@lenards Hey Andy, I have something like this: https://github.com/CCI-MOC/GUI-Frontend/pull/15/files

that adding another view before we actual launch the instance. Do you think it is also a feature that atmosphere would like to have in the troposphere? If does, I can send out this as a PR and for review.

Basic logic is that, after selecting an image and a user click launch, it checks whether the user has an atmosphere project first, and then, check whether the SSH pubkey has uploaded, and finally to the instance launch wizard.

Just a reminder here for Y release, the login modal needs a max height:

I think this is more like a front-end issue. So I decide to put this here. Steve reproduces this issue and this issue appears in my local test atmosphere servers as well

I did what this https://github.com/cyverse/troposphere/tree/master/troposphere/static/theme tells me to do:

1. Having a new folder called themeImages within /theme/
2. Changed mini_logo.png (double-checked that the image has changed)
3. run ./configure

And I have the expected output (all green, successfully configured)

But I cannot any chance the mini_logo and I did a hard refresh/restart web server.

Do I need to run webpack? (I don't think so but I don't know why it is not working)

Thanks!
Lucas

Hi guys, I am wondering if there is a guide for setting variable.ini file?

Right now, we have adopted Stage 0 which is the Strawman level (this included experimental features).

Change to Stage 2 or Stage 3.

https://github.com/cyverse/troposphere/blob/master/troposphere/static/js/components/dashboard/DashboardPage.jsx#L37

and

https://github.com/cyverse/troposphere/blob/master/troposphere/static/js/components/dashboard/DashboardPage.jsx#L45

are identical right?

Does this on purpose?

Thanks!

And L49/L57 as well.

In a parent-child (or container-containee) relationship, you don't want the containee to decide when to render itself. If you don't want a button to be visible, then the parent doesn't include it. The inversion of control allows for the situation where children ignores the parent. The ButtonBar also has an isVisible flag, which it more or less ignores.

Reported by @steve-gregory - it seems that the current version of script/generate_config.py is using a global variable namespace and not utilizing the sections defined in variable.ini to gain namespace isolation.

Investigate the cause and determine if Jinja2 or python config can solve this.

Goal: allow for namespacing in variables.ini.

1. By default, "DISABLE_PUBLIC_AUTH=False" However, when I click on the login, it pops up the loginScreen

2. The modal seems to have some dimensions problem?

Right now, it does not appear that we are making use of caching with the babel loader in webpack.config.js.

It is possible to define a cacheDirectory, which will be the include path if using loaders: ['babel?cacheDirectory'] syntax.

An example:

      {
        test: /\.jsx?$/,
        // Enable caching for improved performance during development
        // It uses default OS directory by default. If you need something
        // more custom, pass a path to it. I.e., babel?cacheDirectory=<path>
        loaders: ['babel?cacheDirectory'],
        // Parse only app files! Without this it will go through entire project.
        // In addition to being slow, that will most likely result in an error.
        include: PATHS.app
     },

source: SurviveJS

Yo there should be a space between "from" and the image name. Screenshot attached.

P.S. waddup

Deployment Ansible SHA is incorrect it just returns the Server sha

If you create a project you must pass a project description. If you pass a description which is whitespace " ". An exception is thrown. We shouldn't require more than a name.

So it complains about the LoginActions is not defined. I think it probably because in the bootstrapper.jsx under the public_site directory: it needs to have:

import actions
actions.LoginActions = require("actions/LoginActions");

I tried to add this two lines, but it next starts complaining about

SplashScreen.jsx?50a0:22 Uncaught TypeError: Cannot read property 'getAll' of undefined(…)

which is from

SplashScreen_getInitialState	@	SplashScreen.jsx?50a0:22 

And about the instance stores. So I don't know what to do next and I make this as an issue but a PR.

Thanks!

webpack dev server will serve up secrets files. We need to use the content assets flag to serve only the troposphere/assets folder. @steve-gregory found this one.

Not sure whether this is in the master branch or not. I am using the u branch.

This looks cool. Just an idea. Not really an issue :)

https://github.com/FormidableLabs/webpack-dashboard

Seems very easy to configure.

From SurviveJS:

If include isn't set, Webpack will traverse all files within the base directory. This can hurt performance! It is a good idea to set up include always. There's also exclude option that may come in handy. Prefer include, however.

Add include to loaders for the source-appropriate locations to improve webpack bundling execution.

Just an idea, maybe support markdown so that syntax on the troposphere would look prettier?

For example:

please edit /root/.ssh/authorized_keys inside the instance...

to become

please edit /root/.ssh/authorized_keys inside the instance...

Just an idea :)

Problem: I would like to be able to see all my running instances in one view, across all projects.

Maybe in a sortable grid? Sort by size, name, machine (image) name, provider, allocation source, status, activity, date started, date ended, ip address... holy shit... That list grew long pretty quickly. That would be cool. Scientists love tables (Excel). Ooh. You could filter by columns as well. And be able to copy & paste it as CSV... Stop me!

Background:

• I was helping Tyson - who has many projects.
• Most of the instances were suspended.
• I was trying to find a running instance - any running instance.
• I was forced to click through all the projects to find one.
• It seems like a tease to show me the number of active and suspended instances on the dashboard, but no way to get to them.
• In the end I found a running instance at the bottom of the truncated log of 'Instance History' at the bottom of the page.

