Web Development with Docker part 3: Deploy to Remote Hosts

If you have not read Part 2, please go here and read it first.

Docker machines are deprecated, please use the Resource Manager in Wappler instead.

In the latest Wappler update we have added the final piece in the Docker integration - Docker Remote Deploy!

Now you can not only fully develop locally with Docker but also deploy to any remote Docker host!
Development and Deploy has never been easier!

Docker Remote Hosts are the machines on your hosting or cloud that you will be deploying your production solutions.

How it works

To be able to deploy to remote hosts, you just have to define them as separate targets in the Wappler Project Settings.

After that you can easily switch to the target and deploy to you with a single click!

Let see how this works

Defining a Remote Docker Target

Defining a Remote Docker Target is works exactly as any other docker target, but you just need to select that the connection type is “remote” instead of local (1) and then enter the Docker Remote settings (2)

Next to the required host IP or name, you can choose to use a secure connection by choosing Use TLS. That is of course the way to go for production hosts as all deploy have to be fully secured.

To be able to connect to such secured host you will need to have the right SSH keys

Follow the Docker guide how to generate those on your remote server.

At the end you will have 3 required files ca.pen, cert.pem and key.pem that are generated and you will have them in a folder that you can choose as “Cert Folder”

If you have a local docker machine as remote you can choose the docker machine folder, see at the end of this guide how to create one.

Performing a Connection Test

After you have entered all your your remote info and you can do a connection test, to see if it all works fine. Just click on the “Test Connection” button:

This will open a popup dialog and try to connect to your remote host. If your remote host is running and the docker engine on it is running and responding, you will see a success message.

Common Connection Errors

Here are some common errors that might occur.

Bad Host

If for example your host is offline, responding or you got bad host name, you will see:

Wrong Port

If you got a wrong docker port:

Security Required

If you try to connect without security, but security is required:

Deploying your Wappler Project to Remote

Deploying your project to the Remote Docker host is really easy. Just click on the Deploy button in the docker toolbar.

The deploy process will kicks in, if you are interested to see what is happening you can expand the panel below

You will see that images are being downloaded and installed on the remote host and also your project files are copied.

If everything goes well you will see that all services are running now.
The green status indicator is now on and also the deploy icon next to the play icon are blue.

image

Updating and Redeploy your Wappler Project

After a while you will probably make some changes to your project and you need to redeploy new version. This is also very easy. Just click on the Deploy button again!

image

This will update your remote services with the new local files that you have and restart them.

Clearing Remote Host

If you want to wipe a remote host clean and do a fresh redeploy, you just have to click on the “shutdown” icon. Do note that any databases on it will be cleaned as well so make sure you have created a backup of their data structure first that will be used for the initial population.

Local Docker Machines for testing / staging

NOTE: The following is a manual procedure, which was later automated and made much more easier to use in Wappler, see:

Some times you need a full Docker machine running but not necessary remote, but just as a local machine for testing or staging environment.

Such local Docker Machine is installed as a virtual machine and behaves exactly as a remote docker host.

Local Docker Machines on Windows

To create a local docker machine on Windows you will be used the Windows 10 Hyper-V that Docker already uses.

You must have a Virtual Switch created first in the Hyper-V Manager, so make sure you follow the instructions to create Virtual Switch first

Then just open a terminal and enter the following command:

docker-machine create --driver hyperv vm1

where vm1 is the name of your virtual machine.
This will kicks in a whole process that will download a special linux image with docker on it, install it, configure docker engine on it and produce you the needed SSH keys.

When it is done, you can check the status with:

docker-machine ls

image

You will see that status of your machine and also its IP address and port that you can use for connecting to it in Wappler.

There are also other docker-machine commands like stop/start or rm (remove) that you can use to manage its status later on.

So now you can enter the connection info in Wappler:

Do note that all Docker machines configuration info, includes its keys are stored in

C:\Users\username\.docker\machine\machines in the machine name (vm1 in our case) as sub folder.

So you have to choose this folder for as cert folder.

Local Docker Machines on Mac

Installing a local docker machine on Mac is a bit difficult than Windows because you have two choices to install your virtual machine there.

You can either use virtualbox from Oracle or use the much better native hyperkit driver.
Docker describes the virtualbox only in their docs, but hyperkit can be better installed.

Install Hyperkit driver

The required hyperkit driver can be easily installed with brew so make sure you go that first.

Then in a terminal enter:

brew docker-machine-driver-hyperkit

Follow the instructions and also perform the two sudo commands at the end to give permissions.

After the hyperkit driver is instaleld you can create the docker machine with:

docker-machine create --driver hyperkit vm1

where vm1 is the name of your virtual machine.
This will kicks in a whole process that will download a special linux image with docker on it, install it, configure docker engine on it and produce you the needed SSH keys.

If the hyperkit driver doesn’t work for you, you can always follow the docker instruction about using the virtualbox

When it is done, you can check the status with:

docker-machine ls

You will see that status of your machine and also its IP address and port that you can use for connecting to it in Wappler.

There are also other docker-machine commands like stop/start or rm (remove) that you can use to manage its status later on.

So now you can enter the connection info in Wappler:

Do note that all Docker machines configuration info, includes its keys are stored in

\Users\username\.docker\machine\machines in the machine name (vm1 in our case) as sub folder.

So you have to choose this folder for as cert folder.

Conclusion

So that is all about remote deployment with Docker. Power all those docker hosts and deploy at wish!

Do supply us with feedback about the process and if everything is clear and easy to use.

Continue to read Part 4 of the full Docker tutorial.

9 Likes

When you want to try this and you are 2 hours away from your laptop.

image

7 Likes

Possible improvement to push remotely to a container registry. Some providers only accept docker deployment via registry. i.e. heroku.

Not going to lie I have no idea what this is after reading it. My guess though is that before docker only worked locally and now it works on any server? So if you’re using managed hosting this is beneficial because it’s essentially a one click set up? Or am I totally off?

@Logan, docker still works locally and will now also work remotely with any server that allows it.

Check with your hosting provider to see if it allows docker.

1 Like

5 posts were split to a new topic: Docker Managing Database

4 posts were split to a new topic: Docker on Digital Ocean droplet - generating certificates

In the above post @George talks about setting up local docker machines on Mac using virtualbox or the native hyperkit driver.

But didn’t the first post in this series already cover setting up a local docker machine? And there was no mention of virtual box or the native hyperkit driver in that post.

I’m curious to know if there is some kind of difference between the local docker mentioned in that first post and the local docker mentioned in this post.

Thanks.

Good question Eddie!

Per default docker is running in its own single virtual machine locally. There you are adding all the containers/services you want.

Remotely you have a server for live deployment that usually runs the same configuration as your local development but then in production mode.

But if you want to test this remote server without having a real server you can build one by having a second local virtual image in docker that has the same configuration.

So this is what I mean by adding a local virtual image with Hyperkit, functioning as dev remote server.

1 Like

The above note references lesson 4 in this series: REVIEW: Docker Part 4 - Deploy in Seconds to the Cloud with Docker Machines

But lesson 4 is about deploying to a live Digital Ocean server…not to a Local Docker Machine using something like hyperkit.

Was the manual process of installing a second local virtual image via hyperkit made easier by Wappler after this tutorial was originally posted? If so, can someone provide me with the link to the post or the doc describing it?

Much appreciated.

1 Like

Can we deploy multiple projects to the same remote host?

Yes you can deploy multiple Wappler docker projects on a single host.

It is only advisable to use external database otherwise multiple docker databases on a single host will conflict and also quickly exhaust your server memory. Like Digital Ocean Managed Databases.