Symptoms
You experience frequent connection drops on the desktop.You see errors similar to:
A network error occurred
Cause
The Horizon API is reporting back a network error which requires investigation on the network level.
Resolution
To resolve the issue, check the following:
Check if there are huge packet losses at the guest OS level. If yes, increase the ring buffer size of the VMware Horizon Connection Server and VMware Horizon Security Server virtual machines. For more information, see:
The output of esxtop show dropped receive packets at the virtual switch (1010071) Large packet loss at the guest operating system level on the VMXNET3 vNIC in ESXi (2039495)
Update Drivers/firmware on the ESXi Host NIC’s. Check that the ESX hosts are appropriately patched for that version.Verify Cpu allocation on the virtual machines. Ensure Cpu for specific virtual machines is not overallocated to physical sockets that is the virtual machine should not have more virtual sockets than the esxi host. Expand cores instead.Check the load balancer configuration between the connection server. The load balancer should not be doing heartbeat on the connection server by doing TCP probes. Instead, they should be issuing HTTP probes (GET /favicon.ico) as specified in the VMware documentation. Ensure that the load balancers are configured for session affinity and the session affinity table should be large enough to accommodate all the clients. Ensure that there is no discrepancy between the TIME_WAIT period for windows with respect to the load balancer. The windows server TIME_WAIT can be adjusted using the registry key: TcpTimedWaitDelay and accepts values from 30 seconds to 300 seconds with the default being 240 (4 minutes). The TIME_WAIT for the load balancer should be configured around the same value.
