Getting Started with Parse Server, Heroku and MongoLabs

Last week, Parse announced they would be shutting down their hosting by January 28th 2017, and like just about every other developer out there, I immediately started playing around with their open sourced Parse Server, and setting it up on Heroku.

For the record, I actually host most of my apps on Flybase (for data), Github Pages or S3 (for frontend) and Heroku for quick node.js apps. I also use Digital Ocean or AWS EC2 for other hosting, but I wanted to demonstrate how to quickly host Parse Server on Heroku first, I’ll show how to set Parse Server up on Digital Ocean and AWS EC2 later in other posts.

Parse Server is actually a pretty nicely written library, and I’ve even forked it to include Flybase support instead of Mongo, as well as having a version that supports multiple apps in one install, but today we’ll look at simply getting Parse Server running on Heroku.


Some Background

Until this week, if you were a Parse user, then you did not need to know how your client side code interacted with your Parse database, and even now, you just need a vague understanding of how it works.

But, there is a lot happening behind the scenes when you host your own Parse Server. So, now that you are going to be hosting it yourself, you should understand how some of it works.

First of all, most of the more difficult work is handled inside the Client side library, and then sent to the server side code as HTTP requests. None of this changes, the only change is where you talk to Parse rather than how you talk to Parse.

That means that all you have to keep in mind when explaining how your self-hosted Parse Server works will be these three sections:

  1. Client code from which you make requests to the server
  2. Server-side code that receives and processes these requests
  3. A Database where the server-side code stores and retrieves data from.

Previously, Parse took care of all of this, but now, we have to setup our mongo database and web server. Luckily, Heroku and Mongolabs work together nicely and almost automatically in this process.

Let’s get started already!

First, You will need to create an account with Heroku, so go create one now if you haven’t already.

Then you’ll need to download the latest version of the Heroku tool-belt here. Once the download has finished you will be able to use heroku commands from the command line. (You’ll see why you want to do this in step 3)

1. Push the button

In your browser, go to: https://github.com/ParsePlatform/parse-server-example, this is a sample node.js app that uses express.js and parse server to

You’ll see a purple button here that says Deploy to heroku.

Press it, press it right now.

2. Prepare the app with Heroku

Heroku’s one button deploys are handy for creating a new app, when you clicked the button in step 1, it took you to a page inside Heroku that asked you to name your app, give it a nice random name, or Heroku will name it for you.

MongoLab was also auto-selected as a necessary add-on, so that’s taken care of for you.

Further down the page, it will give you a list of options:

  1. Keep /parse as your PARSE_MOUNT.
  2. Enter any name you want for APP_ID.
  3. Enter any key you want for the MASTER_KEY.

Now press the deploy for free button, and wait a few minutes.

Once done, you can enter your parse server app by the https://NAME_OF_YOUR_HEROKU_APP.herokuapp.com/parse, if you see a message that says:

I dream of being a web site.

This is placeholder text, and tells you it worked.

3. Modifying your parse server

From the terminal, you want to use the Heroku toolbelt I mentioned before step 1 to download your Heroku’d Parse server and make changes.

Login into heroku using the Heroku toolbelt:

heroku login

Finally, you can clone your new heroku app:

heroku git:clone -a MYPARSEAPPNAME

This will download your parse server app to your local computer inside a folder called whatever you named your heroku app. Replace MYPARSEAPPNAME with the name of your app on Heroku.

If you’ve worked with Parse at all before, then you’re already familiar with having to add a clientKey and appId into the initialization of your client code, to allow your app to communicate with Parse.

Now however, you need to set these keys both in your client code and in the server code.

Open up index.js and set both the clientKey, and appId, to whatever you want

var api = new ParseServer({
  databaseURI: databaseUri || 'mongodb://localhost:27017/dev',
  cloud: process.env.CLOUD_CODE_MAIN || __dirname + '/cloud/main.js',
  appId: 'MYAPPID',
  clientKey: 'myclientKey',
  masterKey: 'myMasterKey'
});

You will then need to go into the initialization block of your code, and set a matching appId.

