You start the saposcol program if you want to begin collecting data on your operating system resources.

Start saposcol no later than when you start the first R/3 instance on a host. A new saposcol will not start if another saposcol is already running.

1. To start saposcol , use either the Operating System Monitor OS Collector option or, at the command line, use the command saposcol.
   
   Normally, saposcol is called without any additional parameters or profiles. This call is equivalent to saposcol -l (where l stands for Launch). For more information, see Controlling SAPOSCOL from the Operating System.
2. If this call finds a saposcol shared memory segment, it takes on the PID of a saposcol , which may already be running, from data stored in shared memory.
3. If no other saposcol is collecting data, this PID equals 0 and saposcol will start, using the existing shared memory segment. The saposcol program will also start if there is no shared memory available.
4. However, if the newly called saposcol recognizes that another saposcol is already collecting data, it will not start.
5. Once saposcol has started, it performs an initialization, setting aside the required amount of shared memory based on such things as the number of available disks, CPUs, and file systems.
6. The program then executes a fork, after which the child process runs in the background and the original program ends.

Error and status messages can be found in the dev_coll file in the directory /usr/sap/tmp . When it starts, saposcol reads the data in the file coll.put . You can change the default path by changing a profile parameter in sapdefault.pfl .

Stopping SAPOSCOL

Operating System Monitor