DEV Community: Ruben Alvarado

MongoDB didn't start because of File permission issue

Ruben Alvarado — Fri, 27 Feb 2026 20:52:10 +0000

The Wired Tiger Turtle in the logs

Originalled published on Web Warrior Toolbox

A Wired Tiger Turtle Story...

While troubleshooting a project, I wanted to run the db locally before deploying my hotfix. I set my pc for that and I mocked a database with some useful data for my test before deploying to production that day.

Two days later a new issue that needed a hotfix came to me and I needed to test again with a local database. I tried to start the db with the usual command systemctl start mongod but I saw it didn't start when I checked with systemctl status mongod.

When I checked status I saw a message saying that it didn't start but i couldn't understand the reason why.🤨

After following multiple suggestions from chatbots and getting lost myself 😵‍💫 I understood that the answer to that question was in a place I didn't use to look at 💬: THE LOGS! 🪵

In Linux you have that place dedicated to record what did your applications say and probably you didn't see or don't remember. Just look at /var/log and then check your respective log.

In my case I found the file /var/log/mongodb/mongod.log

Since it had many lines I just checked the last part of the log with the tail command:

sudo tail -n300 /var/log/mongodb/mongod.log

Still with that I got a wall of text so i decided to fed my chatbot to provide some advice. The chatbot showed that the most frequent text was indeed the most important part of this issue.

[ERROR]: __posix_open_file, 924: /var/lib/mongodb/WiredTiger.turtle: handle-open: open: Permission denied"}}

A wild Wire-tiger-turtle-appeared!🔌🐯🐢

The chatbot suggested that /var/lib/mongodb/WiredTiger.turtle should be owned by mongodb. I noted the file was owned by root. I tryed changing the owner to mongodb and it didn't work. Then tried changing the folder to be owned by mongodb. This last attemp was sucesful when i restarted mongod and check with systemctl status mongod.

What I learned from that story

This story is longer than what I wrote. At the start I didn't know where to look and i received suggestions from the chatbot that lead me nowhere. This is the reason why I want to highlight the importance of logs. They provide some important detail that you can use to feed a chatbot, chat with another team member and to understand yourself what is happening. Take this story as a way to keep checking them when issues arise.

Side story

why did that wiredTiger.turtle file became owned by root? Well it happened that when I did the first hot fix I started the db with sudo mongod --config /etc/mongod.conf

With that command I could use mongodb without problems during that session but I got the problem I tried to tweak again the app two days later.

Let's keep coding, dev-san 🤓 頑張って!

PostGres Database Replication Using Pglogical

Ruben Alvarado — Tue, 24 Feb 2026 20:52:44 +0000

Originally published on Hashnode

This is an example of PostgreSQL database replication using the pglogical extension. In this example, I used Docker containers, with each container having its own database. When data is written to the primary database, the secondary container receives the same data.

👨‍💻A man proudly told his wife, “I’m an SQL DB admin!”

She threw him out of the house…

Because he had one to many relationships💔 — SQL News Network

Start Docker

If you have docker engine use

sudo systemctl start docker
sudo systemctl enable docker // if you want start at boot

Or if you have docker desktop use this in debian Linux (installed from .deb package)

systemctl --user start docker-desktop

Create configuration files

You need to create 2 config files as shown in this folder tree

/<PATH_TO_PROJECT>/docker/
│
├── Dockerfile.pglogical
└── postgres/
    └── docker-compose.yml

Define config files

docker-compose.yml is the file that defines how to run the containers.

cd <RATH_TO_PROJECT>/docker/postgres/ 
nano docker-compose.yml

paste this content for your yml

container names are inside the services part. For these example containers are pg1 and pg2

services:
  pg1:
    build:
      context: ..
      dockerfile: Dockerfile.pglogical      
    container_name: pg1
    command:
      - postgres
      - -c
      - shared_preload_libraries=pglogical
      - -c
      - wal_level=logical
      - -c
      - max_replication_slots=10
      - -c
      - max_wal_senders=10
    environment:
      POSTGRES_USER: postgres
      POSTGRES_PASSWORD: postgres
      POSTGRES_DB: testdb
      POSTGRES_HOST_AUTH_METHOD: trust
    ports:
      - "5433:5432"
    volumes:
      - pg1_data:/var/lib/postgresql/data
    networks:
      - pgnet

