Fwd: [zfs-discuss] zpool kernel panics.
>
> FYI ...
>
> Begin forwarded message:
>
>> From: "James C. McPherson" <James.McPherson at Sun.COM>
>> Date: 12 December 2007 8:06:51 AM
>> To: Edward Irvine <eirvine at tpg.com.au>
>> Cc: ZFS Discussions <zfs-discuss at opensolaris.org>
>> Subject: Re: [zfs-discuss] zpool kernel panics.
>> Reply-To: James.McPherson at Sun.COM
>>
>>
>> Hi Eddie,
>>
>> Edward Irvine wrote:
>>> Each time the system crashes, it crashes with the same error
>>> message. This suggests to me that it is zpool corruption rather
>>> than faulty RAM, which is to blame.
>>> So - is this particular zpool a lost cause? :\
>>
>> It''s looking that way to me, but I''m definitely no
expert.
>>
>>> A number of folks have pointed out that this bug may have been
>>> fixed in a very recent version (nv-77?) of opensolaris. As a
>>> last ditch approach, I''m thinking that I could put the
current
>>> system disks (sol10u4) aside, do a quick install the latest
>>> opensolaris, import the zpool, and do a zpool scrub, export the
>>> zpool, shutdown, swap in the sol10u4 disks, reboot, import.
>>> Sigh. Does this approach sound plausible?
>>
>> It''s definitely worth a shot, as long as you don''t
have
>> to zpool upgrade in order to do it.
OK - this appeared to work: imported the zpool into opensolaris 77,
did a zpool scrub - and no kernel panics. Cool!
But - after reimporting the zpool back into Solaris10u4 (where it
belongs) a zpool scrub still causes a kernel panic - although it
seemed to take a bit longer to panic. Same error message as before -
panic[cpu1]/thread=2a1015c7cc0:
Dec 15 12:49:35 server unix: [ID 361072 kern.notice] zfs: freeing
free segment (offset=423713792 size=1024)
Note that opensolaris 77 and Solaris10u4 are on the same physical
hardware - I''m just booting off different system disks.
>> I pulled your crash dump inside Sun, thankyou, but I haven''t
>> had a chance to analyze it so I''ve passed the details on to
>> more knowledgeable ZFS ppl.
>>
>>
>> James C. McPherson
>> --
>> Senior Kernel Software Engineer, Solaris
>> Sun Microsystems
>> http://blogs.sun.com/jmcp http://www.jmcp.homeunix.com/blog
>>
>>
>
Sigh. This must definitely be a bug.
Eddie