Workload Management (WLM) Rules Fail to Enforce when Greenplum Command Center (GPCC) is Stopped
search cancel

Workload Management (WLM) Rules Fail to Enforce when Greenplum Command Center (GPCC) is Stopped

book

Article ID: 441148

calendar_today

Updated On:

Products

VMware Tanzu Data Suite VMware Tanzu Greenplum VMware Tanzu Greenplum / Gemfire

Issue/Introduction

When the Greenplum Command Center (GPCC) server process is stopped—often to reduce system "chattiness" or overhead—Workload Management (WLM) rules stop being enforced.

This results in the loss of critical controls such as idle session termination and advanced query rules based on real-time performance metrics.

Customers may seek a way to keep WLM active while GPCC is offline.

Cause

The WLM rule engine is not a native component of the Greenplum core database engine. Instead, it is a submodule of the GPCC application. It relies on the GPCC metrics collector to provide the real-time telemetry required to evaluate complex rules. Consequently, if the GPCC service is shut down, the engine that evaluates and executes these rules is also terminated.

Resolution

The WLM rule engine is architecturally embedded within the GPCC server process and cannot run as a standalone service. To enforce workload limits without GPCC, users must utilize native Greenplum Database (GPDB) controls:

  • Concurrency & Memory: Use Resource Groups to enforce concurrency limits and memory ceilings (ALTER RESOURCE GROUP ... SET CONCURRENCY).
  • Statement Timeouts: Configure the statement_timeout GUC to automatically terminate long-running queries.
  • Idle Sessions (GP7): Greenplum 7 supports native idle session management via the idle_session_timeout GUC.
  • Idle Sessions (GP6): There is no native equivalent in GP6; GPCC must remain active to enforce idle session rules.
  • Advanced Rules: Rules based on CPU usage, disk I/O, or spill files require GPCC, as these metrics are only gathered by the GPCC metrics collector extension.