Testing an Active/Passive Datapower Cluster is very easy. Just create exactly the same XML Firewall service on both Datapower A and Datapower B appliances. Specify the front End-Handler IP address as the VIP address created.
Execute the curl command to run the policy where the VIP is the Front Side Handler. Enable the probes on both devices to make sure that the request and response is successful. Switch the active appliance (Datapower A) off or reboot and continue to execute the command. You will see that the passive appliance will take over from Datapower A. When Datapower A is rebooted or switched on again it will take over again.
6 thoughts on “Testing Active/Passive Datapower Cluster”
Using HTTP(S) FSH, I assume we would also use the VIP address as the listening IP?
Correct you will use the VIP address.
In case I want to use both the datapowers in Active-Active self balancing mode, which datapower receives the first request ?
I used the same VIP on datapower 1 and 2. Now I try to test it by sending an SOAP message to the XML firewall with curl using the VIP address. However, the CIP address is not resolvable on my windows machine wehere I ran curl.
Thanks and regards
After a while, the VIP is resolvable on my windows DP client machine. I followed your steps to create active-passive DP cluster (one 1 with 100 as priority and the other 110). The same XML firewall service was created on both with the VIP address (192.168.253.211 for example) as the device IP. I can ping the VIP on windows command window. However, if I run the following command when both datapower are live,
curl –data-binary @test1.xml http://192.168.253.211:2048 -u :
I got: curl: (55) Send failure: Connection was reset
If I bring down one of the datapower, I got the SOAP request message back from the XML firewall service.
I really don’t know why it does not work when both DP are running.
The issue is resolved via Mac addresses folloing IBM technotes.