SDA 3.5 Documentation for SDALOG


sdalog - Generate a report of SDA usage


sdalog -g filename [options]


SDALOG reads the SDA logfile (defined in a HARC file) and generates a report on SDA usage. Note that the logfile read by SDALOG is the special file written by SDA -- not the access log maintained by the Web server software. See the [GENERAL] section in the HARC file document for details on specifying the name of the logfile.

The default output format reports how many times each SDA program or procedure was executed, how many times each dataset (using the name specified in the HARC file) was accessed, and how many of the reported procedures were run in each of the months.

An optional format (used if a ’-c’ option was specified) reports the client IP addresses and how many times each client executed an SDA procedure. This style of report can include the full IP addresses or only the last one, two, or three final segments of the IP addresses. For example, the last segment of the IP address is the major Net domain -- like ’COM’ or ’EDU’. If the ’-c 1’ option is specified, a summary of usage by those major Net domains will be generated.


The following command-line options are recognized. The specification of the name of the log file is required. The other specifications are optional.

-g filename
The specified filename is the pathname of the log file maintained by SDA, if such a file has been specified in a HARC file. (REQUIRED)

-o filename
Output from SDALOG will be written to this file. If this option is not specified, output will be routed to the user’s screen (standard output).

-e filename
The specified filename contains, one per line, a list of client IP addresses (or portions of addresses) to exclude from the report. For example, the IP address used by the local system programmer might be excluded from the usual reports. Note that the IP addresses given in this file are treated as case- insensitive. If ’.Com’ is specified, for example, all log entries containing ’.COM’, or ’.com’ in the IP address field of the log file will be excluded from the report.

-f filter_string
Each line of the logfile is scanned for the presence of the specified string of characters. If the string is not in a line, that line is excluded from the report.
This option can be repeated.

-F filter_string
The specified string of characters is used as a selection filter (as with the ’-f’ option). However, the filter string is case-insensitive. In other words, if the specified filter string is ’gss’, lines containing ’gss’, ’GSS’, or ’Gss’ will all be included in the report.
This option can be repeated.

-c all
The report will list the IP addresses of the client computers used by the SDA users. The number of procedures executed by each client will also be reported.

-c 1
Only the last segment of IP addresses (major Net domains like EDU, COM, NET, etc.) will be reported (instead of the full IP address of each user).

-c 2
The last 2 segments of IP addresses will be reported.

-c 3
The last 3 segments of IP addresses will be reported.

-d number
The specified number is the maximum length, in characters, of a dataset name to process. If the log file is occasionally written incorrectly by the system because of heavy Web usage, some of the fields can be run together and produce what appear to be long dataset names. Setting this option to ’10’, for example, will cause the SDALOG program to skip any log entries with a dataset name longer than 10 characters. This option would only be used if the list of dataset names generated by the program contained run-together names.

-x filename
Write lines with badly formed log entries (if any) into this file. This option is for diagnostic purposes.

Print out a list of options (but do not execute the program)


Basic example
sdalog -g SDAlog -o logreport

Filter for a specific dataset (GSS)
sdalog -g SDAlog -f GSS -o logreport

Get major categories of users
sdalog -g SDAlog -c 1 -o logreport

CSM, UC Berkeley
April 12, 2011