Security Explorer, Capacity overview UI is not displaying or slow with ERROR: No space left on device

Products

• VMware vDefend Firewall with Advanced Threat Prevention

Issue/Introduction

Issue / Introduction

In some environments, Security Explorer, Capacity Overview, and other UI pages in SSP may load very slowly or fail to display. In some cases, partial data may appear after a long delay.

During this issue, the following errors can be observed in the postgresql-ha-postgresql-0 pod logs:

ERROR: could not resize shared memory segment "/PostgreSQL.1103448508" to 67146752 bytes: No space left on device

Additional PostgreSQL logs may indicate that VACUUM is unable to clean up a large number of dead tuples:

LOG: automatic vacuum of table "pace.nsx_config.normalizedcomputeconfig" pages: 0 removed, 537374 remain tuples: 6371597 are dead but not yet removable

Long-running queries against these tables may also be observed:

 

SELECT managerid, metainfo FROM nsx_config.normalizedgroupconfig WHERE metainfo IS NOT NULL; 
SELECT id, metainfo FROM nsx_config.normalizedcomputeconfig WHERE metainfo IS NOT NULL;

Environment

SSP5.0,

SSP5.1

 

Cause

This issue occurs when the PostgreSQL shared memory directory (/dev/shm) becomes full.
As a result:

• PostgreSQL cannot resize shared memory segments

• VACUUM FULL cannot complete

• Dead tuples accumulate

This is commonly seen when:

• Large tables exist (normalizedcomputeconfig, normalizedgroupconfig)

• Vacuum cannot acquire required locks

• Parallel workers increase shared memory pressure

Resolution

SSP 5.1.1 and Later

SSP 5.1.1 includes a partial mitigation to reduce the frequency of this issue.

• This is not a permanent fix. It temporarily reduces memory pressure but does not fully prevent /dev/shm exhaustion. Root cause analysis is still ongoing.

Workaround (SSP 5.0 / 5.1)

• Manual workaround – contact Broadcom ANS Support for assistance if needed.