Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 24 Feb 2022 09:57:14 -0600
From:      Larry Rosenman <ler@lerctr.org>
To:        Alexander Motin <mav@freebsd.org>
Cc:        Freebsd current <freebsd-current@freebsd.org>
Subject:   Re: ZFS PANIC: HELP.
Message-ID:  <5930f3d2b71c0932f903bb5b88a3c87d@lerctr.org>
In-Reply-To: <2bc12c9ee3e6dd71b65079116ff2b845@lerctr.org>
References:  <07c0c9c34b4a4133acab597c96867d27@lerctr.org> <95c7c326-e2f9-6e66-7b97-b9fb2671f4ad@FreeBSD.org> <1b6b2017ba69e6fda1ca237c3016ac61@lerctr.org> <a5f379a6-c2bd-f1e9-a281-bb657f3ccb75@FreeBSD.org> <6182c57bf1859482e72af78037c399e4@lerctr.org> <7abcbe88-446f-44b9-c3a7-0997d3430b57@FreeBSD.org> <2bc12c9ee3e6dd71b65079116ff2b845@lerctr.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On 02/23/2022 9:27 pm, Larry Rosenman wrote:
> On 02/23/2022 9:15 pm, Alexander Motin wrote:
>> On 23.02.2022 22:01, Larry Rosenman wrote:
>>> On 02/23/2022 8:58 pm, Alexander Motin wrote:
>>>> On 23.02.2022 21:52, Larry Rosenman wrote:
>>>>> On 02/23/2022 8:41 pm, Alexander Motin wrote:
>>>>>> Hi Larry,
>>>>>> 
>>>>>> The panic you are getting is an assertion, enabled by kernel built
>>>>>> with INVARIANTS option.  On 13 you may just not have that 
>>>>>> debugging
>>>>>> enabled to hit the issue.  But that may be only a consequence.
>>>>>> Original problem I guess in possibly corrupted ZFS intent log 
>>>>>> records
>>>>>> (or false positive), that could happen so due to use of -F 
>>>>>> recovery
>>>>>> option on `zpool import`, that supposed to try import pool at 
>>>>>> earlier
>>>>>> transaction group if there is some metadata corruption found.  It 
>>>>>> is
>>>>>> not supposed to work 100% and only a last resort.  Though may be 
>>>>>> that
>>>>>> assertion is just excessively strict for that specific recovery 
>>>>>> case.
>>>>>> If as you say pool can be imported and scrubbed on 13, then I'd 
>>>>>> expect
>>>>>> following clean export should allow later import on 14 without -F.
>>>>>> 
>>>>>> On 23.02.2022 21:21, Larry Rosenman wrote:
>>>>>>> 
>>>>>>> 've got my main dev box that crashes on 14 with the screen shot 
>>>>>>> at https://www.lerctr.org/~ler/14-zfs-crash.png.
>>>>>>> Booting from a 13-REL USB installer it imports and scrubs.
>>>>>>> 
>>>>>>> Ideas?
>>>>>>> 
>>>>>>> I can either video conference with shared screen or give access 
>>>>>>> to the console via my Dominion KVM.
>>>>>>> 
>>>>>>> Any help/ideas/etc welcome I really need to get this box back.
>>>>>>> 
>>>>>>> 
>>>>> How can I import the pool withOUT it mounting the FileSystems so I 
>>>>> can export it cleanly on the 13 system?
>>>> 
>>>> Why do you need to import without mounting file systems?  I think 
>>>> you
>>>> may actually wish them to be mounted to replay their ZILs.  Just use
>>>> -R option to mount file systems in some different place.
>>> 
>>> I get the errors shown at:
>>> https://www.lerctr.org/~ler/14-mount-R-output.png
>>> 
>>> Should I worry?  Or do something(tm) here?
>> 
>> This looks weird, but may possibly depend on mount points topology,
>> whether /mnt is writable, etc.  What happen if you export it now and
>> try to import it in normal way on 14 without -F?
> 
> It crashes just after root mount (this is the boot pool and only pool
> on the system),
> seeL
> https://www.lerctr.org/~ler/14-BOOT-Crash.png

Where do I go from here?


-- 
Larry Rosenman                     http://www.lerctr.org/~ler
Phone: +1 214-642-9640                 E-Mail: ler@lerctr.org
US Mail: 5708 Sabbia Dr, Round Rock, TX 78665-2106



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5930f3d2b71c0932f903bb5b88a3c87d>