How do I perform recovery using client’s alternate IP address?

---

I want to perform a network install of a client using a differnet network than is used for daily applications. When trying to perform the network install, I get the mesasge:

“Your host (unknown) is not defined as either a valid client or admin system on host aixrisc1.”

I defined the client on the network administrator using its primary network, but I want the install to take place over the alternate network. I define the “Alternate Networks” address to use in the server record, but the client still cannot talk to the server using that address (which is what I specify in the system installation menus).

ANSWER

---

---

To answer your question, it is best to describe the process of configuring client access permissions on the server. Only client IP addresses with explicit permission may execute commands or access data on a server.

When you define a client using the administrator, the server records are updated to allow access to the client IP address according to standard hostname resolution.

When you have alternate server networks, the server permission record is updated to contain the corresponding client IP address on the alternate network. To do so, each client is told to “contact the server using its alternate IP address and report back the client IP address that the server sees”. This assumes that the client’s routing table is such that the request is routed through the alternate adapter. The client IP address that the server sees is then added to the server’s access configuration file, allowing future access from the client using that (alternate client) address.

Now, to obtain the “alternate client IP”, the client must, of course, be available. If you’re planning to install a new client using the server’s alternate adapter, this may not be the case. So, when configuring the client for network boot/install, we ask for the client’s IP address if you also select to install using the alternate server address. We then include access from that client IP in the server record.

The problem here is that performing recovery of a client that was not previously available on the alternate network. Since you never configured the client for network boot/install, and the client has never been available, the server doesn’t know the client’s alternate IP address (as you defined in the system installation menus), and therefore refuses requests from that address.