Init I/O - Blog

Introducing sss (S3-client)

Tue, 16 Dec 2025 17:00:00 +0100

Make sure to check the repository for the latest version.

sss is yet another S3 client. I respect all the existing clients and sss certainly has its own shortcomings, but some reasons for starting the project:

Having a single deployable binary.
A CLI and config which can handle non-AWS S3 use-cases well (compare).
Get somewhat meaningful errors instead of 'NoneType' is not iterable.
Having a read-only mode (allowing only safe HTTP methods).
Having the S3 feature set I need.
Be able to limit the bandwidth.
Ergonomic usage, e.g. <bucket> get <object> --sse-c-key=<key> (compare).
Hard to misuse (no local file management like mc cp /mydata/ alias-typo/mybucket/mydata).
Writing the bucket name before the operation to switch the operation quickly, e.g. <bucket> policy get and <bucket> policy rm.

» Installation

» Binaries

Binaries are available for all major platforms. See the releases page.

Introducing IP-Auth

Tue, 16 Dec 2025 16:00:00 +0100

ip-auth is a reverse proxy to allow access only by certain IPs. Make sure to check the repository for the latest version.

There are circumstances where properly setup Basic Auth won’t work [1]. IP Auth is a workaround by allowing specific IPs access to the service and proxying the traffic to the original service. Allowed IPs can be specified or dynamically added by passing a Basic Auth login once from any device on the same IP. Everything is stored in memory and will be lost on restarts.

S3 bits

Sun, 02 Feb 2025 15:50:00 +0100

Things might be specific to Hetzner Object-Storage instead of AWS S3.

» Copy Files

» Upload

Using mc:

1

mc cp <LOCAL_PATH> <ALIAS>/<BUCKET>

» Download

Using mc:

1

mc cp <ALIAS>/<BUCKET> <LOCAL_PATH>

» Multipart Uploads

» List

Using mc:

1

mc ls --incomplete <ALIAS>/<BUCKET>

Using aws:

1

aws --profile <PROFILE> s3api list-multipart-uploads --bucket <BUCKET>

» Remove

Using mc:

Kubernetes volumes with rclone (the hacky way)

Sat, 28 Sep 2024 09:00:00 +0200

» Post Updates

2025-11-26: There is now an rclone CSI driver.
2024-09-30: Added a third approach with using nfsmount.
2024-09-30: Added a second approach with using a sidecar container.

» Introduction

Kubernetes has a wide list of various CSI-drivers for mounting remote storage. I was looking for mounting Hetzner Cloud Object Storage, but unfortunately, the given options were not to my satisfaction. There are several options for mounting S3-compatible storage, but few of them offer transparent encryption and if they do, you are stuck with the tool. When I have to be stuck with a tool, why not a very popular one I can use almost everywhere? When I can mount volumes with rclone and its crypt module in Kubernetes, I can use the same settings for mounting the bucket on my other machines, and that was my goal. Unfortunately, there is no rclone CSI-driver yet, so I went with the hacky solution of directly using the host path of the Node.

Introducing jellyctl

Thu, 09 May 2024 09:00:00 +0100

jellyctl is a CLI for managing your Jellyfin server. Make sure to check the repository for the latest version of jellyctl.

As I didn’t want to craft every HTTP call to the API myself, I generated a Go client from the OpenAPI spec. You can find the Go client in the jellyfin-go repository.

» Installation

» Precompiled Binaries

Binaries are available for all major platforms. See the releases page.

» Homebrew

Using the Homebrew package manager for macOS:

Introducing Calculon

Tue, 27 Jun 2023 07:50:00 +0100

Calculon is a basic, open-source, cross-platform, and “note-style” calculator inspired by Numi.

Calculon is written in Java and uses the exp4j library for evaluating the mathematical expressions. It’s developed with the Netbeans IDE and the Swing GUI Builder.
It is a very rudimentary implementation, just enough to solve around 80% of my problem, so it is not very polished right now and the code is far away from best practices.

