[Fedora Base Design WG] Confirmed meeting: Friday, November 8 15:00 UTC on #fedora-meeting IRC

[Phil Knirsch]

Subject says it all, thanks to Jon for making it possible to come on 
Friday at that time as well we now have our first official meeting time 
and day.

The initial meeting agenda items went out with my previous email, but to 
be sure here are recap:

  - The main topics for the meeting agenda from my side will be:
    o Governance discussion & draft (Proposal: Review other WGs 
governance drafts & create our own based on those)
    o Regular meeting times (stay with Friday 15:00 PM or move somewhere 
else?)
    o Base mailing list discussion (yes/no, pros/cons)
    o QE representation discussion (who, how, status, etc)
    o Role of Base Design group discussion ->
    o Start of collecting ideas for Base Design. Proposal: Use wiki
        https://fedoraproject.org/wiki/Base
      Maybe something like a mission statement first that relates to the 
idea that the Base Design is the platform for all other direct Fedora 
products (there can potentially be other separate products based on our 
original products or others consuming Base in the future).
    o Container technology

See you there!

Thanks & regards, Phil

-- 
Philipp Knirsch              | Tel.:  [phone]
Manager Core Services        | Fax.:  [phone]
Red Hat GmbH                 | Email: Phil Knirsch <pknirsch(a)redhat.com&gt;
Wankelstrasse 5              | Web:   http://www.redhat.com/
D-70563 Stuttgart, Germany

[Base] Summary/Minutes from today's Base WG meeting (2013-11-08)

[Dennis Gilmore]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

==============================================================
#fedora-meeting: Fedora Base Design Working Group (2013-11-08)
==============================================================


Meeting started by pknirsch at 15:00:59 UTC. The full logs are available
at
http://meetbot.fedoraproject.org/fedora-meeting/2013-11-08/fedora_base_de...
.



Meeting summary
- ---------------
* Governance discussion & draft  (pknirsch, 15:12:13)
  * LINK: https://fedoraproject.org/wiki/Server/Governance_Charter
    (jwb, 15:17:02)
  * LINK: https://fedoraproject.org/wiki/Server/Governance_Charter looks
    good :)  (haraldh, 15:19:36)
  * AGREED:   (haraldh, 15:20:05)
  * AGREED: Proposal: Take the server governance charter and replace the
    names and present to FESCO for approval  (pknirsch, 15:22:11)
  * ACTION: haraldh  Proposal: Take the server governance charter and
    replace the names and present to FESCO for approval  (pknirsch,
    15:23:36)

* Regular meeting times  (pknirsch, 15:28:03)
  * Keep meeting time at 15:00 UTC for the time being  (pknirsch,
    15:33:00)

* Base mailing list discussion  (pknirsch, 15:33:28)
  * AGREED:   (haraldh, 15:37:11)
  * AGREED: Stay with fedora-devel mailinglist, use [Base] in subject to
    indicate emails around Base Design discussions  (pknirsch, 15:39:33)

* Role of Base Design group discussion  (pknirsch, 15:42:57)
  * LINK:
   
https://fedoraproject.org/wiki/Critical_path_package?rd=Critical_path_pac...
    (dgilmore, 15:45:34)
  * LINK:
   
https://admin.fedoraproject.org/pkgdb/lists/critpath?tg_format=plain&...
    (dgilmore, 15:45:57)
  * LINK: http://fpaste.org/52688/38392758/   (pknirsch, 16:19:54)
  * Base definition: installer, compose tools, minimal install (for some
    definition there), and functionality the majority products want to
    use  (pknirsch, 16:21:30)

* Open Floor  (pknirsch, 16:25:50)
  * jreznik working on a communication plan with other WGs  (pknirsch,
    16:33:25)
  * ACTION: jreznik working on a communication plan with other WGs
    (pknirsch, 16:34:04)

Meeting ended at 16:41:32 UTC.




