On 07/10/2015 01:48 PM, Takeshi Larsson wrote:
Elizabeth K. Joseph <lyz@...> writes:
On Wed, Jul 8, 2015 at 4:06 AM, Takeshi Larsson
<takeshi.p.larsson@...> wrote:
Hows this going? Have you had any thoughts on which way you want to
go?
I'm the "not really a developer" for panopuppet and I was wondering
if there
was something you felt was missing from panopuppet or what it is you
are
"looking" for?
Thanks for taking the time to reach out to us, it's great to hear from
a contributor to a project we're looking at. Decision-wise, right now
we're continuing to use PuppetBoard and we haven't made a move yet.
All we need is a read-only dashboard so anyone in our community can
watch puppet changes land and have the data to debug accordingly to
submit a new patch when things fail. The dashboard is public so anyone
on the internet can view it, which means it has to give useful data,
but can't leak sensitive data. We don't do any configuration via the
dashboard, that's all in git and goes through our standard code review
process.
The main concern about the single developer situation is that this is
the second time we've had to look at changing the dashboard due to
loss of development efforts. It would be great if we could move to a
project that has a stronger community so that when that single
developer has a change in family status/get a new job/focuses
elsewhere, we aren't stuck again.
Hi,
Ofcourse I am interested in all potential parties so replying to the
mailing list seemed like the most responsible thing to do.
I've seen your puppetboard and noticed that you only have 100-200 nodes
and is nothing that puppetboard cannot handle.
PanoPuppet was written because at work we have more than 3000 nodes and
puppetboard could not handle showing that amount of nodes.
It was written to work at my current place of employment to function
with cross function teams with advanced query building to find just the
nodes
you need or are responsible for.
I was trying to make it as "enterprise" ready as possible. So that there
was a solution that could fill the needs of the puppet opensource users.
PanoPuppet much like puppetboard is a read-only dashboard. The data
shown in puppetboard is pretty much what PanoPuppet offers
except that PanoPuppet offers better tools for analytics, debugging and
PuppetDB Query Building.
It seems that your workflow is pretty much the same as it as where I
work. We use hiera and we commit everything for review and have CI's
checking for various errors etc
and also human reviewal before allowing it to reach our master branch.
The single developer project is ofcourse a major factor to your choice
of dashboard. And I am aware of PuppetBoards choice to add more core
developers to their project
that happened several months ago and I have yet to see any kind of major
PR or commits to happen.
I on the other hand get to work on PanoPuppet at work and keep it FOSS
and also I spend a couple of hours at home tinkering with it.
This is to satisfy the needs of my colleagues at work but also some of
my users (most of which have over 10k nodes) using it.
They constantly email me with new feature requests or bugs they have
found.
While I understand the worry about losing the only developer, I can tell
you, I'm 24 years old without kids and not looking to get any for a
couple of years. So as long as the company I work need PanoPuppet or
there are others who need PanoPuppet I will probably be available for a
couple of years and hopefully during that time some more developers
might want to hop on.
I've spoken to Spencer on irc about Puppetboard and while I can consider
helping out Puppetboard I am not sure I will stop developing PanoPuppet
to help out with Puppetboard. Even though Puppetboard has a much bigger
customer base having been available longer than PanoPuppet, PanoPuppet
was not created to take over any other projects and was a way for me to
learn more python and perhaps even decision making based on opinions of
the few PanoPuppet users.
My user base seems to be growing, more people seem to be trying
PanoPuppet out which I think is great.
So the beautiful thing about OpenSource is nobody is going to stop you
taking the effort of bringing a PanoPuppet node online for demo purposes.
If you are serious about looking to get openstack-infra using PanoPuppet
I would first start working on a puppet module that -infra could easily
consume. For example, check out how we do it today with puppetboard[1].
I'm not sure if our puppetdb server is open to the public or not, but it
wouldn't be too hard to get puppetdb up and running locally on your demo
instances.
The project does seem interesting, but as Elizabeth mentioned read-only
mode is the most basic use case I think people use it for.
[1]
https://github.com/openstack-infra/system-config/blob/master/modules/openstack_project/manifests/puppetboard.pp
---
Paul Belanger
_______________________________________________
OpenStack-Infra mailing list
OpenStack-Infra@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-infra
