CentOS virt - Sep 2011 - Having trouble starting more than 8 Xen guests

All,

I've spent quite a few hours trying to figure this out to no avail...  I 
have two machines and the result is the same on both.  Additionally, 
each machine does not support full hardware virtualization...

I am able to create 8 VMs.  When attempting to create ninth VM, I get the
following
error:

Error: Device 51712 (vbd) could not be connected. /etc/xen/scripts/block 
failed; error detected.

To create the VMs, I use the command:

"xm create [vm name]"

My machine has 8 GB RAM and is an "old school" dual opteron box.

"uname -a" is:

Linux centos-host-2 2.6.18-274.3.1.el5.centos.plusxen #1 SMP Wed Sep 7 
06:25:54 EDT 2011 x86_64 x86_64 x86_64 GNU/Linux


To be clear, I can shutdown any one of the 8 VMs and then run the ninth 
"xm create" - that VM will start...

I did some google'ing and from what I cn tell that error arises due to the 
lack of loop back devices (there being 8 by default)...

I've tried both using MAKEDEV (in my /etc/rc.local) to create a number of 
loop back devices and put max_loops in my grub,.conf file.  For example 
here is what I have in grub.conf now:

title CentOS (2.6.18-274.3.1.el5.centos.plusxen)
         root (hd0,0)
         kernel /xen.gz-2.6.18-274.3.1.el5.centos.plus
         module /vmlinuz-2.6.18-274.3.1.el5.centos.plusxen ro root=LABEL=/ 
max_loop=64
         module /initrd-2.6.18-274.3.1.el5.centos.plusxen.img


Due to wrapping, max_loop is actually on the module /vmlinuz- line...

I also tried to put max_loop on the kernel line as well...

As I mentioned above, I also created the loop back devices from 
/etc/rc.local like so:

/sbin/MAKEDEV -d /dev -m 64 loop

And, once I can log in to the machine, I do see everything in /dev (for 
example /dev/loop0 to /dev/loop64)

Has anyone else had this problem?  If so, how did you resolve it????


Thanks ahead of time!!!!!


Scot P. Floess             RHCT  (Certificate Number 605010084735240)
Chief Architect FlossWare  http://sourceforge.net/projects/flossware
                            http://flossware.sourceforge.net
                            https://github.com/organizations/FlossWare

On Sep 18, 2011, at 9:13, "Scot P. Floess" <sfloess at
nc.rr.com> wrote:
> I did some google'ing and from what I cn tell that error arises due to
the
> lack of loop back devices (there being 8 by default)...
> 
> I've tried both using MAKEDEV (in my /etc/rc.local) to create a number
of
> loop back devices and put max_loops in my grub,.conf file.  For example 
> here is what I have in grub.conf now:
> 
> title CentOS (2.6.18-274.3.1.el5.centos.plusxen)
>         root (hd0,0)
>         kernel /xen.gz-2.6.18-274.3.1.el5.centos.plus
>         module /vmlinuz-2.6.18-274.3.1.el5.centos.plusxen ro root=LABEL=/ 
> max_loop=64
>         module /initrd-2.6.18-274.3.1.el5.centos.plusxen.img
> 
> 
> Due to wrapping, max_loop is actually on the module /vmlinuz- line...
> 
> I also tried to put max_loop on the kernel line as well...
> 
> As I mentioned above, I also created the loop back devices from 
> /etc/rc.local like so:
> 
> /sbin/MAKEDEV -d /dev -m 64 loop
> 
> And, once I can log in to the machine, I do see everything in /dev (for 
> example /dev/loop0 to /dev/loop64)
> 
> Has anyone else had this problem?  If so, how did you resolve it????
> 
Actually, you have max_loops=64 in the wrong file.  You want it in
/etc/modprobe.conf as thus:

options loop max_loop=64

Once that is in, a reboot (after shutting down the running VMs) is the quickest
way to activate it.  Take the makedev and grub stuff out...

  -I