fuel-dev team mailing list archive

[Andrew Woodward <xarses@xxxxxxxxx>]

> > The ability to add additional online/custom repositories to a cluster would enable a lot of additional customisation for users.

> But users already have such ability - any deployment engineer can handle it in few minutes by adding needed additional apt/yum repos and installing needed packages on OS nodes, for example, via mcollective. I'm not sure it's worth the effort.

This is besides the point. My expectation here is that this will allow
an easily way to ensure that the deployment is repeatable from given
configuration data. "A person can do this" is error prone,
un-repeatable, and wont scale.


On Fri, Jun 13, 2014 at 1:56 AM, Aleksandr Didenko
<adidenko@xxxxxxxxxxxx> wrote:
> Hi,
>
>
>> The ability to add additional online/custom repositories to a cluster
>> would enable a lot of additional customisation for users.
>
> But users already have such ability - any deployment engineer can handle it
> in few minutes by adding needed additional apt/yum repos and installing
> needed packages on OS nodes, for example, via mcollective. I'm not sure it's
> worth the effort.
>
>
> On Fri, Jun 13, 2014 at 11:33 AM, Jesse Pretorius
> <jesse.pretorius@xxxxxxxxx> wrote:
>>
>> On 13 June 2014 01:18, Andrew Woodward <xarses@xxxxxxxxx> wrote:
>>>
>>> I think this raises some feature gaps we have.
>>> 1) We need to provide the packages with the repos on the ISO.
>>> 2) We should add a field for "install additional packages" in cluster
>>> settings (maybe instead just do #3, but it would be awesome to have
>>> both)
>>> 3) We should add a post provision and post deployment scripts (kinda
>>> cloud init style) so that users can customize their nodes more
>>> efficiently.
>>
>>
>>>
>>> > On Sun, Jun 8, 2014 at 2:14 PM, Andrey Danin <adanin@xxxxxxxxxxxx>
>>> > wrote:
>>>
>>> >>
>>> >> I don't think it's a good idea to manipulate with our base repos on a
>>> >> Master node. Let's develop a standard way to create supplemental
>>> >> repositories and provide them to target nodes. For instance, there
>>> >> should be
>>> >> MCollective/Astute calls to list/create/update additional DEB/RPM
>>> >> repos on a
>>> >> Master node and add/remove them on Target nodes. In this case we will
>>> >> be
>>> >> able to easily plug in such features as VMWare NSX which requires
>>> >> proprietary packages be obtained and installed on the fly.
>>
>>
>> I think that Andrey's idea and Andrew's #2 idea are exactly what's needed.
>> The ability to add additional online/custom repositories to a cluster would
>> enable a lot of additional customisation for users. This may be simple
>> content packages (like a custom image and css for horizon), more complex
>> packages (like some additional horizon dashboard plugins), or custom
>> packages which can be used instead of the patch-on-every-node method
>> described by the Fuel Operations Docs.
>>
>> --
>> Mailing list: https://launchpad.net/~fuel-dev
>> Post to     : fuel-dev@xxxxxxxxxxxxxxxxxxx
>> Unsubscribe : https://launchpad.net/~fuel-dev
>> More help   : https://help.launchpad.net/ListHelp
>>
>
>
> --
> Mailing list: https://launchpad.net/~fuel-dev
> Post to     : fuel-dev@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~fuel-dev
> More help   : https://help.launchpad.net/ListHelp
>



-- 
Andrew
Mirantis
Ceph community