Performing a complete save using the GO SAVE checklist

Use this checklist to perform a complete save operation.

Use the following checklist for menu options 21, 22, and 23 of the GO SAVE command. When appropriate, select the option that you require. If you choose to, you can print system information during the procedure. The Printing system information topic contains detailed instructions on how to print system information if you do not want the Save menu option command to print your system information automatically.

Some of the steps in this checklist might not apply to your system configuration. See Identify optional features that affect your backup for help to determine whether you use optional features in your environment. If you are still unsure how your system is configured, contact your system administrator.

As an alternative to this checklist, use Customizing your GO SAVE backup to produce a set of instructions that is tailored to your save environment.

Attention: If you are using the Hardware Management Console (HMC), you must back up the HMC in addition to using the GO SAVE: Option 21 to obtain a complete save of your system. See Backing up and restoring the HMC.

If you need to restore the entire POWER® blade server, you must back up the Virtual I/O Server, which also contains Integrated Virtualization Manager, in addition to using the GO SAVE: Option 21 command. Virtual I/O Server must be restored on the blade server before you can restore any i5/OS client logical partitions. For saving and restoring Virtual I/O Server, see the PowerVM® Editions Operations Guide.

  1. If you are using software encryption for backup tapes and saving system data (GO SAVE option 21 or 22), set the save/restore master key before you perform the save operation. The save/restore master key is a special purpose master key used to encrypt all the other master keys when saving them in a SAVSYS or GO SAVE operation. For the instructions, see Loading and setting the save/restore master key.
  2. Sign on with a user profile that has *SAVSYS and *JOBCTL special authorities, and also has sufficient authority to list different types of system resources. (The QSECOFR user profile contains all of these authorities.) This ensures that you have the authority that you need to place the system in the necessary state and to save everything.
  3. If you want to exclude virtual images from a full-system save, use one of the following strategies. Virtual images can significantly increase the time it takes to complete an Option 21 save operation, even if the image catalog entries do not contain data:
    • Use the Change Attribute (CHGATR) command to mark the image catalog directory as nonsavable. For example:
      CHGATR OBJ('/MYINFO') ATR(*ALWSAV) VALUE(*NO) 
    • Use the Load Image Catalog (LODIMGCLG) command to make the image catalog ready. Image catalogs with a ready status are omitted from the save operation.
    • In an attended save, you can specify to omit the image catalog directories on the Save Object (SAV) command.
  4. If you have independent auxiliary storage pools (ASPs), make them available if you want them to be included in an Option 21 or 23 save operation.
    Note: If your system includes independent ASPs that are geographically mirrored, it is recommended that you eliminate them from this GO SAVE option by making them unavailable. You must save independent ASPs that are geographically mirrored separate from this GO SAVE operation.

    If the geographically mirrored ASPs remain available during the GO SAVE operation, geographic mirroring is suspended when the system becomes restricted. When you resume mirroring after the save, a complete synchronization is required. Synchronization can be a lengthy process.

  5. If you are operating in a clustered environment and want to save independent ASPs without causing a failover, or you want to save the cluster environment for a node, you must end the device cluster resource group and end clustering before you end subsystems.

    Use the End Cluster Resource Group (ENDCRG) command and the End Cluster Node (ENDCLUNOD) command.

  6. If you have OptiConnect controllers, vary them off before the save operation. You must vary off OptiConnect controllers before ending subsystems and performing a save of the entire system, or before any save that ends the QSOC subsystem. If you do not vary off OptiConnect controllers before ending subsystems, they go into a failed status, the system marks them as damaged, and the system does not save them. For more information, see Networking for logical partitions.
  7. If you have IBM® WebSphere® MQ for Multiplatforms, V6 (5724-H72), you need to quiesce WebSphere MQ, V6 before you save the system. For information about WebSphere MQ documentation, see WebSphere MQLink outside information center (www.ibm.com/software/integration/wmq/library/library53.html) .
  8. If you plan to run the save procedure immediately, make sure that no jobs are running on the system, type the Work with Active Jobs (WRKACTJOB) command.

