Usage with ruby app

[amnesia7]

Hi @ncdc, @rbrower3, @oliviermi

I was wondering if someone could explain how this, or the forked https://github.com/rbrower3/openshift-elasticsearch-cartridge cartridge, works.

I'd like to use a scalable elasticsearch cartridge with my ruby app but once I setup my app using the script supplied in the readme I can add a database cartridge but the ruby framework cartridge is no longer available. Is that something to do with it being considered as a web cartridge or something (preventing another language framework from being applied)?

Do I need to set this up as an app by itself and then put my ruby+postgres cartridges together as a different app that connects to this ES app?

If so, how do you do that? Can the Openshift variables be used for connecting securely between my apps?

I am using the elasticsearch-rails and elasticsearch-model ruby gems (see https://github.com/elasticsearch/elasticsearch-rails) and setting my local connection using:

# Connect to specific Elasticsearch cluster
ELASTICSEARCH_URL = ENV['ELASTICSEARCH_URL'] || 'http://localhost:9200'

Elasticsearch::Model.client = Elasticsearch::Client.new host: ELASTICSEARCH_URL

I hope you can help me. Openshift sounds really good but I'd like to setup my app on the right footing.

Hopefully the answer will be of use to others in the same (or similar) situation as well.

Thanks

[ncdc]

@amnesia7 this cartridge is a web framework cartridge, so it has to live by itself and your ruby app will need to be a separate app.

You can definitely set user environment variables using rhc env set in your ruby app to tell it the URL for your Elasticsearch app.

1 big caveat is that there's currently no security implemented for the Elasticsearch cartridge, so it would be exposed publicly and accessible by anyone knowing the URL. In its current form, this cartridge is essentially a proof of concept. If you or someone else wanted to contribute some additional security features, that would be a welcome addition!

HTH,
Andy

[amnesia7]

Thanks for the quick reply @ncdc

I think it's a little beyond my knowledge to advise on the security issue but hopefully someone else will come forward with a solution(s).

With that in mind I'm just wondering what my best way forward is then in the meantime.

Any thoughts / suggestions so that I can get my app going without limiting its future potential?
I'm a little new to Openshift and don't want to set my app down the wrong route.

Thanks

Col

[ncdc]

@amnesia7 you can use this cartridge in its current form. Create a scalable Elasticsearch app. Create a Ruby (or whatever) app. Connect them together by setting an environment variable in your Ruby app. That should be enough to get you started, right?

[amnesia7]

And hope that, until a more secure option comes along, no-one figures out the url?

[ncdc]

You could try using this plugin, which replaces the stock http server with
a configurable Jetty one: https://github.com/sonian/elasticsearch-jetty. It
includes support for basic authentication.

On Tuesday, April 29, 2014, amnesia7 notifications@github.com wrote:

And hope that, until a more secure option comes along, no-one figures out
the url?

—
Reply to this email directly or view it on GitHub.<
https://ci4.googleusercontent.com/proxy/Y2y94ITNgXn3DSR7YcQWTHXKpOzBbfPytanyYWSOEct5G0JmGKxQ8Dg_hK5Zu7R8GXXVU3wwNslUcZa5Vw8AYd8GHvsYKZ-Pv55DTYbEcv2yOPCPCF_yVGUyYtYuLMROJSuhwzv8W-af4sPcb8aXB_Mg2DhFRppnxgqIlzB4TOr25ytLwcYGLzrm1xMRUg60PYCgZMlf2QEq0SsOULFRUteYZb8M_NWpxprHZyHWUjBmbtkqm83Bic-6Dd71zjclM5TLAMxz5y0NYaaVq1_1OdpN=s0-d-e1-ft#https://github.com/notifications/beacon/354__eyJzY29wZSI6Ik5ld3NpZXM6QmVhY29uIiwiZXhwaXJlcyI6MTcxNDQyMzQ3MiwiZGF0YSI6eyJpZCI6MzEwNzk1OTZ9fQ==--57c0e7f13b8fd83d18f0d1496780ada43c9d8057.gif

[amnesia7]

Thanks, I'll check that out.

I guess the app will live on the normal website domain but the ES app will live on an xxxx-xxxx.rhcloud.com address so if this rhcloud url is complex enough then the only way to find the ES url is via the ruby code connecting to it.

[caruccio]

There is also this version with support for CORS: http://www.emergi.net/elasticsearch/plugins.html

[amnesia7]

Thanks @caruccio.

Once my app has been created using this cartridge script have you any idea how the app would go on for being updated, eg if this cartridge script was updated with any security features?

[amnesia7]

Looks like I might be able to pass the username and password through using the ruby gem (https://github.com/elasticsearch/elasticsearch-ruby/tree/master/elasticsearch-transport#setting-hosts) so I guess I just need to figure out what jetty is and how to apply it.

[amnesia7]

With the support for cors, I guess the Emergi version (http://www.emergi.net/elasticsearch/plugins.html) is probably the one I'd be better going with.

I noticed the plugins.txt file generated by the cartridge script that advises to uncomment any plugins to get them installed. Jetty isn't one of the plugins listed; is that because the ones listed are special or does it only list a select few plugins and I just need to add pmusa/elasticsearch-jetty as a new line in there and then commit and push?

[amnesia7]

Nevermind, I've just seen in the readme (https://github.com/ncdc/openshift-elasticsearch-cartridge#plugins) that I just need to add the plugin to the plugins.txt file, commit and push.

[ncdc]

@amnesia7 do you have any other questions at the moment?

[amnesia7]

@ncdc do I need to just use the searchapp's URL and port numbers as manually set env variables?

Is there any way to better reference the url of another app in my control rather than just manually copying it?

Thanks

[amnesia7]

@ncdc @caruccio I've set the ES url as http://appname-whatever.rhcloud.com:9200 (not applied jetty yet) but calling the search page in the app is timing out (Gateway timed out error 504). I'm just playing with a free account at the moment so the ES app had idled so I had to restart it.
Does that affect whether ES is running or not or does it just start running itself?

I haven't actually managed to confirm it is running yet since creating it so I'm not sure if it actually has run yet or not. How can I check if it is running and do I need to start it manually somehow?

Thanks

[caruccio]

Hi.
AFAIK, an idle app is automatically started once a connection hits openshift, so this is unlikely to be an issue.

Regarding port number (9200), did you changed the cartridge's listening port? The default, externally available, port is 80, not 9200.