|
HP OpenVMS Systemsask the wizard |
|
The Question is:
What is a rooted directory, how do you create one, and when should you use
one.
ie... "sys$common"
The Answer is :
Concealed rooted logical names are used to mask potential (or actual)
differences in the initial component part(s) of the full file
specification.
An example of a traditional physical file name specification and an
equivilent replacement (where DKA100: is assumed to be the local OpenVMS
system disk) using a logical name follows:
DKA100:[SYS0.SYSCOMMON.SYSEXE]
SYS$SYSDEVICE:[SYS0.SYSCOMMON.SYSEXE]
When correctly created, a concealed rooted logical name is largely
indistinguishable from a physical device name when parsing the file.
Unlike a more common logical name that solely references a physical
device, a concealed rooted logical name also references (and conceals)
a root; a directory path.
An example of a concealed rooted logical name follows:
DKA100:[SYS0.SYSCOMMON.SYSEXE]
SYS$COMMON:[SYSEXE]
In this case, SYS$COMMON translates to DKA100:[SYS0.SYSCOMMON.] -- the
trailing period in the directory is significant, and required syntax.
The SYS$COMMON logical name and any other concealed rooted logical name
is marked with /TRANSLATION=CONCEALED. The default logical name
translation scheme will translate back to the physical device name,
to the limit of translations (LNM$C_MAXDEPTH), or until it encounters
a logical name marked as concealed.
As an example of when a concealed rooted logical name might be of
interest, consider a case where multiple group of OpenVMS users have
a scratch disk and wish to share scratch space, but they do not wish
to see the scratch space of other groups. Through the use of a
concealed rooted logical name, each of these groups can all share
the same physical disk, but each will have the appeance of an entirely
private disk:
$ DEFINE/TRANSLATION=CONCEALED/TABLE=LNM$SYSTEM/EXECUTIVE -
USER_SCRATCH DISK$SCRATCH:[GROUP_1.]
As another example, consider a site that is upgrading from the usual
herd of smaller disk drives into a fewer larger disks or into a RAID
configuration. Various applications were designed so that they each
assumed the use of a whole disk, or potentially various versions of
the application are now in use in parallel, and problems with the
directory names used are encountered -- top-level directory names are
no longer unique, and the applications are now colliding. With the
use of concealed rooted logical names, each application can be moved to
a subdirectory lower down in the directory structure, but the application
can still operate unaware; as if it were still working with top-level
directories.
There are various other uses.
|