Forum Discussion

sdhananjeyan_26's avatar
sdhananjeyan_26
Icon for Nimbostratus rankNimbostratus
Aug 30, 2016

Load Balance Across different Datacenters

I have installed BigIp in V-Sphere env.

 

during installation I configures interfaces internal, extertal to two networks in Vsphre 192.168.183.0/24, 192.168.181.0/24 respectively

 

if i create a virtual server having destination in external Ip range(Ex: 192.168.181.3) and with nodes running in internal ip range(ex: 192.168.183.5, 192.168.183.6) it is load balancing as expected. requests are distributed across nodes(snat configured as automap)

 

now i have a subnet in AWS ie.10.1.2.0/24 which is accessible from bigip machine through VPN connectivity(culr is working). Is it possible to create a node for a machine running in this aws subnet and loadbalance them using a virtual server. if i create a node for the machines in that subnet health check is passing but load balancing is not happening.

 

Since i configured 183 network to internal during installation am i restricted to load balance only nodes that are running in that network with IP range 192.169.183.0/24 ??? or is der any way to acheive it.

 

To Summarize: Is it possible to load balance two servers running in different ip range (diff vlans) Is it possible to have Big IP and Servers to be load balanced in different Data Centers or Cloud Providers

 

please share any docs that can help in achieving these things if possible

 

application delivery
BIG-IP
cloud
LTM

2 Replies

Sort By
  • KPS_149915's avatar
    KPS_149915
    Icon for Nimbostratus rankNimbostratus
    Aug 30, 2016

    Basically will not be reliable to do so, but you can try by creating the Nodes of different DC's and associate it in the member of same pool. But you might need to consider the RD, how to route the traffic on other DC Node which dosen't have self-IP configured on this let me know if that helps.

     

  • Henrik_S's avatar
    Henrik_S
    Icon for Nimbostratus rankNimbostratus
    Aug 30, 2016

    Hello, if curl and health check is working, but not server side connections, please debug this further with tcpdump/ssldump to see what's actually going on. Usually the fault lies within missing serverssl profiles (if ssl is used), or the fact that both curl and monitors will use a non-floating self-IP, while serverside connections by default will chose a floating self-IP.

     

    In regards to non-floating Self-IPs, please check that you have a valid route on your internal interface towards 10.1.2.0/24, if not the BIG-IP will actually use the MGMT interface for curl and monitors, and that would most certainly fail for serverside connections.

     

    Another thing to bear in mind is the growing use of SNI, and while curl passes this correctly, a https monitor does not support SNI, so you would have to use a external monitor for this.