Process Terminated by Signal 11 (Segmentation Fault) Due to Memory Management Defect

Products

• VMware Tanzu Greenplum
• VMware Tanzu Greenplum / Gemfire

Issue/Introduction

Users may experience a crash of database connections with a "Process terminated by signal 11" error (Segmentation fault).

This often occurs when specific code paths incorrectly free memory or string resources before processing is complete.

Stack trace in the segment log:

..."PANIC","XX000","Unexpected internal error: Master process received signal SIGSEGV",,,,,,,0,,,,"
1    0x7f45870bd990 libpthread.so.0 <symbol not found> + 0x870bd990
2    0xe5cb31 postgres _ZN4gpmd15CMDFunctionGPDBD1Ev (CMemoryPool.h:240)
3    0xfdb133 postgres _ZN5gpopt8CMDCache5ResetEv (CMemoryPool.h:240)
4    0x116dac5 postgres _ZN9COptTasks12OptimizeTaskEPv (COptTasks.cpp:996)
5    0xe37a1e postgres _ZN4gpos5CTask7ExecuteEv (CTask.cpp:130)
6    0xe38808 postgres _ZN4gpos7CWorker7ExecuteEPNS_5CTaskE (CWorker.cpp:81)
7    0xe371f9 postgres _ZNK4gpos14CAutoTaskProxy7ExecuteEPNS_5CTaskE (CAutoTaskProxy.cpp:302)
8    0xe39224 postgres gpos_exec (CAutoP.h:109)
9    0x1168f71 postgres _ZN9COptTasks7ExecuteEPFPvS0_ES0_ (COptTasks.cpp:270)
10   0x116ae8e postgres _ZN9COptTasks18GPOPTOptimizedPlanEP5QueryP11SOptContext (COptTasks.cpp:1334)
11   0x116e8df postgres _ZN12CGPOptimizer18GPOPTOptimizedPlanEP5QueryPb (CGPOptimizer.cpp:60)
12   0xacd736 postgres optimize_query (orca.c:58)
13   0xac2ad7 postgres standard_planner (planner.c:391)
14   0xac2efb postgres planner (planner.c:344)
15   0xbb96dc postgres <symbol not found> (postgres.c:1033)
"

Environment

Greenplum 7.8.0

Cause

A code defect where Greenplum uses and frees a string/memory resource prematurely. This leads to an invalid memory reference, causing a segmentation fault.

Resolution

The issue is fixed in Greeplum database 7.8.4