Skip to main content

IBM Systems  >   z Systems  >   z/VM  >  

VM Service Tips

Updated: 23 March 2015

Documentation, Planning and Ordering Information

  • For information about z/VM installation information, and installation related problems, check these links:

  • Preventive Service Planning (PSP) Buckets Search (via the System z Technical Help Database)

    This facility provides access to PSP bucket Upgrade and Subset information (for software and hardware topics). When using this facility, be certain to place a check mark in the box for Preventive service planning, under the "Document types" heading. Then, in the "Search for" box, provide the PSP Upgrade and subset names of interest, such as:

      2817DEVICE 2817/ZVM or ZVM630 6304RSU

    Of interest:

    • To review known installation problems for a given z/VM product level, specify the applicable PSP Upgrade name (ZVMvrm) as the only search argument (where, where vrm designates a version, release and modification level). From the "Upgrade/Subset" selections that are presented, select the CP Subset, and then review the information provided under the Installation Information heading.

    • To review service recommendations, check the recommendations listed in the yynnRSU subsets listed for a given release.

  • VMSES/E SSI Cluster Mixed-Release Environment Support (z/VM 6.2)

    Publication changes associated with APAR VM65317 (Support for Mixed-Release Environment in SSI Cluster) are described in the VM65317 APAR MEMO file. (PDF)

  • Corrective Service Order Tips

    When z/VM corrective service (APAR fix updates, provided via a Product Temporary Fix, or PTF) is ordered for multiple components as part of a single order, the lowest Recommended Service Upgrade (RSU) RSU level, that is common to those components, must be specified.

    When separate orders are created on a component-by-component basis, specify the applicable component RSU level for each order.

    To determine the RSU level of a given component, use the VMSES/E SERVICE command:

     service compname status
    
    where compname is that of a known z/VM component (such as CP or CMS).

    The compname operand also can be specified as ALL, which will display the RSU level for all of the components defined in the VM SYSSUF inventory table.

    To determine if an APAR or PTF of interest is applied to a system, add the APAR or PTF number (for example, vm54321 or um23456) after the STATUS keyword of the SERVICE command, as in this example:

     service compname status vm54321
    

    Note:
    For systems that have RSU service installed, it is possible that the APPLIED date reported for a given APAR or PTF can precede its corresponding RECEIVED date, due to the nature of how the z/VM RSU is created.

  • Processing APAR MEMO Files

    For APARs that include a (user) MEMO file, memo information can be found at the bottom of the prodid MEMO file, included as part of the COR service tape.

    For an electronic order, the prodid MEMO can be the prodid MEMO file can be found in the VLSTnnnn (or, ordernum) SHIPDOCS file that is supplied as part of the order.

    An example of an APAR MEMO file is available here: VM62480

    For detailed information about how to process the COR service documentation envelope and extract the MEMO file, consult the z/VM: Service Guide (GC24-6247),

    • Chapter 2: Installing preventive or corrective service
    • "Steps to receive, apply and build preventive or corrective service"
    This documentation also discusses use of the VMUPDAT SYSMEMO command to display APAR memos.

Segment-Related Problems

  • MONDCSS Build Failure due to Active Use

    Attempts to rebuild the MONDCSS saved segment using PUT2PROD (whether on an individual basis, or when all segments are rebuilt using a PUT2PROD SEGMENTS ALL command) often fail because one or more virtual machines continue to actively use this segment during service-related tasks. The messages that follow are indicative of this kind of problem:

     ...
      VMFP2P1965E The command, SEND BLDSEG EXEC VMFBLD PPF SEGBLD
                  ESASEGS SEGBLIST MONDCSS.SEGMENT LINK (ALL NOLOG,
                  failed with return code 8
      VMFP2P2760I PUT2PROD processing completed unsuccessfully for
                  SEGMENTS
      VMFP2P2760I PUT2PROD processing completed unsuccessfully
     ...
    
    When the MONDCSS segment is rebuilt, the segment cannot be in active use by any users (which, for a system that is actively collecting monitor data would be a normal case). To identify which user IDs are using this segment, use the CP QUERY command that follows:
     cp query nss users mondcss
    
    Note:
    APAR VM65651 introduces changes that identify users of the MONDCSS segment, and which alter how this condition is handled by the PUT2PROD command.