    If you plan to schedule the save procedure to run later, send a message to all users informing them when the system will be unavailable.

  9. To perform an attended save of your system, go to step 11.
  10. To perform an unattended save operation, continue with the following steps. An unattended save operation prevents your save operation from stopping because of unanswered messages:
    1. Display the reply list sequence numbers to find what numbers are available for use:
      WRKRPYLE
    2. If MSGID(CPA3708) is not already in your reply list, add it. For xxxx, substitute an unused sequence number from 1 through 9999:
      ADDRPYLE SEQNBR(xxxx) +
               MSGID(CPA3708) +
               RPY('G')
    3. If you are using virtual media for your save media, specify automatic load in the reply list, MSGID(OPT149F), to avoid receiving a message that interrupts the unattended save operation. If necessary, virtual media will use the autoload feature to create additional images with the same capacity as the last image you loaded, provided the disk storage is available.
    4. Change your job to use the reply list and to notify you of any break messages that are sent:
      CHGJOB INQMSGRPY(*SYSRPYL) BRKMSG(*NOTIFY) 
    Note: You can also set up a default so that whenever you select menu options 21, 22, or 23, the system will always use the reply list. To set up the default, select menu option 20 from the Save menu. Specify Yes on the Use system reply list option.
  11. Type GO SAVE at a command prompt to display the Save menu.
  12. Select the option (21, 22, or 23) from the Save menu and press the Enter key.

    A prompt display describes the function of the menu option that you selected.

  13. After reading the Specify Command Defaults prompt display, press the Enter key to continue.
                      Specify Command Defaults
     
    Type choices, press Enter.
     
      Devices . . . . . . . . . . .    TAP01          Names
                                       __________
                                       __________
                                       __________
     
     
    Prompt for commands  . . . . .   Y           Y=Yes, N=No
     
    Check for active file . . . . .  Y           Y=Yes, N=No
    
    Message queue delivery . . . .   *BREAK      *BREAK, *NOTIFY
     
    Start time . . . . . . . . . .   *CURRENT    *CURRENT, time
    
    Vary off network servers . . . . *ALL        *NONE, *ALL
    
    Unmount file systems . . . . . . Y           Y=Yes, N=No
    
    
                               Specify Command Defaults
    
    Type choice, press Enter.
    
    Print system information . . . . N           Y=Yes, N=No
    
    Use system reply list  . . . . . N           Y=Yes, N=No
    
    Spooled file data  . . . . . . . *NONE       *NONE, *ALL
  14. Type your choices for the Devices prompt. You can specify as many as four tape media device names. If you specify more than one device, the system automatically switches to the next tape device when the current tape is full. You can select only one DVD-RAM optical media device, tape media library device, or virtual tape device.

    The first device for options 21 and 22 must be your alternate IPL device. If you are creating media to install on another system, the device must be compatible with the alternate IPL device for that system. This ensures that the system can read the SAVSYS media if you need to restore your Licensed Internal Code and the operating system.

  15. Type your choice for the Prompt for commands prompt. Specify N (No) if you want to run an unattended save operation. Specify Y (Yes) if you want to change the defaults on the SAVxx commands.
  16. Type your choice for the Check for active files prompt. Specify Y (Yes) if you want the system to warn you if active files exist on the save media. The warning you receive gives the following choices:
    • Cancel the save operation.
    • Insert new media and try the command again.
    • Initialize the current media and try the command again.
    Note: If you use DVD-RAM optical media for your save, the system sends inquiry messages to the QSYSOPR message queue when it encounters identical active files. The system sends the inquiry message for each identical active file that it finds.
    Specify N (No) if you want the system to write over any active files on the save media without warning you.
  17. Type your choice for the Message queue delivery prompt. Specify *NOTIFY if you want to do an unattended save operation. This prevents communications messages from stopping the save operation. If you specify *NOTIFY, severity 99 messages that are not associated with the save operation are sent to the QSYSOPR message queue without interrupting the save process. For example, messages that request that a new volume be loaded interrupt the save operation because they are associated with the job. You cannot continue until you reply to these messages.

