Bill Peck | 4 May 21:09 2009
Picon

SetupHOWTO


Hello Everyone,

I have finally edited the SetupHOWTO wiki on beaker.

https://fedorahosted.org/beaker/wiki/SetupHOWTO

I had to fix a few problems with beaker itself before I could complete 
the howto.  Thats my excuse anyway. ;-)

Please check it out and let me know if it works for you.  I'll be 
updating it and adding to it.  I also attached the scripts I'm using 
internally to test beaker.  Hopefully soon we will be to the point where 
beaker can test beaker. :-)  In the mean time if the HOWTO doesn't work 
for you you can look at my test rpm to see how it should work.

Bill Peck | 5 May 20:51 2009
Picon

Beaker status for April

Beaker
----------
- Tested with cobbler 1.6.3
- Switched expire and update scripts to use native python bindings 
again, now that cobbler is fast again.
- Tested s390 support, yay it works.
- Associate kickstarts in update script, this allows for different 
kickstart template for rhel5client vs rhel5server.
- Implemented CSV Export/Import for Systems, Excluded Families, 
InstallOptions, KeyValues, LabIno, SystemGroups and UserGroups.
- Implemented scsidevices and ethdevices.  Along with nostorage and 
noprobe allows for shared storage to work in Beaker.
- Automated system selection is now sorted by the following criteria:
   First you get systems that you are the owner of
   Second you get systems that you have access to via group membership
   Finally any other systems that are shared to everyone.
   This should keep the general pool of systems free
- Only Admins are allowed to share systems to the general pool, all 
others must add a system to a group first.
- Simple search on main page allows quick searching on fqdn.
- Added arch to main page as well
- Added shortcut to types search on left menu
- Implemented per family snippets for main kickstart, %pre and %post.

Plans for May
------------------
- Get proper kerberos working in xmlrpc interface
- Continue testing beaker in beaker.
- Look at the scheduler again.

(Continue reading)

Will Woods | 6 May 23:47 2009
Picon

[PATCH] Fix handling of .py files in rhtslib.spec

Some RPM setups don't generate python .pyc/.pyo files at build time,
so the build shouldn't fail if they don't exist.
---
 rhtslib/rhtslib.spec |   16 ++++------------
 1 files changed, 4 insertions(+), 12 deletions(-)

diff --git a/rhtslib/rhtslib.spec b/rhtslib/rhtslib.spec
index bd77211..657ff15 100644
--- a/rhtslib/rhtslib.spec
+++ b/rhtslib/rhtslib.spec
 <at>  <at>  -36,18 +36,10  <at>  <at>  rm perl/docsjoin
 /usr/share/rhts-library/performance.sh
 /usr/share/rhts-library/dictionary.vim
 /usr/share/rhts-library/virtualX.sh
-/usr/share/rhts-library/python/rlMemAvg.py
-/usr/share/rhts-library/python/rlMemPeak.py
-/usr/share/rhts-library/python/rlMemAvg.pyc
-/usr/share/rhts-library/python/rlMemAvg.pyo
-/usr/share/rhts-library/python/rlMemPeak.pyc
-/usr/share/rhts-library/python/rlMemPeak.pyo
-/usr/share/rhts-library/python/journalling.py
-/usr/share/rhts-library/python/journalling.pyc
-/usr/share/rhts-library/python/journalling.pyo
-/usr/share/rhts-library/python/journal-compare.py
-/usr/share/rhts-library/python/journal-compare.pyc
-/usr/share/rhts-library/python/journal-compare.pyo
+/usr/share/rhts-library/python/rlMemAvg.py*
+/usr/share/rhts-library/python/rlMemPeak.py*
+/usr/share/rhts-library/python/journalling.py*
+/usr/share/rhts-library/python/journal-compare.py*
(Continue reading)

James Laska | 7 May 17:57 2009
Picon

Re: [PATCH] Fix handling of .py files in rhtslib.spec

Thanks for the patch Will!  Looks good to me.  

I'm not sure who is best to apply rhtslib patches at this point, Bill
Peck or Petr Muller?

