-
New Change File While Active
Function. The new Change PF While Active
(CHGPFRSF) command lets you change the layout of a
physical file with minimal disruption. The file can be
in use during the change, except for a brief time at the
end when the new format is installed.
For large files, using the system CHGPF or the ALTER
TABLE commands to change a file layout is not always
practical because these operations run for a long time
during which the file is unavailable.
The new CHGPFRSF command makes changing the layout of large
files practical. The file format is changed in the
background and all record and member changes made
during the update process are preserved.
This new feature is available in both the base and HA
versions of RSF.
-
Faster Role Swaps With Standby
Journals. Having a standby journal in place can speed up
a role swap. New options let you tell RSF whether
to create and manage a standby journal for each library or IFS
directory being replicated.
Here's how it all fits together:
-
The local journal lives on the
production machine where it constantly captures the
changes to be replicated.
-
The remote journal lives on the
backup (target) machine. Any changes that show up in
the local journal are automatically mirrored to the
remote journal. RSF uses the information in the
remote journal to keep the target machine objects in sync.
-
The standby journal also lives on
the target machine. Most of the time, the standby
journal is not used and the the journal changes it
collects are thrown away. However, if a role swap
results in the target machine moving into
production, the standby journal is ready to assume
the role of the local journal.
-
Because the standby journal's main
job is to take over as the local journal in the
event of a role swap, the name and library of the
standby journal on the target machine will match the
name and library of the local journal on the production
machine.
-
What saves time during a role swap
is not just the existence of the standby journal,
but the fact that journaling has already been
started for all of the objects that will use the
journal for replication.
-
New Delete Orphan Objects
Tool. The new Delete Orphan Objects
(DLTOBJRSF) command makes it easy to delete objects in a
library on the backup machine for which there is no
corresponding object on the production machine.
When replicating non-journaled objects, deleting an
object from the production machine does not always
result in the automatic deletion of the corresponding
object on the backup machine. The new DLTOBJRSF
command provides a convenient way to clean up those
orphan objects.
-
New Default Job Name. When setting
the sync attributes for a library, you can now specify
the default job name to use when replicating the library
in batch. Previously, the job name was fixed. Replication jobs for
different libraries with similar names would end up with the same job
name in batch. Now, you can specify the exact job name to
use when replicating any library.
-
New Replication End Option.
The End Synchronization Job (ENDSYNCRSF) command has
been enhanced, allowing you to choose whether to wait
until the replication jobs have ended. Previously,
the ENDSYNCRSF command always returned control
immediately while replication jobs completed in the
background. Now, you can request to wait until all
of the specified replication jobs end. This makes
it easier to schedule another task which depends on all
replication jobs having ended.
-
Flexible Include Options.
When setting library synchronization attributes, you can
specify objects to be included or omitted from
replication. Previously, there was no way to
include all objects of a particular type.
This is now supported. Specify INCLOBJ(*ALL
*DTAARA) to include all data areas, INCLOBJ(*ALL
*USRSPC) to include all user spaces, and so on.
The default, as before, is to include all objects of all
types. This is now represented as INCLOBJ(*ALL
*ALL).
-
Automatic Handling of
Cross-Dependent Logicals. When starting
replication for a library, the entire library must be
refreshed to ensure that both the source and target
versions are identical. To ensure that the library
can restore properly onto the target, any existing
library of the same name must first be cleared. In
the past, problems would arise when logical files in a
different library were based on physical files
in the library being refreshed, preventing the deletion
of some physical files in the target library. Now,
RSF handles this situation automatically. Any
cross-dependent logicals on the target machine are
deleted first and then restored to their original
location after the library being refreshed has been
restored.
-
Automatic Target Release.
When sending objects from one machine to another, the
objects may need to be saved from the source machine and
restored onto the target machine. If the machines
are at different OS release levels, it's important to
specify the correct target release when the objects are
saved.
A new value of *AUTO is now the default for the target
release (TGTRLS) parameter on all RSF SNDxxx commands.
*AUTO tells RSF to poll the target machine and determine
the appropriate target release automatically. For
maximum efficiency, the polling is done once a day and the
value determined is stored for future use.
With this enhancement, RSF send functions can be used
and coded in programs without having to change TGTRLS
values as systems and OS release levels change.
RSF automatically detects the relative OS release levels
and uses the appropriate TGTRLS value.
-
New Options For Non-Journaled
Objects. When replicating a library, you
can now specify whether to look for changes to
non-journaled objects with the same frequency that
journaling progress is checked.
Only database files, data areas and data queues can be
replicated with journaling. All other types of objects
must be replicated using save/restore operations.
For large libraries, where nearly all of the objects are
journaled, looking for changes to non-journaled objects
less frequently can improve performance.
-
Enhanced Spooled File
Commands. The Find Spooled File
(FNDSPLFRSF) command lets you find one or more spooled
files that contain a particular character string.
Now, you can specify the number of matching spooled
files to find before presenting the list, allowing
initial results to be displayed more quickly.
On subsequent searches, you can indicate whether to
start over or continue from where the last search left
off.
The Set Spooled File User Data (CHGSPLFRSF) command now
runs much faster. This command is used to set the
user data for certain RSF job logs to a
value that reflects the purpose of the job that created
the log.
-
New Command Checks IFS
Include/Omit Specs. Creating just the
right generic specification to include or omit a subset
of IFS objects can be tricky. The
new Is IFS Object Omitted? (CHKIOMTRSF) command lets you
test your specifications with specific examples to be
sure the results are what you intended.
-
Easily Change All Sync
Attributes Entries. A new option on the
Set Sync Attributes For a List (SETRSFSA) command makes
it easy to change one or more attributes for all items
in your sync attributes list. Just specify the new
special value of *ATTR for the LIST parameter.
-
New Delete Journal Command.
Now you can easily delete a journal when it's no
longer needed with the new Delete Journal (DLTJRNRSF)
command. Journaling is ended for all objects
using the journal; any remote journals associated with
the journal are detached; then, the journal and all
associated receivers are deleted.
-
Replicate SNADS Configuration.
When replicating system information, you can now include
the SNADS data associated with the Work with Directory
Entries (WRKDIRE) display.
-
Faster Handling of Triggers
and Constraints. In a replication
environment, trigger and constraint information is
extracted periodically and saved in order to be ready
for a role swap. New options on the Extract
Configuration Info (EXTCFGRSF) command make extracting
and restoring trigger and constraint information much
faster. Previously, the information was extracted for all user libraries. Now, you can
specify that the information should only be extracted
for libraries being replicated with journaling.