vROP's DVS data collection takes an unexpectedly long time.

Products

• VMware vCenter Server
• VMware vSphere ESXi

Issue/Introduction

• vROP's data collection takes more than 5 minutes

• DistributedVirtualSwitch#fetchPorts API takes an unexpectedly long time.

• VC 'FetchDVPorts' API response time significantly degraded.

Cause

The API call "DistributedVirtualSwitch#fetchPorts" to fetch ports on the Distributed virtual switch(DVS) is directly related to how many ports on the DVS. A port does not have to be in use to be scanned.

So if the port utilization is 1500 used / 65000 total = 2.3% the scan will iterate through all DVS 65000 ports taking a longer time. 
 

Resolution

Lower the number of total ports to a closer number to the used ports.

If 1500 ports are used you should have around (Used +100 ports = 1600 )

To change the port allocation 

Navigate to the networking tab in vCenter 
Click on the DVS
Under the Summary tab you can see the totals for the entire vSwitch



To get a more specific count on how many unused ports click on the summary tab of each port group 



The total unused ports are 241

Right-click and Edit settings of the port group 

Add the correct number of ports closer to the total used. 


Select Ok

Note: Do this for every port group to reach a more accurate time to scan. 

Workaround:
Exclude DVS's from the vROPS data collection

Additional Information

Impact/Risks:

Slower vROPS data collection for DVS's