Nodes go offline after reboot when using image-based provisioning

[Ryan Moe]

After reboot only eth0 comes up on the node and there is no connectivity to the master node. All of the interface configuration files exist and are correct but the other interfaces never come up.

build_number: "16"
build_id: "2014-12-01_04-35-45"

[Christopher Aedo (docaedo)]

This bug does not occur if you deploy a centos environment. Using image-based deployment, I deployed a single controller and single compute node on a linux server (nodes running under KVM). After I validated the environment is working fine by launching a VM and ensuring network connectivity between the nodes, I rebooted the compute node. After reboot, the compute node WAS able to reach the controller, and had the network was working fine. I also tried rebooting the controller node and after reboot there were no problems with the network.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/139129

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/139129
Committed: https://git.openstack.org/cgit/stackforge/fuel-web/commit/?id=acb4b2a2ecce14bf30eb36c17227aa6c87800da8
Submitter: Jenkins
Branch: master

commit acb4b2a2ecce14bf30eb36c17227aa6c87800da8
Author: Alexander Gordeev <email address hidden>
Date: Thu Dec 4 20:37:06 2014 +0300

    Fix cloud-init metadata re-applying on every single boot

    The actual reason is that cloud-init re-applies its metadata again
    and again. Wiping out all the metadata from datasource resolves
    the issue.

    Change-Id: I0d1912da504df959e36c316ed09effaf1bcb682f
    Closes-Bug: #1398207

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/6.0
Review: https://review.openstack.org/149237

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/149237
Committed: https://git.openstack.org/cgit/stackforge/fuel-web/commit/?id=393195b11567047acee1f5ab808c531655cec79b
Submitter: Jenkins
Branch: stable/6.0

commit 393195b11567047acee1f5ab808c531655cec79b
Author: Alexander Gordeev <email address hidden>
Date: Thu Dec 4 20:37:06 2014 +0300

    Fix cloud-init metadata re-applying on every single boot

    The actual reason is that cloud-init re-applies its metadata again
    and again. Wiping out all the metadata from datasource resolves
    the issue.

    Change-Id: I0d1912da504df959e36c316ed09effaf1bcb682f
    Closes-Bug: #1398207

[Dennis Dmitriev (ddmitriev)]

Released on 6.1:
ISO version: {u'build_id': u'2015-01-22_22-54-45', u'ostf_sha': u'95054878eff472eb6a0d9d185ce65e3105dde632', u'build_number': u'75', u'auth_required': True, u'nailgun_sha': u'b677b29899f9e49532a959948d0ba68a83e41536', u'production': u'docker', u'api': u'1.0', u'fuelmain_sha': u'', u'astute_sha': u'f640780d1c7ebe89e4f343736500945525799ee0', u'feature_groups': [u'mirantis'], u'release': u'6.1', u'release_versions': {u'2014.2-6.1': {u'VERSION': {u'build_id': u'2015-01-22_22-54-45', u'ostf_sha': u'95054878eff472eb6a0d9d185ce65e3105dde632', u'build_number': u'75', u'api': u'1.0', u'nailgun_sha': u'b677b29899f9e49532a959948d0ba68a83e41536', u'production': u'docker', u'fuelmain_sha': u'', u'astute_sha': u'f640780d1c7ebe89e4f343736500945525799ee0', u'feature_groups': [u'mirantis'], u'release': u'6.1', u'fuellib_sha': u'673808c5bd6ff5560df75b74f7f798649d6a0040'}}}, u'fuellib_sha': u'673808c5bd6ff5560df75b74f7f798649d6a0040'}