[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [dm-devel] device mapper not finding all devices



Thanks for the response. I did figured out the issue a few days ago.
This was a CentOS 5.6 system, and 2.6.18 was the newest kernel I could
get for that distribution.  They did just release CentOS 6.0 and I was
trying to prepare for a migration when this issue arose.

I didn't even think I was using dmraid. Turns out that I created a
dmraid set a long time ago and then decided to replace with a spanned
logical volume.  I guess the metadata  was still on the disks and for
some reason it didn't bite me until I upgraded some software (not sure
which package it was) or made some type of change.  I deactivated
dmraid on the drives and the problem was gone.

Thanks!

David


On Tue, Jul 19, 2011 at 11:51 AM, Phillip Susi <psusi cfl rr com> wrote:
> On 7/17/2011 3:03 PM, David Cramblett wrote:
>>
>> Current Software:
>> Linux 2.6.18-238.12.1.el5 #1 SMP Tue May 31 13:23:01 EDT 2011 i686
>> i686 i386 GNU/Linux
>> device-mapper-1.02.55-2.el5
>
> 2.6.18 is now 5 years old; it might be time to upgrade.
>
>> I did some research and the only info I could find was regarding dmsetup:
>>
>> #dmsetup status
>> pdc_eehgddjjp1: 0 976768002 linear
>> pdc_eehgddjj: 0 976772992 linear
>> via_cdeadaheebp1: 0 976768002 linear
>> via_cdeadaheeb: 0 1953546240 striped 2 8:0 8:16 1 AA
>>
>>
>> Looks like something is wrong with the vg0-lvol0 entry above, but I
>> can't figure out how to fix it.  There is also another HDD on each of
>
> You only seem to have 4 entries for two dmraid disks ( which is not well
> supported so you should not be using it unless you need to dual boot with
> windows ) and a partition on each.  It is also quite strange that the pdc
> array is using a linear mapping.  Could it be a degraded mirror?
>
>> the via and pdc controllers that are seen by the system, but not
>> showing up in the "status" list above.  There is also a disk connected
>> to PATA port "/dev/hda" that is part of the lvol.
>
> What disks are supposed to be LVM physical volumes?  The output of pvdisplay
> may prove useful.  An lvchange -ay might also be worthwhile.
>
>> The device-mapper device is being created for the lvol:
>>
>>  /dev/mapper/vg0-lvol0
>
> If it had actually been created, then it should show up in dmsetup status.
>  The system appears to be failing to create it.
>



-- 
David Cramblett


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]