Backing up the control data set

You can use EDGHSKP or EDGBKUP described in Maintaining the control data set to back up the control data set and the journal. Choose the appropriate utility based on your needs. You can start the back up of the control data set or journal at any time. But you cannot start backup if backup is already in progress. You need to schedule backing up and clearing of the journal before it exceeds its threshold using the EDGHSKP utility.

You can request that DFSMSrmm backup processing use DFSMSdss COPY services instead of using DUMP services. Use of COPY services enables the system to exploit the fast replication services of the DASD subsystems, such as FlashCopy V2 data set level flashcopy. COPY services enable an almost instant copy of the DFSMSrmm CDS to be created, which is ready to use without the need to first restore a backup copy.

Note: When you backup the CDS, you have the choice to either create a portable backup of the CDS or to create a copy of the CDS that is usable by DFSMSrmm without first being restored.

Use EDGHSKP to back up both the control data set and the journal when the DFSMSrmm subsystem is active by specifying the BACKUP DD statement and the JRNLBKUP DD statement.
Use the EDGHSKP utility to back up the control data set and to clear the journal data set.
Use EDGBKUP to back up the control data set and journal when the DFSMSrmm subsystem is not active.
Use the EDGBKUP utility to back up the control data set and journal when DFSMSrmm is active. EDGBKUP does not clear the journal data set.

Schedule regular backups for both the control data set and journal. Backing up the DFSMSrmm control data set and the journal as the first step in inventory management is a way for you to create a restore point for the control data set. Schedule back ups of the journal when it reaches the threshold that you define. Use the DFSMSrmm parmlib OPTION command JOURNALFULL operand, described in Defining system options: OPTION. Clearing out the journal data set regularly prevents the journal from becoming full, and thus reduces the risk of losing the updates to the control data set. You can automate the process of backing up the control data set and clearing the journal as described in Steps for automating control data set backup and journal clearing.

You should plan to keep multiple backup copies, so that, should a backup fail, you still have a previous, successful backup from which to recover. Use the current journal with the most recent control data set and journal backups to forward recover to the point of failure.

You can request that DFSMSrmm back up the control data set and the journal, using the access method services (AMS) REPRO command, the DFSMSdss COPY command, or DFSMSdss DUMP command. When you use the access method services REPRO command or DFSMSdss without concurrent copy or without fast replication for backup, DFSMSrmm does not allow updates to the control data set during control data set backup. Backup using DFSMSdss DUMP enables output directly to tape. DFSMSrmm resets the journal data set and discards journal records if the back up completes successfully. It is not necessary to use DFSMSdss concurrent copy when you specify BACKUP(DSS).

When DFSMSrmm is performing an intrusive backup, some processing might wait until the backup completes or some command processing can be interrupted. DFSMSrmm fails RMM TSO subcommand requests that update the DFSMSrmm control data set. For example, if you start backup using the AMS REPRO backup method, command processing can be interrupted. DFSMSrmm functions such as the DFSMSrmm TSO SEARCH subcommands, that only read the control data set, continue uninterrupted.

Example: Issue the RMM ADDRACK subcommand to add 2000 new shelf locations.

RMM ADDRACK RK0000 COUNT(2000)

Example: If backup processing starts before the request completes, DFSMSrmm issues these messages.

EDG3212E REQUEST REJECTED - DFSMSrmm BACKUP
          CURRENTLY IN PROGRESS
EDG3017I THE ERROR OCCURRED WHILE ADDING RACK NUMBER RK0700
EDG3018I 700 RACK NUMBER(S) ADDED

Example: When backup completes, issue the ADDRACK subcommand to add the remainder of the shelves.

RMM ADDRACK RK0700 COUNT(1300)

When a tape is used during inventory management, DFSMSrmm updates the control data set. If a job that opens or closes a tape data set tries to update the control data set during AMS REPRO backup processing, DFSMSrmm waits for up to five minutes. If backup is still in progress, DFSMSrmm issues a write-to-operator message EDG4010D that prompts the operator to retry or cancel each job. If the operator specifies retry, DFSMSrmm retries the job five more times at minute intervals before again issuing a write-to-operator if backup processing has not been completed.

When you use the DFSMSdss DUMP command to request a back up, plan to keep backup copies for both the control data set and the journal. For any backup being restored, a journal backup is also required for forward recovery. For the latest control data set backup, both the latest journal backup and the current journal must be used for forward recovery.

To use DFSMSdss concurrent copy, you must authorize the DFSMSrmm batch inventory management backup job to the RACF profile STGADMIN.ADR.DUMP.CNCURRNT. See z/OS DFSMSdfp Storage Administration for information about protecting DFSMSdss keywords with RACF.

When you use DFSMSdss to back up the control data set and you use concurrent copy or fast replication, DFSMSrmm resets the DFSMSrmm journal only if you also back up the journal data set. The journal and the journal backup are required to forward recover to the latest point in time.

When you use DFSMSdss concurrent copy or fast replication for backup, DFSMSrmm continues to process commands and update the control data set. As an alternative to use of concurrent copy capable hardware, you can use snapshot capable hardware. When you request backup using DFSMSdss, DFSMSdss uses CC-compatible snapshot instead of concurrent copy.