How to have a VIP for 2 servers ( Prod and Dr )

Question

We have prod and dr servers, we would like to have VIP for them. They are not exposed to internet. Any one server will be active for 1st 6 months, and after DR drill DR server will act as prod for next 6 months. Here, we have upstream systems which pushes files ( csv or text or zip ) via SFTP to our servers which would be nearly 200 - 300mb size per day. Currently, during every DR drill these upstream systems need to raise change request to update the IP before DR drill. This take atleast 2 weeks. to resolve this issue we decided to provide a VIP from our end. So that they can use VIP to transfer files via SFTP.

Note: DR server will be up and won't be active. App Services won't be up File transfer is not recommended via SFTP on F5 network.( we are not on F5 ). Both the servers Prod and DR running on vmWare.

We would like to have a VIP for these servers. Need your advice and suggestion.

Thanks in advance. Bala

Answer 1

Bala, I think I understand your question, it's not quite clear what the question is but my perception leads me to believe that you are trying to determine how to load balance the two server nodes.

== > first of all your group will have to acquire an F5 load balancer that is configured in accordance with your network requirements. I am assuming the load balancer is already live on the network. In order to load balance the two servers, you will have to create a pool consisting of the two servers, once the pool has been created, you then create a virtual server and associate the pool with the virtual server. Below are the essential steps required to make this happen. Also note that, the server nodes have to be added to the Nodes in the load balancer(this has to be done first)

Add Node:

Go to local Traffic --- > Nodes -- > Create a. Give the nodes a name b. enter the IP address of the node in the IP field. c. In the Configuration section, select "Node Default" for the Health Monitor leave the rest at their default settings of 1,0,0.

Create Pool:

From the GUI go to Local Traffic -- > Pool --- > Create

a. give the pool a name b. For now use tcp as the monitor (select from the available options) c. In the Resources section Fill in the ff

Load Balancing Method == > Round Robin (traffic distributed in a circular
fashion)

Other options include Least connections, observed, random .... much more, a
good reference which has links to creating pools, vips etc.

https://support.f5.com/kb/en-us/products/big-ip_ltm/manuals/product/ltm_configuration_guide_10_0_0/ltm_pools.html

Create the Virtual Server:

In the GUI, go to --- > Local Traffic --- > Virtual Servers -- > create

a. Give the virtual server a name b. In the Type description Select "Standad" in this case, there are other options, that do not apply to your request at the moment, but I advice you to read up on them they are (forwarding (layer 2), Forwarding IP, Performance HTTP, Performance (layer 4), stateless, Reject, Internal) c. In the source field enter 0.0.0.0/0 d. In the Destination Field Select "Host" and enter the IP address of the vip which is normally the url's dns address. e. Select the service Port, for http traffic select http/80, this could be whatever port your services are being heard on. Note for port 443/https, you will require an ssl certificate. f. In the Configuration section , select Advanced and select the ff:

Protocol: TCP
Protocol Profile (client): tcp
Http Profile: http
Snat : Autosnat

I am assuming you are using Autosnat here, this is much simpler to deal with, otherwise, a snat pool will have to be created.

g. At the bottom under "Resources" In the drop down for the "Default Pool" select the pool you created above.

h Select "Source Address" for the "Default Persistence Profile"

Click finish.

At this point, if the server nodes are live, configured with the appropriate page, access to the appropriate resources should be reachable. There are other criteria such as monitors which can be specifically configured to monitor a particular page, but that is for another session.

I hope I pointed you in the right direction.

Note: You have to determine the type of service and application running on the servers, if the url requires request to return to the same server then in this case