A network shutdown would normally be a process that occurs as part of a scheduled z/OS IPL or swapping of the network from one LPAR to another. A scheduled IPL occurs during a period of little activity on the host and network. Usually, this means it occurs in the evening hours of a weekend.

The reasons for requiring a scheduled IPL are becoming fewer and fewer since hot-swappable devices and components are becoming prevalent on the mainframe. Some maintenance still requires an IPL, however, and there are also organizations that do scheduled IPLs as part of a regular routine. Regardless, if an LPAR is shut down, that means the network components must also be shut down.

A network shutdown might be independent of an LPAR shutdown. For example, in a partial network shutdown, TCP/IP or VTAM might be taken down independently. In such a situation, there might be a VIPA takeover process that might occur.

It could be a dynamic VIPA that moves automatically to a running LPAR in the sysplex. It could be a static VIPA that is moved by the network administrator during the shutdown. Or, it could be an external automation package moving the static VIPA automatically. Client connections might never be aware of the changes occurring.

Because TCP/IP depends upon VTAM, a shutdown of VTAM not only ceases all SNA communications, it also halts all IP communications. TCP/IP automatically detects that VTAM is no longer available and waits for it to be restarted.

This process is normally performed through automation. Note that when halting network VTAM and TCP/IP communications, the network administrator must ensure that access to the console is not also halted!