Comment 5 for bug 1605742

Revision history for this message
Markus Zoeller (markus_z) (mzoeller) wrote :

The Nova periodic stable mitaka test job installs paramiko==1.16.0
    http://logs.openstack.org/periodic-stable/periodic-nova-python27-db-mitaka/9d14b47/console.html#_2016-08-02_06_16_11_180799
This is the upper-constraint since Nov 2015
    https://github.com/openstack/requirements/commit/6bb1357b2a4347a29ca5911499b86de71b92fdc8#diff-0bdd949ed8a7fdd4f95240bd951779c8R212
This works fine.

The openstack-ansible project used paramiko>=1.16.0 and installed 2.0.1
    http://logs.openstack.org/09/342309/8/check/gate-openstack-ansible-dsvm-commit/224f9c0/console.html.gz#_2016-07-22_17_21_57_050330
    https://github.com/openstack/openstack-ansible/commit/9de9f4def3a731563da5778546e7f9f73e2c4214#diff-b4ef698db8ca845e5845c4618278f29aR9
openstack-ansible removed the requirement "paramiko>=1.6.0" later
    https://github.com/openstack/openstack-ansible/commit/b15363c#diff-b4ef698db8ca845e5845c4618278f29aL3
This however is only in openstack-ansible Newton and not in Mitaka:
    https://github.com/openstack/openstack-ansible/blob/stable/mitaka/requirements.txt#L9

Based on ^ I believe this is an issue of the openstack-ansible project which
doesn't cap the upper-constraint of paramiko in its stable/mitaka branch.
I don't see the need to backport anything to Nova's stable/mitaka branch.
I leave this as "incomplete" to get a second pair of eyes from auggy.