Drivers should be using error handler deploying_error_handler() or
cleaning_error_handler() from conductor.utils for handling the failures
of deploy or clean step instead of directly calling process_event('fail').
These handlers perform necessary cleaning before moving the node to failed
state.
Conflicts:
ironic/drivers/modules/ilo/management.py
ironic/tests/unit/drivers/modules/ilo/test_management.py
ironic/tests/unit/drivers/modules/ilo/test_raid.py
NOTE(stendulker): The conflict is due to not having following changes
in Train:
9fb4074bfed927dd0d0a
Change-Id: I5d9b6831137cdef171a9929f144956f741390953
(cherry picked from commit 76a6fca8d7)