Installing Unicorn

So we've got engine X forwarding our apps traffic to Puma, 0:00 now let's get it working with unicorn instead. 0:03 So I'm going to quit out of the rail server here. 0:06 Unicorn is distributed as a Ruby gem, so 0:10 to install it am going to edit our applications gem file. 0:12 Down here at the bottom I'm going to add a new production group. 0:18 So that this only gets installed on our production server. 0:22 So group symbol production as well as block. 0:26 And in this block we will specify the gems we want to install in production so 0:33 we want to install the gym unicorn. 0:36 And we're gonna use the total walk of syntax here to lock that 0:42 two minor version 5.2, that's the current one at the time of recording. 0:46 Okay Ctrl all to write that out for in the filename Ctrl+x to exit and 0:52 now we need to run bundle install to actually install it. 0:56 It will download Unicorn and all its dependencies. 1:04 And when it returns to the prompts, Unicorn will be installed. 1:09 Now, we need to set up a configuration file for Unicorn. 1:12 Just like with Engine X, we're going to set it up within our config subdirectory. 1:15 So we use the nano editor and will create a new file in the config directory, 1:20 we're going to name it unicorn.rb. 1:25 We name the other file nginx.conf because it was in nginxs format but 1:30 unicorn configuration files are plain ruby files so we give this one a .rb extension. 1:36 Okay, so we've got a few directives that we need to set up here first. 1:42 The first is going to be the working directory option working_directory. 1:45 Again this is basically a Ruby method call we're going to pass at string and 1:54 we'll specify the path to our Rails app. 1:59 So that's home/deploy/guestbook. 2:01 We're gonna need that route a few more times, so I'm gonna copy that here. 2:11 Then we need to specify a file that unicorn will write out its process ID to 2:18 so that we can shut the server down if we need to. 2:22 So I'm going to use the pid directive and 2:25 specify that it should to a file within our guest book app directory. 2:27 Within a PID's sub directory and we're gonna name that fur 2:33 unicorn.pid. 2:38 Next we have a couple options for logging. 2:44 So I'm going to say that errors should be logged to 2:48 a particular path with stderr_path. 2:51 We're going to have it write to a file in our guestbook app directory again, 2:55 within a unicorn sub directory. 3:02 We'll name that unicorn.log And 3:07 then we'll use the STD out path directive to have it 3:12 write standard output to the same location to the same log file. 3:15 Okay, this next part is crucial. 3:21 So engine X is going to communicate with unicorn and Gen X 3:23 is one process running on our Linux server and the Unicorn server is another process. 3:28 In Linux processes communicate with each other through sockets which is kind of 3:35 like a file except that one process writes into this virtual file and 3:39 another process reads out of it. 3:43 So we're going to give unicorn the listen directive and 3:45 tell it where it should listen for requests. 3:49 And we're gonna give it the name of the socket to listen on. 3:53 So we'll create a string that points to a path in the tmp root directory. 3:55 And we're going to name that socket unicorn.guestbook.sock. 4:02 Okay, so there's the socket that unicorn is going to listen on will Set up ngenx to 4:10 write to that socket later. 4:14 Then we need to specify how many unicorn processes it should start up. 4:18 For right now we're just going to say worker_processes and 4:22 we'll just give it a number 2 for right now. 4:29 And finally we should specify how long it should give one of those processes to 4:32 respond before it terminates it. 4:36 Be that with the timeout directive and we'll set that to 30 seconds for 4:38 right now. 4:42 Okay, let's save the file an exit. 4:47 Okay and with that configuration in place we can try running the unicorn. 4:50 We're going to want to do that in the production environment, so 4:53 I'll specify rails equals production. 4:56 Since we installed the unicorn through, 5:05 bundler we're going to want to run it using bundlers So we use bundle exec. 5:06 The executable is named unicorn. 5:12 And next we're going to use the -c command line flag to specify a configuration 5:17 file it should load. 5:21 We're going to want to use the config directory unicorn.rb And 5:25 finally we'll use dash capital D to specify that it should run as a demon, 5:31 that is, it should continue running in the background. 5:35 If we try running this, we'll see an error. 5:39 It says that the directory for the P I D file isn't writeable and 5:43 that's because we forgot to create that directory, so let's do that now. 5:47 So, we'll create a, we'll make a directory and pids. 5:52 Here within our rails app directory. 5:56 If we try running it again and 6:00 will fail again because it says the directory for unicorn.log isn't writable. 6:02 We forgot to create a unicorn subdirectory as well. 6:06 So, add directory unicorn. 6:10 Okay let's try one more time. 6:14 And this time it kicks off successfully. 6:19 We've got unicorn running now but 6:22 if we try to visit our Rails app we'll see 502 Bad Gateway again. 6:24 This is because engine x isn't set up to communicate with unicorn via the socket we 6:27 told unicorn to listen on so we're going to need to fix that now. 6:31 So I'm going to go back into the nginx configuration file. 6:36 config/nginxx.conf. 6:41 And specify a socket that nginx should write requests out to. 6:45 We're going to use the upstream directive. 6:50 We're going to name this upstream location unicorn. 6:54 And then options for 6:58 that upstream location are here within brackets, curly brackets. 7:00 And here we're going to use the server directive and 7:04 we're going to specify the Unix socket that it should write out to. 7:08 So unix:/tmp/unicorn.guestbook.sock. 7:12 Which is the same path that we specified within the unicorn configuration. 7:22 And here at the end of the line we're going to add one more option, 7:28 we're going to say fail underscore timeout equals zero. 7:33 That fail timeout equals zero causes engine X to always retry request if 7:40 it fails. 7:44 We need this because a unicorn worker might get shut 7:45 down by the master process if it takes too long to respond and 7:47 we don't want to lose the request so that will cause engine X to retry it. 7:51 And now instead of forwarding requests that it gets onto local host port 3000, 7:56 we want to get it to forwarded on to our upstream location. 8:00 So we're going to have it forward it to another HTTP server. 8:04 But we're going to have it go to the unicorn upstream location. 8:09 So we use the same name as our upstream directive http://unicorn. 8:12 Okay write that out save it, and 8:17 as always we need to restart engine x. 8:22 It will want our sudo password. 8:28 And now if we go to our host IP and try to reload the root path. 8:35 We'll see our Rails app. 8:39 Nginx is successfully communicating with Unicorn using the Unix socket. 8:41