[Xen-users] Re: scsi streamer inside a XEN VM crashes the VM

[Christian Tramnitz <chris.ace@xxxxxxx>]

Rupert wrote:
> Hello,
>
> i have a superloader3 with a DLT-S4 drive connected to my dom0,
> I blocked the device for dom0 and added it to one VM.
> We use Xen 3.3 on a CentOS 5.3 System, the VM is located on a LVM device.
>
> Everything worked fine for 3 days, but now when I start my backups 
> script the whole VM crashes!
> At the new reboot it seems to hang for ages on this line:
>
> Starting udev: Disabling IRQ #16
>
> I did these commands to block the device on dom0:
>
> modprobe pciback
> echo -n '0000:03:04.0' > /sys/bus/pci/drivers/aic79xx/unbind
> echo -n '0000:03:04.0' > /sys/bus/pci/drivers/pciback/new_slot
> echo -n '0000:03:04.0' > /sys/bus/pci/drivers/pciback/bind
>
> than I added  pci=['03:04.0'] the VM configfile.
>
> Strangely when I start tar from the command live it exits with an error 
> but does not crash the
> machine
>
> tar: /dev/st0: Wrote only 4096 of 10240 bytes
> tar: Error is not recoverable: exiting now
>
> During a reboot the messages is flooded with these messages:
> [...]
>
> did we do something wrong here?
>
>
>
>
> thnx
>
> Rupert

Did you blacklist the aic79xx driver on dom0? While unbinding and 
rebinding to pciback may work well in general I can think of problems if 
the driver gets loaded twice (i.e. initializing /loading a card specific 
rom).
Do you see any error messages releated to permissive mode? This would be 
most likely on dom0 while domU is loading the driver or starting the tar.

If that doesn't help you may want to try pvscsi. Not sure if this is 
incorporated into the CentOS kernels yet, but its generally less 
intrusive (i.e. pci-passthrough blocks migration).

Best regards,
   Christian


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