Re: [openstack-dev] [Fuel][Agent] Moving Fuel Agent to a separate repo

2015-01-27 Thread Vladimir Kozhukalov
Mike,

You are absolutely right about our current priorities for 6.1 and this
thread is not about immediate action.

But just to be fair, moving Fuel Client to a separate repo was a priori
much more complicated procedure because it is tested together with nailgun.
For Fuel Agent we just need to create a separate repo (30 minutes) and make
changes for fuel-main (30 minutes). There is nothing to worry about because
it is completely independent. And taking into account our recent activities
about integrating it with Ironic I would say there is no reason to postpone
this until 7.0

Alexander,

As for fuel_agent_ci, it is not used currently on the regular basis, so I
think we need it to be moved to the same separate repo and then one day in
the future maybe we will use it for functional testing.



Vladimir Kozhukalov

On Tue, Jan 27, 2015 at 1:39 AM, Roman Prykhodchenko m...@romcheg.me wrote:

 I think the idea is not to work on it right at this moment but to accept
 the general idea of fuel-agent being moved somewhere it can be alone. I’m
 not sure there is one single approach for separating a component from the
 common repository because each of them has their own use-cases and
 requirements so for every single one of them there is a need to do the same
 job as we’ve done for Fuel Client.

 Thar said I’d like to note that only by having a clear specification of
 how work- data- and test-flaws have to be changed after the component is
 put to its own repository it will be possible to judge on the time frame
 and the number of resources required to accomplish this task.


 - romcheg

  26 січ. 2015 о 20:39 Mike Scherbakov mscherba...@mirantis.com
 написав(ла):
 
  -1 to make changes now
  +1 to Alexandra
 
  Let's finish fuel-client first. Also, it is about prioritization. We
 have many things to be resolved in 6.1 (e.g. package the rest of the stuff
 which not yet packaged into RPM/DEB; split repos openstack/fuel/linux,
 etc.), and fuel agent in particular has pretty low priority to me in 6.1.
 
  In examples I have provided, which are essential for 6.1, we are
 experiencing lack of hands. Let's see if we can focus our work on those
 items and many other essential things, and come to this question later.
 
  On Mon, Jan 26, 2015 at 10:29 PM, Aleksandra Fedorova 
 afedor...@mirantis.com wrote:
  It seems that we have general agreement about the idea, but to make it
 happen we need much more detailed proposal.
 
  Even with python-fuelclient it is not quite clear right now, which
 version of nailgun should be used to test it, and the opposite: which
 version of fuelclient we have to use in iso builds. We also don't handle it
 in the build system very well right now, as we use git hashes, and not
 fixed versions, or packages.
 
  Maybe we should complete the python-fuelclient transformation first and
 see how it is going to work for us?
 
  On Jan 26, 2015 8:59 PM, Roman Prykhodchenko m...@romcheg.me wrote:
  Vladimir,
 
  As a fuel-separatist I give this initiative a big +1 because of the
 following advantages I can see:
 
   - Git is designed for keeping smaller single-compoent repos, keeping
 everything to one repo is a discouraged pattern
   - Having a separate -core group that will only contain active core
 reviewers for fuel-agent project so getting core-reviews will be easier.
   - It makes possible to re-use some of the existing jobs in OpenStack CI
   - Making independent releases becomes possible
 
  AFAIK fuel-agent is positioned as an independent provisioning tool which
 will not be exclusively used by Fuel. There is a work in progress to
 integrate it with Ironic. Integrating it to any other provisioning system
 should also be possible then. From that perspective putting it into its own
 repo also brings the following advantages:
 
   - Connecting 3rd party CI will be possible
   - Getting involved for the new folks will be much easier
 
 
  - romcheg
 
 
   26 січ. 2015 о 17:42 Vladimir Kozhukalov vkozhuka...@mirantis.com
 написав(ла):
  
   Fuelers,
  
   As most of you might know we have a bunch of projects inside fuel-web
 repo which are not directly related to Fuel Web application. Some of them
 are tested together and it seemed we could end up with a set of
 incompatibility issues if we separated them and stopped tracking their
 versions on the git level (instead of release level).
  
   Recent activities about separating Fuel Client from Nailgun (api) make
 me think we are mature enough to move all other not related project out of
 fuel-web repo and bring them together not earlier than on the stage of
 system/functional testing.
  
   Next step would be moving out Fuel Agent project. The reason is that
 it is independent and potentially could be used even out of Fuel because
 its data parsing mechanism is implemented so as to be agnostic to the data
 format. Some people could be potentially interested in using it
 independently with their own data format. It is tested together with other
 

Re: [openstack-dev] [Fuel][Agent] Moving Fuel Agent to a separate repo

2015-01-26 Thread Mike Scherbakov
-1 to make changes now
+1 to Alexandra

Let's finish fuel-client first. Also, it is about prioritization. We have
many things to be resolved in 6.1 (e.g. package the rest of the stuff which
not yet packaged into RPM/DEB; split repos openstack/fuel/linux, etc.), and
fuel agent in particular has pretty low priority to me in 6.1.