Thanks,
James

On Wed, 2009-05-06 at 17:47 -0400, Will Woods wrote:
> Some RPM setups don't generate python .pyc/.pyo files at build time,
> so the build shouldn't fail if they don't exist.
> ---
>  rhtslib/rhtslib.spec |   16 ++++------------
>  1 files changed, 4 insertions(+), 12 deletions(-)
> 
> diff --git a/rhtslib/rhtslib.spec b/rhtslib/rhtslib.spec
> index bd77211..657ff15 100644
> --- a/rhtslib/rhtslib.spec
> +++ b/rhtslib/rhtslib.spec
>  <at>  <at>  -36,18 +36,10  <at>  <at>  rm perl/docsjoin
>  /usr/share/rhts-library/performance.sh
>  /usr/share/rhts-library/dictionary.vim
>  /usr/share/rhts-library/virtualX.sh
> -/usr/share/rhts-library/python/rlMemAvg.py
> -/usr/share/rhts-library/python/rlMemPeak.py
> -/usr/share/rhts-library/python/rlMemAvg.pyc
> -/usr/share/rhts-library/python/rlMemAvg.pyo
> -/usr/share/rhts-library/python/rlMemPeak.pyc
> -/usr/share/rhts-library/python/rlMemPeak.pyo
> -/usr/share/rhts-library/python/journalling.py
(Continue reading)

James Laska | 7 May 17:59 2009
Picon

Re: Beaker status for April

Great summary!

On Tue, 2009-05-05 at 14:51 -0400, Bill Peck wrote:
> Plans for May
> ------------------
> - Get proper kerberos working in xmlrpc interface

This is probably a long ways off, but do you have any thoughts on how
the authentication in beaker can be modular?  Meaning, is this something
we could plug into FAS or openid when the time comes for a hosted
instance in Fedora?

Thanks,
James
Bill Peck | 7 May 18:13 2009
Picon

Re: Beaker status for April

James Laska wrote:
> Great summary!
>
> On Tue, 2009-05-05 at 14:51 -0400, Bill Peck wrote:
>   
>> Plans for May
>> ------------------
>> - Get proper kerberos working in xmlrpc interface
>>     
>
> This is probably a long ways off, but do you have any thoughts on how
> the authentication in beaker can be modular?  Meaning, is this something
> we could plug into FAS or openid when the time comes for a hosted
> instance in Fedora?
>
> Thanks,
> James
>   
It should be able to be modular.

I'm developing the app with two entry points for authentication:
1) coming in via the web browser
2) coming in via xmlrpc

Both entry points then use the identity module in turbogears to 
authenticate the user with a slight modification if you do kerberos auth.

Kerberos auth for the web interface relies on apache doing the negotiate 
with the browser, When it gets to Turbogears identity module we are 
already authenticated and I ask cherrypy for remote_user and if that 
(Continue reading)

James Laska | 7 May 18:51 2009
Picon

Re: Beaker status for April

On Thu, 2009-05-07 at 12:13 -0400, Bill Peck wrote:
> James Laska wrote:
> > Great summary!
> >
> > On Tue, 2009-05-05 at 14:51 -0400, Bill Peck wrote:
> >   
> >> Plans for May
> >> ------------------
> >> - Get proper kerberos working in xmlrpc interface
> >>     
> >
> > This is probably a long ways off, but do you have any thoughts on how
> > the authentication in beaker can be modular?  Meaning, is this something
> > we could plug into FAS or openid when the time comes for a hosted
> > instance in Fedora?
> >
> > Thanks,
> > James
> >   
> It should be able to be modular.
> 
> I'm developing the app with two entry points for authentication:
> 1) coming in via the web browser
> 2) coming in via xmlrpc
> 
> Both entry points then use the identity module in turbogears to 
> authenticate the user with a slight modification if you do kerberos auth.
> 
> Kerberos auth for the web interface relies on apache doing the negotiate 
> with the browser, When it gets to Turbogears identity module we are 
(Continue reading)

