This file contains late-breaking notices regarding:
1.
The
Model Server of
SAS Cost and Profitability Management 8.4 is written in C++. It uses SAS/ACCESS
to Oracle, SAS/ACCESS to PostgreSQL, or SAS/ACCESS to ODBC to make connections
to external databases. This change requires you to specify connection
information again on data import/export wizards.
2.
SAS
Cost and Profitability Management 8.4 supports importing and exporting model
data from Microsoft Office applications. However, users should be aware that the
Office applications have size restrictions that may limit their usefulness for
managing large amounts of data. The points below list the limitations of which
SAS is aware, although others may exist:
1) Excel 97-2003 Workbook
(*.XLS) does not allow more than 65536 rows or more than 256 columns in a table
or worksheet.
2) Excel 2007 Workbook (*.XLSX) does not allow more than
1048576 rows or more than 256 columns in a table or worksheet. The Microsoft
support web site mentions that the column limit for *.XLSX is 16384, but SAS has
only created up to 256 columns by default.
3) Access supports a virtually
unlimited number of rows but only allows 255 columns in an Access database
table. In addition, an Access table is limited to 1 GB of storage, and an Access
database (*.MDB) is limited to 2 GB of storage for all of its tables.
3.
The
default configuration of Microsoft SQL Server 2008 and 2012 allows SQL Server to
consume all available RAM on the local computer. This behavior can leave other
processes "starved" for memory if other components are also installed on the SQL
Server machine. You might want to change this setting so that SQL Server does
not consume more than 50% of the available RAM on a shared computer, and no more
than 80% of the available RAM on a dedicated database computer. This setting can
be changed in SQL Server Management Studio on the Memory configuration page in
the Properties dialog box for your SQL Server instance.
4.
If
you run the SAS Deployment Wizard again to reconfigure an existing SAS
deployment or to install additional SAS software, you must run the SAS
Deployment Wizard with the "Custom" prompting level to ensure that it recognizes
and reconfigures all of the web application servers that are installed on that
computer. In addition, you must restart the Object Spawner on SAS Cost and
Profitability Management Server after reconfiguration so that it can select new
values.
1.
If
you are planning to perform an in-place database migration from a previous
version of the SAS Activity-Based Management solution, it is important to back
up your existing database.
2.
With SAS Cost and Profitability Management 8.4, the mid-tier server no longer
launches long-running operations in a separate process. Except for cube
generation, the execution of all long-running operations is now performed on the
Model Server. However, the mid-tier server is still responsible for initiating
and managing the operations on the Model Server.
3.
If
you are installing SAS Cost and Profitability Management 8.4 and any of the
language packs, you must make sure that the locale of your language pack, your
operating system, and your Adobe Acrobat Reader are compatible with your Base
SAS language configuration (if applicable). A locale incompatibility might cause
the SAS Cost and Profitability Management 8.4 import, export, and reporting
mechanisms to incorrectly handle locale-specific characters or to fail to
operate entirely.
4.
When
using SAS Cost and Profitability Management with SAS Metadata Server on
non-English systems, the SASV9.CFG file, which is typically located in
"C:\Program Files\SASHome\SASFoundation\9.4\nls\langcode
"
(where langcode is the language code for your installation) must have a
line added of the form "validvarname=any". This addition is necessary to ensure
that the column names in registered tables appear with correctly localized text.
If this option is not added to the .CFG file, column names may appear with
default names such as _COL1_, _COL2_, etc., which causes subsequent operations,
such as cube generation, to fail.
1.
Oracle
11g has a new feature that enables data recovery after queries. This is
accomplished by creating "dummy" tables in your Oracle schema that hold the
necessary information to roll back changes. These tables have names of the
following form:
"BIN$+HQXPwmySkCtpdSzeQKKgA==$0".
They are neither used by
nor created by SAS Cost and Profitability Management. However, they may become
very large, as they contain enough data to roll back large database updates that
were made by made by SAS Cost and Profitability Management. To prevent this
unnecessary usage of disk space on your database server, you can disable this
Oracle data recovery feature for an entire Oracle instance. Disabling the
feature will not impact SAS Cost and Profitability Management in any way.
2.
If
you are using the Oracle 12c client on the Model Server to make a connection to
the Oracle server, you may encounter the following error at the time of cube
generation:
The cube cannot be made because the input data set does
not contain any logical observations
There is a known bug in Oracle
12c client software installation, which is missing ORANFSODM12.DLL or ORAODM12.dll. These missing DLLs
cause Oracle bulk loader to fail (Doc ID 1593348.1, Doc ID 1936068.1). You can copy these DLLs from
the server installation to client or obtain them from Oracle Support.
It is also observed that some installations may have only ORAODM12.dll. If you rename this file as ORANFSODM12.DLL then this issue is resolved, as described here: https://community.oracle.com/thread/3652577
3.
A
known issue with Oracle’s JDBC driver version 12.0.1.2 is described here: https://community.oracle.com/thread/3682300
You
might encounter a “maximum open cursors exceeded” error with that driver. The
workaround is to use version 12.0.1.1 or 11.1.0.7.0 or another version that does
not have this issue.
1.
A
PostgreSQL database will perform more slowly on large models than either
Microsoft SQL Server or Oracle databases. This happens because PostgreSQL does
not have an external "bulk load" interface that is optimized for loading large
amounts of data.
1.
When
you are importing data from a Microsoft Access database or a Microsoft Excel
spreadsheet, the file must be visible and accessible from the client. Previous
versions of SAS Cost and Profitability Management required that the file be
visible from the server. However, SAS has changed the way this import process
operates to accommodate changes in Microsoft's support strategy for the Office
Suite.
2.
Accounts
with dimension intersections No <Dimension1> x No <Dimension2> x ...
No <DimensionN> can be exported but cannot be imported.
3.
Deleting
period and scenario data might take a long time when a large amount of data must
be deleted.
4.
Do
not set the monitor display font size to anything larger than 96 DPI.
5.
Users
of the Accessibility capabilities of SAS Cost and Profitability Management often
use the JAWS screen-reading program. There is a known bug in JAWS causing it to
occasionally lock input fields of HTML pages, preventing users from entering
data. If this happens while you are using JAWS, press Alt+N to re-enable
screen input.
6. If an end user who is using the IWA login launches a
long-running operation, SAS recommends that the user not close the
client while the long-running operation is in progress. The operation might fail if the client is closed.
The
following steps are required in order to push the data successfully to SAS LASR
Analytic Server from SAS Cost and Profitability Management
8.4.
1.
SAS
Cost and Profitability Management Cube creator and Modeler groups should be
members of the Visual Analytics Data Administrators group.
2.
Add
SAS Cost and Profitability Management Cube creator and Modeler groups to the
folder /Products/SAS Visual Analytics Administrator
with the
following permissions:
1.
ReadMetaData
2.
WriteMemberMetaData
3.
Read
4.
Write
5.
Delete
6.
Create
If
any of the above permissions are not granted successfully, the Regenerate Fact
and Cube operation fails with the following message: "ERROR: Metadata Server
denied access to operation."