Re: Group vars for tasktotron- prod stg and dev

[David Shier]

Thanks, Tim. I'll skip clients  and add resultsdb* to my list. I hope to get to poke
these tonight. Built a new box last night so still moving in. :)

Sent using Boxer

On Dec 2, 2015 10:42 AM, Tim Flink <tflink(a)redhat.com&gt; wrote:

On Wed, 2 Dec 2015 10:22:39 -0500
David Shier <davidjshier(a)gmail.com&gt; wrote:

>  Tim-
>  
>  Thanks for the detailed reply! I will review and make the adjustments.
>  
>  In the list there are entries for the below that are flagged as
>  needing group vars updates:
>  
>  taskotron-dev
>  taskotron-dev-client-hosts
>  taskotron-dev-clients
>  taskotron-prod
>  taskotron-prod-clients
>  taskotron-stg
>  taskotron-stg-clients
>  
>  I planned on updating all of those, but figured I would start with
>  this piece. If there are other pieces that are not in the list let me
>  know and I will add them to my list of vars to update. 
resultsdb-* are also part of taskotron.

I'd suggest skipping the taskotron-*-clients playbooks, they're not
long for this earth and unless there's a huge urgency in getting 100%
coverage with csi vars, I don't think it'd be worth the time.

Tim

>  Dave Shier / odin2016
>  
>  
>  On 12/02/2015 10:14 AM, Tim Flink wrote:
>  > On Thu, 26 Nov 2015 17:17:46 -0500
>  > David Shier <davidjshier(a)gmail.com&gt; wrote:
>  >
>  > Comments inline for the dev patch, stg and prod have similar
>  > comments.
>  >
>  > +csi_security_category: Moderate
>  > +csi_primary_contact: #fedora-admin infrastructure(a)fedoraproject.org
>  > +csi_purpose: Run automated tasks, such as builds on fedora
>  > components and report results of those task runs.
>  >
>  > Taskotron does not run builds. I suppose it could but that would
>  > likely overlap with Koji and/or copr.
>  >
>  > A better way to put it would be:
>  > Runs automated tasks against Fedora components triggered by fedmsgs
>  > and reports results of those tasks to resultsdb.
>  >
>  > The difference between the three systems is:
>  >
>  >   * dev is where we try out new features, often before they're
>  > ready for general consumption
>  >
>  >   * stg is very close to prod and is used to verify that changes
>  > work before they're deployed to prod
>  >
>  >   * prod is the production system which most folks interface with
>  >
>  > +csi_relationship: |
>  > +
>  > +The system is made up of several components:
>  > +
>  > +- trigger
>  > +
>  > +- task execution system
>  > +  this is a master/slave system, currently using buildbot
>  > +
>  > +- results storage (in resultsdb)
>  > +
>  > +- mirror task git repos
>  >
>  > I'm a little fuzzy on what's supposed to be in this field. The
>  > taskotron-{dev,stg,prod} playbooks are responsible for the following
>  > parts of the Taskotron system:
>  >
>  >   * run the trigger program which listens for fedmsgs which kick off
>  >     tasks
>  >
>  >   * run the buildmaster
>  >
>  >   * host the static landing page at taskotron.fp.o
>  >
>  > +- requires access to the git mirror(s), builders (koji, copr) and
>  > +resultsdb
>  >
>  > This is not quite right. A more accurate description of the
>  > dependencies would be:
>  >
>  > requires access to kojihub, bodhi, fedmsg hub, dist-git and the
>  > upstream task repositories.
>  >
>  > Are we planning to fill in the CSI vars for the other Taskotron
>  > related playbooks or is this meant to describe the system in its
>  > entirety?
>  >
>  > Thanks for getting this started,
>  >
>  > Tim
>  >
>  >
>  > _______________________________________________
>  > infrastructure mailing list
>  > infrastructure(a)lists.fedoraproject.org
>  > http://lists.fedoraproject.org/admin/lists/[e-mail]...
>   
_______________________________________________
infrastructure mailing list
infrastructure(a)lists.fedoraproject.org
http://lists.fedoraproject.org/admin/lists/[e-mail]...