  • HELPSEG Build Failure due to Missing Files

    There are various conditions that can cause HELPSEG segment build errors. In some instances, the cause might be readily apparent, while for others, this may not be the case. For example, the following (or similar) set of messages, in the PUT2PROD message log ($VMFP2P $MSGLOG):

     ...
     ST:BLDSEG  : VMFBDS1965E The command, VMFSETUP, failed with return code 8
     ST:          when issued with the argument(s):
     ST:BLDSEG  : ZVM CMS (LINK LOG NONE NOCONS RETAIN A B D S Y
     ST:BLDSEG  : VMFBLD1851E (1 of 1) VMFBDSEG completed with return code 100
     ST:        : Some objects were not built
     ...
     SV:VMFP2P1965E The command, CP SEND BLDSEG EXEC VMFBLD PPF SEGBLD ESASEGS
     SV:            SEGBLIST HELPSEG.SEGMENT LINK (SERVICED NOLOG, failed with
     SV:            return code 8
     ST:VMFP2P2760I PUT2PROD processing completed unsuccessfully for SEGMENTS
    
    were produced because the CMS MAINT 990 minidisk was empty. This disk is populated with zCMS files as part of the z/VM system deliverable, so should not be empty. To repopulate the disk, one could use the VMFBLD command shown here:
     vmfbld ppf servp2p cms|cmssfs dmsbl990 * (all
    
    Afterward, the following steps would be used to complete the
    1. Invoke the ZCMSBLD exec, to create a new zCMS nucleus map.
    2. Receive the generated nucleus map file from the virtual reader the CMS BASE minidisk (MAINT 3B2) as this file: ZCMS NUCMAP
    3. Issue this VMFBLD command to resolve outstanding build requirements from the original (failed) build:
       vmfbld ppf servp2p cms|cmssfs (serviced
      

  • Recovery of the Default Segment Spool Files (NSS and DCSS)

    For z/VM 6.2.0, and later, consult "Chapter 4: Running stand-alone builds" of the z/VM: Service Guide (GC24-6247) for specific steps to rebuild the default segment spool files.

    To recover the CMS or GCS segments, first use the commands cited in "Rebuilding the CP, CMS, and GCS nucleus", followed by those in "Rebuilding segments".

    For z/VM 5.4.0 or z/VM 6.1.0, use the following steps:

    1. Complete the steps cited in "Rebuild the Saved Segments", as part of "Step 12. Place Your Serviced Components into Production" in Chapter 3 of the (540/610) z/VM: Service Guide (GC24-6117/GC24-6232).

      Notes:

      • Instead of using individual VMFBLD commands to rebuild each segment, the following command can be used to rebuild all of the CMS and LE segments:
         vmfbld list segbld esasegs zvmsegs blddata (all
        
      • If segments are built on an individual basis, the ALL option must be specified in place of the SERVICED option, shown for the VMFBLD command cited by this documentation.
    2. Issue these commands, to create new CMS and GCS segments, and to save them:
       service cms bldnuc
       service gcs bldnuc
       put2prod
      
    3. With the above steps complete, the MONDCSS, PERFOUT and ZCMS segments remain to be built. Use the following steps to build these segments.

      Here, z/VM 5.4 prodid values (5VMCMS40 and 5VMPTK40) are used. For z/VM 6.1, the respective values would be 6VMCMS10 and 6VMPTK10.

      • For the MONDCSS and PERFOUT segments, invoke the VMFSGMAP utility:
         vmfsgmap
        
        For MONDCSS:
        1. From the "VMFSGMAP - Segment Map" panel, use: F10 (Add Obj)
        2. The "Add Segment Definition" panel is displayed. On this panel:
          1. Supply MONDCSS for the OBJNAME field
          2. Supply 5VMCMS40 CMS for the PRODID field
          3. Press F10 (Seginfo)

            (This should retrieve the correct segment definition from the 5VMCMS40 PRODPART file); a message that reports the segment is already defined can be safely dismissed).

          4. Press F5 (Map), to return to the main VMFSGMAP panel
          5. Press F5 (File) to complete MONDCSS processing
        For PERFOUT:

        Repeat the above steps, but supply PERFOUT for the OBJNAME field, and 5VMPTK40 PERFTK for the for the PRODID field

        After exiting the VMFSGMAP utility, use the VMFBLD command to build the MONDCSS and PERFOUT segments (These segments are "data" segments, so this build step formally defines the segments):

         vmfbld ppf segbld esasegs segblist (serviced
        

      • For ZCMS:

        Consult the z/VM z/Architecture CMS (PDF) documentation for detailed instructions.

Miscellanous Service Problems

  • Handling a Missing PTF Requisite During SERVICE Processing

    If the VMSES/E SERVICE command fails due to a missing PTF requisite, use the steps that follow recover from this problem and to allow a SERVICE RESTART to complete.

    Note:
    The example commands that follow are based on the application of service that involves electronic media (that is, a SERVLINK envelope file — service_env_filename).

    For tape media, one would omit the ENV options of the VMFREC command, and the SERVICE RESTART command required would be that cited by the VMF2310W message at the time of the original SERVICE command failure.