Repository: github.com/sj14/calculon

Go bits

Thu, 22 Jun 2023 07:50:00 +0100

» Cross compile using Zig

When you need a CGO enabled build and want to cross compile, you can use Zig for effortlessly cross compiling the C part.

Example for compiling to amd64 linux:

1

env CGO_ENABLED=1 GOOS=linux GOARCH=amd64 CC="zig cc -target x86_64-linux" CXX="zig c++ -target x86_64-linux" go build -o output_binary main.go

References:

» VS Code debugging

Add the following to your user settings (or just the configurations object inside the .vscode/launch.json file in the root of your project) :

Introducing Kubedump

Sun, 16 Apr 2023 07:20:00 +2000

kubedump is a tool for dumping manifests from your Kubernetes clusters. It’s highly inspired by the available shell scripts (1, 2) and just a Go implementation with similar behaviour I wrote mostly for fun. A clear drawback of the Go implementation is the enormous binary size for such small functionality. Currently, the code is not even 200 lines of code but due to the imports from the Kubernetes ecosystem, the binary is around 30 MB. Nevertheless, having a binary instead of a shell script might have other benefits, e.g. no other required dependencies and compared to the scripts it seems to be very fast.

Replacing Docker Desktop with Podman Desktop on macOS

Wed, 15 Feb 2023 15:20:00 +0100

» Post Updates

2023-02-17: There is an official documentation about migrating from Docker to Podman Desktop which I missed when searching for it. So instead of reading this post, you should checkout Migration from Docker to Podman Desktop first.

» Setup

Install Podman Desktop, e.g. with homebrew:

1

brew install podman-desktop

Install the Podman Helper for compatibility with tools using the Docker socket.

$tool bits

Sun, 28 Aug 2022 13:52:00 +0100

Same procedure as last time? Same procedure, just for stuff which won’t require an own post (yet).

» ADB (Android Debug Bridge)

USB Devices are connected automatically.

Connect device via network:

1

adb connect <IP:PORT>

List connected devices:

1

adb devices -l

Sideload:

1

adb sideload <file>.zip

» curl

Save with remote filename (short -O):

MySQL bits

Sat, 20 Aug 2022 06:59:00 +0100

Recently, I had to examine a performance issue of a MySQL query. Here are some notes with helpful stuff I’ve used. All the commands might work with MariaDB too, but this was not tested by me.

Update 2023-05-18: I found Understanding database Indexes in PostgreSQL by Paweł Dąbrowski which contains similar and more detailed information but for PostgreSQL.

» Common

Currently, my preferd mySQL GUI management tool on macOS: Sequel Ace.
Alternative cross-platform and cross-database GUI tools: DbGate, Jailer
Use backticks (`) e.g. `table-name` when the name contains “unusual” characters such as -.

» Describe table

Get an overview of the table:

Bash bits

Mon, 08 Aug 2022 18:06:00 +0100

Another post with stuff I can’t remember or might forget again 🤷‍♂️

» Bash shebang

Specify the explicit path:

1

#!/bin/bash

Let the environment decide:

1

#!/usr/bin/env bash

Explanation: Alec Bennett

» Fail fast … and more

Best added directly after the shebang.

1
2
3
4


set -o xtrace # output every line which gets executed
set -o nounset # fail if an unset variable is used
set -o errexit # fail if a command exits with non-zero
set -o pipefail # fail if a command in a pipe fails

or in short:

Python bits

Wed, 13 Jul 2022 10:35:00 +0100

» pip

Remove all packages to get a somewhat clean system (e.g. when you forgot to use a virtual environment, like I did…):

1

pip freeze > requirements.txt && pip uninstall -r requirements.txt -y

or without creating a requirements.txt.

1

pip uninstall -y -r <(pip freeze)

Credits to Giorgos Myrianthous

» venv

Create a virtual environment:

1

python3 -m venv venv/

Activate the environment:

Introducing Confible

Sun, 10 Jul 2022 10:17:00 +0100

» Introduction

Many people backup and sync their configs by just copying the whole dotfiles. Unfortunately, I’m not a big fan of this because then I backup a lot of lines I do not care about. Usually, I do only add a few lines to a config so that it fits my preferences. Otherwise, I like using the default settings the tools come with. So, I tried to find a tool which would fit my preferences but was unlucky. Confible was born. It can adjust config files and execute commands. Check the repository for the latest version and readme.