There should not be SelectMenu2 and SelectMenu, this commit (89efcfc) resurrected the original select menu during a hotfix.

The old select menu allowed you to pass a selector function, which would be used to identify the selected element. The new select menu required that you passed it the selected element in addition to the list elements, with the caveat that the selected must be in the list (or it doesn't even make sense to be selected).

The new approach has issues, because the selected element doesn't exist as the same object in the passed list.

In #506, we introduced a login form hosted within Troposphere instead of "deferring" to a authentication service provider (like a compatible OAuth service: CAS, etc).

It has been reported that you cannot logout.

Logout should be a removal of the auth-cookie and a (potential) redirection.

How to reproduce:

send a PR to branch v27 (maybe master as well) and you will see the CI fails due to a permission error.

Solution:

use: group: deprecated-2017Q3

Not sure whether this is a "global" issue yet.

I followed this https://github.com/cyverse/troposphere#quick-feedback
and I still cannot start the dev server:

(troposphere_giji-v29-changes) root@giji-test:/opt/dev/troposphere# npm run serve -- --host 128.31.25.251 --port 8080 --https  --cert /etc/ssl/certs/128.31.25.251.fullchain.crt --key /etc/ssl/private/128.31.25.251.key

> [email protected] serve /opt/dev/troposphere
> webpack-dev-server --hot --env development --content-base troposphere/assets "--host" "128.31.25.251" "--port" "8080" "--https" "--cert" "/etc/ssl/certs/128.31.25.251.fullchain.crt" "--key" "/etc/ssl/private/128.31.25.251.key"

  build [==                  ] 10%events.js:183
      throw er; // Unhandled 'error' event
      ^

Error: listen EADDRNOTAVAIL 128.31.25.251:8080
    at Object._errnoException (util.js:1022:11)
    at _exceptionWithHostPort (util.js:1044:20)
    at Server.setupListenHandle [as _listen2] (net.js:1334:19)
    at listenInCluster (net.js:1392:12)
    at doListen (net.js:1501:7)
    at _combinedTickCallback (internal/process/next_tick.js:141:11)
    at process._tickCallback (internal/process/next_tick.js:180:9)
    at Function.Module.runMain (module.js:686:11)
    at startup (bootstrap_node.js:187:16)
    at bootstrap_node.js:608:3
npm ERR! code ELIFECYCLE
npm ERR! errno 1
npm ERR! [email protected] serve: `webpack-dev-server --hot --env development --content-base troposphere/assets "--host" "128.31.25.251" "--port" "8080" "--https" "--cert" "/etc/ssl/certs/128.31.25.251.fullchain.crt" "--key" "/etc/ssl/private/128.31.25.251.key"`
npm ERR! Exit status 1
npm ERR!
npm ERR! Failed at the [email protected] serve script.
npm ERR! This is probably not a problem with npm. There is likely additional logging output above.

npm ERR! A complete log of this run can be found in:
npm ERR!     /root/.npm/_logs/2018-02-22T18_09_27_659Z-debug.log

I checked the log but there is no meaningful msg to me there. Any help/idea would be much appreciated.

My /etc/nginx/locations/tropo.conf:

(troposphere_giji-v29-changes) root@giji-test:/opt/dev/troposphere# cat /etc/nginx/locations/tropo.conf
# NOTE: This file was automatically generated by troposphere/configure
# to make changes to this file, update your variables.ini
# and re-run troposphere/configure

location /assets {
    # Production Mode: By default,
    # Troposphere serves static files by saving
    # necessary staticfiles and webpack output to:
    #     `/opt/dev/troposphere/troposphere/assets`
    # alias /opt/dev/troposphere/troposphere/assets;

    # Developer mode: For easy development on the GUI
    # Webpack-dev-server can also be used to host assets
    # and refresh the clients page as content is updated
    # To enable webpack-dev-server:
    # - Ensure that the output of webpack-dev-server includes the line:
    #       `Project is running at https://128.31.25.251:8080/`
    # - uncomment the 'proxy_pass' line below.
    # - comment out the 'alias' line above

    proxy_pass https://128.31.25.251:8080;

}

location ~^/(application|maintenance|login|globus_login|oauth2.0/callbackAuthorize|logout|forbidden|version|cf2|tropo-admin|tropo-api|web_shell|web_desktop|allocations|guacamole) {
    proxy_set_header Host $host;
    proxy_set_header X-Real-IP $remote_addr;
    uwsgi_pass unix:///tmp/troposphere.sock;
}

location ~^/cas/(oauth2.0|service) {
    uwsgi_pass unix:///tmp/troposphere.sock;
}

The current loader definitions in webpack.config.js use a perfectly valid but not overall easy to read style with query string arguments.

   { test: /\.woff$/ , loader: "url?limit=10000&mimetype=application/font-woff" },
   { test: /\.woff2$/, loader: "url?limit=10000&mimetype=application/font-woff2" },

source: webpack.config.js

An example of another approach might be to define the url loader as:

  { 
    test: /\.woff$/ , 
    loader: "url",
    query: {
      limit: 10000,
      mimetype: "application/font-woff"
    }
  }