Unbalanced storage controller traffic and Round Robin multipathing configuration for ESXi hosts

Products

• VMware vCenter Server

Issue/Introduction

Monitoring tools (such as Zabbix) indicate that iSCSI storage traffic is unbalanced across array controllers. Traffic heavily utilizes Controller 1, while Controller 2 experiences minimal utilization, despite both controllers reporting as healthy.

Symptoms

• Storage traffic is unbalanced across controllers (e.g., Controller 1 is heavily utilized while Controller 2 is idle).
• Running command esxcli storage nmp device list shows that the LUN is restricted to a single working path using the Most Recently Used path selection policy:
  naa.####
   Device Display Name: QSAN iSCSI Disk (naa.####)
   Storage Array Type: VMW_SATP_ALUA
   Storage Array Type Device Config: {implicit_support=on; explicit_support=off; explicit_allow=on; alua_followover=on; action_OnRetryErrors=off; {TPG_id=0,TPG_state=AO}{TPG_id=1,TPG_state=AO}}
   Path Selection Policy: VMW_PSP_MRU
   Path Selection Policy Device Config: Current Path=vmhba66:C0:T0:L1
   Path Selection Policy Device Custom Config:
   Working Paths: vmhba66:C0:T0:L1
   Is USB: false
• Furthermore, from the command esxcfg-mpath -b output confirms that the storage devices have multiple active paths available, but they are not being actively utilized for I/O.

Environment

• VMware vSphere 7.x / 8.x / 9.x

Cause

This issue occurs because the VMware Native Multipathing (NMP) configuration is utilizing the Most Recently Used (VMW_PSP_MRU) Path Selection Policy instead of a load-balancing policy. Because MRU does not actively load balance across available paths, the ESXi host sends all traffic to the active paths on a single controller, leaving the active paths on the secondary controller idle.

Cause validation 
Run the command  esxcli storage nmp device list to verify the Path Selection Policy ,in below example PSP is currently set to Most Recently Used (VMW_PSP_MRU) and only a single working path is active per LUN.
   naa.#############
   Device Display Name: QSAN iSCSI Disk (naa.############)
   Storage Array Type: VMW_SATP_ALUA
   Storage Array Type Device Config: {implicit_support=on; explicit_support=off; explicit_allow=on; alua_followover=on; action_OnRetryErrors=off; {TPG_id=0,TPG_state=AO}{TPG_id=1,TPG_state=AO}}
   Path Selection Policy: VMW_PSP_MRU
   Path Selection Policy Device Config: Current Path=vmhba66:C0:T0:L1
   Path Selection Policy Device Custom Config:
   Working Paths: vmhba66:C0:T0:L1
   Is USB: false

Resolution

Change the VMware Native Multipathing (NMP) Policy from Most Recently Used (VMW_PSP_MRU) to Round Robin (VMW_PSP_RR). The Round Robin policy uses an automatic path selection algorithm that rotates I/O across all available active paths, distributing storage traffic evenly across both controllers.

Execute the following steps in the vSphere Client:

1. Log in to the vSphere Client.

2. Select the target ESXi host.

3. Navigate to Configure > Storage > Storage Devices.

4. Select the target Storage Device from the list.

5. In the bottom pane, click the Properties tab.

6. Scroll down to Edit Multipathing Policies.

7. Change the Path Selection Policy from Most Recently Used (VMW_PSP_MRU) to Round Robin (VMW_PSP_RR).

8. Click Save.

Additional Information

For more details on managing storage paths, refer to the Broadcom TechDocs: Change the Path Selection Policy