Can't connect Weaviate instance to Weaviate playground (receiving panick's)

[StudioSpindle]

This is a follow-up on the discussion in the Slack general channel.

The service seems to be restarting once you try to make a connection.

Steps to reproduce:

• Installing the Weaviate Playground on your local machine Install Manual
• Installing a Weaviate as described in the docs of Weaviate
• In the root folder for Weaviate Playground run $ yarn start
• In the root folder of the Weaviate instance run $ docker-compose up -d && docker-compose logs -f weaviate

Then wait for it to finish. To verify if it's finished correctly, try the following in a curl or browser:

• Verify whether the Weaviate is working: http://localhost:8080/weaviate/v1/meta <- This should show an empty Weaviate
• Then verify this URL: http://localhost:8080/weaviate/v1/graphql <- This should give a 405 response "method GET is not allowed, but [POST] are"

If both are true then:

• Open the Chrome browser on http://localhost:3000/. This opens the index page for the playground
• Then in the Weaviate URL paste: http://localhost:8080/weaviate/v1/graphql.

Here is where the error occurs.

It produces one of two things:

1. After clicking on connect immediately the error is thrown: "Network error: Failed to fetch". If you then in another tab try to open http://localhost:8080/weaviate/v1/graphql again that suddenly isn't working anymore.
2. Other times it starts loading the app. This hangs for quite some time. Finally it comes back and throws and

Debugging info:

As stated in the Weaviate documentation logs can be used to see the type of error. This is a snippet of the feedback containing the 'panicks':

Docker was running:

What fixed it for me:

Using another sandboxed Weaviate URL which in the local Playground did work.

The solution was to add 127.0.0.1 weaviate001 to the hosts file on the local machine and flushing the DNS. ($ sudo nano /etc/hosts and $ sudo killall -HUP mDNSResponder. Depending on the OSX you're running this command could differ)

And then restart the docker instance ($ docker-compose down) and then bringing it up again ($ docker-compose up).

[etiennedi]

closed through #878