ESXi Ramdisk /tmp is 100% full

Products

• VMware vSphere ESXi
• VMware vSphere ESXi 8.0

Issue/Introduction

This article provides troubleshooting steps when the ESXi /tmp ramdisk reaches 100% utilization. A full /tmp partition prevents vMotion operations, causes HA agent failures, and restricts access to the VMware Host Client

Symptoms:

• vMotion fails to initiate.
• HA agents stop responding.
• Unable to log into the VMware Host Client.
• The image profile on the ESXi host shows 'Unknown-no profile defined'.
• vmkwarning.log contains: WARNING: VisorFSRam: 203: Cannot extend visorfs file ... because its ramdisk (tmp) is full.

Environment

VMware vSphere ESXi

Cause

The /tmp partition is a memory-backed ramdisk with limited capacity. It fills when excessive logs, stale patch artifacts, or temporary process files (e.g., vim-cmd*.txt from NSX-T or auto-backup.sh scripts) accumulate.

Resolution

1. Verify the current space utilization: 

   vdf -h | grep tmp

2. Identify the largest files or directories consuming space: 

   ls -lSh /tmp | head -n 20

3. Analyze the results:
   • FDM VIB artifacts: Remove stale VIB installation files found in /tmp.
   • NSX-T exporter: If vim-cmd*.txt files are present, restart the service: /etc/init.d/nsx-exporter restart and remove the files.
   • Generic files: Remove confirmed obsolete log files or misplaced files manually using rm.
4. Verify space restoration: 

   vdf -h | grep tmp

Note: A host reboot will clear the /tmp ramdisk temporarily, but identifying the root cause is necessary to prevent recurrence.

Additional Information

• If the issue persists, contact Broadcom Support for further analysis: Contact Broadcom Support.