GPSS Excessive Log Generation: "throttle for job already closed" Warning

Products

• VMware Tanzu Data Intelligence
• VMware Tanzu Data Suite
• VMware Tanzu Data Suite
• VMware Tanzu Greenplum
• VMware Tanzu Greenplum / Gemfire

Issue/Introduction

Users may experience excessive log generation and high disk space consumption in Greenplum Stream Server (GPSS) due to a flood of warning messages. The GPSS logs rapidly fill with the following warning, often recording multiple entries per millisecond:

YYYYMMDD HH:MM:SS.ms,[ID],[PID],warning,throttle for job:[job_name] already closed
YYYYMMDD HH:MM:SS.ms,[ID],[PID],warning,throttle for job:[job_name] already closed
YYYYMMDD HH:MM:SS.ms,[ID],[PID],warning,throttle for job:[job_name] already closed

Environment

 

Version: Greenplum Stream Server (GPSS) 1.X

 

Cause

The log entry does not indicate an issue with rate or interval-based throttling logic. In this context, the word "throttle" refers to GPSS's internal throttle flow buffer node, which sits between the source reader and the batch/gpfdist pipeline.

When running certain jobs (such as RabbitMQ jobs), there are multiple code paths that call a close function on this same throttle buffer. The first call successfully closes the buffer; however, every subsequent call falls into a fallback branch in the code and generates the warning log.

Note: This warning is entirely harmless to the data loading path and data integrity. It is simply a verbose logging behavior that clogs the logs.

Resolution

GPSS does not support toggling or filtering log generation at the individual line level. To resolve this issue, please use one of the following approaches:

Workaround (For GPSS 1.X)

To prevent these warnings from consuming excessive disk space, change the GPSS logging level to error. This will suppress the harmless warning messages while still capturing critical system errors.

Permanent Fix

Upgrade to GPSS 2.0 or a later version. These redundant warning message was successfully removed in GPSS 2.0.