Raul102801
Technical User
I am trying to setup HACMP/XD over a WAN with IBM's help. I am supposed to assist with the network portion and can't seem to quite understand the whole process so I'll try to explain the best I can.
We have 2 IBM Pseries running AIX 5, this 2 servers will be located in different states as part of a dissastr recovery plan. each server has 7 interfaces and each server is in a different subnet.
site A Site B
12.1.1.2 12.1.2.2
12.1.1.3 12.1.2.3
12.1.1.4 12.1.2.4
12.1.1.5 12.1.2.5
12.1.1.6 12.1.2.6
12.1.1.7 12.1.2.7
12.1.1.8 12.1.2.8
There is also what is called Service IP label 192.168.12.12 which is a virtual address bonded to 12.1.1.2 and is the address that the oracle clients will look for. If Site A fails, the service IP label will become active on site B automatically on 12.1.2.2 so the oracle clients will keep working on a different server with no interruption of service or at least very little. Here is my question...
I don't understand how the oracle clients can get to the service IP label since it is on a network that does not exist within my switch. To prove the point I try to ping it from an address on a different vlan 12.1.3.3 with no success. Only if I add the oracle client PC to the same vlan and assign a static ip within the same range than 12.1.3.3 will I ping that address. IBM claims that is the way the system needs to be setup but noone can tell me how it is going to work...
am I missing something?
We have 2 IBM Pseries running AIX 5, this 2 servers will be located in different states as part of a dissastr recovery plan. each server has 7 interfaces and each server is in a different subnet.
site A Site B
12.1.1.2 12.1.2.2
12.1.1.3 12.1.2.3
12.1.1.4 12.1.2.4
12.1.1.5 12.1.2.5
12.1.1.6 12.1.2.6
12.1.1.7 12.1.2.7
12.1.1.8 12.1.2.8
There is also what is called Service IP label 192.168.12.12 which is a virtual address bonded to 12.1.1.2 and is the address that the oracle clients will look for. If Site A fails, the service IP label will become active on site B automatically on 12.1.2.2 so the oracle clients will keep working on a different server with no interruption of service or at least very little. Here is my question...
I don't understand how the oracle clients can get to the service IP label since it is on a network that does not exist within my switch. To prove the point I try to ping it from an address on a different vlan 12.1.3.3 with no success. Only if I add the oracle client PC to the same vlan and assign a static ip within the same range than 12.1.3.3 will I ping that address. IBM claims that is the way the system needs to be setup but noone can tell me how it is going to work...
am I missing something?