What Is a Dump Job?

Dump database (Dump database) refers to the process of catastrophic errors (such as corrupted recovery log) through the database administrator DBA to copy the entire database to tape or another disk to save.

Refers to the process by which the DBA copies the entire database to tape or to another disk

These alternate data texts are called backup copies or backup copies.

In the case of catastrophic errors, such as corrupted recovery logs, you can use this command as the first step in dumping, reinitializing, and reloading the server database. You cannot use the DSMSERV RESTORE DB command. No other server activity is allowed during DSMSERV DUMPDB processing.

After DSMSERV DUMPDB processing is complete, perform the following steps:

Issue the DSMSERV LOADFORMAT command to reinitialize the database and restore the log.
Issue the DSMSERV LOADDB command to reload the database.
If necessary, issue the DSMSERV AUDITDB command to locate and correct any database inconsistencies.

note:

A message at the end of the output of the DSMSERV LOADDB command indicates whether the database must be audited. If the server becomes unresponsive when the server is stopped, the DSMSERV AUDITDB command is not required.

Note: The DSMSERV DUMPDB command only supports the use of manual libraries (LIBTYPE = MANUAL in the DEFINE LIBRARY command).

You must have a device configuration file that includes a definition for this device class and any required libraries and drivers. You must also specify the name of the device configuration file by using the DEVCONFIG option in the server options file. If you previously included the DEVCONFIG option in the server options file and then started the server, the device configuration file should be available. If the device profile has been lost or never created, do one of the following:

If the server can be restarted, restart the server and issue the BACKUP DEVCONFIG command.
If the server cannot be restarted, manually create a device configuration file with an editor. For information on how to manually create a device profile, see the Administrator's Guide.

Before you dump the database, consider the following:

Before you dump the database to sequential access media, estimate how much media you need. If the server cannot be started, use the size of the existing database volume.
If the server can be started, use these steps to estimate its size:

Issue the QUERY DB command.
From the output of the QUERY DB command, multiply the number of pages used by 4096 bytes to determine the space occupied by the database.
Divide the product of step 2 by the capacity of a single media to estimate the number of media required.
If possible, and if needed, use the LABEL LIBVOL command to label other tape volumes.

When dumping, the use of scratch volumes ensures that space is available to store the database. Regardless of whether a dedicated or scratch volume is used, the label name and sequence of each mounted volume is recorded during the dump. The volumes used by the DSMSERV DUMPDB command are not recorded in the sequential volume history information.
If there are not enough pre-labeled private or scratch volumes to dump the database, you will need to install and start a second server and use it to label the tape volume (using the LABEL LIBVOL command).
The server recovery log is not accessed during the dump process. Therefore, database entries that were not written to the database at the time of the dump are not recorded. During recovery after a catastrophic failure, the latest database updates may not be recovered. ^[2]

note:

When the database is mounted or restored, the server-to-server communication verification token is changed. The authentication token is a property of the database and it is not stored in the database itself. Part of the token is the date and time the database was installed. For those servers that have been defined for server-to-server communication, you must issue the UPDATE SERVER command with FORCESYNC = YES.

Dump database parameters

DEVclass
Specifies the device class name associated with the volume to which the database dump will be written.
Scratch
Specifies whether scratch volumes are available for dumping the database. The default is YES.
Yes
You can use scratch volumes. If a series of volumes is included on the VOLUMENAMES parameter, scratch volumes are used only if there is insufficient space on the specified volume to dump the database. If the device type associated with the specified device class is FILE, the file name of the temporary volume is generated based on a time stamp.
No
You cannot use scratch volumes. You must include a series of volumes on the VOLUMENAMES parameter to include all database data.
VOLumenames
Specifies the volume to be used for the dump database. This parameter is optional, but must be specified if SCRATCH = NO. If this parameter is not specified and SCRATCH = YES is specified or assumed, a scratch volume is used.
Tivoli Storage Manager does not record the use of volumes by DSMSERV DUMPDB commands in the volume history file. Therefore, you must record the used volume names and specify them in an absolutely identical order in future DSMSERV LOADDB commands.
Possible values are:
volume_name
The name of one or more volumes to be used for the dump database. You can specify multiple volume names, separating each volume name with a comma, without spaces. They will be used in the order in which they are listed.
FILE: file_name
Contains a series of file names for the volumes to be used for dumping the database. In the file, each volume name is entered on a separate line. The volumes are listed in the order they are used.

Dump database example

Task 1

Dump the database to previously labeled tapes named DB0001, DB0002, and DB0003:

Stop the server.
Make sure that the DEVCONFIG option is specified in the server options file. The device configuration file must contain the drive definitions required for device classes, libraries, and dump operations.
Issue the DSMSERV DUMPDB command.

command

dsmserv dumpdb devclass = tapeclassvolumenames = db0001, db0002, db0003

Task 2

Use the tape dump database listed in a file named ADSM.VOLLIST.

command

dsmserv dumpdb devclass = tapeclass volnames = file: adsm.vollist