On your client side, the only code change you’ll need to make is to specify your new server address. This is the address of your Heroku app that hosts your server side code. You can find it in the settings of your app on your Heroku Dashboard. It is just a URL of the form

https://NAME_OF_YOUR_HEROKU_APP.herokuapp.com/parse

Your client code would now look something like this (for javascript):

Parse.initialize('APP_ID', 'CLIENT_KEY');
Parse.serverURL = 'https://NAME_OF_YOUR_HEROKU_APP.herokuapp.com/parse';
var Test = Parse.Object.extend('TestObject');
var test = new Test();
test.save({'animal': true, 'type': 'dog'},{
    success: function () {
        console.log('success');
    },error: function() {
        console.log('error');
    }
});

The final step is to update your Heroku app, since you changed the appId and added a clientKey.

To do that, run these three commands in the terminal:

$ git add --all .
$ git commit -m "My first commit"
$ git push heroku master

When you change code locally, it isn’t updated on Heroku automatically, so you have to run those three commands to:

  1. Add any new files
  2. Commit all changed, added or removed files so the server knows what changed
  3. Upload the changes to Heroku

Also, where you see My first commit, it helps to update that to reflect any changes you’ve made.

4. FINISHED

You can now use your client side Parse code just as you always did, and works as it did before.

Now instead of connecting to a Parse run server, your client code is connecting directly to your Heroku app, which saves all your data in your MongoDB database.

Your queries all work the same as before.

About Cloud Code

you may have noticed this line:

  cloud: process.env.CLOUD_CODE_MAIN || __dirname + '/cloud/main.js',

This tells your server where your cloud code is located. Cloud code is a feature from Parse that serves as short functions that you can call quickly via HTTP requests such as this one:

curl -X POST \
    -H "X-Parse-Application-Id: myAppId" \
    -H "Content-Type: application/json" \
    -d '{}' \
    https://NAME_OF_YOUR_HEROKU_APP.herokuapp.com/parse/functions/hello

When you migrate from Parse to a self-hosted Parse Server, you can also move over any cloud code you’ve set up so that you just call it when you want to work with it.

The default, included cloud code function is:

Parse.Cloud.define('hello', function(req, res) {
    res.success('Hi');
});

Which means when you make the HTTP request above, you get a return of Hi, where hello is the name of the function you created. You can have as many function as you want, and they all serve different purposes.


It is worth mentioning that several features that made Parse really useful, will not be available in your parse server right away.

The biggest feature that is missing is Push Notifications. But as it happens, it’s pretty easy to add Push Notification support to node.js, and I’ll cover in a later tutorial just how to add Push to your new Parse Server.

Background jobs was another feature missing, but you can look at queue systems or cron jobs to replace them easily. I also plan to write about migrated background jobs to a queued jobs setup.

I also recommend installing the free New Relic add-on inside your heroku app for handy analytics.

Is Heroku really free?

Yes and no, when you first start out, it’s free, but when you are ready to actually use your Parse Server apps then you’ll want to look at the $7 hobby plan at least, the reason for this is because free heroku apps are required to sleep after 30 minutes of inactivity and must sleep for 6 hours in a 24 hour period. This is fine for development, but when ready to actually make use of it, you want to upgrade to the hobby plan for starters.

As I introduce you to posts about deploying Parse Server to other setups such as Digital Ocean and Amazon’s AWS instances, you’ll get a feel for which pricing plans you prefer.

Still Confused?

If you have questions about Parse Server, setting up push notifications, moving your data from Parse to MongoDB or Flybase, then just feel free to get in touch at anytime and I’ll be happy to answer any questions or lend a hand or two.

Use Data McFly to build a real-time Todo app

If you like to keep organized, then you want a todo app. In this tutorial, we’re going to build a simple real-time todo app.

This app will consist of four functions:

  1. Add new items to do
  2. Mark items as done
  3. Mark items as not done (undo)
  4. Delete items

This is simple functionality, but it is exactly what a todo app is all about, and can run anywhere as this will consist of an html file, a javascript file and some very basic css.

All you need to build your todo app is a free Data McFly account. There’s no server-side programming needed for this app.

This was a quick, fun tutorial to write. Showing how to build a simple todo list app using Data McFly.