  pg2:
    build:
      context: ..
      dockerfile: Dockerfile.pglogical
    container_name: pg2
    command:
      - postgres
      - -c
      - shared_preload_libraries=pglogical
      - -c
      - wal_level=logical
      - -c
      - max_replication_slots=10
      - -c
      - max_wal_senders=10
    environment:
      POSTGRES_USER: postgres
      POSTGRES_PASSWORD: postgres
      POSTGRES_DB: testdb
      POSTGRES_HOST_AUTH_METHOD: trust
    ports:
      - "5434:5432"
    volumes:
      - pg2_data:/var/lib/postgresql/data
    networks:
      - pgnet

networks:
  pgnet:
    driver: bridge

volumes:
  pg1_data:
  pg2_data:

docker-compose.yml has a limitation: It doesn’t let you install software inside the system images!

Since pglogical is not included in the images we need to create Dockerfile.pglogical to specify that we want to add pglogical to the image.

FROM postgres:16

RUN apt-get update \
 && apt-get install -y postgresql-18-pglogical \
 && rm -rf /var/lib/apt/lists/*

Start the containers

Use this command to start the containers

cd <ROUTE_TO_DOCKER_YML>
docker compose up -d

Check that containers are running

docker ps

you will get something like this

CONTAINER ID   IMAGE          COMMAND                  CREATED      STATUS         PORTS                                         NAMES
c61ef5d393cc   postgres-pg1   "docker-entrypoint.s..."   3 days ago   Up 4 minutes   0.0.0.0:5433->5432/tcp, [::]:5433->5432/tcp   pg1
b055528784d5   postgres-pg2   "docker-entrypoint.s..."   3 days ago   Up 4 minutes   0.0.0.0:5434->5432/tcp, [::]:5434->5432/tcp   pg2

Where each line is a container you have running

Get pglogical running in each container

Follow this instructions for pg1 and pg2 the examples shown are also for pg1

Install pglogical in each container using the following command

docker exec -it pg1 psql -U postgres -d testdb

Enable the extension in each container

CREATE EXTENSION pglogical;

The use this sql command to check installed extensions in each container

testdb=# \dx

example output when pglogical is installed

                   List of installed extensions
   Name    | Version |   Schema   |          Description           
-----------+---------+------------+--------------------------------
 pglogical | 2.4.6   | pglogical  | PostgreSQL Logical Replication
 plpgsql   | 1.0     | pg_catalog | PL/pgSQL procedural language
(2 rows)

SELECT pglogical.create_node(
    node_name := 'pg1',
    dsn := 'host=pg1 port=5432 dbname=testdb user=postgres password=postgres'
);

Create table in the provider database

On pg1

Connect postgres

SELECT pglogical.create_replication_set(
    'demo_set',
    replicate_insert := true,
    replicate_update := true,
    replicate_delete := true,
    replicate_truncate := true
);

Create demo table & baseline data

On pg1:

CREATE TABLE demo_events (
    id SERIAL PRIMARY KEY,
    message TEXT,
    created_at TIMESTAMP DEFAULT now()
);

INSERT INTO demo_events (message)
SELECT 'baseline row ' || g
FROM generate_series(1,5) g;

Add the table to the replication set:

SELECT pglogical.replication_set_add_table(
    'demo_set',
    'demo_events'
);

Create subscriptor

On pg2:

SELECT pglogical.create_subscription(
    subscription_name := 'sub_pg1',
    provider_dsn := 'host=pg1 port=5432 dbname=testdb user=postgres password=postgres',
    replication_sets := ARRAY['demo_set'],
    synchronize_structure := true,
    synchronize_data := true
);

Wait a few seconds, then verify:

SELECT * FROM demo_events;

You should see the 5 baseline rows on pg2.

Finally if you try to insert data in the table from pg1 (primary db) it will be successful but if you try from pg2 (replica db) you will get a error-only error. which shows that replication is working as expected.

INSERT INTO demo_events (message)
VALUES ('inserted new row');

Enjoy! 🎉 You now have your replicated databases working. Feel the power of synchronized data with no more differing versions! 🔄💪

Docker Desktop and LM Studio installation

Ruben Alvarado — Wed, 18 Feb 2026 20:44:23 +0000

Originally published on Hashnode.

