[Linux-cluster] cleaning up LVM cluster after disk loss

[Terry]

On Thu, Feb 23, 2012 at 8:05 PM, Digimer <linux at alteeve.com> wrote:

> On 02/23/2012 08:53 PM, Terry wrote:
> > We had a disk loss where we lost 4 clustered volumes.  Any commands I
> > run give me errors similar to:
> > /dev/vg_data01h/lv_data01h: read failed after 0 of 4096 at
> > 6597069701120: Input/output erro
> >
> > How do I just clean all of the broken pv's, lv's, and vg's out and start
> > fresh?  I have tried removing and it reports reports not found.
> >
> > Thanks!
>
> I make no claim to expertise, but I *think* you want:
>
> dmsetup remove /dev/vg_data01h/lv_data01h
>
> I use this for cleaning up VG/LVs when I kick out a USB based PV by
> accident.
>
> --
> Digimer
> E-Mail:              digimer at alteeve.com
> Papers and Projects: https://alteeve.com
>


Thanks for this.  It cleaned them up fine and I was able to move forward
with that one.  However, I am trying to recreate the PVs and pvdisplay is
giving me these errors which has me concerned:

  Found duplicate PV e32c4FWG0wWpikEDk1O76SfLQ1n1NQoR: using /dev/sdag not
/dev/sdq
  Found duplicate PV e32c4FWG0wWpikEDk1O76SfLQ1n1NQoR: using
/dev/mapper/eql-0-8a0906-18a0eb801-d0f000000054e679-omadvnfs01-data01d not
/dev/mapper/eql-0-8a0906-b560eb801-857006ad19b4f46a-omadvnfs01-data01c
  Found duplicate PV 0XsTBwrfBBEHJRuS1QoslkwKN0JJ9sBx: using /dev/sdc not
/dev/sdb
  Found duplicate PV IprT4Sf1idGpLXJErPpzwywUpNRScX8H: using
/dev/mapper/eql-0-8a0906-ef991aa01-5984c8d590748372-omadvnfs01-data01ap1
not /dev/mapper/eql-0-8a0906-ac70eb801-541006ad1984f46a-omadvnfs01-data01bp1

The last two volumes are data01b and data01c which are referenced in these
errors.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20120224/9dfa15ae/attachment.htm>