[Bug?] App Search Dashboard Fails To Load After Restarting The Server

[The-Big-K]

I'm trying to install app search on ubuntu 18.04 and was able to successfully setup. Here's an overview of my setup:

1. Installed Elasticsearch and App Search 7.4.2 via apt
2. Added xpack.security.enabled: true to elasticsearch.yml
3. Generated automatic passwords by running bin/elasticsearch-setup-passwords auto and noted down the passwords.
4. Added the following to app-search.yml:
   allow_es_settings_modification: true
   app_search.auth.source: standard
   elasticsearch.username: elastic
   elasticsearch.password: <password generated in step #3>
5. Ran app-search by running bin/app-search

I then noted down the username and default password generated
username: app-search@example.com
password: hnrxxam9khgtq38u

I was able to login to the app-search console. I however then stopped the app-search by pressing CTRL+C and tried to restart with by running bin/app-search.

The server did start; but now I could not get to the dashboard again. The console output was as follows, while the chrome got into endless waiting for page state -

**app-server.1 |** [2019-11-02T13:20:08.019+00:00][6800][2268][app-server][INFO]: [c1d9ba72-f40e-4925-b5f0-56c1e9744054] Started GET "/" for 49.36.27.83 at 2019-11-02 13:20:08 +0000
**app-server.1 |** [2019-11-02T13:20:08.044+00:00][6800][2268][action_controller][INFO]: [c1d9ba72-f40e-4925-b5f0-56c1e9744054] Processing by LocoTogo::HomeController#index as HTML
**app-server.1 |** [2019-11-02T13:20:08.047+00:00][6800][2268][action_controller][INFO]: [c1d9ba72-f40e-4925-b5f0-56c1e9744054] Parameters: {"host"=>"139.59.11.80:3002", "protocol"=>"http"}
**app-server.1 |** [2019-11-02T13:20:08.164+00:00][6800][2268][action_controller][INFO]: [c1d9ba72-f40e-4925-b5f0-56c1e9744054] Redirected to http://139.59.11.80:3002/login
**app-server.1 |** [2019-11-02T13:20:08.180+00:00][6800][2268][action_controller][INFO]: [c1d9ba72-f40e-4925-b5f0-56c1e9744054] Completed 302 Found in 130ms
**app-server.1 |** [2019-11-02T13:20:08.234+00:00][6800][2250][app-server][INFO]: [eb2e7e67-bfa5-4349-aaad-1add172b44ac] Started GET "/login" for 49.36.27.83 at 2019-11-02 13:20:08 +0000
**app-server.1 |** [2019-11-02T13:20:08.245+00:00][6800][2250][action_controller][INFO]: [eb2e7e67-bfa5-4349-aaad-1add172b44ac] Processing by LocoTogo::SessionsController#login as HTML
**app-server.1 |** [2019-11-02T13:20:08.248+00:00][6800][2250][action_controller][INFO]: [eb2e7e67-bfa5-4349-aaad-1add172b44ac] Parameters: {"host"=>"139.59.11.80:3002", "protocol"=>"http"}
**app-server.1 |** [2019-11-02T13:20:08.295+00:00][6800][2250][action_view][INFO]: [eb2e7e67-bfa5-4349-aaad-1add172b44ac] Rendered eui_icons/_close.html (0.4ms)
**app-server.1 |** [2019-11-02T13:20:08.306+00:00][6800][2250][action_view][INFO]: [eb2e7e67-bfa5-4349-aaad-1add172b44ac] Rendered eui_icons/_lock.html (0.9ms)
**app-server.1 |** [2019-11-02T13:20:08.317+00:00][6800][2250][action_view][INFO]: [eb2e7e67-bfa5-4349-aaad-1add172b44ac] Rendered loco_togo/app/views/loco_togo/sessions/login.html.rb (47.4ms)
**app-server.1 |** [2019-11-02T13:20:08.320+00:00][6800][2250][action_controller][INFO]: [eb2e7e67-bfa5-4349-aaad-1add172b44ac] Completed 200 OK in 71ms (Views: 52.2ms)

I can confirm that I've repeatedly reproduced this issue. Can someone from Elastic team confirm that this is an issue? Or is there something wrong with my configuration?

[The-Big-K]

Update 1: I've sent the diagnostics folder to support at swiftype dot com [#00405240]. I hope someone will look into this issue and respond.

Update 2: Here's an interesting observation I made. After restarting the server, if I just let chrome try to access the dashboard - it takes Chrome about 20-25 minutes of 'waiting' to fetch the page. I'm not sure what really happens in those ~25 minutes; because console doesn't output anything when the page actually loads. What could possibly be taking ~25 minutes to complete in the backend?

[orhantoy]

Thanks for reporting this @The-Big-K

From the logs it seems like the response times are in the milliseconds, so the 20-25 minute waiting time seems weird. Wondering if this can be a browser related thing; do you experience the same behavior with other browsers than Chrome?