Add internal subnet for dhcp server instead of using existing one to avoid IP shortage and conflicts

[Xicheng Chang]

questions for dhcp start and end ip address during compass installation provides default value based on cidr calculated from IP info on the specified NIC or default eth0, however, the default value 100-250 probably will have conflicts with existing IP plan. in Tsinghua, 10.1.0.* is pretty used up, there is great possibility that conflicts will occur.( Although in Tsinghua 10.1.* has a 16 bit mask the default value use 10.1.0.* ). Months ago i suggested allocating a dedicated internal subnet for dhcp pool only when installing compass to avoid conflict with customer's existing network.
Also on documents, those values should be printed as examples instead of default, so users will know they need to change it rather than pressing enter all the way down.

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/135801
Committed: https://git.openstack.org/cgit/stackforge/compass-adapters/commit/?id=f639a5d67f00cc282258107cb5ff592a895067cf
Submitter: Jenkins
Branch: dev/experimental

commit f639a5d67f00cc282258107cb5ff592a895067cf
Author: Xicheng Chang <email address hidden>
Date: Wed Nov 19 16:50:32 2014 -0800

    Separate cobbler IP from compass IP in cobbler snippets

    Compass IP now is isolated so that during PXE process,
    a pxe-server IP is available and DHCP will have its own
    subnet and cobbler server owns the vip for that subnet.
    This only partially closes bug #1393957, further change
    will be maide in compass-core. So no "Closes-Bug" will
    be added in the message.
    Partial-bug: #1393957
    Change-Id: I0d693f2958a33b5a093728dccf063e2114c7406b

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: dev/experimental
Review: https://review.openstack.org/136453

[OpenStack Infra (hudson-openstack)]

Change abandoned by Xicheng Chang (<email address hidden>) on branch: dev/experimental
Review: https://review.openstack.org/136453
Reason: seems impossible to pass ci. abandon it for now.

[OpenStack Infra (hudson-openstack)]

Change abandoned by Shuo Yang (<email address hidden>) on branch: master
Review: https://review.openstack.org/189386
Reason: obviously incorrect.

[OpenStack Infra (hudson-openstack)]

Change abandoned by Shuo Yang (<email address hidden>) on branch: master
Review: https://review.openstack.org/189401
Reason: still not the right merge strategy.