Guardian Programmer's Guide

Using the File System
Guardian Programmer’s Guide — 421922-014
2 - 9
Location Independent Disk-File Names
Note that a process descriptor always contains a node name and a sequence number.
It never contains qualifiers.
Location Independent Disk-File Names
Location independent disk-file names are supported by the NonStop Storage
Management Foundation (SMF) product, which is designed to help automate system
storage-management tasks. Location independent naming means that a disk file has
both an external, or logical, name and an internal, or physical, name.
Normally, a disk file’s name indicates the location of the file. For example, the file
\SYS99.$BIGVOL.MYSUBVOL.MYFILE would designate a file located on the
subvolume MYSUBVOL, on the volume $BIGVOL, on the node \SYS99. However, if
this file were managed by the SMF subsystem, its location would be independent of
the name, except for the name of the node.
The SMF subsystem controls the mapping of the external name of a file to the internal
name. This allows the internal name, which identifies the file’s physical location to the
disk process, to change when a file is moved to a different location, while the external
name remains the same to applications and to users. The mapping function is
transparent to applications and to users.
The external name of a file managed by the SMF subsystem follows the normal syntax
for a disk file name; you cannot tell that it is an SMF external name by looking at it.
You can perform any normal operation on the file by using its external name.
However, there are restrictions against directly accessing an SMF file by its internal
name. Also, information requests based on the internal name are disallowed unless
explicitly asked for. (For example, wild-card searches either by TACL commands, such
as the command FILEINFO $VOL.*.*, or by calls to the FILENAME_FIND* procedures,
do not return information about files contained in the ZYT* and ZYS* subvolumes,
which are reserved for SMF internal files; ZYS* and ZYT* must be specified to get
information on internal files that they contain.)
For more information on the SMF product and how to use it, refer to the NonStop
Storage Management Foundation User’s Guide.
Passing File Names to Processes
There are two ways in which you can pass file names to a process:
•
Using a CLASS MAP DEFINE (or other DEFINE CLASS that passes file names)
•
Using the startup sequence of messages
Either of these methods allows you to use the same program to access different files
without changing your program code.
These concepts are introduced below. For simplicity, early sections of this guide refer
to file names directly, not by DEFINE name or by reference to the startup sequence.