I have followed the following tutorial: https://devcenter.heroku.com/articles/queuing-ruby-resque and it doesn't state anything about AUTH with Redis.
When I try to open the resque-web to manage the workers, I get an Internal Error, also when trying this, inside the Heroku rails console:
irb(main):001:0> Resque.queues
I get this:
Redis::CannotConnectError: Error connecting to Redis on 127.0.0.1:6379 (ECONNREFUSED)
So, I am wondering what should I do to make this work correctly. DO I need to set up any kind of auth? If so, where? I am using Redis Cloud addon.
Thanks
It looks like you're not initializing resque. To do so, you need something like the following:
# config/initiazlizers/resque.rb
uri = URI.parse ENV["REDISCLOUD_URL"]
Resque.redis = Redis.new host:uri.host, port:uri.port, password:uri.password
This initializer reads your URL from the Heroku environment variable that was set by the rediscould add-on, parses it, and passes it to Resque so it connects and authenticates with the redis server.
Related
Before, I added code to a file called config/initializers/remote_publishers.rb which set up a connection to RabbitMQ using the Bunny gem on server startup.
However, this is now also executed when running rails c, rails g model SomeModel foo:integer, rails db:migrate etc.
For this app, the RabbitMQ-connection only makes sense when rails is started using rails s(erver).
What is the proper way to conditionally execute this code? Is there a way to see if Rails is starting as server, or only as task-runner?
What web server are you using? On Puma, for example, you can use
on_worker_boot do
# Establish RabbitMQ connection
end
Another possibility might be to check if defined?(Rails::Server) in your initializer: this should only be true when running in the context of the web server.
Im using redis in on of my projects and have a initialiser with redis = Redis.new and using the redis gem. Problem is that if Redis is not running I cannot do anything like a simple database migration.
Is there a more elegant way to handle using redis so that my application throws an error instead of just not working when redis is not running?
Im using process monitoring to keep redis running correctly and monitor its memory/cpu but still if it does not run all things break and look for a more elegant way. if possible
EDit:
This is my initialiser
$redis = Redis.new
heartbeat_thread = Thread.new do
while true
$redis.publish("heartbeat","thump")
sleep 30.seconds
end
end
at_exit do
# not sure this is needed, but just in case
heartbeat_thread.kill
$redis.quit
end
Why do you create a connection to redis before doing any request ?
It seems that you dont need that and I suggest you do connect to redis only when you actually have to send some request to it.
I am not very familiar with redis ruby client but I am pretty sure you can achieve this easily.
I have to use websockets in my rake task and for that I changed my event.rb to
config.synchronize = true
# Uncomment and edit to point to a different redis instance.
# Will not be used unless standalone or synchronization mode
# is enabled.
config.redis_options = {:host => 'localhost', :port => '3000'}
and when I start my rails server I get this error:
! Invalid request
Exiting
/usr/local/rvm/gems/ruby-1.9.3-p194#socialmail/gems/redis-3.0.4/lib/redis/connection/synchrony.rb:115:in `read': Got 'Protocol error, got "H" as reply type byte' as initial reply byte. If you're in a forking environment, such as Unicorn, you need to connect to Redis after forking. (Redis::ProtocolError)
What am I doing wrong?
Thanks
Hey thanks for the question, I personally couldn't get a rake task of mine to post to a websocket channel I had open on my rails server. Your synchronize command helped (along with starting a Redis server locally).
Your problem through - seems like you're pointing to 3000. Is that your rails server or the Redis instance? If you're running it locally, I'd omit that line.
in ruby on rails console 'net/http' works, but in controller it doesn't and gives timeout error.
require 'net/http'
uri = URI('http://localhost:3000/api_json.json')
json = Net::HTTP.get(uri)
parsed_json = ActiveSupport::JSON.decode(json)
Most likely you're using default Webrick server, that serves one request a time. So, from console it works fine, but fails when you try to call it from controller (when the Webrick worker is already busy).
You can try to setup and run another server like unicorn or thin, or run two Webrick instances on different ports:
rails server
rails server -p 3001
and go to localhost:3001
#dimuch's solution might have solved your issue, but it might help someone facing similar situation. I will explain the issue, and the solution in detail (extension of #dimuch's solution).
Issue:
You might have a controller like some:"/test_controller/test_method", and you might want to call a method in a controller, like /api/v1/some_test_api, and facing error like Completed 500 Internal Server Error in 60004.4ms
[27580c5c46770812c550188346c2dd3e] [127.0.0.1] [/xauth_test/sanity_oauth_login]
Timeout::Error (Timeout::Error):
Solution:
As said by #dimuch, "
Most likely you're using default Webrick server, that serves one request a time.....". 1. You need to run the application on different ports, like
rails s -p 3000, and rails s -p 3001, then make the request from 3001.
If you face an issue like "A server is already running. Check /tmp/pids/server.pid. Exiting", then try running rails s -p 3001 -P PROCESS_ID.
2. Use other server's like Unicorn, or Puma.
Note: If you want it for just testing purpose in local, then I would suggest to go with the first solution, which is easy and simple. I am sorry for poor English, and I found most of solutions from other stack overflow pages, and websites, which I am attaching (links for refs) below, and sorry if I missed some one or some thing to refer. Hope this helps someone.
Refs:
For running multiple instances:
Running multiple instances of Rails Server
Similar errors and way they are handled:
Rails HTTParty Getting Timeout::Error
Faraday timeout error with omniauth (custom strategy)/doorkeeper
Strange Timeout::Error with render_to_string and HTTParty in Controller Action
Configuring Unicorn &Puma:
http://vladigleba.com/blog/2014/03/21/deploying-rails-apps-part-3-configuring-unicorn/
https://github.com/puma/puma
I have been playing around with Cramp to do some real time pushing of information in an app. Its all working great locally but when I push off to heroku I seem to be having issues with the ports.
I have a socket set up in cramp which inherits from websocket
class LiveSocket < Cramp::Websocket
and I also have a cramp action called home which basically just renders some erb for the home page
class HomeAction < Cramp::Action
in my route file I set up the following and also a static file server
Rack::Builder.new do
puts "public file at #{File.join(File.dirname(__FILE__), '../public')}"
file_server = Rack::File.new(File.join(File.dirname(__FILE__), 'public'))
routes = HttpRouter.new do
add('/').to(HomeAction)
get('/LiveSocket').to(LiveSocket)
end
run Rack::Cascade.new([file_server, routes])
end
Then on the client end the javascript connects to
var ws = new WebSocket("ws://<%= request.host_with_port %>/LiveSocket");
As I say locally it all works. We connect and start receiving notifications from the server. On heroku we run thin on the Cedar stack and have a profile which looks like
web: bundle exec thin --timeout 0 start -p $PORT
When I load up the site the page itself loads fine but on trying to connect the websocket I get an error which says
servername.herokuapp.com Unexpected response code: 200
I am guessing this has something to do with how heroku routes its requests but I do know that you can run a node.js websocket server on heroku so figure there must be a way to get this working too.
Thanks in advance for any help.
cheers
stuart
I don't think Heroku supports websockets :( http://devcenter.heroku.com/articles/http-routing#the_herokuappcom_http_stack