Tasks performed either from vCenter or directly from the ESXi hosts are slow

Products

• VMware

Issue/Introduction

The address the performance issue related to tasks.

Symptoms:

• Any tasks performed either from vCenter or when connected directly to the ESXi hosts are slow
• Hostd cpu usage is 100% when checking in esxtop
• CPU usage is 100% when running  below command and opening of html -  https://esxip/mob/?moid=networkSystem&doPath=networkInfo

         vim-cmd hostsvc/advopt/update Config.HostAgent.plugins.solo.enableMob bool true

• Running below command   shows 

          time vim-cmd hostsvc/net/info 
          real 1m 10.91s
          user 0m 0.38s
          sys 0m 0.00s      

• hostd.log show below messages

2019-10-01T13:00:52.567Z warning hostd[3193604] [Originator@6876 sub=PropertyProvider opID=2dfb8fd3 user=dcui] It took 136717325 microseconds, 00:02:16.717325 (hh:mm:ss.us), to get property network for vim.Datacenter:ha-datacenter
2019-10-01T13:03:09.255Z warning hostd[3193604] [Originator@6876 sub=PropertyProvider opID=2dfb8fd3 user=dcui] It took 136685348 microseconds, 00:02:16.685348 (hh:mm:ss.us), to get property childEntity for vim.Folder:ha-folder-network
2019-10-01T13:05:26.440Z warning hostd[3193604] [Originator@6876 sub=PropertyProvider opID=2dfb8fd3 user=dcui] It took 137166500 microseconds, 00:02:17.166500 (hh:mm:ss.us), to get property network for vim.ComputeResource:ha-compute-res
2019-10-01T13:07:42.239Z warning hostd[3193604] [Originator@6876 sub=PropertyProvider opID=2dfb8fd3 user=dcui] It took 135794524 microseconds, 00:02:15.794524 (hh:mm:ss.us), to get property network for vim.HostSystem:ha-host
 

Note: The preceding log excerpts are only examples. Date, time and environmental variables may vary depending on your environment

Cause

This issue occurs due to many large number of  unused portgroups in vSwitch.

Resolution

This issue is resolved in VMware vSphere ESXi 6.7 Patch ESXi670-202004002

To download ESXi patch refer to How to Download ESXi, vCenter Server Patches in Customer Connect

Workaround:
To workaround this issue

1. Reduce the number of port groups on vSphere Standard Switch(vSS).

Below steps will cleanup the unused portgroups in vSwitch.
    To manually delete the portgroup from a console or SSH session:

1. List the available vSwitches with their associated portgroups:
   
   # esxcfg-vswitch --list
    
2. Verify that the portgroup you want to remove exists on the vSwitch:
   
   # esxcfg-vswitch --check-pg="PG Name" vSwitchX
   
   Replace vSwitchX with the vSwitch number where the affected port group was built.
    
3. Delete the portgroup with:
   
   # esxcfg-vswitch --del-pg="PG Name" vSwitchX

2. Use VGT(Virtual Guest VLAN tagging ) to add VLAN tag in the guest VM, other than in the vSwitch. For more information refer to VLAN configuration on virtual switches, physical switches, and virtual machines

Additional Information

Impact/Risks:
We will be reducing the portgroups to lesser number.