Python astral library ported to Go

Mon, 19 Apr 2021 18:13:00 +0200

Calculations for the position of the sun and moon.

This is a Go port of the Python astral package. Check out the latest Go version on Github.

The astral package provides the means to calculate the following times of the sun:

dawn
sunrise
noon
midnight
sunset
dusk
daylight
night
twilight
blue hour
golden hour
~~rahukaalam~~ TODO

plus solar azimuth and elevation at a specific latitude/longitude. It can also calculate the moon phase for a specific date.

Kubernetes bits

Mon, 15 Feb 2021 22:44:00 +0100

» Post Updates

2023-03-22: Added bootstrap manifest.
2022-09-21: Added ephemeral debug containers.
2022-02-23: Added contexts and Service FQDN.
2022-01-24: Added logs with stern section.
2021-12-01: Added Wireshark section.
2021-10-09: Added Kubescape section.
2021-04-21: Replaced ctr with crictl for finding the PID.

» Switching Contexts

Credits to Sarasa Gunawardhana

List all contexts:

1

kubectl config view -o jsonpath='{.contexts[*].name}' | tr " " "\n"

Get current context:

Deploy a free Kubernetes cluster in the cloud

Sat, 30 Jan 2021 14:17:00 +0100

» Post Updates

2022-08-20: The k3os project seems to be dying and currently I would not recommend it for new cluster installations. Personally, I’ve switched to Talos.
2021-09-17: Added a new section about securing web applications with TLS.
2021-07-08: Oracle now offers 4 Ampere (ARM) instances with cumulated memory of 24GB. So, multi-main clusters are now possible and resources are less of a problem.

» Introduction

For installing a completely free Kubernetes cluster, you can use the free instances of Oracle Cloud. The free tier is of course very low-end but it’s enough for some testing and playing around without always running something like minikube, kind or k3d on your local machine.

Docker bits

Sun, 13 Sep 2020 13:08:00 +0200

» Post Updates

2023-03-01: Added Nixery as useful project.
2022-05-16: Added example for mounting the current directory into a docker container.
2021-06-12: Added --all --volumes --force args to docker system prune and removed separate commands for removing all images and volumes.

» Shell into a fresh container

Handy for some quick testing.

» Debian

1

docker run -it debian /bin/bash

» Alpine

1

docker run -it alpine /bin/ash

» Mount the current directory

1

docker run -it --volume $(pwd):/hostdir alpine /bin/ash

» Hostname of the host

When you are inside a docker container, you can use host.docker.internal for connections to the host machine.

Quick Tip: Simulate a bad network connection on macOS (and Linux)

Fri, 19 Jun 2020 17:11:00 +0200

» Introduction

Yesterday, I got a hint from a colleague to fake a poor network connection (for debugging an issue) with tc (man page):

1

tc qdisc add dev eth0 root netem loss 30% delay 500ms 300ms

This works with Linux and will add a delay and packet loss to the eth0 interface.

I was curious if there is someting similar for macOS, and there is: pfctl and dnctl. This following commands are mainly copy-pasted from the posts from John Ruble and Sal Ferrarello. Check their posts and the man pages of the commands for more details.

Quick Tip: Hiding changes of specific folders in GitLab merge requests

Fri, 17 Apr 2020 20:24:00 +0200

When you store third-party dependencies (such as a vendor folder for Go, node_modules for npm, or venv for Python) in your repositories and don’t want changes in the dependencies to show up in your merge requests, you can collapse all diffs for specific folders.

All you have to do is creating a .gitattributes file with the following content:

1

vendor/* -diff

Just replace vendor with the folder which contains the changes you want to ignore. This is only applied when the .gitattributes file is already merged in your default branch (e.g. master). Otherwise, it will still show the diffs in all merge requests!

Working with CSV datasets in Go

Tue, 24 Mar 2020 20:00:39 +0100

See the repository for the current version!

