The maximum file size supported by the system depends on the OS. Therefore, you must set the system to support large files by using one of the methods described below:

• Set the supported maximum file size to a value that is greater than the maximum file size used by the Datareplicator processes.
• Do not set a maximum value (set it to be unlimited).

For details about the OS commands required for changing the maximum value and how to use them, see the OS and shell documentation.

Note that in AIX, the default file size is 1 GB. To check the maximum file size supported by each OS, use the limit or ulimit command. For details, see the applicable OS and shell documentation.

Note:

In the AIX edition, the maximum file size used by a process that is started from inetd is the same as for the administrator privilege (superuser). Because the extraction master process is started from inetd, use the procedure described in the table below to change the maximum value for the administrator privilege (superuser). For details about the communication settings, see 2.5.1 Setting up a source Datareplicator's communications environment.

Step	Setting	Command
1	Add the service name and port number that are used for communication between extraction master processes and extraction node processes.	# vi# /etc/services
2	Set the communication environment for inetd.	# vi# /etc/inetd.conf
3	Set the maximum file size for the administrator privilege (superuser).	# vi# /etc/security/limits
4	Terminate inetd.	# stopsrc -s inetd
5	Start inetd.	# startsrc -s inetd

#

Any text editor can be used.

To store a queue file with a size of 2 GB or greater in a Datareplicator file system area, set 2 GB or a greater value in the hdsfmkfs -n command. Note that the maximum permissible size depends on the OS; check the following table for the applicable value.

Table 6-23 Maximum size of Datareplicator file system area when large files are used

Datareplicator type	Condition		Maximum size of Datareplicator file system area (MB)
HP-UX edition	Large files not used	Regular file	2,047
		Character special file
	Large files used	Regular file	131,071
		Character special file
Solaris edition	Large files not used	Regular file	2,047
		Character special file
	Large files used	Regular file	1,048,575
		Character special file
AIX edition#	Large files not used	Regular file	2,047
		Character special file
	Large files used	Regular file (JFS)	65,411
		Regular file (JFS2)	1,048,575
		Character special file
Linux edition	Large files not used	Regular file	2,047
		Character special file
	Large files used	Regular file	1,048,575
		Character special file

#

When you use large files, the permitted maximum size of a Datareplicator file system area is different between JFS and JFS2. Therefore, use the smit command to specify the necessary settings so that large files can be used in the target file system.

To handle large files using link names, you can specify only one nesting link. If more than one link is nested, as shown in the figure below, the target file cannot be created as a large file.

In this example, use Linkname2, not Linkname1.