In examples I have provided, which are essential for 6.1, we are
experiencing lack of hands. Let's see if we can focus our work on those
items and many other essential things, and come to this question later.

On Mon, Jan 26, 2015 at 10:29 PM, Aleksandra Fedorova 
afedor...@mirantis.com wrote:

 It seems that we have general agreement about the idea, but to make it
 happen we need much more detailed proposal.

 Even with python-fuelclient it is not quite clear right now, which version
 of nailgun should be used to test it, and the opposite: which version of
 fuelclient we have to use in iso builds. We also don't handle it in the
 build system very well right now, as we use git hashes, and not fixed
 versions, or packages.

 Maybe we should complete the python-fuelclient transformation first and
 see how it is going to work for us?
 On Jan 26, 2015 8:59 PM, Roman Prykhodchenko m...@romcheg.me wrote:

 Vladimir,

 As a fuel-separatist I give this initiative a big +1 because of the
 following advantages I can see:

  - Git is designed for keeping smaller single-compoent repos, keeping
 everything to one repo is a discouraged pattern
  - Having a separate -core group that will only contain active core
 reviewers for fuel-agent project so getting core-reviews will be easier.
  - It makes possible to re-use some of the existing jobs in OpenStack CI
  - Making independent releases becomes possible

 AFAIK fuel-agent is positioned as an independent provisioning tool which
 will not be exclusively used by Fuel. There is a work in progress to
 integrate it with Ironic. Integrating it to any other provisioning system
 should also be possible then. From that perspective putting it into its own
 repo also brings the following advantages:

  - Connecting 3rd party CI will be possible
  - Getting involved for the new folks will be much easier


 - romcheg


  26 січ. 2015 о 17:42 Vladimir Kozhukalov vkozhuka...@mirantis.com
 написав(ла):
 
  Fuelers,
 
  As most of you might know we have a bunch of projects inside fuel-web
 repo which are not directly related to Fuel Web application. Some of them
 are tested together and it seemed we could end up with a set of
 incompatibility issues if we separated them and stopped tracking their
 versions on the git level (instead of release level).
 
  Recent activities about separating Fuel Client from Nailgun (api) make
 me think we are mature enough to move all other not related project out of
 fuel-web repo and bring them together not earlier than on the stage of
 system/functional testing.
 
  Next step would be moving out Fuel Agent project. The reason is that it
 is independent and potentially could be used even out of Fuel because its
 data parsing mechanism is implemented so as to be agnostic to the data
 format. Some people could be potentially interested in using it
 independently with their own data format. It is tested together with other
 Fuel components during system testing only.
 
 
 
  Vladimir Kozhukalov
 
 __
  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 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




-- 
Mike Scherbakov
#mihgen
__
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] [Fuel][Agent] Moving Fuel Agent to a separate repo

2015-01-26 Thread Aleksandra Fedorova
It seems that we have general agreement about the idea, but to make it
happen we need much more detailed proposal.

Even with python-fuelclient it is not quite clear right now, which version
of nailgun should be used to test it, and the opposite: which version of
fuelclient we have to use in iso builds. We also don't handle it in the
build system very well right now, as we use git hashes, and not fixed
versions, or packages.

Maybe we should complete the python-fuelclient transformation first and see
how it is going to work for us?
On Jan 26, 2015 8:59 PM, Roman Prykhodchenko m...@romcheg.me wrote:

 Vladimir,

 As a fuel-separatist I give this initiative a big +1 because of the
 following advantages I can see:

  - Git is designed for keeping smaller single-compoent repos, keeping
 everything to one repo is a discouraged pattern
  - Having a separate -core group that will only contain active core
 reviewers for fuel-agent project so getting core-reviews will be easier.
  - It makes possible to re-use some of the existing jobs in OpenStack CI
  - Making independent releases becomes possible

 AFAIK fuel-agent is positioned as an independent provisioning tool which
 will not be exclusively used by Fuel. There is a work in progress to
 integrate it with Ironic. Integrating it to any other provisioning system
 should also be possible then. From that perspective putting it into its own
 repo also brings the following advantages:

  - Connecting 3rd party CI will be possible
  - Getting involved for the new folks will be much easier


 - romcheg


  26 січ. 2015 о 17:42 Vladimir Kozhukalov vkozhuka...@mirantis.com
 написав(ла):
 
  Fuelers,
 
  As most of you might know we have a bunch of projects inside fuel-web
 repo which are not directly related to Fuel Web application. Some of them
 are tested together and it seemed we could end up with a set of
 incompatibility issues if we separated them and stopped tracking their
 versions on the git level (instead of release level).
 
