Memory usage of vmdird is high in a VCHA environment

Products

• VMware vCenter Server

Issue/Introduction

• A Memory Exhaustion alarm is triggered on the vCenter Server.
  
  
• Executing the following command via SSH to the VCSA reveals high memory usage by the vmdird process:
  
  

  # ps -auwwx --sort -%mem
  USER         PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
  vmdird     10921  4.3 10.9 25651700 2351512 ?    Sl   Feb21 143:06 /usr/lib/vmware-vmdir/sbin/vmdird -L /var/log/vmware/vmdird/vmdird.log -l 0 -f /usr/lib/vmware-vmdir/share/config/vmdirschema.ldif -m standalone

 

• The affected environment is using vCenter HA (VCHA).
  Note: This is not vSphere HA.

Cause

Periodic health checks from the vcha service cause a memory leak in the vmdird service.

Resolution

Currently, there is no permanent resolution for this issue.

Workaround:

Restarting the vmdird service temporarily mitigates the issue.

1. Put VCHA in Maintenance Mode to avoid unnecessary failovers:
   
   1. Select the vCenter Server from the top of the vSphere Client inventory tree.
   2. Navigate to Configure > Settings > vCenter HA.
   3. Click EDIT, select Maintenance Mode, and click OK.
2. Restart the vmdird service:
   1. SSH into the VCSA.
   2. Execute the following command:
      service-control --restart vmdird
3. Re-enable VCHA:
   1. Select the vCenter Server from the top of the vSphere Client inventory tree.
   2. Navigate to Configure > Settings > vCenter HA.
   3. Click EDIT, select Enable vCenter HA, and click OK.