    Specify *BREAK if you want to be interrupted for severity 99 messages that require a reply.

  18. Type your choice for the Start time prompt. You can schedule the start of the save operation up to 24 hours later. For example, assume that the current time is 4:30 p.m. on Friday. If you specify 2:30 for the start time, the save operation begins at 2:30 a.m. on Saturday.
    Note:
    1. The system uses the Delay Job (DLYJOB) command to schedule the save operation. Your workstation will be unavailable from the time you request the menu option until the save operation is completed.
    2. Make sure that your workstation is in a secure location. Your workstation remains signed on, waiting for the job to start. If the system request function is used to cancel the job, your workstation displays the Save menu. The workstation remains signed on with your user profile and your authority.
    3. Make sure that the value for the QINACTITV system value is *NONE. If the value for QINACTITV is other than *NONE, the workstation will vary off in the amount of time specified. If you changed the value to *NONE, write the old value down.
    4. If you specify a delayed start and want your save operation to run unattended, be sure you have done the following:
      • Set up the system reply list.
      • Specified *NONE on QINACTITV system value.
      • Specified *NOTIFY on message queue delivery.
      • Specify *NOTIFY for any break messages.
      • Responded N to the Prompt for commands prompt.
      • Responded N to Check for active files.
  19. Type your choice for the Vary off network servers prompt. If you use integrated servers, you can optionally vary off the network server descriptions before beginning the save procedure. Examples of network servers include running Windows or Linux operating systems using IBM Extended Integrated Server Support for i5/OS, or running Linux or AIX® in a guest partition.

    Select one of the following options to specify which hosted network servers need to be varied off before starting the save operation and varied on after completing the save operation:

    *NONE
    Does not vary off network servers. No data is saved for the network servers because the system is in a restricted state. Saving the individual objects requires an active connection to the system.
    *ALL
    Varies off all network servers before starting the save operation. The save operation takes less time but the network server data is not saved in a format that allows restoration of individual objects. You can only restore all of the data from the network servers.
  20. Type your choice for the Unmount file system prompt. Specify Y (Yes) if you want to allow all dynamically mounted file systems to be unmounted. This step enables you to save user-defined file systems (UDFSs) and their associated objects.
    Note: After the save operation is completed, the system does not attempt to remount the file systems.

    Specify N (No) if you do not want all dynamically mounted file systems to be unmounted. The file system attributes for the UDFS are saved for all UDFSs that are saved when mounted. To rebuild these mounted file UDFSs during a restore, you must specify the RBDMFS(*UDFS) parameter on the RST command.

  21. Type your choice for the Print system information prompt. Specify Y (Yes) if you want to print the system information. The system information might be useful for disaster recovery. The "Printing system information" topic explains how to print your system information manually without using the automatic GO SAVE command menu option function.
  22. Type your choice for the Use system reply list prompt. Specify Y (Yes) if you want to use the system reply list when the system sends an inquiry message.
  23. Type your choice for the Spooled file data prompt. Specify whether this backup should save spooled file data for output queues that are saved. The possible choices are:
    *NONE
    No spooled file data is saved.
    *ALL
    For each output queue that is saved, all available spooled file data on the output queue is saved.
    Note: Saving spooled files might require more save media and will take additional time.
  24. Press the Enter key. If you chose a later start time, your display shows message CPI3716. The message tells when the save operation was requested and when it will start. You cannot use the display until the save operation is completed. The input-inhibited indicator might appear. You have completed the steps for setting up the save operation.

    If you did not choose a later start time, continue with step 25. If the value for QSYSOPR message queue delivery is *BREAK with a severity level of 60 or lower, you must respond to the End Subsystem (ENDSBS) messages. This is true even if you plan to run an unattended save operation specifying a start time of *CURRENT.

