[Xen-API] vm export bug in XCP 0.5

[George Shuklin <george.shuklin@xxxxxxxxx>]

Good day.

vbd, used by control domain for vm export does not destroyed if vm
export is canceled.

Steps to reproduce (may include some unnecessary steps):

a) Create vm with vdi on lvmoiscsi storage, and host suspend-sr on
lvmoiscsi-storage
b) Start VM on slave
c) suspend VM
d) start export to file (in my case it was NFS share) via cli directly
from slave host where VM was resident before suspend.
e) interrupt export about 20-30 sec after start by pressing Ctrl-C
f) Try to start VM (vm-start), got: 
xe vm-resume vm=test
The server failed to handle your request, due to an internal error.  The
given message may give details useful for debugging the problem.
message: Failure("The VDI 4813431c-56ee-4a68-811a-b15b69a20e57 is
already attached in RO mode; it can't be attached in RW mode!")

Yes, VBD was attached to control domain:

uuid ( RO)             : b10b03c1-79f1-3406-aea1-46a13bb040aa
          vm-uuid ( RO): d4e93255-cbc0-43a6-95cc-936c9dcdd79c
    vm-name-label ( RO): Control domain on host: cvt-xh3
         vdi-uuid ( RO): 4813431c-56ee-4a68-811a-b15b69a20e57
            empty ( RO): false
           device ( RO): xvda



_______________________________________________
xen-api mailing list
xen-api@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/mailman/listinfo/xen-api