Stop saposcol only if you want to restart the host or if you want to swap to a different version of saposcol itself.

1. Stop saposcol with the command saposcol -k.

A new saposcol is started which stops the active saposcol . The process that is collecting data should then stop working after one second.

2. The new saposcol connects to the shared memory.
3. From the shared memory, it determines the PID of the saposcol that is collecting data, and the status of the active flag.
4. If the new saposcol finds a valid PID, then a stop flag is set in the shared memory. When the old saposcol finds this flag, it resets the active flag, and deletes the PID from the shared memory. If this does not happen within one second, the new saposcol sends a signal to the background process to stop working.
5. The new saposcol checks this several times against the active flag in the shared memory.

If necessary, the process clears the data in the shared memory itself.

6. Before the old saposcol stops, it writes the data in the shared memory to the file coll.put in the home directory of saposcol .

The file coll.put is imported when the host is restarted, so that the summarized data in the shared memory is available to R/3. For example, if saposcol is stopped at 12:03 and restarted at 14:49, then the data from hours before 12 is still available to the R/3 System. To avoid confusion, invalid data for hours 12 to 14 is not displayed in the recent hours overview in the operating system monitor.

The time for which a shared memory segment exists depends on the operating system. On a UNIX operating system, it is retained until saposcol deletes it. On Windows NT, the shared memory is deleted by the operating system if there is no process connected to it.

Starting SAPOSCOL

Controlling SAPOSCOL from the R/3 System

Displaying Collector Data