Source: http://blog.flybase.io/2015/02/06/todo-app

Build a real-time SMS call center with Data McFly and Twilio

Do you want to know one of the beautiful things about Data McFly? It integrates really easily with other services.

In this article, we are going to walk through using Data McFly and Twilio together to build a real-time SMS call center.

This could be used as a customer help desk where customers send a text message for help and an agent sends a reply back from their web browser.

The actual phone work will be handled by Twilio, and Data McFly will store the data and display the chats in real-time. We’ll use node.js to send and receive the text messages and an HTML frontend to handle the actual chatting.

I decided this week to give myself the task of building a real-time SMS call center that used Data McFly for the data handling and real-time communication, and Twilio for the phone work.

The end result turned out pretty well.

Source: http://blog.flybase.io/2015/01/29/sms-contact-twilio

Upgrade an ASUS Chromebox – more RAM, bigger SSD, and Ubuntu

ASUS makes a pretty handy Chromebox, and it’s handy not just because it’s running ChromeOS, it’s handy because of everything you can do to the box itself.

The ASUS Chromebox is easily upgradeable, and capable of running just about any linux distribution.

The model I picked up, the M004U has the following specs:

  • Celeron 2955U (1.4GHz) 64 bit Dual core processor with 2MB L3 Cache
  • 2GB DDR3 1600 RAM with 2 slots
  • 16GB SSD HDD
  • 802.11 b/g/n dual-band wireless, Bluetooth 4.0, and gigabit ethernet
  • four USB 3.0 ports
  • Intel HD Graphics with support for dual displays
  • an SD/MMC card reader

This post will cover how to go about upgrading the hardware, and get it up and running in a dual boot mode to ChromeOS and Chrubuntu.

Here are the parts I used:

  • ASUS CHROMEBOX-M004U Desktop PC Celeron 2955U (1.4GHz) 2GB DDR3 16GB SSD HDD from Amazon
  • Crucial 8GB Kit (4GBx2) DDR3/DDR3L 1600 MT/s (PC3-12800) CL11 SODIMM 204-Pin 1.35V/1.5V Memory For Mac CT2C4G3S160BM from Amazon
  • MyDigitalSSD SC2 Super Cache 2 42mm SATA III 6G M.2 NGFF M2 SSD Solid State Drive (128GB) From Amazon

I also had a wireless Logitech k360 keyboard and mouse I could use, as well as a Samsung monitor for a display.

Getting into the Hardware

First off, you don’t need to make any kind of system image before modifying the unit, you can simply remove the installed SSD and RAM and insert your new parts and go from there. In the instructions below, you’ll make a USB stick that can be used to restore the disk to a factory fresh image of Chrome OS.

To do this, first remove the four rubber feet on the bottom of the unit. You may need to use a small flat head screwdriver to help get them off. Don’t place them anywhere sticky side down!

You’ll then find four screws underneath where the feet were, remove them. Then carefully remove the bottom of the unit. This may be a little difficult, so you may try and slip a small screwdriver in the vents to help. Don’t scratch it up or use force.

You should now be taking a look inside, and marvelling at the relative simplicity of the device!

You’ll see the RAM chip(s), which can be easily removed by opening the clips on the edges and lifting outward.

cb1

If you are upgrading the SSD, you’ll need a tiny screwdriver to remove the screw holding in the wireless/bluetooth card, which is easily identified by the two wires that connect to it. You may remove this card entirely, or do like I did and just remove the screw holding it in and it should give you enough room to install the SSD. Underneath this is the stock Sandisk SSD, which is held in by one screw on it’s far edge, so remove the screw, and pull out the SSD.

Place your new SSD in the slot, and put the tiny screw back in to hold it. Now put the wireless/bluetooth card back in place, and put it’s screw back in as well.

Now put the bottom of the case back on, put the screws back in, and put the feet back on. You’re done with the hardware upgrade!

Getting into ChromeOS

Now you’ll need to install Chrome OS onto your new empty SSD. To do this, visit this page and follow the instructions for your operating system to create a USB stick;

https://chrome.google.com/webstore/detail/chromebook-recovery-utili/jndclpdbaamdhonoechobihbbiimdgai?utm_source=gmail

This tool is a chrome extension but it creates the USB stick beautifully, just choose your chromebox model and it will install the recovery disk for you.

