What is worker , dyno and zero-downtime deploys in heroku - ruby-on-rails

These three terms have given a lot of importance in understanding different app server in heroku tutorials but I can't understand the purpose and definition of these three terms.
Can anybody have info about that. Kindly share
Thanks

The Heroku reference guide has a lot of information on all of this, and lots more, but in answer to your question;
A dyno is effectively a small virtual server instance set up to run one app (it's behind an invisible load balancer, so you can have any number of them running side-by-side). You don't need to worry about the server admin side of things, as it just takes your source code from a Git push and runs it.
A worker is a type of dyno, usually designed to process tasks in the background (in contrast to a web dyno, which just serves web pages. For example, Rails has ActiveJob, which plugs into something like Resque or Sidekiq, completes tasks which would slow down the web interface if it has to complete them, like sending e-mails, or geocoding addresses.
Zero-downtime deploys is really marketing speak for "if you push your code, it will wait until the new version is up and running before swapping the web interface to use it". It means you don't need to do anything, and your web app won't go offline while it is switched over to.

Related

Ruby on Rails on few servers

I have a big application. One of the part of this is highload processing with user files. I decide to provide for this one dedicate server. There will be nginx for distribution content and some programs (non rails) for processing files.
I have two question:
What better to use on this server? (Rails or something else, maybe Sinatra)
If I'll use Rails how to deploy? I can't find any instruction. If I have one app and two servers how to deploy it and delegate task for each other?
ps I need to authorize user on both servers. In Rails I use Devise.
You can use Rails for this. If both servers will act as a web client to the end user then you'll need some sort of load balancer in front of the two servers. HAProxy does a great job on this.
As far as getting the two applications to communicate with each other, this will be less trivial than you may think. What you should do is use a locking mechanism on performing the tasks. Delayed_job by default will lock a job in the queue so that any other works will not try and work on the same job. You can use callbacks from ActiveJob to notify the user via web sockets whenever their job is completed.
Anything that will take time or calling an external API should usually be placed into a background processing queue so that you're not holding up the user.
If you cannot spin up more than the two servers, you should make one of them the master or at least have some clear roles of the two servers. For example, one server may be your background processing and memcache server while the other is storing your database and handles your web sockets.
There are a lot of different ways of configuring the services and anything including and beyond what I've mentioned is opinionated.
Having separate servers for handling tasks is my preference as it makes them easier to manage from a Sys Admin perspective. For example, if we find that our web sockets server is hammered, we can simply spin up a few more web socket servers and throw them into a load balancer pool. The end user would not be negatively impacted from your networking changes. Whereas, if you have your servers performing dual roles outside of your standard Rails installation, you may find yourself cloning and wasting resources. Each of my web servers usually also perform background tasks on low-intermediate priority queues while a dedicated server is left for handling mission critical jobs.

