Re: [openstack-dev] [ironic] Removal of the ipminative / pyghmi driver

2017-03-10 Thread Pavlo Shchelokovskyy 
Hi all,

small note for future - we could re-evaluate possibility of testing this
driver on gates with VMs when qemu version with BMC support (plus
[python-]libvirt support for such functionality) is generally available in
distros.

In this case we could stop using virtualbmc and use qemu natively. Not sure
how valuable such testing would be, but it won't for sure be pyghmi talking
to pyghmi.

Cheers,

Dr. Pavlo Shchelokovskyy
Senior Software Engineer
Mirantis Inc
www.mirantis.com

On Thu, Mar 9, 2017 at 8:40 PM, Dmitry Tantsur  wrote:

> On 03/09/2017 07:19 PM, Jay Faulkner wrote:
>
>> Hi all,
>>
>> The ipminative driver Is currently an anomaly in ironic’s tree, despite
>> the driver being initially deprecated in Newton[1], and   our desire to
>> drop them reiterated on the mailing list in December[2], it was has not
>> been removed from the tree prior to the Ocata release.
>>
>> At the PTG the ironic team had a short discussion about the ipminative
>> (aka pyghmi) driver — the conclusion was that unless third party CI was run
>> against the driver, we would be forced to follow through on the deprecation
>> and remove it. Testing in upstream CI, against VirtualBMC, was mostly
>> rejected due to both the ipminative driver and virtualbmc using the same
>> python ipmi library (pyghmi), and therefore not being a valid test case.
>> Additionally, further adding urgency to the removal, several active ironic
>> contributors who have tested ipminative drivers in real-world environments
>> have reported them as unstable.
>>
>> The promise of a native python driver to talk to ipmi in ironic is great,
>> but without proper testing and stability, keeping it in-tree does more harm
>> to ironic users than good — in fact, there’s very little indication to a
>> deployer using ironic that the driver may not work stably.
>>
>> Therefore, I’m giving the mailing list a two week warning — unless
>> volunteers come willing to run third party CI against the ipminative
>> drivers in the next two weeks, I will be submitting a patch to remove them
>> entirely from the tree. The driver could then be moved into
>> ironic-staging-drivers by any interested contributors.
>>
>
> Thanks for writing it Jay. Indeed, we spent too much time waiting for
> someone to overtake this driver. I'm very much in favor of removing it, if
> somebody does not step up *right now*.
>
>
>
>> -
>> Jay Faulkner
>> OSIC
>>
>> Related-bug: https://bugs.launchpad.net/ironic/+bug/1671532
>>
>> [1] https://docs.openstack.org/releasenotes/ironic/newton.html
>> [2] http://lists.openstack.org/pipermail/openstack-dev/2016-Dece
>> mber/108666.html
>> 
>> __
>> OpenStack Development Mailing List (not for usage questions)
>> Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscrib
>> e
>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>>
>>
>
> __
> OpenStack Development Mailing List (not for usage questions)
> Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>
__
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

Re: [openstack-dev] [ironic] Removal of the ipminative / pyghmi driver

2017-03-09 Thread Dmitry Tantsur 

On 03/09/2017 07:19 PM, Jay Faulkner wrote:

Hi all,

The ipminative driver Is currently an anomaly in ironic’s tree, despite the 
driver being initially deprecated in Newton[1], and   our desire to drop them 
reiterated on the mailing list in December[2], it was has not been removed from 
the tree prior to the Ocata release.

At the PTG the ironic team had a short discussion about the ipminative (aka 
pyghmi) driver — the conclusion was that unless third party CI was run against 
the driver, we would be forced to follow through on the deprecation and remove 
it. Testing in upstream CI, against VirtualBMC, was mostly rejected due to both 
the ipminative driver and virtualbmc using the same python ipmi library 
(pyghmi), and therefore not being a valid test case. Additionally, further 
adding urgency to the removal, several active ironic contributors who have 
tested ipminative drivers in real-world environments have reported them as 
unstable.

The promise of a native python driver to talk to ipmi in ironic is great, but 
without proper testing and stability, keeping it in-tree does more harm to 
ironic users than good — in fact, there’s very little indication to a deployer 
using ironic that the driver may not work stably.

Therefore, I’m giving the mailing list a two week warning — unless volunteers 
come willing to run third party CI against the ipminative drivers in the next 
two weeks, I will be submitting a patch to remove them entirely from the tree. 
The driver could then be moved into ironic-staging-drivers by any interested 
contributors.


Thanks for writing it Jay. Indeed, we spent too much time waiting for someone to 
overtake this driver. I'm very much in favor of removing it, if somebody does 
not step up *right now*.




-
Jay Faulkner
OSIC

Related-bug: https://bugs.launchpad.net/ironic/+bug/1671532

[1] https://docs.openstack.org/releasenotes/ironic/newton.html
[2] http://lists.openstack.org/pipermail/openstack-dev/2016-December/108666.html
__
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev




__
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

[openstack-dev] [ironic] Removal of the ipminative / pyghmi driver

2017-03-09 Thread Jay Faulkner 
Hi all,

The ipminative driver Is currently an anomaly in ironic’s tree, despite the 
driver being initially deprecated in Newton[1], and   our desire to drop them 
reiterated on the mailing list in December[2], it was has not been removed from 
the tree prior to the Ocata release.

At the PTG the ironic team had a short discussion about the ipminative (aka 
pyghmi) driver — the conclusion was that unless third party CI was run against 
the driver, we would be forced to follow through on the deprecation and remove 
it. Testing in upstream CI, against VirtualBMC, was mostly rejected due to both 
the ipminative driver and virtualbmc using the same python ipmi library 
(pyghmi), and therefore not being a valid test case. Additionally, further 
adding urgency to the removal, several active ironic contributors who have 
tested ipminative drivers in real-world environments have reported them as 
unstable.

The promise of a native python driver to talk to ipmi in ironic is great, but 
without proper testing and stability, keeping it in-tree does more harm to 
ironic users than good — in fact, there’s very little indication to a deployer 
using ironic that the driver may not work stably.

Therefore, I’m giving the mailing list a two week warning — unless volunteers 
come willing to run third party CI against the ipminative drivers in the next 
two weeks, I will be submitting a patch to remove them entirely from the tree. 
The driver could then be moved into ironic-staging-drivers by any interested 
contributors.

-
Jay Faulkner
OSIC

Related-bug: https://bugs.launchpad.net/ironic/+bug/1671532

[1] https://docs.openstack.org/releasenotes/ironic/newton.html
[2] http://lists.openstack.org/pipermail/openstack-dev/2016-December/108666.html
__
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev