ETERNUS SF XL-DATA/MV V13 User's Guide

D.2 The Trace Information

XL-DATA/MV collects the following trace information:

• Transfer command trace information
• Daemon or service trace information
• Trace information when accessing the data transfer volume

The trace information is stored to one of the trace files on the following directory.

Solaris Server	/var/opt/FJSVmftf/data
HP-UX Server	/var/opt/FHPmftf/data
Linux Server	/var/opt/FJSVmftf/data
IA Server	The installation directory$data (default = C:$Program Files$Datamover$data) (Note: "$" means a backslash.)

The trace information is stored in the two trace files (apilog1, apilog2) using the following cyclic operations:

1. Stores to the first file.

2. When the size of the first file exceeds the following values, stores to the second file.
       UNIX Server:  6MB
       IA Server  : 60MB
   

3. When the size of the second file exceeds the size as mentioned above, overwrites the first file to store.
   These operations are repeated.

The transfer command and daemon/service both use the same output file in a local system.
The size of one trace file is 6MB for a UNIX server and 60MB for an IA server.
The amount of trance information that can be stored in a single trace file is equivalent to approximately 10GB of data transfer information.
Since the two trace files are operated cyclically, a total of approximately 20GB of data transfer trace information can be stored.

---

Free Space for Storage Directory

The directory for storing the trace files must have the following capacity or more:
    UNIX Server:  12MB
    IA Server  : 120MB
If there is insufficient free space, an insufficient area error message is output. Also, all trace information cannot be stored.
This storage directory is used also when extracting the Error-log output file. Refer to "The Error-log Information" for details on the Error-log information.

---

Collecting the Trace Information

The trace information can be put on each system. Collect the trace information on each system to determine the cause of an error.

Collecting the Trace Information on the Client System

1. Use the transfer command for trace ended with "deb" as shown below insted of the original transfer command.
   If operating the command with a shell script, replace the command name in the shell script.
   

   mftfsend -f /export/home/a/data1 ...	==>	mftfsenddeb -f /export/home/a/data1 ...
   mftfrecv -t /export/home/a/data2 ...	==>	mftfrecvdeb -t /export/home/a/data2 ...

   The options are all the same.

2. Executing the transfer command for trace creates a trace file on the client system.

3. When the data transfer process terminates with the problem, the trace information has been stored.
   If operating with the shell script, restore the shell script modified in step 1.

4. Send the trace information to a Fujitsu system engineer (SE).

Collecting the Trace Information on the Server System

The server system means the daemon on a UNIX server and the service on an IA server.

[Collecting the Trace Information of the Daemon]

1. Stop the normal daemon (mftfd) running for the target data transfer volume.

2. Start the trace daemon (mftfddeb), which output the trace information, for the target data transfer volume.
   

   mftfd -start /dev/rdsk/c2t0d0s2

   ==>

   mftfddeb -start /dev/rdsk/c2t0d0s2

   The options are the same as the normal daemon.

3. As requesting data transfer on the client system, the trace information is collected on the server system.

4. Stop the daemon, when the data transfer process terminates with the problem,

5. Send the trace information to a Fujitsu system engineer (SE).

[Collecting the Trace Information of the Service]

1. Stop the XL-DATA/MV service.

2. Specify the "-D device name" in the startup mode file (startup.ini file).
   Specify the name of the device targeted for the trace.

OPTION= -D device name

Example: OPTION= -m -D PHYSICALDRIVE1

3. Start the XL-DATA/MV service.
   

   *	The trace information only for the device specified by -D in step 2 could be collected.

4. As requesting data transfer on the client system, the trace information is collected on the server system.

5. Stop the service, when the data transfer process terminates with the problem,

6. Delete the "-D device name" specified in step 2.

7. Send the trace information to a Fujitsu system engineer (SE).