RabbitMQ with EventMachine and Rails

we are currently planning a rails 3.2.2 application where we use RabbitMQ. We would like to run several kind of workers (and several instances of a worker) to process messages from different queues. The workers are written in ruby and are laying in the lib directory of the rails app.
Some of the workers needs the rails framework (active record, active model...) and some of them don't. The first worker should be called every minute to check if updates are available. The other workers should process the messages from their queues when messages (which are send by the first worker) are present and do some (time consuming) stuff with it.
So far, so good. My problem is, that I only have little experiences with messaging systems like RabbitMQ and no experiences with the rails interaction between them. So I'm wondering what the best practices are to get the two playing with each other. Here are my requirements again:
Rails 3.2.2 app
RabbitMQ
Several kind of workers
Several instances of one worker
Control the amount of workers out of rails
Workers are doing time consuming tasks, so they have to be async
Only a few workers needs the rails framework. The others are just ruby files with some dependencies like Net or File
I was looking for some solution and came up with two possibilities:
Using amqp with EventMachine in a new thread
Of course, I don't want my rails app to be blocked when a new worker is created. The worker should run in another thread and do its work asynchronously. And furthermore, it should not start a new instance of my rails application. It should only require the things the worker needs.
But in some articles they say that there are some issues with Passenger. And another fact that I don't like is, that we are using webbrick for development and we ought to include workarounds for that too. It would be possible to switch to another webserver like thin, but I don't have any experience with that either.
Using some kind of daemonizing
Maybe its possible to run workers as a daemon, but I don't know how much overhead this would come up with, or how I can control the amount of workers.
Hope someone can advise a good solution for that (and I hope I made myself clear ;)
It seems to me that AMQP is a big shot to kill your problem. Have you tried to use Resque? The backed Redis database has some neat features (like publish/subscribe and blocking list pop) which make it very interesting as a message queue, and Resque is very easy to use in any Rails app.
The workers are daemonized, and you decide which worker of your pool listens to which queue, so you can scale each type of job as needed.
Using EM reactor inside a request/response cycle is not recommended, because it may conflict with an existing event loop (for instance if your app is served by thin), in any case you have to configure it specifically for your web server, OTOS it may be interesting to have an evented queue consumer, if your jobs have blocking IO and are not processor-bound.
If you still want to do it with AMQP, see Starting the event loop and connecting in Web applications and configure for your web server accordingly. Or use bunny to push synchronously in the queue (and whichever job consumer you deam useflu, like workling for instance)
we are running slightly different -- but similar technology stack.
daemon kit is used for eventmachine side of the system... no rails, but shared models (mongomapper & mongodb). EM is pulling messages off the queues, and doing whatever logic is required (we have ruleby in the mix, but if-then-else works too).
mulesoft ESB is our outward-facing message receiver and sender that helps us deal with the HL7/MLLP world. But in v1 of the app, we used some java code in ActiveMQ to manage HL7 messages.
the rails app then just serves up stuff for the user to see -- again, using the shared models.

How should I build a VPS that will host multiple small Rails applications?

Up until now, I've always built my VPS dedicated to running just a single application in multiple instances, mostly with Unicorn. That way I could set up the whole environment to fit perfectly for that one specific application and be happy with it.
But now I need to build a VPS, that will host multiple small Ruby applications. Some of them will be Rails and some Sinatra. They will have basically zero traffic (under 100 visits per day), which means I don't even need multiple instances of a single app.
I don't really have experience with other servers than unicorn + nginx, but what I think I need would look something like this.
request to app1, gets loaded into memory and serves the request
request to app2, gets loaded into memory and serves the request
request to app3, there is not enough free memory
app1 gets killed before the app3 is booted to serve the request
I know this isn't an exactly perfect scenario, but imagine having a 10 or 20 small apps on a single server, where each app gets 5 hits a day. They don't exactly need to be up and running at all times.
As far as I know, Heroku does this with their free tier, where Dynos get killed after some idle time, and then they get loaded back up when a request comes in. That's basically what I need to do on my own server.
I would recommend using Apache + Passenger. Passenger by default loads the application only when you need it, e.g. the first request will take a bit longer (actually as long as it takes to load your framework).
If the application is idle for some predefined time, it will be removed from memory.
Setup is very easy and adding new applications is just adding one line in your apache configuration.

