Load balancers perform passive health checks to monitor failures during client connections and mark servers causing consistent failures as DOWN.
Passive health check monitors client traffic going through the load balancer for failures. For example, if a pool member sends a TCP Reset (RST) in response to a client connection, the load balancer detects that failure. If there are multiple consecutive failures, then the load balancer considers that server pool member to be temporarily unavailable and stops sending connection requests to that pool member for some time. After some time, the load balancer sends a connection request to check if the pool member has recovered. If that connection is successful, then the pool member is considered healthy. Otherwise, the load balancer waits for some time and tries again.
Passive health check considers the following scenarios to be failures in client traffic.
- For server pools associated with Layer 7 virtual servers, if the connection to the pool member fails. For example, if the pool member sends a TCP RST when the load balancer tries to connect or perform a SSL handshake between load balancer and the pool member fails.
- For server pools associated with Layer 4 TCP virtual servers, if the pool member sends a TCP RST in response to client TCP SYN or does not respond at all.
- For server pools associated with Layer 4 UDP virtual servers, if a port is unreachable or a destination unreachable ICMP error message is received in response to a client UDP packet.
Server pools associated to Layer 7 virtual servers, the failed connection count is incremented when any TCP connection errors, for example, TCP RST failure to send data or SSL handshake failures occur.
Server pools associated to Layer 4 virtual servers, if no response is received to a TCP SYN sent to the server pool member or if a TCP RST is received in response to a TCP SYN, then the server pool member is considered as DOWN. The failed count is incremented.
For Layer 4 UDP virtual servers, if an ICMP error such as, port or destination unreachable message is received in response to client traffic, then it is considered as DOWN.
- With admin privileges, log in to NSX Manager.
- Select .
- Enter a name and description for the passive health monitor.
- Configure the values to monitor a service pool.
You can also accept the default active health monitor values.
Option Description Fall Count Set a value when the consecutive failures reach this value, the server is considered temporarily unavailable. Timeout Period Set the number of times the server is tested before it is considered as DOWN.For example, when the consecutive failures reach the configured value 5, that member is considered temporarily unavailable for 5 seconds. After this period, that member is tried again for a new connection to see if it is available. If that connection is successful, then the member is considered available and the failed count is set to zero. However, if that connection fails, then it is not used for another timeout interval of 5 seconds.
- Click OK.
What to do next
Associate the passive health monitor with a server pool. See Add a Server Pool for Load Balancing in Manager Mode.