[rhelv6-beta-list] My first experiences with RHEL6 beta

[Nico Kadel-Garcia]

On Sat, Jun 12, 2010 at 2:26 PM, Justin Clift <justin at salasaga.org> wrote:
> On 06/13/2010 03:40 AM, Nico Kadel-Garcia wrote:
> <snip>
>>
>> LVM is also a bad, bad, bad idea for virtualized environments, which
>> can use the virtualization host's tools to generate snapshots and for
>> which probing the filesystem of the guest for analysis or data
>> recovery is seriously encumbered by needing to manipulate LVM inside
>> the guest's image files. And ye gods, if two guests have identical
>> naming schemes you're in for a world of hurt.
>
> As a possibly useful thought, LVM usage with virtualized environments is
> something I'm getting into now.  Writing up and revising things as I go,
> presently into the Fedora "Virtualization Guide", which then gets
> rebranded/tweaked into the the RHEL "Virtualization Guide".
>
> (Very) draft snapshot here.  The LVM stuff is just rough notes at the moment
> and has a long way to go:
>
>
> http://justinclift.fedorapeople.org/Virtualization_Guide/chap-Virtualization-Storage_Pools-Storage_Pools.html#sect-Virtualization-Storage_Pools-Creating-LVM

Cool. The documentation seem sreasonable. But why would anyone create
a storage pool on a raw device, /dev/sdc in the documentation you
describe? Wouldn't you normally do this on a file image, or on a
partition, to provide more detail for anyone else trying to look at
your configurations?  Raw devices can't themselves be easily probed or
resized, or the data on them correctly reported. And too many tools
attempt to poke such devices and report them as "not formatted, what
would you like me to do with this, please let me partition it, pretty
please pretty please pretty please I'll give you a lollipop!!!!" The
clamoring is so insistent, especially for network based iSCSI or
external USB devices that may be passed around among multiple
operating systems, that I'd far rather put a partition on them to give
some indication of that the "device" is for. Unless you use LVM on the
raw storage device to re-allocate partitions for your guests, which I
suppose is doable, but then why not simply use an LVM primary
partition that takes up the whole disk so you can identify what the
heck it is if you probe it with tools like fdisk?

>
> (Chris Curran is the author of 99.9% of the Virtualization Guide, I'm only
> interested in the Storage side of things at present.)
>
> Chris recently pointed out "guestfish", a tool for accessing guest
> filesystems:
>
>  http://libguestfs.org/guestfish.1.html
>
> Haven't tried it yet, so no idea of it's capabilities, but apparently it's
> not bad.
>
> Any interest in trying it out, seeing if it's practical, and then writing up
> content about it for the Virtualization Guide?
>
> (hoping for a yes here)
>
> :)
>
> Regards and best wishes,
>
> Justin Clift

I'm afraid I'm running RHEL 6 beta in a Virtualbox guest environment
right now, and don't have a spare Linux box to use as a Virtualbox or
other virtualization technology to poke this. I'm doing some KVM
testing under RHEL 5: if I can spare cycles, I'll take a look but no
promises right now.