Action Items
- ------------
* haraldh  Proposal: Take the server governance charter and replace the
  names and present to FESCO for approval
* jreznik working on a communication plan with other WGs




Action Items, by person
- -----------------------
* haraldh
  * haraldh  Proposal: Take the server governance charter and replace
    the names and present to FESCO for approval
* jreznik
  * jreznik working on a communication plan with other WGs
* **UNASSIGNED**
  * (none)




People Present (lines said)
- ---------------------------
* pknirsch (195)
* dgilmore (105)
* haraldh (69)
* jreznik (68)
* jwb (57)
* masta (15)
* zodbot (9)
* kmacleod (1)
* sgallagh (1)
* Southern_Gentlem (1)
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v2.0.22 (GNU/Linux)
>
> iQIcBAEBAgAGBQJSfRaIAAoJEH7ltONmPFDRLtUP/A+1Gs7Mp5aQ9dRAhMLE9uUe
> CXawhmK3QNrJy6E3cA33ECF3TSjAi/Roq3gDCtyE9hIzT4AcLbZVUj4JwcCQzhkE
> Mr9TVE3JXjfVHk3OFWPLaW41q71TDLo8rpb9mEpjR+ZzeW86FqGKz/PCsZjImUjf
> XROLmA7ybr+5T2UoSpC1LoBkk4V7A3uX902thV9WCC+WdbBCLHKLHkwamEXXM4ds
> zOMoYMK3fHD31Vz9SVk1wLn9dkaIlFNwgT6aXzqMubhJ8V65vdLuZ0hvjtwXk5sd
> blGb0BNlDoOaD6kU6JBF9/zBX6MhiXyyQCsThtWvo3jshzthLiDNU2he4PW8kzST
> wRgMu4oL87igMdNskJQpugOyc7kIaHA9hoYuy3NH9PJL460hbU9/NVf0EHBPEvUI
> Z3KrtZpyROl76P9ohX1olz2Vk3V/c55qBCVQ96fDlCcxASDwdwiLi03gQOGfZQ3G
> 7HkNbtNkJhE1yGt76BOvYrRsgf2NR1xu59A9cUZy+eUKZF6v+MmF5u5qisJbBGLN
> WeBfwAzb6YZ35pc/20SXec5veEXGCf0r7OyNgOFVUfnordAy9pgk9u/IP/xqGhfL
> j0qQl27/X32EKNOQy7Uf13NeRBsld/uIvDT1+bgmMuUnt/RsfZy2yZPDhGAgALqJ
> eaCw0DbJqqitFCoA365k
> =B1/8
> -----END PGP SIGNATURE-----
>

[Base] Summary/Minutes from today's Base WG meeting (2013-11-08)

[Bill Nottingham]

Dennis Gilmore (dennis(a)ausil.us) said: 
>  Meeting started by pknirsch at 15:00:59 UTC. The full logs are
> available
>  at
>
> http://meetbot.fedoraproject.org/fedora-meeting/2013-11-08/fedora_base_de...
Apologies for not making the meeting. Some comments

>    * LINK: http://fpaste.org/52688/38392758/   (pknirsch, 16:19:54)
>    * Base definition: installer, compose tools, minimal install (for some
>      definition there), and functionality the majority products want to
>      use  (pknirsch, 16:21:30) 
My concern with this, as a base design, is that the current implementation
of anaconda leads to a dependency tree that is *not* small; it's much larger
than the minimal install, as it extends to storage tools, NetworkManager,
graphical toolsets, and other items. (You could even arguably pull in X and
in the future wayland.)

It's not a *bad* thing to have as a base design - I could make the argument
that changes in that base (NM, X, etc.) are all things that should be
coordinated as a layer under the separate products. But it does give a wider
area of coverage than may be expected.

(Also, stabilizing the build dependencies therein. Wheee.)

Bill

[Base] Summary/Minutes from today's Base WG meeting (2013-11-08)

[Matthias Clasen]

On Fri, 2013-11-08 at 10:51 -0600, Dennis Gilmore wrote:

>    * LINK: http://fpaste.org/52688/38392758/   (pknirsch, 16:19:54)
>    * Base definition: installer, compose tools, minimal install (for some
>      definition there), and functionality the majority products want to
>      use  (pknirsch, 16:21:30)
>   
Hey, I read your minutes, and discussed some of my thoughts on this with
my workstation wg peers. Basically, I think that defining a 'base' as a
particular set of packages (minimal install, or some variant thereof)
does not really provide us what we need to build one or more products.'

What we really need as a base is a definition of the apis that are
guaranteed to be stable and that the products and applications can rely
on. Packages can to some extent serve as a proxy for that, but they are
really just an implementation detail of how the product is put
together. 

For example, when you talk about an OS being systemd-based, the
important part is not that systemd is part of the minimal install, but
the fact that applications can rely on org.freedesktop.login1 and so on
being available on the session bus, and that installing a unit file
in /usr/lib/systemd/system is the proper way to make a service known to
the system, etc.

