[Ovirt-devel] poweroff in vm causes the host to become unavailable

[Scott Seago]

Justin Clacherty wrote:
> Ian Main wrote:
>>> I'm also still having problems with storage not working.  On reboot 
>>> if I start a VM that was working it complains that it can't find the 
>>> storage.  The only way I've been able to fix this is to delete all 
>>> the VMs, delete all the attached storage then add the storage and VM 
>>> again, a little painful.
>>>     
>
> Are you able to point me in some direction to debug this problem?  I 
> have 4 VMs set up all on separate iscsi partitions.  On restart I get 
> the following error in the task log.
>
> Unable to find volume 8.0.0.0 attached to pool 
> 192.168.50.37-iqn.2009-3.san0:services-3260.
>
> Replace the 8.0.0.0 with whatever lun it's decided to call itself for 
> each of the VMs.  One of the VMs is on a LVM volume in one of the 
> iscsi partitions.  The rest just have the entire iscsi partition 
> exported and all partitioning is done in the VM.  The VM that is on 
> the LVM partition starts every time.  The other three have the error 
> message above.  I'm not sure if this is just coincidence.  The only 
> way to get the other three to start is to go through the above process 
> which is a bit tedious.  ie. delete all three VMs, delete all three 
> storage servers.  Wait a bit, add the storage and VMs back one by one.
>
> Cheers,
> Justin.
It's possible that this is related to the existing problem we've got 
with iscsi -- rescanning the target results in a _new_ lun listed in the 
db even though there's only one lun active. It seems that each rescan 
results in a new lun name 8.0.0.0, rescan shows 9.0.0.0, etc.

I don't know if what you're seeing here is related to this problem or 
not -- I know that Ian is currently working to track down the rescan bug 
-- once there's a fix there it could help this issue too.

Scott