UFS Journaling gone south

[Mike Barnard]

Am I on the wrong mailing list with this question... should I bump it over
to another... if so, which one, freebsd-geom?



On Wed, May 19, 2010 at 2:43 PM, Mike Barnard <mike.barnardq at gmail.com>wrote:

>
>
> On Mon, May 17, 2010 at 8:38 PM, Mike Barnard <mike.barnardq at gmail.com>wrote:
>
>>
>>
>> On Mon, May 17, 2010 at 2:47 PM, Mike Barnard <mike.barnardq at gmail.com>wrote
>>>
>>>
>>>
>>> Prior to rebooting, I checked and I did have da0s1f.journal and
>>> da0s1g.journal and the da0s2d and da0s2e devices, after the reboot and the
>>> panic, I check /dev and the seem to have reverted back to da0s1f and
>>> da0s1g... and da0s2d and da0s2e have also gone AWOL... :-w
>>>
>>> fortunately, this is a new installation on which I want to test journal
>>> devices. I'll do this again, reserving some sectors on both partitions.
>>>
>>>
>> I must admit, this has me hands down. I have entered custom newfs flags on
>> sysinstall (-r 2, 4, and 8).
>>
>> With a test on these sector sizes to be reserved at the end of the disk,
>> gjournal label da0s1f da0s2d still says that I need to use -f to overwrite.
>> I do so and end up with journal devices, but when I reboot... the journal
>> devices and the journal providers are missing.
>>
>> I have even done -J flag on the newfs options in sysinstall but I end up
>> with the same results... no journal devices are created and I end up with a
>> kernel panic and into single usermode. There are no journal devices created.
>> Mounting the disks fails with: warning: GJOURNAL flag on fs but no gjournal
>> provider below.
>>
>> Am I doing all this wrong, Am I missing something?
>>
>>
> I have redone this setup with FreeBSD 8.0 STABLE but I get the same
> results. My da0s1f.journal and da0s1g.journal disappear and I get this:
>
> mount: /dev/da0s1g.journal : No such file or directory
> mount: /dev/da0s1f.journal : No such file or directory
> mount: /dev/da0s1g.journal : No such file or directory
>
> While in single usermode, an ls /dev still shows that da0s2* are not
> present.
>
> Is there a particular reason I'd have the journal devices vanish? Is there
> a solution for this?
>
>
>
>
> --
> Mike
>
> Of course, you might discount this possibility, but remember that one in
> a million chances happen 99% of the time.
> ------------------------------------------------------------
>



-- 
Mike

Of course, you might discount this possibility, but remember that one in
a million chances happen 99% of the time.
------------------------------------------------------------