16.11.6 Operating archivable multi-chunk tables (when the multi-node function is used)
This section explains how to operate archivable multi-chunk tables when using the multi-node function.
If you are using the multi-node function, you need to perform operation specific to the multi-node function for the following operations relating to archivable multi-chunk tables:
-
Setting archive directories
-
Archiving chunks
-
Unarchiving chunks
Operations other than the preceding operations are the same as when not using the multi-node function. For details, see 11.4 Performing operations on multi-chunk tables.
- Organization of this subsection
(1) Setting archive directories
Make sure that the archive directory can be referenced using the same absolute path from all nodes.
For example, when using the distributed file system NFS, use the following procedure to prepare the archive directory:
- Procedure:
-
-
Prepare the NFS server, and then create an archive directory on the NFS server.
-
Export the archive directory you created in step 1.
Perform this operation on the NFS server.
-
Mount the directory exported in step 2 so that the paths are the same on all nodes.
Perform this operation on all nodes.
-
The following figure shows an example system configuration using NFS:
|
|
![[Figure]](GRAPHICS/F0000933.GIF)
(2) Archiving chunks
If you are using the multi-node function, the only node that can execute the adbarchivechunk command for archiving chunks is the master node. Slave nodes cannot execute the adbarchivechunk command.
For information about the adbarchivechunk command, see adbarchivechunk (Archive Chunk) in the manual HADB Command Reference.
(3) Unarchiving chunks
If you are using the multi-node function, the only node that can execute the adbunarchivechunk command for unarchiving chunks is the master node. Slave nodes cannot execute the adbunarchivechunk command.
For information about the adbunarchivechunk command, see adbunarchivechunk (Unarchive Chunk) in the manual HADB Command Reference.