Category: DEVOPS
-
When Automation Becomes the Enemy: Lessons from AWS’s October 2025 DynamoDB Outage
The AWS post-mortem reveals a significant service disruption from October 19-20, caused by a race condition within their DNS management. This incident underscores the complexity of automation and the necessity of manual override procedures. It highlights the cascading failures from a single issue and the importance of mapping service dependencies and preparing for unexpected incidents.
-
Optimize LLM Performance on Mac and Ubuntu
This post discusses optimizing performance with large language models using Ollama on macOS with Apple’s M1/M2 chips and dual NVIDIA 2080 Ti GPUs on Ubuntu. It provides installation steps and GPU acceleration tips, highlights alternative tools, and outlines how to implement OpenAI-compatible APIs efficiently, maximizing hardware performance for local inference.
-
Cloud vs Open-Source: Comparing AWS, Azure, GCP Services
Modern enterprises must choose between cloud-native services from providers like AWS, Azure, and Google Cloud, and open-source tools that offer more control. This guide discusses key domains including Machine Learning, Data Engineering, DevOps, and Business Intelligence, comparing strengths, weaknesses, and use cases for each solution to help inform decisions.
-
Ansible Command Cheat Sheet
📦 Core Commands Command Description Example ansible Run ad-hoc commands with modules ansible all -i inventory.yaml -m ping ansible-playbook Execute a playbook ansible-playbook -i inventory.yaml site.yaml ansible-doc View module or plugin documentation ansible-doc copy ansible-config Display current configuration ansible-config list ansible-inventory Inspect or validate your inventory ansible-inventory -i inventory.yaml –list ansible-vault Encrypt/decrypt sensitive files ansible-vault…
-
Chrome on macOS Cannot Access Router — Full Analysis and Fix
A user experienced connectivity issues with accessing a local router on Chrome for macOS, encountering “ERR_ADDRESS_UNREACHABLE” while Safari works fine. Troubleshooting involved reinstalling Chrome, resetting settings, and command-line flags. The root cause was Chrome’s security mechanisms and user data directory conflicts. Cleaning the directory resolved the issue effectively.
-
A Fast Start with Ansible: Managing Your Servers Like a Pro
Ansible provides an agentless method to manage infrastructure, including cloud and physical servers. This guide covers project setup, secure password management with Vault, and execution of playbooks. It emphasizes the importance of inventory structure, common CLI commands, and troubleshooting tips, enabling users to automate their environments effectively.
-
How to Fix SSH “Client_loop: send disconnect: Broken pipe” Error
The Client_loop: send disconnect: Broken pipe error in SSH occurs due to inactivity, leading to connection loss. To resolve it, configure SSH to send keep-alive messages by modifying the server’s SSH daemon settings or using client command-line options. These changes help maintain uninterrupted SSH access and enhance productivity.
-
Practical Guide: Debugging Kubernetes Nodes and Managing Docker in Custom Setups
This guide addresses troubleshooting Kubernetes Node NotReady issues, configuring SSH and hostnames, fixing kubectl connection errors, and setting up a private Docker registry. It also covers accessing NodePort services and suggests advanced practices like using TLS for Docker registries. Additional learning resources are recommended for skill enhancement.
-
Running Local Docker Images in Minikube: A Quick Guide
Minikube allows you to run a single-node Kubernetes cluster locally, making it ideal for testing and development. However, Kubernetes typically pulls images from remote registries, which can be cumbersome when working with local Docker images. This guide explores two efficient methods to use your local Docker images within a Minikube cluster. Method 1: Load Local…
-
Updating Deprecated Commands in “Kubernetes in Action”
The post reviews the book “Kubernetes in Action” by Marko Luksa, highlighting deprecated commands in Kubernetes as it evolves. It updates commands for managing ReplicaSets instead of ReplicationControllers, emphasizes using deployments for scaling and updates, and suggests using declarative management for better resource management in Kubernetes.
Johngai’s Tech Forge 