There will of course be APIs that are product-specific. Eg, the server
product might define a remote management api that is not relevant for a
workstation. And desktop applications running on the workstation need to
know which session services they can rely on.


Just my 2 cents, Matthias

[Base] Summary/Minutes from today's Base WG meeting (2013-11-08)

[Matthew Miller]

On Fri, Nov 08, 2013 at 03:21:04PM -0500, Matthias Clasen wrote:
>  Hey, I read your minutes, and discussed some of my thoughts on this
> with
>  my workstation wg peers. Basically, I think that defining a 'base' as a
>  particular set of packages (minimal install, or some variant thereof)
>  does not really provide us what we need to build one or more products.'
>  
>  What we really need as a base is a definition of the apis that are
>  guaranteed to be stable and that the products and applications can rely
>  on. Packages can to some extent serve as a proxy for that, but they are
>  really just an implementation detail of how the product is put
>  together.  
+1 to this. In fact, I think we'll *need* to do it this way in order to be
able to innovate at both layers.

-- 
Matthew Miller  ☁☁☁  Fedora Cloud Architect  ☁☁☁  <mattdm(a)fedoraproject.org&gt;

[Base] Summary/Minutes from today's Base WG meeting (2013-11-08)

[Phil Knirsch]

On 11/09/2013 08:26 PM, Matthew Miller wrote:
>  On Fri, Nov 08, 2013 at 03:21:04PM -0500, Matthias Clasen wrote:
> > Hey, I read your minutes, and discussed some of my thoughts on this with
> > my workstation wg peers. Basically, I think that defining a 'base' as a
> > particular set of packages (minimal install, or some variant thereof)
> > does not really provide us what we need to build one or more products.'
> >
> > What we really need as a base is a definition of the apis that are
> > guaranteed to be stable and that the products and applications can rely
> > on. Packages can to some extent serve as a proxy for that, but they are
> > really just an implementation detail of how the product is put
> > together.
>
>  +1 to this. In fact, I think we'll *need* to do it this way in order to be
>  able to innovate at both layers.
>  
I absolutely like the idea and i think we brought that up during our 
meeting on Friday actually and were all more or less in favor of it.

In order to get there though the Base team will need very detailed input 
from the other WGs so that we can then find the smallest common 
denominator of a common stable API and/or ABI that Base will then 
provide for any given release for all products.

Thanks & regards, Phil

-- 
Philipp Knirsch              | Tel.:  [phone]
Manager Core Services        | Fax.:  [phone]
Red Hat GmbH                 | Email: Phil Knirsch <pknirsch(a)redhat.com&gt;
Wankelstrasse 5              | Web:   http://www.redhat.com/
D-70563 Stuttgart, Germany

[Base] Summary/Minutes from today's Base WG meeting (2013-11-08)

[Jaroslav Reznik]

----- Original Message -----
>  On Fri, 2013-11-08 at 10:51 -0600, Dennis Gilmore wrote:
>  
>  >   * LINK: http://fpaste.org/52688/38392758/   (pknirsch, 16:19:54)
>  >   * Base definition: installer, compose tools, minimal install (for some
>  >     definition there), and functionality the majority products want to
>  >     use  (pknirsch, 16:21:30)
>  > 
>  
>  Hey, I read your minutes, and discussed some of my thoughts on this with
>  my workstation wg peers. Basically, I think that defining a 'base' as a
>  particular set of packages (minimal install, or some variant thereof)
>  does not really provide us what we need to build one or more products.'
>  
>  What we really need as a base is a definition of the apis that are
>  guaranteed to be stable and that the products and applications can rely
>  on. Packages can to some extent serve as a proxy for that, but they are
>  really just an implementation detail of how the product is put
>  together.
>  
>  For example, when you talk about an OS being systemd-based, the
>  important part is not that systemd is part of the minimal install, but
>  the fact that applications can rely on org.freedesktop.login1 and so on
>  being available on the session bus, and that installing a unit file
>  in /usr/lib/systemd/system is the proper way to make a service known to
>  the system, etc.
>  
>  There will of course be APIs that are product-specific. Eg, the server
>  product might define a remote management api that is not relevant for a
>  workstation. And desktop applications running on the workstation need to
>  know which session services they can rely on. 
I agree, I mentioned it in the meeting - we should define it more in the
terms of interfaces over packages. As an example I can use sendmail we had
a long discussion about a few month before. If we define some sort of MTA
has to be part of the base design, there's no reason to argument, if it
has to be sendmail or any other solution. Just something has to be there.
On the other hand - at least in the beginning, basing it on packages is
not a bad way how to kick it off. But one day we should go back and take
a look on the packages, take a look on what these packages provide and
try to set default expectation of APIs we have. That's pretty similar
to the initial proposal from the last FUDCon, I'll try to dig out the
etherpad...

