[Xen-users] xl stack problems on CentOS6 XEN4

["Hildebrand, Nils (BIT II 9)" <Nils.Hildebrand@xxxxxxxxxxx>]

I am currently in the process of replacing my SLES11 SP3 Dom0s with CentOS6 XEN4 (at the moment: proof of concept with a test-cluster).

 

I want to use the new xl-stack, instead of the old xm-stack, too.

 

So far I stumbled across a number of minor glitches which could be worked around, I got DomUs with CentOS6 and CentOS 5 PV up and running with drbd-disk-backend (primary/secondary mode).

 

But now I am stuck at the problem that pygrub does not seem to know about drbd-devices.

 

I could work around this during startup of a DomU by writing a wrapper-script, that will switch drbd to primary before startup and changes to /dev/drbd/by-res before starting “xl create”.

 

My disk-spec looks something like “disk = [ ‘drbd:vmname,xvda,w’ ] which worked very well on SLES11 SP3 (no workarounds needed there).

 

Now I tried to reboot a DomU by issuing (shutdown –r within the DomU).

 

What happens:

After shutdown the DomU is destroyed. xl knows how to handle disk-resources of type drbd with block-drbd and switches the disk to secondary mode.

 

But then xl create kicks in, first by starting up pygrub – thus failing to get to the initrd/kernel of the DomU.

 

I see two possible solutions here:

a)    make pygrub drbd-aware (like SLES11 SP3 does?)

b)    xl has to start up drbd in primary mode before pygrub kicks in?

c)    a mixture of a) and b)?

 

I hope you can help

 

 

Kind regards

 

Nils Hildebrand

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxx
http://lists.xen.org/xen-users