Now, to get started:

  1. Start your Chromebook.
  2. When the “Chrome OS is missing or damaged” screen appears, insert the USB flash drive or SD card you created into the USB port or SD card slot on your Chrome device.
  3. Wait for the Chromebook to boot up from the flash drive.
  4. Follow the instructions that appear on the screen.
  5. On successful installation of the Chrome operating system, you will be prompted to remove the USB flash drive or SD card.
  6. Remove the USB flash drive or SD card when prompted, and your Chromebook will automatically restart.
  7. You should now be able to start your Chromebook as normal.

Now you are ready to do the install of Ubuntu if you so desire. If you just want a fast ChromeOS box, you are done!

Getting into Chrubuntu

Make sure you have ethernet plugged in, or access to a wireless access point so you can download the necessary files.

Find the developer switch on your machine (it’s in a little hole on the back of the left side next to the SD slot). Very carefully press a small paper clip end into this hole and press the power button and the remove the paperclip. When it boots, press ctrl-D to turn off OS verification mode. It should prompt you to turn on Developer mode by using the paper clip again without pressing the power button. It will now boot into developer mode, and take several minutes to prepare itself.

When it has booted, hit Ctrl-D and it will come up to the setup screen, make sure your correct networking is selected. Select continue and accept the agreement.

At the sign in screen press Ctrl-Alt-F2 to get into a shell.

Login with user chronos and hit enter, …it should not ask for a password.

Become root using the command;

sudo bash

Then issue the command to boot into the development bios by default;

chromeos-firmwareupdate --mode=todev

(if you don’t do this it will erase Ubuntu on next boot. You can undo this by booting again to the shell and issuing the command; chromeos-firmwareupdate --mode=normal)

Now type the command;

reboot

When it has booted, hit Ctrl-D and it will come up to the login screen, and now press Ctrl-Alt-F2 to get into a shell.

Login with user chronos again and hit enter, it should not ask for a password.

At this point you can download custom update scripts that will install the appropriate kernel, and a choice of several flavours of linux.

To use install Chrubuntu, type the command;

curl -L -O http://goo.gl/s9ryd > s9ryd

And after it has downloaded the script, type the command;

sudo bash s9ryd

It will prompt you to choose the appropriate size in Gigabytes you wish to give to ubuntu, I gave it 100

Let the machine reboot (still into ChromeOS), it will say it’s repairing itself and reboot. Now choose the language and networking again, and at the login prompt press Ctrl+Alt+f2 to get a shell and give the username chronos (no password)

You’ll need to run the script again, so type the command;

curl -L -O http://goo.gl/s9ryd > s9ryd

And after it has downloaded the script again, you can install by using parameters, where;

  • default (ubuntu-desktop on x86, xubuntu-desktop on arm)
  • kubuntu-desktop
  • lubuntu-desktop
  • xubuntu-desktop
  • edubuntu-desktop
  • ubuntu-standard (no GUI installed)

So to install the default you would enter;

sudo bash s9ryd

or to install ubuntu-desktop lts, type;

sudo bash s9ryd ubuntu-desktop lts

Now it will run for quite a long time, depending on your connection speed. It will prompt you for locale and other settings, most users can just hit enter.

In fact, once the download starts, I highly suggest you go grab a sandwich or something, anything is better than watching the download count slowly roll around.

When it is finished, it will prompt you to reboot by hitting Enter.

You should now verify that everything is as you like it. You can switch between ChromeOS and Chrubuntu by using the following commands;

To make it always boot Ubuntu, go into Terminal and type;

sudo cgpt add -i 6 -P 5 -S 1 /dev/sda

To make boot to Chrome;

sudo cgpt add -i 6 -P 0 -S 1 /dev/sda

(if, for some reason your keyboard doesn’t work on boot, try reconnecting it)

The default Ubuntu user is user with password also set to user. Please change the password!

You’re done! You can stop here unless you need to always boot to Ubuntu, and want to write some scripts to switch envronments, or want to change the default user.