Zero downtime on Heroku

Is it possible to do something like the Github zero downtime deploy on Heroku using Unicorn on the Cedar stack?
I'm not entirely sure how the restart works on Heroku and what control we have over restarting processes, but I like the possibility of zero downtime deploys and up until now, from what I've read, it's not possible
There are a few things that would be required for this to work.
First off, we'd need backwards compatible migrations. I leave that up to our team to figure out.
Secondly, we'd want to migrate the db right after a push, but before the restart (assuming our migrations are fully backwards compatible, this should not affect anything)
Thirdly, we'd want to instruct Unicorn to launch a new master process and fork some workers, then swap the PIDs and gracefully shut down the old process/workers
I've scoured the docs but I can't find anything that would indicate this is possible on Heroku. Any thoughts?
I can't address migrations, but the part about restarting processes and avoiding wait time:
There is an beta feature for heroku called preboot. After a deploy, it boots your new dynos first and waits a while before switching traffic and killing the old ones:
https://devcenter.heroku.com/articles/labs-preboot/
I also wrote a blog post that has some measurements on my app's performance improvements using this feature:
http://ylan.segal-family.com/blog/2012/08/27/deploy-to-heroku-with-near-zero-downtime/
You might be interested in their feature called preboot.
Taken from their documentation:
This feature provides seamless deploys by booting web dynos with new code before killing existing web dynos.
Some apps take a long time to boot up, and this can cause unacceptable delays in serving HTTP requests during deployment.
There are a few caveats:
You must have at least two web dynos to use this feature. If you have your web process type scaled to 1 or 0, preboot will be disabled.
Whoever is doing the deployment will have to wait a few minutes before the new code starts serving user requests; this happens later than it would without preboot (but in the meanwhile, user requests are still served promptly by old dynos).
There will be a short period (a minute or two) where heroku ps shows the status of the new code, but user requests are still being served by old code.
There is much more information about it, so refer to their documentation.
It is possible, but requires a fair amount of forward planning. As of Rails 3.1 there's three tasks that need carrying out
Upload the new code
Run any database migrations
Sync the assets
Uploading code and restarting is fairly straightforward, the main problem lies with the other two, but the way round them is the pretty much the same.
Essentially you need to:
Make the code compatible with the migration you need to run
Run the migration, and remove any code written specifically for it
For instance, if you want to remove a column, you’ll need to deploy a patch telling ActiveRecord to ignore it first. Only then you can deploy the migration, and clean up that patch.
In short, you need to consider your database and the code compatability an work around them so that the two can overlap in terms of versioning.
An alternative to this method might be to have two versions of the application running on Heroku at the same time. When you deploy, switch the domain to the other version, do the deploy, and switch it back again. This will help in most instances, but again, database compat is an issue.
Personally, I would say that if your deployments are significant to require this sort of consideration, taking parts of the application offline are probably the safest answer. By breaking up an application into several smaller applications can help mitigate this and is a mechanism that I use regularly.
No - this is currently not possible using Unicorn on Heroku cedar. I've been bugging Heroku about this for weeks.
Here was Heroku Support's reply to my email on March 8, 2012:
Hi, you could enable maintenance mode when doing a deploy, at least your users would see a maintenance page instead of an error, and also request queue wouldn't build up.
We're definitely aware this is a pain and we're working to offer rolling / zero-downtime deploys in the future. We have no ETA to announce, though.

Best way to run rails with long delays

I'm writing a Rails web service that interacts with various pieces of hardware scattered throughout the country.
When a call is made to the web service, the Rails app then attempts to contact the appropriate piece of hardware, get the needed information, and reply to the web client. The time between the client's call and the reply may be up to 10 seconds, depending upon lots of factors.
I do not want to split the web service call in two (ask for information, answer immediately with a pending reply, then force another api call to get the actual results).
I basically see two options. Either run JRuby and use multithreading or else run several regular Ruby instances and hope that not many people try to use the service at a time. JRuby seems like the much better solution, but it still doesn't seem to be mainstream and have out of the box support at Heroku and EngineYard. The multiple instance solution seems like a total kludge.
1) Am I right about my two options? Is there a better one I'm missing?
2) Is there an easy deployment option for JRuby?
I do not want to split the web service call in two (ask for information, answer immediately with a pending reply, then force another api call to get the actual results).
From an engineering perspective, this seems like it would be the best alternative.
Why don't you want to do it?
There's a third option: If you host your Rails app with Passenger and enable global queueing, you can do this transparently. I have some actions that take several minutes, with no issues (caveat: some browsers may time out, but that may not be a concern for you).
If you're worried about browser timeout, or you cannot control the deployment environment, you may want to process it in the background:
User requests data
You enter request into a queue
Your web service returns a "ticket" identifier to check the progress
A background process processes the jobs in the queue
The user polls back, referencing the "ticket" id
As far as hosting in JRuby, I've deployed a couple of small internal applications using the glassfish gem, but I'm not sure how much I would trust it for customer-facing apps. Just make sure you run config.threadsafe! in production.rb. I've heard good things about Trinidad, too.
You can also run the web service call in a delayed background job so that it's not hogging up a web-server and can even be run on a separate physical box. This is also a much more scaleable approach. If you make the web call using AJAX then you can ping the server every second or two to see if your results are ready, that way your client is not held in limbo while the results are being calculated and the request does not time out.

Resources