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

Re: [Xen-users] Re: Accessing an LV domU


  • To: "Alain BARTHE" <alain.barthe@xxxxxxxx>
  • From: Simone <dezmodue@xxxxxxxxx>
  • Date: Mon, 7 Apr 2008 00:18:08 +0100
  • Cc: xen-users@xxxxxxxxxxxxxxxxxxx, xen-users-bounces@xxxxxxxxxxxxxxxxxxx
  • Delivery-date: Sun, 06 Apr 2008 16:18:50 -0700
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:references; b=A6jCX0H1Cn9jK9BHKWulFaxOKyHGxom6lgOjIehSMT48cc7Q3G69gqVE3s8BWF5EyaxDoLLZLe6vPCbiiUI+PWbxYums6O13o8Wp4o/jPHlY4rQiT6wGG5dhH2q7+IWMSuSDqMwxD1CiSH7EQh02X8KM9etl6G2/ywq76SkesKM=
  • List-id: Xen user discussion <xen-users.lists.xensource.com>

Thanks all for your help, it is all working as expected now :)

Simone

On Tue, Apr 1, 2008 at 8:12 AM, Alain BARTHE <alain.barthe@xxxxxxxx> wrote:

Hi Simone,

kpartx is our friend :

        kpartx -a /dev/XenVG/qanew

Then look in /dev/mapper, you should see a new entry named something like "qanewpa", then

        mount /dev/mapper/qanewpa /opt/test

should work. If so, do the same with your swap.

Cheers.

Alain

xen-users-bounces@xxxxxxxxxxxxxxxxxxx a écrit sur 31/03/2008 17:02:42 :


> Hi again,
>
> I am afraid I will have to ask fro more advice. I have followed the
> suggestions, basically:
>
> Created an LV /dev/XenVG/qanew and /dev/XenVG/qanew_swap (no
> filesystem created on it)
>
> in the conf file for the VM I have:
>
> disk = [ 'phy:XenVG/qanew,xvda,w', 'phy:XenVG/qanew_swap,xvdb,w' ]
>
> I create only one partition xvda1 mounted as / and filesystem ext3
> while xvdb1 is used as swap, and GRUB is installed on MBR
>
> Once the installation is done and the VM is not on, I try to mount
> /dev/XenVG/qanew on /opt/test but it complains I need to specify
> filesystem, if I run mount -t ext3 .... it complains it is not an
> ext3 filesystem.
>
> I have been on the net trying to understand where i am going wrong
> but can't seem to find any valid solution.
>
> Any further help would be much appreciated.
>
> Thanks,
>
> Simone
>

> On Mon, Mar 31, 2008 at 11:19 AM, Simone <dezmodue@xxxxxxxxx> wrote:
> Thanks all for your replies.
>
> Considering the trouble I would have to go through and the fact that
> I don't need snapshotting within the VM, I am moving to a simpler
> configuration as suggested.
>
> Thanks again,
>
> Simone

>

> On Mon, Mar 31, 2008 at 2:49 AM, Jayson Charles Vantuyl <
> jvantuyl@xxxxxxxxxxxxxx> wrote:

> It's not so unusual.  It just wasn't easy to come across before
> widespread virtualization.

>
> Nesting LVM setups like this CAN be done, but generally shouldn't.

>
> Essentially, LVM works by detecting the physical volumes that are
> part of the LVM.  Since you are setting up an LVM that is
> effectively on another LVM device, the detection can get weird--
> especially if you name an inner LVM volume group the same name as an
> external LVM volume group.  Let's just say that mixing LVM data from
> the inner and outer devices can happen, and then things can get really broken.

>
> The easiest way to do this (and it's not so easy) is to set up LVM
> to scan your other LVM devices using a second lvm.conf.  Essentially
> you configure both copies of LVM (the internal one and the external
> one) in separate files with careful limits on which devices may be
> scanned for PVs.

>
> A generally better solution is just to use LVM on the outside and
> pass through the LVs to look like local disks.  The downside is that
> it requires cooperation from the Dom0 to resize and snapshot; and
> that FS extensions currently require rebooting the VM to pick up the
> size changes.  The upside is the LVM is really simple to manage if
> you just do it outside, and the volume is available to the Dom0 for
> mounting like you described.

>
> Unless you absolutely need snapshotting normally inside or live
> resizing, I wouldn't go through the trouble of running LVM inside of
> LVM.  It is fraught with peril.

>
> I am aware that you are going to have to reinstall your systems,
> but, if they are doing anything important, you should avoid things
> that can fatally confuse LVM (like nested LVM setups).  While you're
> at it, don't partition the disks inside of the DomUs either.  Just
> pass through the partitions directly.

>
> Good luck.

>
> On Mar 30, 2008, at 4:34 PM, Simone wrote:

> I guess I am trying to do something unusual :)
>
> Any thoughts?
>
> Simone
>

> On Fri, Mar 28, 2008 at 6:47 PM, Simone <dezmodue@xxxxxxxxx> wrote:
> Hi list,
>
> I am experiencing with xen on centos5 and so far everything is going
> well, I am really pleased. The domU have a dedicated LV
> (/dev/vg1/xenVM1, /dev/vg1/xenVM2 etc) and the guest OS is Centos4.
> At guest install time I have choosen to use LVM so that also inside
> the guest I have /dev/vg0/root, /dev/vg0/tmp etc. Is there a way to
> mount and edit the guest filesystem to customize files etc?
> The idea would be to have a guest template that can be cloned and
> then edited to generate new VMs.
>
> Thanks, have a good weekend
>
> Simone
>

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

>
>

> --
> Jayson Vantuyl
> Systems Architect
> Engine Yard
> jvantuyl@xxxxxxxxxxxxxx
> 1 866 518 9275 ext 204
>
> _______________________________________________
> Xen-users mailing list
> Xen-users@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-users

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

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.