  Recent activities about separating Fuel Client from Nailgun (api) make
 me think we are mature enough to move all other not related project out of
 fuel-web repo and bring them together not earlier than on the stage of
 system/functional testing.
 
  Next step would be moving out Fuel Agent project. The reason is that it
 is independent and potentially could be used even out of Fuel because its
 data parsing mechanism is implemented so as to be agnostic to the data
 format. Some people could be potentially interested in using it
 independently with their own data format. It is tested together with other
 Fuel components during system testing only.
 
 
 
  Vladimir Kozhukalov
 
 __
  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 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] [Fuel][Agent] Moving Fuel Agent to a separate repo

2015-01-26 Thread Sebastian Kalinowski
+1 I'm all for separating it.

2015-01-26 17:52 GMT+01:00 Alexander Gordeev agord...@mirantis.com:

 Hello Vladimir,

 totally +1 for separating Fuel Agent out of fuel-web.

 what will happen with fuel_agent_ci ?

 On Mon, Jan 26, 2015 at 7:42 PM, Vladimir Kozhukalov
 vkozhuka...@mirantis.com wrote:
  Fuelers,
 
  As most of you might know we have a bunch of projects inside fuel-web
 repo
  which are not directly related to Fuel Web application. Some of them are
  tested together and it seemed we could end up with a set of
 incompatibility
  issues if we separated them and stopped tracking their versions on the
 git
  level (instead of release level).
 
  Recent activities about separating Fuel Client from Nailgun (api) make me
  think we are mature enough to move all other not related project out of
  fuel-web repo and bring them together not earlier than on the stage of
  system/functional testing.
 
  Next step would be moving out Fuel Agent project. The reason is that it
 is
  independent and potentially could be used even out of Fuel because its
 data
  parsing mechanism is implemented so as to be agnostic to the data format.
  Some people could be potentially interested in using it independently
 with
  their own data format. It is tested together with other Fuel components
  during system testing only.
 
 
 
  Vladimir Kozhukalov
 
 
 __
  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 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] [Fuel][Agent] Moving Fuel Agent to a separate repo

2015-01-26 Thread Alexander Gordeev
Hello Vladimir,

totally +1 for separating Fuel Agent out of fuel-web.

what will happen with fuel_agent_ci ?

On Mon, Jan 26, 2015 at 7:42 PM, Vladimir Kozhukalov
vkozhuka...@mirantis.com wrote:
 Fuelers,

 As most of you might know we have a bunch of projects inside fuel-web repo
 which are not directly related to Fuel Web application. Some of them are
 tested together and it seemed we could end up with a set of  incompatibility
 issues if we separated them and stopped tracking their versions on the git
 level (instead of release level).

 Recent activities about separating Fuel Client from Nailgun (api) make me
 think we are mature enough to move all other not related project out of
 fuel-web repo and bring them together not earlier than on the stage of
 system/functional testing.

 Next step would be moving out Fuel Agent project. The reason is that it is
 independent and potentially could be used even out of Fuel because its data
 parsing mechanism is implemented so as to be agnostic to the data format.
 Some people could be potentially interested in using it independently with
 their own data format. It is tested together with other Fuel components
 during system testing only.



 Vladimir Kozhukalov

 __
 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] [Fuel][Agent] Moving Fuel Agent to a separate repo

2015-01-26 Thread Roman Prykhodchenko
Vladimir,

As a fuel-separatist I give this initiative a big +1 because of the following 
advantages I can see:

 - Git is designed for keeping smaller single-compoent repos, keeping 
everything to one repo is a discouraged pattern
 - Having a separate -core group that will only contain active core reviewers 
for fuel-agent project so getting core-reviews will be easier.
 - It makes possible to re-use some of the existing jobs in OpenStack CI
 - Making independent releases becomes possible

AFAIK fuel-agent is positioned as an independent provisioning tool which will 
not be exclusively used by Fuel. There is a work in progress to integrate it 
with Ironic. Integrating it to any other provisioning system should also be 
possible then. From that perspective putting it into its own repo also brings 
the following advantages:

 - Connecting 3rd party CI will be possible
 - Getting involved for the new folks will be much easier


- romcheg


 26 січ. 2015 о 17:42 Vladimir Kozhukalov vkozhuka...@mirantis.com 
 написав(ла):
 
 Fuelers,
 
 As most of you might know we have a bunch of projects inside fuel-web repo 
 which are not directly related to Fuel Web application. Some of them are 
 tested together and it seemed we could end up with a set of  incompatibility 
 issues if we separated them and stopped tracking their versions on the git 
 level (instead of release level).
 
