Re: [linux-lvm] vgremove fails to remove empty VG (v2.02.98)

From: Marian Csontos <mcsontos@redhat.com>
To: LVM general discussion and development <linux-lvm@redhat.com>
Cc: Joschi Brauchle <joschi.brauchle@tum.de>
Subject: Re: [linux-lvm] vgremove fails to remove empty VG (v2.02.98)
Date: Thu, 24 Jan 2013 11:44:53 +0100	[thread overview]
Message-ID: <510110A5.9010103@redhat.com> (raw)
In-Reply-To: <5100EC4B.50306@tum.de>

On 01/24/2013 09:09 AM, Joschi Brauchle wrote:
> Thanks for the detailed instructions, here are (some of) the missing
> gaps :-)
>
>>> I get the following error with lvm2.02.98:
>>> -----------
>>> #/sbin/vgdisplay --units k -v 'system'
>>>
>>> --- Volume group ---
>>> VG Name system
>>> System ID
>>> Format lvm2
>>> Metadata Areas 1
>>> Metadata Sequence No 13
>>> VG Access read/write
>>> VG Status resizable
>>> MAX LV 0
>>> Cur LV 0
>>> Open LV 0
>>> Max PV 0
>>> Cur PV 1
>>> Act PV 1
>>> VG Size 104669184.00 KiB
>>> PE Size 4096.00 KiB
>>> Total PE 25554
>>> Alloc PE / Size 0 / 0 KiB
>>> Free PE / Size 25554 / 104669184.00 KiB
>>> VG UUID Qs0uaZ-L88n-tC7T-fZQj-sgJY-2yRX-Dlt2Cr
>>>
>>> --- Physical volumes ---
>>> PV Name /dev/sda3
>>> PV UUID wzGjXi-F2lT-d2d4-l5a6-EmfM-W838-YxaFIh
>>> PV Status allocatable
>>> Total PE / Free PE 25554 / 25554
>>> -----------
>>>
>>> VG system is empty, but still:
>>> -----------
>>> # /sbin/vgremove 'system'
>>> -----------
>>> failes with status code 5, which means 'can't remove volume group
>>> containing logical volumes' (from man page).
>>>
>> 0. Which distro?
> openSUSE 12.3 Beta 1 - Problem happens during installation, where an
> existing VGroup shall be removed.
>>
>> 1. lvmdump -ma
> attached (hope that's ok)
>
>> 2. vgremove -vvvv &> vgremove-dirty.out
> http://pastebin.com/cmek1bYi
>
>> 3. ps aux | grep lvmetad
> Not running, no processes found. Skipping the remaining steps...
>>
>> 3.1 if so, the issue is likely to go away after restarting it. If you
>> `gcore` or `kill -SEGV` it to get a core dump it may help us.
>> 3.2 do another lvmdump here, to get "updated" data after lvmdump restart
>> 3.3 retry 'vgremove -vvvv &> vgremove-reloaded.out'
>> 4. pastebin the vgremove-dirty.out and we will be able to tell more then.

Any chance the FS containing /etc/lvm is mounted read-only?

If so it may fail. There is a bug as no error is reported in that 
situation. See backup section of lvm.conf - set backup and archive to 
point to a writable FS.

If not, could you send a tarball with /etc/lvm: `tar czf lvmdump-etc 
/etc/lvm/`

Joining us on IRC: #lvm @ freenode guarantees faster answers.

-- Marian

>
>
> _______________________________________________
> linux-lvm mailing list
> linux-lvm@redhat.com
> https://www.redhat.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/