» Installation

1

go get -u github.com/sj14/csvutil

» Examples

For all options please check the godoc. The examples ignore all error handling!

» Create a new dataset

All datasets should contain a header for further processing.

1
2
3
4
5
6
7
8


records := [][]string{
 {"first_name", "last_name", "username"},
 {"Rob", "Pike", "rob"},
 {"Ken", "Thompson", "ken"},
 {"Robert", "Griesemer", "gri"},
}

ds := csvutil.New(records)

» Add a new row

1

ds.AddRow([]string{"my first name", "my last name", "my nick"})

1
2
3
4
5


first_name,last_name,username
Rob,Pike,rob
Ken,Thompson,ken
Robert,Griesemer,gri
my first name,my last name,my nick

» Add multiple rows

1
2
3
4
5


ds.AddRows([][]string{
 {"my first name 0", "my last name 0", "my nick 0"},
 {"my first name 1", "my last name 1", "my nick 1"},
 {"my first name 2", "my last name 2", "my nick 2"},
})

1
2
3
4
5
6
7


first_name,last_name,username
Rob,Pike,rob
Ken,Thompson,ken
Robert,Griesemer,gri
my first name 0,my last name 0,my nick 0
my first name 1,my last name 1,my nick 1
my first name 2,my last name 2,my nick 2

» Add a new column at index 1

1
2


ds.AddCol([]string{"column_headline", "my row 1", "my row 2", "my row 3"}, 1)
ds.Write(os.Stdout)

1
2
3
4


first_name,column_headline,last_name,username
Rob,my row 1,Pike,rob
Ken,my row 2,Thompson,ken
Robert,my row 3,Griesemer,gri

» Extract Column

1
2


lastNames, _ := ds.ExtractCol("last_name")
fmt.Println(lastNames)

1

[last_name Pike Thompson Griesemer]

» Rename Column

1

ds.RenameCol("username", "nick")

1
2
3
4


first_name,last_name,nick
Rob,Pike,rob
Ken,Thompson,ken
Robert,Griesemer,gri

» Delete Column

1

ds.DeleteCol("first_name")

1
2
3
4


last_name,nick
Pike,rob
Thompson,ken
Griesemer,gri

» Modify Column

1
2
3


addRowNumber := func(val string, i int) string { return fmt.Sprintf("%v (%v)", val, i) }
ds.ModifyCol("first_name", addRowNumber)
ds.Write(os.Stdout)

1
2
3
4


first_name,last_name,username
Rob (1),Pike,rob
Ken (2),Thompson,ken
Robert (3),Griesemer,gri

» Write the dataset

1

ds.Write(os.Stdout)

1
2
3
4


first_name,last_name,username
Rob,Pike,rob
Ken,Thompson,ken
Robert,Griesemer,gri

» Write options

1

ds.Write(os.Stdout, csvutil.Delimiter('|'), csvutil.UseCLRF(true))

1
2
3
4


first_name|last_name|username
Rob|Pike|rob
Ken|Thompson|ken
Robert|Griesemer|gri

» Get dataset as [][]string

1

fmt.Println(ds.Raw())

1

[[first_name last_name username] [Rob Pike rob] [Ken Thompson ken] [Robert Griesemer gri]]

» Compare two (raw) datasets

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12


a := [][]string{
 {"a", "b", "c"},
 {"d", "e", "f"},
}

b := [][]string{
 {"g", "h", "i"},
 {"j", "k", "l"},
}

result := csvutil.Equals(a, b)
fmt.Println(result)

1

false

Using reCAPTCHA with Go

Sat, 21 Mar 2020 17:54:39 +0100

See the repository for the current version!

» Installation

1

go get -u github.com/sj14/recaptcha

» Usage

The result will always return an error when any kind of failure occured, thus only checking the error is sufficient.

