RHEV VM paused due to storage I/O problem.

Issue

• VM persistently transitions to paused state in RHEV.

• The event log in the RHEV-M Admin Portal contains;

  VM vm03 has paused due to storage I/O problem.
  

• The qemu log for this VM contains several instances of the following error;

  block I/O error in device 'drive-virtio-disk2': Input/output error (5)
  

• The vdsm logs on the host that the VM was started on contains eio errors;

  libvirtEventLoop::INFO::2013-11-01 05:30:04,377::libvirtvm::2519::vm.Vm::(_onAbnormalStop) vmId=`90833f68-83b4-4216-afdb-e5c068c673a7`::abnormal vm stop device virtio-disk2 error eio
  

• qemu-img check shows that the qcow2 image has leaked clusters;

  # qemu-img check /rhev/data-center/ec759bb4-834b-11e1-943e-52540027efbf/b0c17a6d-c5a5-4646-b4b5-edd85bc658db/images/34f13ccb-75a0-4877-87a2-7fe2eca09e5b/795dbdb2-e577-456d-b97f-7aa028fc00df
  
  Leaked cluster 1393307 refcount=1 reference=0
  Leaked cluster 1393308 refcount=1 reference=0
  Leaked cluster 1393309 refcount=1 reference=0
  Leaked cluster 1393310 refcount=1 reference=0
  
  28906 errors were found on the image.
  Data may be corrupted, or further writes to the image may corrupt it.
  
  68825 leaked clusters were found on the image.
  This means waste of disk space, but no harm to data.
  
  16916 internal errors have occurred during the check.
  
  An error has occurred during the check: Success
  The check is not complete and may have missed error.
  #