Environment
Vertica Analytics Platform 12.0.4.-18Situation
A production six‑node cluster running Vertica 12.0.4‑18 in Enterprise mode on Linux 8.8 experienced a severe performance degradation. Vertica‑reported CPU and memory usage on all nodes remained below 50%. As part of initial remediation, Vertica was stopped on node 1 and the host was rebooted, but the issue persisted.
A simple diagnostic test showed that, despite identical hardware metrics (CPU, memory, network, I/O, and load), the query SELECT 1 required approximately 1 second on node v_ops_node0002 (10.218.0.17), compared to the usual sub‑10 ms response time. A full node restart and host reboot eventually resolved the condition.
Scrutinize data can be provided for further analysis.
