Analyzing the Server Log |
Overview of the SLTOOL2 Sample Program |
SLTOOL2 reads the SAS data file produced by SLTOOL1 and creates a group of SAS data sets. Usually, SLTOOL2 is executed during the same SAS session as the SLTOOL1 program.
The libref SLOGDATA is associated with a SAS library by the LIBNAME statement in SLTOOL0.SAS.
The data sets created by SLTOOL2 are stored in the SAS library and are associated with the libref SLOGDATA. It is most efficient to run SLTOOL0.SAS one time to create the data files in the library SLOGDATA, and then run multiple analysis programs that access the library SLOGDATA. With large server logs, creating the data sets in the library SLOGDATA can take quite a long time and should be done only one time for each server log.
SLTOOL2 creates data sets with names that include either INFO or SUM. INFO files contain observations that record specific SAS/SHARE activities, such as each time a server is started and stopped. SUM files present the total number of times a specific activity occurred, such as the total number of connections made to a server.
The following sections describe each data set that is created by SLTOOL2.
SLOGDATA.SERVINFO |
The SERVINFO data set records the server name and the times at which it was started and stopped. You can use this information to write a descriptive header on a report that relates to that server.
SLOGDATA.CONNINFO |
The CONNINFO data set contains one observation for each time a user connects to the server and one observation for each time a user disconnects from the server.
From this data set, you can obtain a list that shows who connected to a server, how long each user remained connected, or how many times each user connected to the server. You can also chart the simultaneous number of connections to a server over a period of time, which shows peaks and valleys in the number of users who access data through a specific server.
SLOGINFO.CONNSUM |
The CONNSUM data set contains only one observation with one variable that stores the total number of connections to this server.
SLOGDATA.TASKINFO |
The TASKINFO data set contains one observation for each creation of a mirror resource environment and one observation for each termination of a mirror resource environment.
The name of a resource environment in this data set corresponds to the name of a SAS procedure or a window that is used to access data through the server. From this data set, you can obtain a list of those SAS procedures and windows and the length of time each procedure or each window remained active.
SLOGDATA.LIBINFO |
The LIBINFO data set contains one observation for each time a user accesses a SAS library and one observation for each time a user releases a SAS library.
From this file, you can determine how many times each library was accessed through the server and the length of time that each library was accessed. You should use the physical name for the library because each library can be referred to by different librefs at various times. To obtain a list of the libraries accessed through a server, use SLOGDATA.PHYSINFO.
SLOGDATA.PHYSINFO |
The PHYSINFO data set contains a list of the physical names that correspond to the libraries that were accessed through the server.
SLOGDATA.ENGSUM1 |
The ENGSUM1 data set contains a list of the engines that were used to access SAS libraries through the server.
SLOGDATA.MEMINFO |
The MEMINFO data set contains an observation for each time a SAS library member is opened, reopened, closed, renamed, repaired, or deleted.
From this data set, which usually is very large, you can derive a list of members for each library that has been accessed through the server; the length of time and how many times each member was accessed; whether each member was created, read, or updated; and the number of simultaneous users of each member over a period of time.
You should use the physical name for the library because each library can be referred to by different librefs at various times.
SLOGDATA.OBJINFO |
The OBJINFO data set contains an observation for each time a SAS catalog entry is opened, closed, renamed, deleted, aliased, or has its directory information or options changed.
This data set is similar to SLOGDATA.MEMINFO, but it contains information for catalog entries instead of members of SAS libraries.
SLOGDATA.IDXINFO |
The IDXINFO data set creates an observation for each time a user creates or deletes an index through the server.
Because creating an index tends to be expensive, this data set is probably most useful as a warning signal. Also, because indexes can be very helpful for SAS data sets that are accessed concurrently, having a list of indexes that were deleted during a server's session can also be a warning signal.
SLOGDATA.DIRINFO |
The DIRINFO data set contains an observation for each time the directory of a SAS library or the directory of a SAS catalog is opened or closed through the server.
SLOGDATA.IDXSUM |
The IDXSUM data set contains only one observation with two variables. One variable counts how many indexes were created through the server; the other variable counts how many indexes were deleted through the server.
SLOGDATA.ACCTINFO |
The ACCTINFO data set contains one observation for each accounting message that is written to the server's log after a user disconnects. You must specify the LOG= option in the PROC SERVER statement to collect this data.
Copyright © 2007 by SAS Institute Inc., Cary, NC, USA. All rights reserved.