Tag: #고가용성
Kubernetes, DevOps and MLOps, written from things we ran and measured ourselves · 2 posts
Why Two Servers Cannot Achieve HA 2: What a Single Witness Changes and Remaining Pitfalls
Adding a witness VM with no data (2 vCPUs, 2 GiB) to two data nodes enables automatic failover. When we isolated a node, PostgreSQL promoted a new master in 22.6 to 32.7 seconds, and Redis Sentinel in 6.5 to 7.0 seconds,
2026-10-30 · 26 min read #고가용성#HA#PostgreSQL#Redis#OpenSearchWhy 2 Servers Are Not Enough for HA 1: Practical Measurements for PostgreSQL, Redis, OpenSearch, ClickHouse, and SeaweedFS
It might seem that with 2 servers, if one dies the other will take over, but in reality, that is not the case. All five services—PostgreSQL (Patroni + etcd), Redis (Sentinel), OpenSearch, ClickHouse (Keeper), and Seaweed
2026-10-30 · 25 min read #고가용성#HA#PostgreSQL#Redis#OpenSearch