Open Stack

-----Original Message-----
From: gordon chung <gord at live.ca>
Reply: OpenStack Development Mailing List (not for usage questions) <openstack-dev at lists.openstack.org>
Date: June 14, 2016 at 06:51:08
To: openstack-dev at lists.openstack.org <openstack-dev at lists.openstack.org>
Subject:  Re: [openstack-dev] [ceilometer] [stable] Re: [Openstack-stable-maint] Stable check of openstack/ceilometer failed

> i don't know if anyone is looking at this -- i'm not sure where this test is even run. i usually  
> let mriedem yell at me but i take it he has bigger things on his plate now :)

Heh, I just looked through these quickly yesterday. I suspect Matt would have worked his way around to these.

> this seems like a pretty simple fix from the error output[1]. i guess my question is: should  
> the correct fix be to cap oslo.utils? based on the error, the issue seems to be total_seconds  
> method was removed. this was deprecated in Mitaka[2], so i don't think it should've been  
> removed from Liberty. as this is an easy fix, i'm pretty indifferent if we decide to fix  
> this rather than slow down progress. the original purpose of this method (based on commit  
> messsage) seems to be related to py2.6. i don't know if this is still an issue.

I wonder why more projects aren't seeing this in stable/liberty. Perhaps, ceilometer stable/liberty isn't using upper-constraints? I think oslo.utils 3.2.0 (https://github.com/openstack/requirements/blob/stable/liberty/upper-constraints.txt#L202) is low enough to avoid this if you're using constraints. (It looks as if the total_seconds removal was first released in 3.12.0 https://github.com/openstack/oslo.utils/commit/8f5e65cae3aaf8d0a89d16d8932c266151de44f7)

I'll let soemone else determine if the right answer is capping stable/liberty.

Cheers,
--  
Ian Cordasco