When you define your network resources for XRF, you should consider
the following situations:
- If you allow dynamic CDRSCs and trial and error rerouting, you
do not need to predefine a CDRSC for the USERVAR name or CDRSCs for
the active and backup application program.
Note: Referencing dynamic
CDRSCs with USERVARs can cause situations in which session initiations
fail intermittently, because dynamically
defined CDRSCs can be deactivated and deleted by VTAM®, depending on the dynamic CDRSC timer value
(CDRSCTI) and whether there are any active sessions.
- If you do not allow dynamic CDRSCs to be created in the network,
you need to predefine a CDRSC for the USERVAR name, a CDRSC for the
active application program, and a CDRSC for the backup application
program. If you do predefine a CDRSC with the USERVAR name, you should
not specify the cross-domain resource manager (CDRM) name because
the USERVAR may not always be managed from the same SSCP. However,
you should define the CDRSC under the network in which the USERVAR
resides to reduce the amount of time for the ADJSSCP search process.
The following applies regardless of whether you use USERVARs with
XRF:
- You can use USS tables, interpret tables, or both with USERVAR
translation; the USS and interpret processing is done first and USERVAR
translation is done using the output of those processes.
- You are not required to define interpret table entries for terminals
to use the USERVAR function. If you have already done so, however,
you do not have to remove them.
- If the application is cross-domain, the USERVAR name must be defined
as a CDRSC using either static or dynamic definitions and must be
unique within the network.
- The OLU host of a session initiation and the host that is resolving
the USERVAR name must support name translation. For VTAM 4.1 and earlier, the TRANSLAT start option
must include USERVAR. In VTAM 3.4.2 and later, RACALIAS must include USERVAR in ISTRACON.