  25. Start of changeIf you responded Y to the system prompt, Prompt for commands, the End Subsystem display appears after TCP has ended. Type any changes and press the Enter key. While the system is ending subsystems, you see the following messages. You must respond to them if the QSYSOPR message queue is set to *BREAK with a severity level of 60 or lower. Each message appears at least twice. Press the Enter key to respond to each message.
    1. CPF0994 ENDSBS SBS(*ALL) command being processed
    2. CPF0968 System ended to restricted condition

    If you responded N to the Prompt for commands prompt, skip to step 27.

    End of change
  26. When the system is ready to perform each major step in the save operation, you are shown the prompt display for that step. The time between prompt displays might be quite long.
    Note: If independent ASPs are available, you will see additional prompts for options 21 and 23, as described in the Saving independent ASPs topic.

    For option 21 (Entire system) these prompt displays appear:

    ENDSBS SBS(*ALL) OPTION(*IMMED)
    SAVSYS 
    SAVLIB LIB(*NONSYS) ACCPTH(*YES) 
    SAVDLO DLO(*ALL) FLR(*ANY)
    SAV DEV('/QSYS.LIB/media-device-name.DEVD') +
        OBJ(('/*') ('/QSYS.LIB' *OMIT) +
            ('/QDLS' *OMIT)) +
        UPDHST(*YES) 
    STRSBS SBSD(controlling-subsystem)

    For option 22 (System data only) these prompt displays appear:

    ENDSBS SBS(*ALL) OPTION(*IMMED)
    SAVSYS 
    SAVLIB LIB(*IBM) ACCPTH(*YES) 
    SAV DEV('/QSYS.LIB/media-device-name.DEVD') +
        OBJ(('/QIBM/ProdData') +
            ('/QOpenSys/QIBM/ProdData')) +
        UPDHST(*YES) 
    STRSBS SBSD(controlling-subsystem)
    For option 23 (All user data) these prompt displays appear:
    ENDSBS SBS(*ALL) OPTION(*IMMED)
    SAVSECDTA 
    SAVCFG
    SAVLIB LIB(*ALLUSR) ACCPTH(*YES) 
    SAVDLO DLO(*ALL) FLR(*ANY)
    SAV DEV('/QSYS.LIB/media-device-name.DEVD') +
         OBJ(('/*') ('/QSYS.LIB' *OMIT) +
             ('/QDLS' *OMIT) +
             ('/QIBM/ProdData' *OMIT) +
             ('/QOpenSys/QIBM/ProdData' *OMIT)) +
         UPDHST(*YES)  
    STRSBS SBSD(controlling-subsystem)
    Type your changes at each prompt display and press the Enter key.

  27. When the system sends a message that asks you to load the next volume, load the next media and respond to the message. For example, if the message is the following, load the next volume and then enter R to try again (C cancels the operation):
    Device was not ready or next volume was
    not loaded (C R)
    Attention: If an unrecoverable media error occurs during the SAVLIB procedure, see Recovering from a media error during a SAVLIB operation.
  28. Mount all other user-defined file systems at this point if you unmounted them for the save operation.
  29. Change the QINACTITV system value back to its original value. You wrote this value down in step 18 c.
  30. When the save operation is completed, print the job log. It contains information about the save operation. Use it to verify that the operation saved all objects. Type one of the following:
    DSPJOBLOG * *PRINT
    Or
    SIGNOFF *LIST

    You have completed the save operation. Make sure that you mark all of your media and store it in a safe, accessible place.

  31. If you ended clustering before running the save operation, restart clustering on the save node from a node where clustering is already active.

    For more information, see the online help for Cluster Resource Services, or see i5/OS Clusters technology.

  32. Restart the device cluster resource group to enable resiliency.
  33. When your independent ASP was saved, the Qdefault.UDFS was unmounted, if you chose to unmount file systems. In order to use the independent ASP again, remount Qdefault.UDFS. Do this step for each independent ASP that you saved.
    MOUNT TYPE(*UDFS) MFS('/dev/iasp_name/Qdefault.UDFS') MNTOVRDIR('/iasp-name')