Xen users - Jan 2009 - pygrub CentOS domU - unable to mount lvm partition form dom0 -

Hi,

am using Xen 3.3.1 ontop of Debian etch x64.

Most of the domU are custom made or downloaded from jailtime.org.
Am using custom 2.6.18 kernel and slowly migrating to PV-OPS kernel -
2.6.26 and higher.

Specially for CentOs installation am using network install via xen
image and than domU boot via pygrub with original RH/CentOS xenified
kernel.

Everything runs fine but suddenly yesterday I have realized something
that really annoys me.

Am using lvm based partitions for all domU. For xenified an pv-ops
kernels there is no problem mounting the lvm domU partition from dom0.
But when trying pygrub CentOS installed domU - it goes wrong:

mount /dev/vg/centospy /mn/lvm/
mount: you must specify the filesystem type

mount -t ext3 /dev/vg/centospy /mn/lvm/
mount: wrong fs type, bad option, bad superblock on /dev/vg/x1hd,
       missing codepage or other error
       In some cases useful info is found in syslog - try
       dmesg | tail  or so

in dmesg:
VFS: Can''t find ext3 filesystem on dev dm-0.


fdisk -l /dev/vg/x1hd

Disk /dev/vg/centospy: 32.2 GB, 32212254720 bytes
255 heads, 63 sectors/track, 3916 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

       Device Boot      Start         End      Blocks   Id  System
/dev/vg/centospy   *           1        3916    31455238+  83  Linux


During the CentOS install annaconda is reporting something like "loop
partition" detected and have to reinitialize disk.



Please can somene help me out how to mount this partition from dom0?


Br

Peter

_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users

On Sat, Jan 31, 2009 at 5:05 AM, Peter Braun <xenware@gmail.com>
wrote:
>       Device Boot      Start         End      Blocks   Id  System
> /dev/vg/centospy   *           1        3916    31455238+  83  Linux
Shouldn''t it be /dev/vg/centospy1 ?
I''m getting the feeling you''re trying to anonymize LV name and
change
it on the output line but forgot to do the same on the command line
(and vice-versa for the mount command)
> During the CentOS install annaconda is reporting something like "loop
> partition" detected and have to reinitialize disk.
If you let anaconda reinitialize disk (and thus creating a new
partition table), there will be a partition on top of that LVM, which
fdisk SHOULD show by adding the the number "1" or "p1" to
the LV name
> Please can somene help me out how to mount this partition from dom0?
Try using kpartx. Something like
kpartx -av /dev/vg/centospy (or /dev/vg/x1hd, whatever your LV is)
mount /dev/mapper/centospy1 or x1hd1 (or whatever new block device you
see on /dev/mapper)
unmount afterwards, and don''t forget to delete dev mappings with kpartx
-dv

Regards,

Fajar

_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users

2009/1/30 Fajar A. Nugraha <fajar@fajar.net>:
> On Sat, Jan 31, 2009 at 5:05 AM, Peter Braun <xenware@gmail.com>
wrote:
>>       Device Boot      Start         End      Blocks   Id  System
>> /dev/vg/centospy   *           1        3916    31455238+  83  Linux
>
> Shouldn''t it be /dev/vg/centospy1 ?
> I''m getting the feeling you''re trying to anonymize LV
name and change
> it on the output line but forgot to do the same on the command line
> (and vice-versa for the mount command)
mount /dev/vg/centospy1 /mnt/lvm
mount: special device /dev/vg/centospy1 does not exist

mount /dev/vg/centospy/1 /mnt/lvm
mount: special device /dev/vg/centospy/1 does not exist
       (a path prefix is not a directory)
>
>> During the CentOS install annaconda is reporting something like
"loop
>> partition" detected and have to reinitialize disk.
>
> If you let anaconda reinitialize disk (and thus creating a new
> partition table), there will be a partition on top of that LVM, which
> fdisk SHOULD show by adding the the number "1" or "p1"
to the LV name
>
>> Please can somene help me out how to mount this partition from dom0?
>
> Try using kpartx. Something like
> kpartx -av /dev/vg/centospy (or /dev/vg/x1hd, whatever your LV is)
> mount /dev/mapper/centospy1 or x1hd1 (or whatever new block device you
> see on /dev/mapper)
> unmount afterwards, and don''t forget to delete dev mappings with
kpartx -dv
Unfortunately kpartx is only for debian lenny or sid and I dont wannt
to mess my dom0 :(
> Regards,
>
> Fajar
>
_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users

On Sat, Jan 31, 2009 at 5:45 AM, Peter Braun <xenware@gmail.com>
wrote:
> 2009/1/30 Fajar A. Nugraha <fajar@fajar.net>:
>> On Sat, Jan 31, 2009 at 5:05 AM, Peter Braun <xenware@gmail.com>
wrote:
>>>       Device Boot      Start         End      Blocks   Id  System
>>> /dev/vg/centospy   *           1        3916    31455238+  83 
Linux
>>
>> Shouldn''t it be /dev/vg/centospy1 ?
>
> mount /dev/vg/centospy1 /mnt/lvm
> mount: special device /dev/vg/centospy1 does not exist
What I meant was, shouldn''t the output of fdisk -l has "1" on
it (e.g.
centospy1 instead of cenotspy) ? If yes, then you need to create the
device mappings first with kpartx, and THEN mount it (it will be on
/dev/mapper, NOT /dev/vg)

> Unfortunately kpartx is only for debian lenny or sid and I dont wannt
> to mess my dom0 :(
Then you''re out of luck.
Well there''s a hard way to do it, actually. Something like

mount -o offset=32256 /dev/vg/centospy /mnt/lvm

To get the offset (in my example 32256), do "fdisk -lu" and multiply
the starting sector (usually 63) with 512

Regards,

Fajar

_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users