Linux High Availability: Keepalived or Pacemaker? The Definitive Decision Guide

Brought to you by DevXOps — https://devxops.tech

Still confused about Keepalived vs Corosync + Pacemaker? You are not alone. This animated masterclass walks you through everything an SRE or DevOps engineer needs to know about Linux high availability -- from architecture deep dives to real-world decision frameworks.

What you'll learn:

What Keepalived and Corosync + Pacemaker actually are and how they differ
Why each tool exists -- the problems they were built to solve
Keepalived architecture: VRRP, three-process daemon, Netlink, IPVS, BFD
Corosync + Pacemaker architecture: CIB, CRMd, LRMd, STONITH, votequorum
Failover flows compared -- VIP migration vs ordered service recovery
Where each fits in production -- network edge vs service orchestration
The key differences: VIP switch vs service orchestrator
Two-node cluster dangers and the split-brain problem
qdevice / qnetd as an external quorum tie-breaker
A concrete decision framework with real-world examples
Common SRE mistakes that cause production incidents
Recommended production patterns for stateless and stateful HA
Modern alternatives: cloud LBs, Kubernetes, Patroni, app-native HA
Latest developments: Keepalived 2.3.x, Pacemaker 3.0.x, Corosync security patches
Key concepts: Virtual Router Redundancy Protocol (VRRP), Virtual IP (VIP), Gratuitous ARP, IP Virtual Server (IPVS), Bidirectional Forwarding Detection (BFD), Corosync totem, votequorum, Cluster Information Base (CIB), Cluster Resource Management Daemon (CRMd), Local Resource Manager Daemon (LRMd), STONITH (Shoot The Other Node In The Head), fencing, quorum, qdevice, qnetd, resource agents, ordering constraints, colocation constraints

Resources:

Keepalived Documentation: https://www.keepalived.org/doc/software_design.html
ClusterLabs Pacemaker: https://projects.clusterlabs.org/w/projects/pacemaker/
Red Hat HA Documentation: https://docs.redhat.com/en/documentation/red_hat_enterprise_linux/9/html/configuring_and_managing_high_availability_clusters/
RFC 5798 (VRRP v3): https://datatracker.ietf.org/doc/html/rfc5798
Corosync qdevice: https://github.com/corosync/corosync-qdevice
DevXOps -- Practical DevOps and SRE content. Subscribe for more: https://devxops.tech
Timeline:
0:00 Hook: Database Down, VIP Never Moved
0:55 What Are These Tools?
2:30 Why Do They Exist?
3:55 Keepalived Architecture Deep Dive
5:35 Corosync + Pacemaker Architecture Deep Dive
7:20 Failover Flow Inside Each Tool
9:00 Where Each Fits in Production
10:30 Why They're Confused and Key Differences
11:50 Two-Node Clusters and Split-Brain
13:10 Decision Framework and Real-World Examples
14:40 Common SRE Mistakes and Production Patterns
16:05 Modern Alternatives and Latest Developments
17:20 Recap and Close

Hashtags
#Keepalived #Pacemaker #Corosync #LinuxHA #HighAvailability #DevOps #SRE #VRRP #STONITH #Failover #Clustering

Видео Linux High Availability: Keepalived or Pacemaker? The Definitive Decision Guide канала DevXOps Tech