Jaroslav

>  
>  
>  Just my 2 cents, Matthias
>  
>  --
>  devel mailing list
>  devel(a)lists.fedoraproject.org
>  https://admin.fedoraproject.org/mailman/listinfo/devel
>  Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

[Base] Summary/Minutes from today's Base WG meeting (2013-11-08)

[Matthias Clasen]

Forgot one thing - the installer question.

I think it already came up tangentially in some of the workstation prd
discussions, but I think we should at least be open to the idea that the
different products need to be able to define their own install
experience.

That would imply that the 'installer' you mention as part of base should
perhaps be just a shared installer backend which can support different
frontends. I know the anaconda team has been working towards
establishing such a separation, with kickstart being the interface
between frontend and backend. I don't know how far that got though, so
maybe this is more of a long-term goal.

Just wanted to mention it.

[Base] Summary/Minutes from today's Base WG meeting (2013-11-08)

[Daniel J Walsh]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 11/08/2013 11:51 AM, Dennis Gilmore wrote:
>  -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
>  
>  ============================================================== 
>  #fedora-meeting: Fedora Base Design Working Group (2013-11-08) 
>  ==============================================================
>  
>  
>  Meeting started by pknirsch at 15:00:59 UTC. The full logs are available 
>  at 
>
> http://meetbot.fedoraproject.org/fedora-meeting/2013-11-08/fedora_base_de...
>
>   
Also I apologize for not being there, I was at a conference.

A few other things, I would like to see broken out would be sort of a lower
level definition for containers, based on the docker model.  IE Can we remove
the need for systemd/kernel from the new container group?

Base > Minimal Install > Container Install.

Since Containers were put off til later, we can discuss then.

Dreaming big would be to specify base language of English and then define new
ways for packaging to add other languages as needed.

Is remote access to the Base Instance required?  sshd?

Is Login required?

Audit?

Firewall?
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.15 (GNU/Linux)
> Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
>
> iEYEARECAAYFAlJ/fpgACgkQrlYvE4MpobNGZQCgnsBRqryJAoMDVMsGUOmOsB3e
> nbsAoOZrshFBzUx5izVzkKgGEPkZwEO2
> =YlEC
> -----END PGP SIGNATURE-----

[Base] Summary/Minutes from today's Base WG meeting (2013-11-08)

[Harald Hoyer]

On 11/10/2013 01:39 PM, Daniel J Walsh wrote:
>  A few other things, I would like to see broken out would be sort of a
> lower
>  level definition for containers, based on the docker model.  IE Can we remove
>  the need for systemd/kernel from the new container group?
>  
>  Base > Minimal Install > Container Install.
>  
>  Since Containers were put off til later, we can discuss then. 
What kind of containers? App containers or OS containers?

Isn't systemd a required addition for OS containers?
I think it has to manage cgroups and d-bus if kdbus kicks in.

[Base] Summary/Minutes from today's Base WG meeting (2013-11-08)

[Daniel J Walsh]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 11/11/2013 10:31 AM, Harald Hoyer wrote:
>  On 11/10/2013 01:39 PM, Daniel J Walsh wrote:
> > A few other things, I would like to see broken out would be sort of a
> > lower level definition for containers, based on the docker model.  IE Can
> > we remove the need for systemd/kernel from the new container group?
> > 
> > Base > Minimal Install > Container Install.
> > 
> > Since Containers were put off til later, we can discuss then.
>  
>  What kind of containers? App containers or OS containers?
>  
>  Isn't systemd a required addition for OS containers? I think it has to
>  manage cgroups and d-bus if kdbus kicks in.
>   
Application/docker containers.  Does not need kernel or systemd.  Just running
httpd within a container.
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.15 (GNU/Linux)
> Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
>
> iEYEARECAAYFAlKBDOkACgkQrlYvE4MpobM8MwCggEWv5erkIK98llk9IH4cJC61
> zoYAnjNrx9NVvQzczChmYa8sy/rEfBUt
> =npKI
> -----END PGP SIGNATURE-----