Boot to Chrubuntu, and Script switching back and forth

To always boot into chrubuntu, at the Google sign in screen press Ctrl-Alt-F2 to get into a shell.

Login with user chronos and hit enter, it should not ask for a password.

type the commands;

sudo cgpt add -i 6 -P 5 -S 1 /dev/sda
sudo reboot

Just wait for it to restart, and it will boot up into your alternate OS

To make it boot to Chrome every time;

sudo cgpt add -i 6 -P 0 -S 1 /dev/sda
sudo reboot

Note: don’t hit Ctrl-D anymore at the OS Security screen, just wait for it to come up… this for some reason didn’t work on the next boot, it did however work from then onward. It also boots up amazingly fast.

WARNING: if you switch the developer switch back off YOU WILL LOSE YOUR UBUNTU INSTALL and go back to a completely reset ChromeBox.

You may want to do yourself a favor, and set up some scripts, so that you can boot to chrome from the terminal in Ubuntu by typing chromeos, and to Ubuntu by doing Ctrl+Alt+F2 and typing ubuntu. To do this do the following steps.

When booted into Ubuntu, find your .bashrc file for the default account user, and add this to the bottom using the vi editor; (you can look up how to use vi on the web, but to put it simply, you would type vi .bashrc, scroll to the bottom, and on a empty line press the “a” key to add text, type in the line below, and when done hit escape, then type :wq to save and exit from vi)

alias chromeos='sudo cgpt add -i 6 -P 0 -S 0 /dev/sda;sudo reboot'

After you have saved the file exit the terminal by typing;

exit

now open a terminal again, and type the command;

chromeos

It should prompt for a password, you should enter user…(we’ll change it later)

Wait for it to boot after beeping, and do a Ctrl+Alt+F2 to get to the chrome shell, login with user chronos, and enter the command;

sudo vim .profile

And edit the file (hit the a key to get into insert mode just like in vi) and add:

alias ubuntu='sudo cgpt add -i 6 -P 5 -S 1 /dev/sda;sudo reboot'

Press esc to return to ‘command mode’ Press shift + z then shift + z again to save the file and return to the terminal

Type exit and press enter.

Now in Ubuntu, you can type in a terminal the command; chromeos …and you will boot into chrome …and in Chrome, you can hit Ctrl-Alt-F2 to get the shell, login as chronos, and type ubuntu to boot into ubuntu!

You won’t have to do anything when the dive is rebooted or powered up, just wait a few seconds and it will beep a couple of times, and then boot up into whatever OS was running previously.

Update: I’ve updated this post with a different RAM package, as the original RAM that was being sold got updated to only support 1.5V, and the Chromebox requires 1.35V, so the Crucial RAM listed here now is 1.35V, which works fine with this Chromebox 🙂

Wasabi Coleslaw

Wasabi Coleslaw

This light and tangy slaw combines a crisp mix of cabbage, daikon, carrots and radishes with a bright dressing of rice vinegar and spicy wasabi.

What you need

  • 1/4 cup unseasoned rice vinegar
  • 1 teaspoon low-sodium soy sauce
  • 2 teaspoons wasabi paste
  • 1 teaspoon honey
  • 4 cups very thinly sliced cabbage , preferably Napa
  • 2 cups shredded peeled daikon radish (from about 10 ounces)
  • 2 large carrots, shredded
  • 6 radishes , trimmed and thinly sliced

How to make it

  1. In a large bowl, whisk together vinegar, soy sauce, wasabi and honey.
  2. Add cabbage, daikon, carrots and radishes, and toss to combine.
  3. Refrigerate at least 15 minutes before serving to allow flavors to blend.
  4. Taste before serving and adjust flavors if necessary.

Maple Lemon Chicken

Maple Lemon Chicken

My daughter and I made this for dinner tonight and it was tasty.

One great thing with this dish is it’s easy to modify, swap the maple syrup for honey and you have a tasty honey lemon chicken dish, for example.

