Today when I cloned a new project and ran yarn install, the below error popped up. yarn install v1.22.17 [1/4] Resolving packages... [2/4] Fetching packages... error Command failed. Exit code: 128 Command: git Arguments: ls-remote --tags --heads git://github.com/WearyMonkey/ngtemplate-loader.git Directory: /home/xxx/source/github.com/xxx/xxx/app Output: fatal: unable to connect to github.com: github.com[0: 220.28.17.148]: errno=Connection timed out The solution is to execute the below command under the cloned git repository. git config --global url."https://".insteadOf git://

I will show a simple example in this post. Let’s say the program called cli-example. package main import ( "flag" "fmt" "os" ) func main() { command := os.Args[1] os.Args = append(os.Args[:1], os.Args[2:]...) switch command { case "version": fmt.Println("1.1.1") } } The 1st argument will always be the binary name, so the command starts from the 2nd argument. If your app also needs the subcommand, then extract it from the 3rd argument, like docker image ls. ...

Containerizing the application is very popular nowadays and it makes the deployment so easy. This post will show how to create a Dockerfile for golang application. There are basically two steps in the Dockerfile: Build the golang program source code Containerize the golang app binary # Build source code FROM golang:1.18-alpine AS build WORKDIR /go/src/github.com/app COPY . . RUN go mod tidy RUN CGO_ENABLED=0 GOOS=linux go build -a -o main . # Containerize the binary FROM alpine COPY --from=build /go/src/github.com/app/main . EXPOSE 3500 CMD ["/main"] Exporting the port 3500 is because golang app listens to 3500. Here is just an example. ...

In Golang, when a package is imported, by default the codebase of the package’s default branch will be used by the caller project. However, sometimes we want to import a package’s specific branch codebase. For example, the specific version fixed a bug where we can’t wait until the repository maintainer merges the pull request. So in this post, I will walk you through how to specify a package version in your Golang project. Let me take an example to help explain better. ...

A quick reference guide for commonly used Dokku commands. 1. Create and list app # Create app dokku apps:create <app_name> # List app dokku apps:list 2. Set domains # Set app's domain dokku domains:set <app_name> app.example.com # Set global domain dokku domains:set-global example.com 3. Join app to a specific network # List all networks dokku network:list # Add an app to a specific network dokku network:set <app_name> initial-network <target_network> 4. Show the app’s configuration dokku config:show <app_name> 5. Manully restart an app dokku ps:rebuild <app_name> 6. Bind app container to all interface bind to all interfaces (0.0.0.0), so the container can be accessed by external request. However, the host port will be randomly assigned. ...

About installing the dokku, I have written another post before: How to install Dokku with Portainer. So I will assume that you have Dokku installed in your server in this post. 1. Environment Setting Dokku: version: 0.27.5. HTTP port: 80, HTTPS port: 443, SSH port: 5022 IP of remote server where the Dokku is installed: 166.23.44.25 (example) Domain: example.com Name of App that will be deployed by Dokku: blog Port of App: 3000 2. SSH Configuration Generate ssh key public file id_rsa.pub ssh-keygen -o We will name the file as dokku_rsa, then you will see two files are generated: dokku_rsa and dokku_rsa.pub. ...

When I ran up Dokku container, I map the host port 8080 to Dokku container port 80. With this configuration, all external requests to Dokku have to reach the host port 8080 first and then they can reach the app deployed by Dokku. In Dokku container, Dokku server still listens to the port 80, 443, 22. Therefore, the port mapping from host to Dokku container looks like 8080:80 8443:443 5022:22 Again in Dokku container, the app’s nginx should listen to port 80. It’s because external requests will hit host port 8080, then will be converted to Dokku container port 80. Afterwards, Dokku app’s nginx can detect the requests and route them to the matching app according to the domain name. ...

We can easily figure out how to set up Dokku by following the instructions on Dokku website. But for me, I like to host services in docker container as it will bring a clearer structure when it comes to host multiple services in a single machine. In this post, I will assume that you already had some idea about what Dokku is. If not, you can find more information here. In the title, I also mentioned another tool called Portainer. Basically it is a management tool of docker container and kubernete cluster with UI. But it is actually much more than that. (I feel I am digging a bit hole while introducing more and more tools). ...

Sometimes we need to build Dockerfile to the docker image and push it to registry in CI pipeline. For instance, my development environment is windows/amd64, but my VPS is linux/arm64. At the end, I want to host the service in my VPS. Thus, it will be easy to build the docker image inside the CI pipeline. In this post, I will show how to configure the pipeline to build and push docker image in DroneCI, which is a self-service Continuous Integration platform for busy development teams. ...

Let’s say we have a reverse proxy. The incoming request is HTTP, the target server hosts HTTPs APIs with self-signed TLS certificates. It is very common that we change the URL host or path with reverse proxy, so that the request can be directed a different API. However, in this case, we also need to add the self-signed certificates in reverse proxy. The below example will show how to achieve it. ...