----- Original Message -----
From: "Nick Coghlan" <ncoghlan(a)redhat.com>
To: beaker-devel(a)lists.fedorahosted.org
Sent: Tuesday, July 16, 2013 11:21:47 AM
Subject: Re: [Beaker-devel] Tool for Inventory gathering
On 07/16/2013 10:02 AM, Amit Saha wrote:
>
>
> ----- Original Message -----
>> From: "Nick Coghlan" <ncoghlan(a)redhat.com>
>> To: beaker-devel(a)lists.fedorahosted.org
>> Sent: Thursday, July 11, 2013 5:40:10 PM
>> Subject: Re: [Beaker-devel] Tool for Inventory gathering
>>
>> On 07/11/2013 02:12 PM, Dan Callaghan wrote:
>>> The GitHub approach works best if upstream is on GitHub as well though,
>>> so maybe we should ask that before anything else...
>>
>> Bringing the list up to date with today's IRC discussion: Amit has
>> contacted Lyonel offering our assistance in finishing the migration to
>> GitHub (there is already a repo there, but nothing in it).
>
> Lyonel emailed me this morning saying that he has got everything up now
> here:
>
https://github.com/lyonel/lshw
>
>
>>
>> If that happens, then I think Dan's plan will work nicely - we'll
>> maintain beaker-lshw as a fork on GitHub and submit pull requests
>> upstream for any changes we make, while packaging beaker-lshw for our
>> own use to ensure we maintain control over our release cycles.
>
> So, looks like we are set to go.
Yay!
I've been thinking further about this based on our IRC discussion the
other day, and I'm thinking it makes sense to do a few things in the
0.15 time frame (i.e. by the end of August):
1. Create our own lshw fork as beaker-lshw, that we keep close to
upstream lshw (aside from the different name) and submit pull requests
on GitHub whenever we change anything
I submitted a pull request earlier today [1] to add support for reading CPU info
on s390x systems. It is from my *own* fork of lshw. Either, I can keep doing this
and build the beaker-lshw packages from this, or fork it with beaker-project.org's
account and submit pull requests from there, whichever is more favorable.
[1]
https://github.com/lyonel/lshw/pull/1
So basically what you are suggesting is create a project out of the task, and maintain
it separately. This project will use beaker-lshw along with the other tricks.
As a good to have goal, this package should hopefully be even suitable for use standalone
outside of Beaker.
3. Change our inventory task to depend on beaker-systemscan and just
upload the scan results to the Beaker server rather than containing the
scanning logic directly
So, we will have two new packages to distribute from our repos:
* beaker-lshw
* beaker-systemscan
I have so far ran lshw based task on s390x and PPC64 systems besides x86_64
with the goal of making sure the XML parsing code is robust, since the the information
exported by the Kernel and hence lshw varies. To this effect, made a few fixes yesterday
to the parsing code. I haven't yet gone around doing a thorough comparison of smolt
and lshw output.
At some point of time, I also need to fix [2].
[2]
https://bugzilla.redhat.com/show_bug.cgi?id=896302
I will keep this thread updated.
Best,
Amit.
--
Amit Saha <
http://echorand.me>
Infrastructure Engineering and Development
Red Hat, Inc.