What you need

  • 2 tablespoons tomato paste
  • 2 tablespoons lemon juice
  • 2 tablespoons water
  • 1 tablespoon maple syrup
  • 1/2 teaspoon salt
  • 1 pound chicken thighs, cut into 2″ chunks
  • 1/4 teaspoon salt
  • 1/4 teaspoon black pepper
  • 2 tablespoons all-purpose flour
  • 1 teaspoon extra virgin olive oil
  • 1 clove of garlic, minced

How you make it

  1. In a small bowl, whisk together the tomato paste, lemon juice, water, maple syrup and 1/2 teaspoon of salt.
  2. Sprinkle the chicken pieces with 1/4 teaspoon of salt and pepper and then dust with the flour.
  3. Heat a pan over medium heat until hot and then add the olive oil and garlic, swirling the oil and garlic around the pan. If you are using skinless chicken, add an extra teaspoon of olive oil.
  4. Add the chicken in a single layer and fry undisturbed until browned on one side (about 5 minutes). If the chicken starts spattering too much, remove the pan from the heat and use a paper towel and tongs to sop up some of the extra oil.
  5. Flip the chicken over and fry the second side until the chicken is mostly cooked through (another 3 minutes).
  6. Remove the pan from the heat and use paper towels to sop up as much oil as possible.
  7. Return the pan to the heat and add the sauce. Turn up the heat and cook the sauce down until there is almost no liquid remaining and the chicken is coated with a thick layer of glossy red sauce.

Cheesy Baked Spaghetti Squash

Chili-lime chicken, beans, spaghetti squash and cheese… This is a tasty combination that we love to make.

What you need

  • 1 (3-ounce) cooked Chicken Breast, cut into chunks
  • 1 medium Spaghetti Squash (about 4 cups cooked)
  • 1 tablespoon Olive Oil
  • 1 cup Kidney Beans, drained and rinsed
  • 1 cup fresh Salsa
  • 3 tablespoons fresh Coriander, chopped
  • 1/2 cup Shredded Cheddar Cheese
  • Sea Salt & Pepper
  • 1 tablespoon Chili Flakes, optional

How you make it

  1. Preheat your oven to 425°F and line your baking sheet with parchment.
  2. Cut the squash in half, rub with inside with olive oil and sprinkle sea salt & pepper. Place face down on the baking sheet & bake for about 25-30 minutes. The edges on the inside should be a bit browned and shreds should form easily with a fork.
  3. Meanwhile, toss the chili flakes with the cheeses and stir the fresh coriander into the salsa.
  4. When the squash is ready, take it out and let it stand for 5 minutes. Turn the oven to the broiler setting. Fully shred the squash, leaving the shell intact. You will be stuffing it!
  5. Transfer the spaghetti squash to a mixing bowl. Stir in the salsa, followed by the beans and chicken. Split the mixture in half and stuff each squash shell. Top with the shredded cheese.
  6. Place them back on the same baking sheet & broil until bubbly and browned. About 3-4 minutes.

Fish Cakes

I’m originally from Newfoundland, and fish cakes are a staple we grew up with alongside another traditional dish called fish and brews.

You can actually make this with just about any type of fish, but I prefer to use Cod, or Sole, as white fish tends to take on the flavours you cook it with.

What you need

  • 1 ½ pounds salt dried cod
  • ¼ cup butter
  • 1 small chopped onion
  • 6 cups mashed potato
  • 1 beaten egg
  • 2 tbsp dried savoury (optional)
  • 1/2 tsp black pepper
  • 1/2 tsp paprika
  • 1/2 tsp garlic powder

How to make it

  1. Simmer the cod in boiling water for about 15 minutes.
  2. Drain the water off the fish and allow the fish to cool to almost room temperature.
  3. When the fish is cool, flake it apart with a fork into small pieces.
  4. In a sauté pan, melt the butter over medium heat.
  5. Add the onions and cook until they are softened.
  6. Add the flaked fish along with the mashed potato, egg, pepper, paprika, garlic and savoury.
  7. Mix together until well combined, then form into small cakes and roll in flour.
  8. Fry the fish cakes in canola oil over medium heat until golden brown on both sides.

This is a traditional way to make it, you can also add a little hot sauce into the mixture for flavour if you wanted to.