NSX-T Load balancer stats from edge nodes are not consistent

Products

• VMware
• VMware NSX Networking

Issue/Introduction

Symptoms:

• Current session counter of Virtual-Server or pool is greater than the real number.

For example, if you see the current number of sessions for VIP 10.40.14.34:8443 is 716 but if you actually count the session number using the session-table it is 102, which is far less than the stats output.

​​​​​edge01> get load-balancer <UUID> virtual-server <UUID> stats
Virtual Server
UUID                            : <UUID>
Display-Name                    : lb-pks-<UUID>-virtual-server
VIP                             : TCP 10.40.14.34:8443
Type                            : L4
Sessions                        :
    (Cur, Max, Total, Rate)     : (716, 749, 140784, 0)
Bytes                           :
    (In, Out)                   : (7863722100, 3133046605)
Packets                         :
    (In, Out)                   : (22142738, 10807216)


edge01> get load-balancer <UUID> session-tables | count 10.40.14.34
Number of lines that match pattern '10.40.14.34': 102

Environment

VMware PKS 1.x
VMware NSX-T Data Center

Cause

If an old pool member has been deleted, stats are not cleared from the Virtual-Server counter and the Pool counter

Resolution

This issue is resolved in VMware NSX-T Data Center 2.5, available at VMware Downloads.

Additional Information

To be alerted when this article is updated, click the subscribe button [cid:[email protected]] . For more information on KB subscription features, see the Knowledge Base Article FAQs: How to Subscribe to VMware Knowledge Base Articles (76417).