Tencent Cloud Developer
Jan 14, 2021 · Databases
PostgreSQL Master-Slave Replication Performance Optimization: Solving Drop Table Bottleneck
The article explains how massive DROP TABLE operations in PostgreSQL master‑slave replication trigger costly buffer‑invalidation loops that cause severe lag, and describes extracting this step into a separate subprocess with a shared hash table, cutting replication lag from over 400 GB to about 10 MB—a 30,000‑fold speedup.
Database InternalsDatabase ReplicationMaster-Slave Replication
0 likes · 7 min read