Hi everyone!
On one of our meetings [1] we agreed on keeping *ED states (DELETED,
INSPECTED, CLEANED, etc) as no-ops for now. Since then, however, the
inspection patch [2] got several comments from cores requesting removal
of INSPECTED state. That was done by Nisha.
Today we decided to approve [2] and bring the discussion here. We can
always add the state later, and blocking this patch again would be a bit
unfair IMO. We'll create a follow-up to [2] if we decide that we need
INSPECTED state.
So now, are we keeping/adding these *ED states to our state machine? I
personally agree with what was discussed on the meeting, namely:
1. Keep *ED states
2. Make them no-ops, so that code that does INSPECTING -> MANAGEABLE
right now will do INSPECTING -> INSPECTED -> MANAGEABLE instead.
Having INSPECTED is also useful for distinguish between OOB case (when
inspect_hardware returns after having everything done) and in-band case
(when inspect_hardware returns after initializing inspection). We could
use return value of inspect_hardware being INSPECTING or INSPECTED.
WDYT?
Dmitry.
[1]
http://eavesdrop.openstack.org/meetings/ironic/2015/ironic.2015-02-09-17.00.log.html
starting 17:47
[2] https://review.openstack.org/#/c/147857/
__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