You can self-host LLMs to automate tasks on your system like retrieving information from a website, writing a database, interacting with note apps, etc. To set up a PC for these automations you can install Docker and LM Studio. This document provides step-by-step instructions for installing and configuring these tools.🤓

🐳Install docker desktop in Ubuntu Linux🐧

Docker Desktop provides pre-packaged MCP servers to interact with web tools, databases, and other systems. It integrates seemlesly with LM Studio, which runs and manages your local LLMs.

Before installing and running these tools, it is important to ensure that your system supports hardware virtualization, which is required for efficient container and virtual machine execution.

First check that you have kmv module loaded to linux kernel with this command

lsmod | grep kvm

You should get a response like this that shows that kvm is running

08:14:19
kvm_intel             487424  18
kvm                  1425408  17 kvm_intel
irqbypass              12288  1 kvm

Download latest DEB package. In my case I used this lnk

Docker Desktop DEB package Download Link

Install the package using apt. Replace <route_to_download_folder> with the route where you have your downloads.

 sudo apt-get update
 cd <route_to_download_folder>
 sudo apt-get install ./docker-desktop-amd64.deb

Sucess you now have Docker Desktop!🤵

🎧Install LM Studio

LM Studio lets you run large language models locally on your PC, including models such as LLaMA, Mistral, and DeepSeek. Running models locally ensures full privacy and allows you to work offline.

To install LM Studio, download the application from the official website:

https://lmstudio.ai/

Follow these commands to install app image. Replace <lmstudio_folder> and <download_folder> with the respective routes in your PC.

mkdir <lmstudio_folder>
cd <lmstudio_folder>
mv <download_folder>/LM-Studio-0.3.36-1-x64.AppImage .
sudo chmod +x LM-Studio-0.3.36-1-x64.AppImage
./LM-Studio-0.3.36-1-x64.AppImage -appimage-extract

After extraction is complete, navigate to this folder inside <lmstudio_folder> and set chrome-sandbox to be executed as root from a normal user (no sudo command requiered for execution)

cd squashfs-root
sudo chown root:root chrome-sandbox
sudo chmod 4755 chrome-sandbox

Now enable LM Studio to be found when you press windows (super key) in Ubuntu. Follow these commands to create the desktop file.

cd ~/.local/share/applications/
sudo nano lmstudio.desktop

Paste these lines to the desktop file. Replace <USER_NAME> with your username in your PC

[Desktop Entry]
Name=LM Studio
Type=Application
Exec=/home/<USER_NAME>/lmStudio/squashfs-root/lm-studio
Icon=/home/<USER_NAME>/lmStudio/squashfs-root/lm-studio.png
Terminal=false

Press the super key and type “lm studio”. You can access the application without knowing its exact location now.🤩

References to Explore

Ubuntu | Docker Docs

Learn how to install, launch, and upgrade Docker Desktop on Ubuntu. This quick guide will cover prerequisites, installation methods, and more.

docs.docker.com

Set New Branch in Git

Ruben Alvarado — Wed, 08 Oct 2025 18:35:59 +0000

If you want to add a new feature to an app you have cloned. You can create a new branch for that feature following these git commands.

git switch -c <branch-name> // creates the branch
git add <file-or-folder>     // or use '.' to stage all changes
git commit -m "<commit-message>"
git push --set-upstream origin <branch-name>

You can also simply try git push, in that case you will get a warning that the branch doesn't exist in remote yet and you receive a suggestion using the --set-upstream parameter. Instead of memorizing the last command with the parameter you can copy and paste the command that the terminal suggests.

And that it. Use this snippet to preserve working code in master and your new code in your feature or bug fix branch.

🤓Happy coding😎

🚀 How to Upload Environment Variables to GitLab CI/CD

Ruben Alvarado — Sat, 13 Sep 2025 21:28:24 +0000

Managing secrets and configuration in GitLab pipelines doesn’t have to be messy. Here’s a quick guide to securely upload your environment variables and integrate them into your .gitlab-ci.yml workflow.

In Gitlab goto left side bar, Settings, CI/CD
Go to variables section, click add variable

From right side bar select select visible and deselect protect variable
In Key write the name of the file that will store all the variables
In value enter the key and values of your variables (values are written without ")
Make sure that your yml script adds the env vars file
.gitlab-ci.yml

Autotest:
  stage: test
  image: node:22
  before_script:
    - cat "$ENV_VARS" | tr -d '\r' > .env

You’ve now securely injected environment variables into your GitLab CI/CD pipeline. No more hardcoding secrets or juggling config files—just clean, maintainable automation.