XMM-Newton
Science Analysis System: Users' Guide
Next: 2.8 SAS input and output files
Up: 2 Analysis of XMM data: an overview
Previous: 2.6 Running a SAS task from the command line
2.7 General settings affecting all SAS tasks
The behaviour of a task is modified by the following environment variables:
- SAS_PATH is a list of colon-separated directories that contain SAS
packages. It should at minimum contain the toplevel directory where the SAS
is installed.
- SAS_CCF points to the directory containing the Current Calibration
File (CCF) or contains the name of the Calibration Index File. If a value is
provided via the command line option -i (see above) this takes precedence
over any environment specification. This variable is only relevant to tasks
making use of the Calibration Access Layer.
- SAS_CCFFILES has the same meaning as the command line option -f.
If a list of CCF constituents is given both via -f and SAS_CCFFILES the
former takes precedence over the latter. This variable is only relevant to
tasks making use of the Calibration Access Layer.
- SAS_ODF points to a directory containing
an ODF or to a SAS summary file (*SUM.SAS). If a name is
specified via -o and SAS_ODF is defined the former takes precedence over
the latter.
This environment variable is only relevant to tasks making use of the ODF Access Layer.
- SAS_TRACER determines the debug level of the libraries.
- SAS_MEMORY_MODEL determines how the internal memory is used by the data
access (through the Data Access Layer, DAL).
There are basically two options for the memory model: High Memory (high)
and HighLow Memory (highlow or low) model. In the high model
every time a dataset is opened it is kept entirely in memory and all subsequent
operations are performed on the memory-loaded dataset. In the low
model the data is loaded into memory only when
their access is needed (when opening a dataset only their attributes are kept
in memory). While the high model should bring higher performance, its use
implies normally a high memory consumption, which can lead to swapping,
producing a poorer performance. Therefore for machines with less than about
1GB RAM it is recommended
to use the low model.
The default value for SAS_MEMORY_MODEL is high
- SAS_VERBOSITY determines the debug level of the task. The value ranges
from 0 to 10 with increasing verbosity level. The default value is 1, a value of
4 is recommended for getting a very communicative SAS, larger values should be
only used for debugging purposes.
SAS tasks produce error messages at levels called message, warning, error
and fatal.
- At message level, a message is reported to the user and processing
continues.
- At warning level, a warning message is reported to the user and
processing continues.
- At error level, an error message is reported and current operation is
aborted. Control may return to the calling program which can take an
appropriate action.
- At fatal level, a fatal error is reported to the user and all processing
is aborted. A fatal error is generated if the internal state of the program is
disrupted, for example if an invalid value is found in a variable.
Error messages are characterised by a layer and a verbosity level. The layer
indicates the layer in the system where the message is coming from (e.g.
application, user interface, application library or system library). The
verbosity level determines whether or not the message will be reported. The
user can specify the verbosity level. If the verbosity of a message is large
enough compared to the system verbosity level, the message will be reported.
The verbosity level differs for the different layers.
Next: 2.8 SAS input and output files
Up: 2 Analysis of XMM data: an overview
Previous: 2.6 Running a SAS task from the command line
European Space Agency - XMM-Newton Science Operations Centre