0

I'm evaluating a VE LTM Trial, 25 Mbps, BIG-IP 12.1.1 Build 2.0.204 Hotfix HF2
It's running on Hyper-V on Windows Server 2012R2.

When I run ping from the Hyper-V console window of the LTM VM I can measure the following times:

ping -I 172.27.50.1 172.27.50.151 = 7 ms .. 30 ms (pinging from the LTM internal static self-IP to another VM attached to the same Virtual Switch)

ping -I 172.27.50.1 172.27.50.161 = 7 ms .. 30 ms
(pinging from the LTM internal static self-IP to another VM reached through the external network, through a physical switch)

ping -I 172.27.50.1 172.27.51.1 < 1 ms (pinging from the LTM internal static self-IP to the LTM external static self-IP)

ping -I 172.27.50.1 172.27.52.1 < 1 ms (pinging from the LTM internal static self-IP to the LTM management address)

ping -I 172.27.50.1 172.27.51.51 = 2 ms .. 4 ms (pinging from the LTM internal static self-IP to any of the configured LTM Virtual Servers)

pings between the two devices over the HA VLAN are even higher: tens of ms !

I reserved what I judge to be the recommended amounts of vCPU and memory to the LTM VE. I have also disable Virtual Machine Queues in the PhyNICs and in the LTM VNICs.

Has someone suggestions of configurations to check/change, or troubleshooting procedures to reveal the cause of the high latencies above ?

Zamoth
  • 23
  • 5
  • A packet capture running at both ends (with clocks synced) will reveal whether this is a local processing delay or delay on the network between them. – Jason Rahm Jan 06 '17 at 18:07

1 Answers1

1

My experience says that you'd be better off running F5 VE under VMware Player/Workstation/ESX. I'm running a VE under VMware ESX and even from a remote location connected by 600M I'm getting <1ms ping responses.

I'm not going to join the ESX - Hyper-V war/discussion but F5 VE seems to get along quite well with VMware products. You might look and see how LTM improves your app performance and not concentrate on ping response times.

nutcase
  • 80
  • 7