Koci Martin (koca | 11 May 08:29 2009
Picon

Re: [PATCH] Fix handling of .py files in rhtslib.spec

CC'ing Petr Muller.
koca

James Laska wrote:
Thanks for the patch Will! Looks good to me. I'm not sure who is best to apply rhtslib patches at this point, Bill Peck or Petr Muller? Thanks, James On Wed, 2009-05-06 at 17:47 -0400, Will Woods wrote:
Some RPM setups don't generate python .pyc/.pyo files at build time, so the build shouldn't fail if they don't exist. --- rhtslib/rhtslib.spec | 16 ++++------------ 1 files changed, 4 insertions(+), 12 deletions(-) diff --git a/rhtslib/rhtslib.spec b/rhtslib/rhtslib.spec index bd77211..657ff15 100644 --- a/rhtslib/rhtslib.spec +++ b/rhtslib/rhtslib.spec <at> <at> -36,18 +36,10 <at> <at> rm perl/docsjoin /usr/share/rhts-library/performance.sh /usr/share/rhts-library/dictionary.vim /usr/share/rhts-library/virtualX.sh -/usr/share/rhts-library/python/rlMemAvg.py -/usr/share/rhts-library/python/rlMemPeak.py -/usr/share/rhts-library/python/rlMemAvg.pyc -/usr/share/rhts-library/python/rlMemAvg.pyo -/usr/share/rhts-library/python/rlMemPeak.pyc -/usr/share/rhts-library/python/rlMemPeak.pyo -/usr/share/rhts-library/python/journalling.py -/usr/share/rhts-library/python/journalling.pyc -/usr/share/rhts-library/python/journalling.pyo -/usr/share/rhts-library/python/journal-compare.py -/usr/share/rhts-library/python/journal-compare.pyc -/usr/share/rhts-library/python/journal-compare.pyo +/usr/share/rhts-library/python/rlMemAvg.py* +/usr/share/rhts-library/python/rlMemPeak.py* +/usr/share/rhts-library/python/journalling.py* +/usr/share/rhts-library/python/journal-compare.py* /usr/share/rhts-library/perl/deja-summarize /usr/share/rhts-library/test/README /usr/share/rhts-library/test/coverageTest.sh _______________________________________________ Beaker-devel mailing list Beaker-devel <at> lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/beaker-devel


Petr Muller | 11 May 14:31 2009
Picon

Re: [PATCH] Fix handling of .py files in rhtslib.spec

On Thu, 2009-05-07 at 11:57 -0400, James Laska wrote:
> Thanks for the patch Will!  Looks good to me.  
> 
> I'm not sure who is best to apply rhtslib patches at this point, Bill
> Peck or Petr Muller?

That would be me, although not necessarily - Petr Splichal, Ales Zelinka
or Jan Hutar know BeakerLib internals just as good as I do :)

> On Wed, 2009-05-06 at 17:47 -0400, Will Woods wrote:
> > Some RPM setups don't generate python .pyc/.pyo files at build time,
> > so the build shouldn't fail if they don't exist.

Thanks, the *.pyc assumption was incorrect of course. I've applied it.

Petr

Bill Peck | 22 May 20:26 2009
Picon

git re-org


Hello Everyone,

If you have looked at the git repo lately you may have noticed a big 
restructure.  We had many sub-packages in beaker that all had their own 
way of doing things.  I found some great Makefile rules from another 
Fedora project called python-slip.  These help with git, scm, python, 
rpmspecs, etc..  So I modified our Makefiles to take advantage of them.

This let us stop using python setup.py sdist as a way to build our 
source archive.  This was just a bad idea and created lots of problems.

So what do we have now?

At the top level we have a common spec file called beaker.spec which 
lists the following sub-packages:

beaker-server                - This used to be medusa-server
beaker-lab-controller   -   This used to be medusa-lab-controller
beaker-client               - This is our client interface to 
beaker-server.  This relies on python-kobo which isn't released quite yet.
beaker-lib                   - This used to be rhtslib.

All of these can be built from the top level directory with the 
following commands:

make archive <- This will check that you don't have any outstanding 
commits waiting.
                           Also checks for any incoming changes from master
                           Pushes changes to master
                           Tags based on version in beaker.spec
                           Finally creates a tarball using the 
git-archive command
make srpm/rpm <- Just like before it will build an SRPM or RPM.
make push <- This will push the archive to fedorahosted.org

For local development in-between releases you can use snaparchive 
instead of archive and it will only make an archive local without 
tagging or pushing to master.  This is good for verifying things work 
before pushing.

I've removed any reference to medusa in the repo,  Everything will be 
referenced by beaker.  The scheduler components in Logan will be merged 
into server.  More on that in another email.

Email me if you have any questions!


Gmane