Open Stack

Sent from my iPad

Begin forwarded message:

From: Rochelle Grober <rochelle.grober at huawei.com<mailto:rochelle.grober at huawei.com>>
Date: March 24, 2015 at 19:14:48 PDT
To: Shamail <itzshamail at gmail.com<mailto:itzshamail at gmail.com>>
Subject: RE: [Product] Prevent another "Cinder drivers fiasco"

The list of removed drivers can be found here:
https://review.openstack.org/#/q/status:merged+project:openstack/cinder+branch:master+topic:cinder-driver-removals,n,z

Of these, the removals in contention are discussed in these threads:
Zadara (Oracle):  http://lists.openstack.org/pipermail/openstack-dev/2015-March/059501.html
Hauwei:  http://lists.openstack.org/pipermail/openstack-dev/2015-March/059451.html http://lists.openstack.org/pipermail/openstack-dev/2015-March/059453.html and description of the Huawei issue here: http://lists.openstack.org/pipermail/openstack-dev/2015-March/059571.html
NetApp FC:  http://lists.openstack.org/pipermail/openstack-dev/2015-March/059477.html with explanation here:  http://lists.openstack.org/pipermail/openstack-dev/2015-March/059718.html

Of the other removed drivers, the only comments I saw on them was for the Microsoft driver and the developer stated that they had issues getting an external address for posting of results. And they stated they would be in L as soon as it opened. I can't find that email.

>From the Huawei perspective, there were numerous issues ranging from language and culture to availability of Cinder devs on IRC during the China workday.  There was also a big issue in that the responsible developer within Huawei had changed, but the places to notify the Cinder PTL of this were multiple, of differing ages and information, and very much in error.  To the Cinder PTL's credit, he reached out to any and every Huawei employee he met to try and get the information to the responsible engineer.  Unfortunately, that was finally successful just after the Philly Ops meeting and by then, the schedules all around were set.

I think all of the vendor engineers who asked for FFEs have been working diligently at least since Paris to try to meet the requirements.  No, it's not January or even Atlanta, but the means of communication in OpenStack do not  meld well with the means within large corporations.  Especially corporations who are used to heavy attacks on their internal networks from various nefarious individuals and organizations.

Anyway, here are the email threads.  How can Product WG make this better for next time and for the next time new processes get rolled out to vendors?

--Rocky



-----Original Message-----
From: Shamail [mailto:itzshamail at gmail.com]
Sent: Tuesday, March 24, 2015 18:36
To: Stefano Maffulli
Cc: product-wg at lists.openstack.org<mailto:product-wg at lists.openstack.org>
Subject: Re: [Product] Prevent another "Cinder drivers fiasco"

Hi Stefano,

I have caught up on the issue at hand.  First off, I will say that we were not impacted by this change BUT that has a significant amount to do with how our Cinder teams laid out (internally) the critical nature of adhering to 3rd party CI for driver contributions after Atlanta.

The communication seemed to be there; I am not certain if the "cut-off" date/release was well-defined but this could simply be  due to the fact that I was partially removed from the process and didn't notice it.

I am wondering if the bigger issue is the support model development teams have inside their own organizations?  For example, I wonder if the developers raised the "alarm" internally but could not find the right process (or group to work with) to get a publicly accessible server deployed within their organization or other similar requirements for 3rd party CI?

If that is indeed the issue then we might be able to help by laying out the requirements, ramifications, and deadlines openly (via the same mechanism as long-term roadmap) and also sharing these items within our own companies.

My personal opinion is that the issue may have been internal propagation (and ability to take action) of an external (community) requirement.  I could be completely off base, but I am glad you raised this topic here.  We should try to help prevent a repeat in the future, however we best can as a group.

I, too, would love to hear from someone that was impacted to determine root-cause and how we can prevent this.

Thanks,
Shamail





On Mar 24, 2015, at 5:47 PM, Stefano Maffulli <stefano at openstack.org<mailto:stefano at openstack.org>> wrote:

TL;DR If you have drivers in Cinder that are being removed from Kilo
please let me know your side of the story.

You may have noticed that a few drivers have been removed from Cinder
tree and there are very little chances that they will be included back
for Kilo. Also, the documentation team has removed the documentation of
such drivers from their tree as a consequence.

Cinder's PTL and others have expressed their view on the mailing list
and I think this time it's not just about lack of communication and
clarity from developer's side. Something else, much deeper is going off
the rails and I think this group can help to figure out what that is.

I'm sure some here have been impacted by this decision: let's talk about
how to avoid to get to this point in the future.

/stef


_______________________________________________
Product-wg mailing list
Product-wg at lists.openstack.org<mailto:Product-wg at lists.openstack.org>
http://lists.openstack.org/cgi-bin/mailman/listinfo/product-wg

_______________________________________________
Product-wg mailing list
Product-wg at lists.openstack.org<mailto:Product-wg at lists.openstack.org>
http://lists.openstack.org/cgi-bin/mailman/listinfo/product-wg