» reCAPTCHA v2

See reCAPTCHA v2 docs

» HTML

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12


<html>
 <head>
 <script src="https://www.google.com/recaptcha/api.js" async defer></script>
 </head>
 <body>
 <form action="?" method="POST">
 <div class="g-recaptcha" data-sitekey="<YOUR_SITE_KEY>"></div>
 <br/>
 <input type="submit" value="Submit">
 </form>
 </body>
</html>

» Go

1
2
3
4
5


result, err := recaptcha.VerifyV2(recaptchaSecret, httpRequest)
if err != nil {
 log.Fatalf("recaptcha failed: %v\n", err)
}
// result is not necessary to check, only required for more details

» reCAPTCHA v3

See reCAPTCHA v3 docs

Installing FFmpeg on Mac

Sat, 14 Mar 2020 13:44:00 +0100

Check the repository for the current version of the script.

» Why?

brew install ffmpeg installs too many dependencies.
This script downloads the static binary and moves it to /usr/local/bin/, thats all.

» Usage

Latest stable:

1

curl -s https://raw.githubusercontent.com/sj14/playground/master/ffmpeg-mac/install.sh | bash -s

Latest snapshot:

1

curl -s https://raw.githubusercontent.com/sj14/playground/master/ffmpeg-mac/install.sh | bash -s snapshot

» Updating

For updating to the latest stable or snapshot build, just run the desired command again, it will overwrite the previously installed version.

Introducing review-bot.com

Mon, 17 Feb 2020 21:30:00 +0100

Based on the CLI tool review-bot, I created a new SaaS product review-bot.com. It will allow you to send reminders without hosting anything yourself or dealing with the command line and cron jobs.

That’s all folks, just check the website for more information :=)

Get the SonarQube Quality Gate status for the current commit

Sat, 10 Aug 2019 14:30:00 +0100

Update 2021-03-05: Simplified Code

Update 2021-01-11: Added authentication

When you use SonarQube and want to fail a pipeline if the Quality Gate result for the current git commit is not “OK”, this script might be helpful for you.

You can always find the latest version in my git repository. Open an issue or a pull request for questions and enhancements.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56


#!/usr/bin/env bash

# Description
# ===========
# Get the SonarQube Quality Gate status of a specific project for the current git commit.
# When the Quality Gate analysis is "OK", return 0, otherwise, return 1.
# Well suitable for usage in a build pipeline.
#
# Requirements
# ============
# git, curl, jq
#
# Usage
# =====
# Execute this script `./sonar-status.sh <host> <auth_token> <project_key>` and check for the return code.
# Example: `./sonar-status.sh http://localhost:9000 d7981b9c1333bef3ff83502cea6f6cd6ce4d70f6 test`

set -euf -o pipefail

SONAR_HOST=$1
AUTH_TOKEN=$2 # project key token or user token (see https://docs.sonarqube.org/latest/user-guide/user-token/)
PROJECT_KEY=$3

GIT_COMMIT=$(git rev-parse HEAD)


retry=0
while [ $retry -lt 6 ]; do
 # get latest analysis id to corresponding git commit hash
 analysis_id=$(curl -s -u "$AUTH_TOKEN": "$SONAR_HOST"/api/project_analyses/search?project="$PROJECT_KEY" | jq '.analyses[] | select(.revision == "'"$GIT_COMMIT"'") | .key' | head -n 1)

 # remove double quotes from analysis id (e.g. "AWv6wb07Y5FuS8wxa-xk" -> AWv6wb07Y5FuS8wxa-xk)
 analysis_id=$(echo "$analysis_id" | tr -d "\"\`'")

 # get quality gate status of this anlysis
 analysis_status=$(curl -s -u "$AUTH_TOKEN": "$SONAR_HOST"/api/qualitygates/project_status\?analysisId="$analysis_id" | jq '.projectStatus.status')

 # break this loop when analysis was found,
 # otherwise, try again (next loop iteration)
 if [ "$analysis_status" != null ]
 then
 break
 fi

 echo "Retry $retry: $PROJECT_KEY"
 (( retry=retry+1 ))
 
 sleep 10s
done

# print and return status depending on result
printf "%s \t %-35s \t $SONAR_HOST/dashboard?id=%s\n" "$analysis_status" "$PROJECT_KEY" "$PROJECT_KEY"
if [ "$analysis_status" != \""OK\"" ]
 then
 exit 1
fi