<div dir="ltr"><div dir="ltr">Hi,<br></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Dec 13, 2019 at 3:14 PM Julia Kreger <<a href="mailto:juliaashleykreger@gmail.com">juliaashleykreger@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On Fri, Dec 13, 2019 at 4:24 AM Dmitry Tantsur <<a href="mailto:dtantsur@redhat.com" target="_blank">dtantsur@redhat.com</a>> wrote:<br>
><br>
> Hi,<br>
><br>
> On Thu, Dec 12, 2019 at 10:30 PM Julia Kreger <<a href="mailto:juliaashleykreger@gmail.com" target="_blank">juliaashleykreger@gmail.com</a>> wrote:<br>
>><br>
>> Greetings fellow ironicans and stackers!<br>
>><br>
>> I'd like some thoughts on the `ibmc` driver[1] that was added early<br>
>> this year. Specifically it merged into ironic with an operating<br>
>> Third-Party CI, however soon after a series of unfortunate events<br>
>> occurred where the Third-Party CI stopped reporting sometime in early<br>
>> May, and shortly their after restrictions were put into place[2]. With<br>
>> that being said, I feel like we should go ahead and remove this driver<br>
>> since it does not seem to be maintained nor tested.<br>
><br>
><br>
> Do you mean deprecate and remove in V or just straight remove? I'd follow the regular procedure. Otherwise, it seems that we have no other options :(<br>
<br>
Either... Kind of why I emailed the mailing list. :) I'd lean for<br>
deprecated and take that route, but maybe someone has interest in<br>
keeping it working<br>
<br>
><br>
> An interesting topic to discuss is how to avoid such situations when a driver only lives 1-2 cycles..<br>
<br>
Knowing a preference may exist for us to avoid the word support, I<br>
feel like it was a preview that didn't pan out. I don't feel that fear<br>
of this happening again should change any of our patterns moving<br>
forward though, since that would be the exact opposite of what is<br>
needed to encourage adoption.<br></blockquote><div><br></div><div>We could make all new hardware types starting in some sort of "tech preview" state, so enabling them would require something like enable_tech_preview_hardware_types=True in ironic.conf. It would be similar to the staging area in Linux. The reason is my (completely ungrounded) assumption that if a driver has survived 2 cycles, it has more chances of surviving longer.<br></div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<br>
We also have some drivers that do the minimal, some that are engaged,<br>
and some that are kind of like they have a warp drive attached. I'd<br>
love for us to merge the coffee driver as a POC interface to help<br>
people understand mechanics, fwiw. Not that we can have a coffee<br>
machine wired into CI.<br>
"launched" "has_impulse" "has_warp_drive"?<br></blockquote><div><br></div><div>We have ironic-staging-drivers specifically for this ;)</div><div><br></div><div>Dmitry<br></div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<br>
-Julia<br>
<br>
><br>
> Dmitry<br>
><br>
>><br>
>><br>
>> -Julia<br>
>><br>
>><br>
>> [1]: <a href="https://review.opendev.org/#/c/639288/" rel="noreferrer" target="_blank">https://review.opendev.org/#/c/639288/</a><br>
>> [2]: <a href="https://www.federalregister.gov/documents/2019/05/21/2019-10616/addition-of-entities-to-the-entity-list" rel="noreferrer" target="_blank">https://www.federalregister.gov/documents/2019/05/21/2019-10616/addition-of-entities-to-the-entity-list</a><br>
>><br>
<br>
</blockquote></div></div>