<html>
<head>
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#333333">
<br>
<div class="moz-cite-prefix">On 2013/06/12 00:00, James Slagle
wrote:<br>
</div>
<blockquote
cite="mid:CAHV77z8BaNKVs5YT65v=+_u3mDb4nF6_2nKHS8qU3=Rwv_qoyQ@mail.gmail.com"
type="cite">
<pre wrap="">Jarda,
This stuff looks really great. I like the flow quite a bit. One thing
that stuck out to me though is that I didn't see anything around
troubleshooting a Deployment. Since stuff does occassionally go wrong
:), it might be nice to mock up how troubleshooting that might look.
I think we already have the logs in Horizon, but bubbling that up to
the user on the deployment screen might be nice. Or, perhaps links
back to the logs if something failed.
In the pdf for deployment[1], slide 13, there's shown a node that
failed. But, I don't see much there in terms of getting at how/why it
failed.
I know I'm probably thinking way far out here :), but there was some
good discussion at the summit on Heat making it easier to debug stack
failures, adding retries, etc. I know that's stuff still early, but
it might make sense to think about hooking into that once available.
[2]. There was also an idea of adopting a stack [3]. This might be a
way for folks who already have a deployment to bring it under the
control of Tuskar so to speak.
Anyway, it doesn't look like anything in the mockups would make it
difficult to hook into these new API's once avaiable, but I just
thought I would mention them. Thanks!
[1] <a class="moz-txt-link-freetext" href="http://people.redhat.com/~jcoufal/openstack/tripleo/2013-12-03_tripleo-ui_03-deployment.pdf">http://people.redhat.com/~jcoufal/openstack/tripleo/2013-12-03_tripleo-ui_03-deployment.pdf</a>
[2] <a class="moz-txt-link-freetext" href="https://blueprints.launchpad.net/heat/+spec/troubleshooting-low-level-control">https://blueprints.launchpad.net/heat/+spec/troubleshooting-low-level-control</a>
[3] <a class="moz-txt-link-freetext" href="https://blueprints.launchpad.net/heat/+spec/adopt-stack">https://blueprints.launchpad.net/heat/+spec/adopt-stack</a>
</pre>
</blockquote>
<br>
James,<br>
<br>
this is great note and very important part what we need to focus. I
completely agree with you.<br>
<br>
At the moment I tried to get something up for designing and
deploying a cloud - to make sure that this is good direction to go.
And that we can actually start delivering.<br>
<br>
Most of the pages what you see at the moment are overview pages -
their purpose is to show general information and to help to notice
that 'something' went wrong.<br>
<br>
To figure out what exactly went wrong, I expect to get into bigger
details a bit later (through notifications, logs, node detail
pages).<br>
<br>
We definitely should keep above listed blueprints in our minds to
think about how to hook them for our needs - they look very useful.
I will keep them also in mind when designing troubleshooting in more
details.<br>
<br>
Thanks a lot this is great help<br>
-- Jarda<br>
</body>
</html>