Tracing the watch daemon

You can start traces for only the watch daemon (dsmwatchd).

Procedure

This procedure relates to a static trace setting. 
The setting is applied to all subsequent daemon starts.

Stop the HSM daemons (but not the watch daemon) with the following command: /usr/bin/dsmmigfs stop.
Add trace options to the dsm.opt client options file as described in Enabling a Space Management client trace.
Stop and restart the watch daemon.
- On Linux® systems, stop and restart the watch daemon by using the following commands:
```
systemctl stop HSM  
systemctl start HSM
```
- On AIX® systems, stop the watch daemon with the following command:
```
kill -15 dsmwatchd_pid
```
  The watch daemon is restarted by the next system init process.
Remove the trace options from the dsm.opt file.
If you do not remove the trace options, other daemons are traced when you restart other daemons.
Start all HSM daemons other than the watch daemon by using the following command: /usr/bin/dsmmigfs start
Because the dsm.opt file contains no tracing information, the other daemons are not traced.

You can also execute a procedure for a dynamic trace setting. 
The setting is applied to the currently running daemon process.

Use the command dsmmigfs control to enable dynamic tracing. See dsmmigfs control for details.

The space management client captures data from the watch daemon.

You can stop tracing the watch daemon in two steps: