The behaviour of python gethostname differs from Linux to Windows

[Luis Fernández Álvarez]

The standard behaviour of the 'gethostname' function in Python differs from Linux to Windows. A common Linux configuration will return the fully qualified name, while a Windows one will return only the host name.

This problem leads to an inconsistent node naming in deployments that mix windows and linux nodes.

To make things more homogeneus among windows and linux compute nodes, it is proposed to use 'getfqdn' as default function instead of 'gethostname'. This function is more predictable in all cases.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/13636

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/13636
Committed: http://github.com/openstack/nova/commit/5dd1553cca7f7e62eebce75e1d936fc211b239ec
Submitter: Jenkins
Branch: master

commit 5dd1553cca7f7e62eebce75e1d936fc211b239ec
Author: Luis Fernandez Alvarez <email address hidden>
Date: Tue Sep 25 17:33:59 2012 +0200

    Replaced default hostname function from gethostname to getfqdn

    Fixes bug 1055503

    The standard behaviour of the 'gethostname' function in Python differs from
    Linux to Windows. A common Linux configuration returns the FQDN, while a
    Windows one returns only the host name.

    To resolve inconsistent node naming in deployments that mix windows and
    Linux, it is proposed to use 'getfqdn' as default function instead of
    'gethostname'. This is function is more predictable in all cases.

    Change-Id: I3164d9a36df2b8484bbf9a57879c31fa0e342503

[Michael Still (mikal)]

I have had to revert this because it broke grizzly upgrades. I'm going to work on another fix though.

(The upgrade breakage was because services ended up being registered twice, as well as the host column of the instances table being wrong).

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/24177

[Michael Still (mikal)]

Retargeting this because of risk and the existence of a work around (setting the host flag to something explicit).

[Russell Bryant (russellb)]

and now it's probably too late to mess with for havana ...

[Alessandro Pilotti (alexpilotti)]

Hi guys, should we put this bug on track for Icehouse RC1 or at this point directly for J1?

[Sean Dague (sdague)]

Marking as low as the preponderance of mixed os clusters is small.

[Bogdan Dobrelya (bogdando)]

Please note, that HA deployments with Pacemaker & Corosync rely on "crm_node -n" instead. And this one could differ from getfqdn as well.

[Sean Dague (sdague)]

I feel like this is firmly into the won't fix category. All the fixes so far have caused issues.