 Recent activities about separating Fuel Client from Nailgun (api) make me 
 think we are mature enough to move all other not related project out of 
 fuel-web repo and bring them together not earlier than on the stage of 
 system/functional testing.
 
 Next step would be moving out Fuel Agent project. The reason is that it is 
 independent and potentially could be used even out of Fuel because its data 
 parsing mechanism is implemented so as to be agnostic to the data format. 
 Some people could be potentially interested in using it independently with 
 their own data format. It is tested together with other Fuel components 
 during system testing only.
 
 
 
 Vladimir Kozhukalov
 __
 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



signature.asc
Description: Message signed with OpenPGP using GPGMail
__
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] [Fuel][Agent] Moving Fuel Agent to a separate repo

2015-01-26 Thread Roman Prykhodchenko
I think the idea is not to work on it right at this moment but to accept the 
general idea of fuel-agent being moved somewhere it can be alone. I’m not sure 
there is one single approach for separating a component from the common 
repository because each of them has their own use-cases and requirements so for 
every single one of them there is a need to do the same job as we’ve done for 
Fuel Client.

Thar said I’d like to note that only by having a clear specification of how 
work- data- and test-flaws have to be changed after the component is put to its 
own repository it will be possible to judge on the time frame and the number of 
resources required to accomplish this task.


- romcheg

 26 січ. 2015 о 20:39 Mike Scherbakov mscherba...@mirantis.com написав(ла):
 
 -1 to make changes now
 +1 to Alexandra
 
 Let's finish fuel-client first. Also, it is about prioritization. We have 
 many things to be resolved in 6.1 (e.g. package the rest of the stuff which 
 not yet packaged into RPM/DEB; split repos openstack/fuel/linux, etc.), and 
 fuel agent in particular has pretty low priority to me in 6.1.
 
 In examples I have provided, which are essential for 6.1, we are experiencing 
 lack of hands. Let's see if we can focus our work on those items and many 
 other essential things, and come to this question later.
 
 On Mon, Jan 26, 2015 at 10:29 PM, Aleksandra Fedorova 
 afedor...@mirantis.com wrote:
 It seems that we have general agreement about the idea, but to make it happen 
 we need much more detailed proposal.
 
 Even with python-fuelclient it is not quite clear right now, which version of 
 nailgun should be used to test it, and the opposite: which version of 
 fuelclient we have to use in iso builds. We also don't handle it in the build 
 system very well right now, as we use git hashes, and not fixed versions, or 
 packages.
 
 Maybe we should complete the python-fuelclient transformation first and see 
 how it is going to work for us?
 
 On Jan 26, 2015 8:59 PM, Roman Prykhodchenko m...@romcheg.me wrote:
 Vladimir,
 
 As a fuel-separatist I give this initiative a big +1 because of the following 
 advantages I can see:
 
  - Git is designed for keeping smaller single-compoent repos, keeping 
 everything to one repo is a discouraged pattern
  - Having a separate -core group that will only contain active core reviewers 
 for fuel-agent project so getting core-reviews will be easier.
  - It makes possible to re-use some of the existing jobs in OpenStack CI
  - Making independent releases becomes possible
 
 AFAIK fuel-agent is positioned as an independent provisioning tool which will 
 not be exclusively used by Fuel. There is a work in progress to integrate it 
 with Ironic. Integrating it to any other provisioning system should also be 
 possible then. From that perspective putting it into its own repo also brings 
 the following advantages:
 
  - Connecting 3rd party CI will be possible
  - Getting involved for the new folks will be much easier
 
 
 - romcheg
 
 
  26 січ. 2015 о 17:42 Vladimir Kozhukalov vkozhuka...@mirantis.com 
  написав(ла):
 
  Fuelers,
 
  As most of you might know we have a bunch of projects inside fuel-web repo 
  which are not directly related to Fuel Web application. Some of them are 
  tested together and it seemed we could end up with a set of  
  incompatibility issues if we separated them and stopped tracking their 
  versions on the git level (instead of release level).
 
  Recent activities about separating Fuel Client from Nailgun (api) make me 
  think we are mature enough to move all other not related project out of 
  fuel-web repo and bring them together not earlier than on the stage of 
  system/functional testing.
 
  Next step would be moving out Fuel Agent project. The reason is that it is 
  independent and potentially could be used even out of Fuel because its data 
  parsing mechanism is implemented so as to be agnostic to the data format. 
  Some people could be potentially interested in using it independently with 
  their own data format. It is tested together with other Fuel components 
  during system testing only.
 
 
 
  Vladimir Kozhukalov
  __
  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 Development Mailing List (not for usage questions)
 Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe