rebuild fails with ilo* drivers

[Ramakrishnan G (rameshg87)]

Instance rebuild fails with iscsi_ilo and agent_ilo drivers. After the node is set to reboot with the deploy ramdisk, the node boots from the instance again and never boots to deploy ramdisk. Eventually the rebuild fails with the timeout on staying at wait-call-back state.

[Ramakrishnan G (rameshg87) (rameshg87)]

This is because of ilo_boot_iso (if it exists in instance_info) is always attached to the node while it is powered on . On a rebuild, the ilo_boot_iso exists on the node (which is the boot iso of the previous deloyed image). Hence when the deploy driver tries to power o n the node after attaching deploy_iso - the instance boot_iso gets re-attached again and causes the failure.

https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/ilo/power.py#L64-L66

We should be attaching the boot_iso only if the instance is active.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/167317

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/167317
Committed: https://git.openstack.org/cgit/openstack/ironic/commit/?id=be8e17f09543716283e0885d550ba39fbf219b4b
Submitter: Jenkins
Branch: master

commit be8e17f09543716283e0885d550ba39fbf219b4b
Author: Ramakrishnan G <email address hidden>
Date: Tue Mar 24 16:38:01 2015 +0000

    Attach ilo_boot_iso only if node is active

    Currently iLO drivers attach instance_info/ilo_boot_iso
    (if it exists) to the node everytime it powers on the node.
    This commit changes the ilo_boot_iso to be attached
    only if the node is in active state.

    This is for scenarios like instance rebuild where
    instance_info/ilo_boot_iso might exist, but it shouldn't
    be attached while powering on the node (during rebuild,
    the node should boot from deploy_iso again to deploy the
    new image).

    Closes-bug: 1435959
    Change-Id: I0fb0b0c016e52f88807d13c45c03127b7efdf4fb