Fedora xen - Jun 2007 - can''t create domU after failed install

Hello Everyone,

I''m trying to use virt-install for the first time to create a
paravirtualized CentOS 5 domU on a Fedora Core 6 dom0.  Here''s the
command I used:

virt-install -n pbx -r 256 --vcpus=1 -m 00:16:3E:66:94:5D --vnc krs -p
-l http://blizzard/mirror/centos/5/os/i386

I''m then prompted to enter the disk I want to use, at which point I
enter "/dev/vg00/asterisk" (a LV I previously created).  At this
point,
the install appears to start, but eventually fails, giving this reason:

Starting install...
ERROR:  Domain named pbx already exists!

There was a "pbx" config file in /etc/xen, but even after deleting it,
I
get the same error when I try to recreate the domU.

I assume I can clear this from "memory", but how?  Rebooting dom0
doesn''t help, so I''m assuming virt-install is pulling that
info from
somewhere.

Regards,

Ranbir

-- 
Kanwar Ranbir Sandhu
Linux 2.6.20-1.2944.fc6 i686 GNU/Linux 
21:11:04 up 5 days, 11:59, 5 users, load average: 0.71, 0.64, 0.54

On Thu, Jun 14, 2007 at 09:29:10PM -0400, Kanwar Ranbir Sandhu
wrote:
> Hello Everyone,
> 
> I''m trying to use virt-install for the first time to create a
> paravirtualized CentOS 5 domU on a Fedora Core 6 dom0.  Here''s the
> command I used:
> 
> virt-install -n pbx -r 256 --vcpus=1 -m 00:16:3E:66:94:5D --vnc krs -p
What is this ''krs'' arg you''re providing              
^^^^^^ ?
> -l http://blizzard/mirror/centos/5/os/i386
> 
> I''m then prompted to enter the disk I want to use, at which point
I
> enter "/dev/vg00/asterisk" (a LV I previously created).  At this
point,
> the install appears to start, but eventually fails, giving this reason:
BTW, you can use    --file /dev/vg00/asterisk     to specify the disk
on the command line.
> Starting install...
> ERROR:  Domain named pbx already exists!
> 
> There was a "pbx" config file in /etc/xen, but even after
deleting it, I
> get the same error when I try to recreate the domU.
Try  ''virsh undefine pbx'' to ensure all trace is removed.

Regards,
Dan.
-- 
|=- Red Hat, Engineering, Emerging Technologies, Boston.  +1 978 392 2496 -=|
|=-           Perl modules: http://search.cpan.org/~danberr/              -=|
|=-               Projects: http://freshmeat.net/~danielpb/               -=|
|=-  GnuPG: 7D3B9505   F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505  -=|

On Fri, 2007-06-15 at 03:22 +0100, Daniel P. Berrange
wrote:
> > virt-install -n pbx -r 256 --vcpus=1 -m 00:16:3E:66:94:5D --vnc krs -p
> 
> What is this ''krs'' arg you''re providing         
^^^^^^ ?
That''s me being gay.  I thought I could specify an argument to
"--vnc"
to get the vnc connection on my local desktop (I ssh''ed to the dom0
machine to start a remote install).  Eventually I figured out what the
--vnc option was for, and removed the "krs" (my laptop''s
hostname).
> BTW, you can use    --file /dev/vg00/asterisk     to specify the disk
> on the command line.
Yep - started to do that after reading the man page again.
> > There was a "pbx" config file in /etc/xen, but even after
deleting it, I
> > get the same error when I try to recreate the domU.
> 
> Try  ''virsh undefine pbx'' to ensure all trace is removed.
Is that in the man page?  Anyway, I noticed that the pbx domU was
actually running - that is, the install had started, but since I had
ssh''ed to the dom0 machine, DISPLAY wasn''t set.  So, the vnc
session
couldn''t start.  It was waiting for me to connect to
"localhost:5900".

I destroyed the pbx domU, forwarded vnc ports over ssh, restarted the
install, and connected to "localhost:5900" on my laptop.  I proceeded
with the install at that point.  Yeah!

Only hang up: the graphical install failed, which forced me to do a text
install via the vnc session.  I''m not sure why that happened.  Any
ideas?

Regards,

Ranbir

-- 
Kanwar Ranbir Sandhu
Linux 2.6.20-1.2944.fc6 i686 GNU/Linux 
22:53:51 up 5 days, 13:42, 5 users, load average: 0.72, 0.59, 0.61

On Thu, Jun 14, 2007 at 10:55:33PM -0400, Kanwar Ranbir Sandhu
wrote:
> On Fri, 2007-06-15 at 03:22 +0100, Daniel P. Berrange wrote:
> > > virt-install -n pbx -r 256 --vcpus=1 -m 00:16:3E:66:94:5D --vnc
krs -p
> > 
> > What is this ''krs'' arg you''re providing    
^^^^^^ ?
> 
> That''s me being gay.  I thought I could specify an argument to
"--vnc"
> to get the vnc connection on my local desktop (I ssh''ed to the
dom0
> machine to start a remote install).  Eventually I figured out what the
> --vnc option was for, and removed the "krs" (my laptop''s
hostname).
> 
> > BTW, you can use    --file /dev/vg00/asterisk     to specify the disk
> > on the command line.
> 
> Yep - started to do that after reading the man page again.
> 
> > > There was a "pbx" config file in /etc/xen, but even
after deleting it, I
> > > get the same error when I try to recreate the domU.
> > 
> > Try  ''virsh undefine pbx'' to ensure all trace is
removed.
> 
> Is that in the man page?  Anyway, I noticed that the pbx domU was
> actually running - that is, the install had started, but since I had
> ssh''ed to the dom0 machine, DISPLAY wasn''t set.  So, the
vnc session
> couldn''t start.  It was waiting for me to connect to
"localhost:5900".
> 
> I destroyed the pbx domU, forwarded vnc ports over ssh, restarted the
> install, and connected to "localhost:5900" on my laptop.  I
proceeded
> with the install at that point.  Yeah!
> 
> Only hang up: the graphical install failed, which forced me to do a text
> install via the vnc session.  I''m not sure why that happened.  Any
> ideas?
Not enough RAM probably - try increasing to 300-400 MB just for the install
time.

Dan.
-- 
|=- Red Hat, Engineering, Emerging Technologies, Boston.  +1 978 392 2496 -=|
|=-           Perl modules: http://search.cpan.org/~danberr/              -=|
|=-               Projects: http://freshmeat.net/~danielpb/               -=|
|=-  GnuPG: 7D3B9505   F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505  -=|