<div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote">On Wed, Feb 17, 2016 at 5:55 AM, Sean Dague <span dir="ltr"><<a href="mailto:sean@dague.net" target="_blank">sean@dague.net</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5">On 02/17/2016 08:42 AM, Doug Hellmann wrote:<br>
> Excerpts from Victor Stinner's message of 2016-02-17 14:14:18 +0100:<br>
>> Le 17/02/2016 13:43, Henry Gessau a écrit :<br>
>>> And it looks like eventlet 0.18.3 breaks neutron:<br>
>>> <a href="https://bugs.launchpad.net/neutron/+bug/1546506" rel="noreferrer" target="_blank">https://bugs.launchpad.net/neutron/+bug/1546506</a><br>
>><br>
>> 2 releases, 2 regressions in OpenStack. Should we cap eventlet version?<br>
>> The requirement bot can produce patches to update eventlet, patches<br>
>> which would run integration tests using Nova, Keystone, Neutron on the<br>
>> new eventlet version.<br>
>><br>
>> eventlet 0.18.2 broke OpenStack Keystone and OpenStack Nova<br>
>> <a href="https://github.com/eventlet/eventlet/issues/296" rel="noreferrer" target="_blank">https://github.com/eventlet/eventlet/issues/296</a><br>
>> <a href="https://github.com/eventlet/eventlet/issues/299" rel="noreferrer" target="_blank">https://github.com/eventlet/eventlet/issues/299</a><br>
>> <a href="https://review.openstack.org/#/c/278147/" rel="noreferrer" target="_blank">https://review.openstack.org/#/c/278147/</a><br>
>> <a href="https://bugs.launchpad.net/nova/+bug/1544801" rel="noreferrer" target="_blank">https://bugs.launchpad.net/nova/+bug/1544801</a><br>
>><br>
>> eventlet 0.18.3 broke OpenStack Neutron<br>
>> <a href="https://github.com/eventlet/eventlet/issues/301" rel="noreferrer" target="_blank">https://github.com/eventlet/eventlet/issues/301</a><br>
>> <a href="https://bugs.launchpad.net/neutron/+bug/1546506" rel="noreferrer" target="_blank">https://bugs.launchpad.net/neutron/+bug/1546506</a><br>
>><br>
>> FYI eventlet 0.18.0 broke WSGI servers:<br>
>> <a href="https://github.com/eventlet/eventlet/issues/295" rel="noreferrer" target="_blank">https://github.com/eventlet/eventlet/issues/295</a><br>
>><br>
>> It was followed quickly by eventlet 0.18.2 to fix this issue.<br>
>><br>
>> Sadly, it looks like bugfix releases of eventlet don't include a single<br>
>> bugfix, but include also other changes. For example, 0.18.3 fixed the<br>
>> bug #296 but introduced "wsgi: TCP_NODELAY enabled by default" optimization.<br>
>><br>
>> IMHO the problem is not the release manager of eventlet, but more the<br>
>> lack of tests on eventlet, especially on OpenStack services.<br>
>><br>
>> Current "Continious Delivery"-like with gates do detect bugs, yeah, but<br>
>> also block a lot of developers when the gates are broken. It doesn't<br>
>> seem trivial to investigate and fix eventlet issues.<br>
>><br>
>> Victor<br>
>><br>
><br>
> Whether we cap or not, we should exclude the known broken versions.<br>
> It looks like getting back to a good version will also require<br>
> lowering the minimum version we support, since we have >=0.18.2<br>
> now.<br>
><br>
> What was the last version of eventlet known to work?<br>
<br>
</div></div>0.18.2 works. On the Nova side we had a failure around unit tests which<br>
was quite synthetic that we fixed. I don' know what the keystone issue<br>
turned out to be.<br>
<span class="HOEnZb"></span></blockquote><div><br></div><div>I believe the keystone issue was a test specific issue, not a runtime issue. We disabled the test.<br></div><div>--Morgan <br></div></div></div></div>