    • Order the missing requisites that have been identified
    • Log on as the maintenance user ID (MAINTvrm, or its equivalent)
    • Use the commands that follow to receive the PTFs for the missing requisites (contained in a requisite_env_filename SERVLINK file).

      For these commands, ppfname and compname values that pertain to the affected component(s) must be supplied. Pertinent values can be obtained from a review of the SERVICE command message log ($VMFSRV $MSGLOG). In most instances, the PPF name required will be SERVP2P, and the component name will be that of a known z/VM component (such as CP or CMS).

      If the missing requisite order contains PTFs for more than one component, repeat these commands for each, as needed.

       vmfsetup ppfname compname ( link
       vmfrec ppf ppfname compname ( env requisite_env_filename
      
    • With receipt of the missing requisite PTFs complete, restart the SERVICE processing that failed:
       service restart service_env_filename
      

  • OSA/SF MODULE Rebuild Commands (for z/VM 6.3)

    Note: This problem is resolved with application of the 6304 RSU.

    Customers using OSA/SF on z/VM 6.3 need to use the commands that follow, to rebuild the parts that were incorrectly supplied at an older service level.

    For the listed commands:

    • If a PPF override has been created for the SERVP2P PPF, substitute the override PPF name in place of SERVP2P
    • If OSA has been installed to Shared File System (SFS) directories, use OSASFS instead of OSA as the component name.
    1. Log on to the MAINT630 user ID
    2. Issue these commands:
       vmfsetup servp2p osa|osasfs
       vmfbld ppf servp2p osa|osasfs ioabltcp * (all
       vmfbld ppf servp2p osa|osasfs ioablmod * (all
       service osa build
       put2prod osa
      
    3. For a multi-member SSI cluster, log off MAINT630 (from the current member system). Then, log on MAINT630 on each additional member, and repeat PUT2PROD OSA command. Recycle the OSA/SF servers on each system that has been updated

    Additional Notes:

    Common symptoms, indicative of the need to rebuild affected OSA/SF parts (as just described), include:

    • An IOACMD failure with results similar to the following:
       ...
       IOACMD: ******************************************************
       IOACMD: Query completed with the following result
       IOAE033E APPC failed attempting to make connection with OSA/SF
       IOACMD: ******************************************************
       ...
      
    • An OSASF abend reported (and, likely producing a dump), from having issued the IOACMD command:
       ...
       OSASF abending: CMS Abend code 000000E0 generated at location 80045B02
       ...
       DMSABE141T Operation exception occurred at 800FA894 in routine IOACMAIN
       ...
      

  • PUT2PROD LOADBFS Processing Errors (Associated with TCP/IP)

    If LOADBFS processing errors are reported by the PUT2PROD (or, TCP2PROD) command for the TCP/IP component, review the information for the LOADBFS Error DTCPRD3038E topic, provided as part of the TCP/IP for z/VM Service Hints and Tips page.

  • Message HCPLMN107E for the 6VMRAC20 651 Minidisk With Application of APAR VM65320 PTF

    If the following messages are received when the PTF for RACF APAR VM65320 is applied:

     ...
     VMFSET2204I Linking 6VMRAC20 651 as 651 with link mode MR
     HCPLNM107E 6VMRAC20 0651 not linked; not in CP directory
     VMFSET1965E The command, CP LINK 6VMRAC20 651 651 MR "", failed with
                 return code 107
     VMFSET2760I VMFSETUP processing completed unsuccessfully
     VMFSUI1965E The command, VMFSETUP, failed with return code 100 when
                 issued with the argument(s): SERVP2P RACFSFS (LINK
                 NOPROMPT RETAIN D
     ...
    
    use the information supplied in the UMEMO section of the $PTFPART file (for the PTF that is associated with this APAR) to resolve this problem. The UMEMO states:
     ************************************************************
     ****           IMPORTANT CP DIRECTORY CHANGE           *****
     ************************************************************
     The 6VMRAC20 user ID entry in the CP user directory needs to
     have a new MDISK statement for a 651 minidisk added to it with a
     size of 1 cylinder. The 651 disk must be CMS formatted before
     the PTF for this APAR is applied. If the SERVICE command for the
     PTF fails because the 651 minidisk does not exist then you can
     create and format the minidisk and then issue SERVICE RESTART to
     finish the PTF application.
    
    The VMFSRV2310W message (in the SERVICE command message log, $VMFSRV $MSGLOG) will cite the SERVICE restart command that is necessary to restart SERVICE processing.

  • Message DMSSLG278E Reported by CMS Applications that use SET LANGUAGE

    If DMSSLG278E Unable to set requested language AMENG; AMENG forced is received during service processing, review the How to handle DMSSLG278E message page for more information and circumventive actions.

  • DMSUPD174W Sequence error IRAQVS MACRO during VMFBLD Processing

    The following error messages might be received during SERVICE processing, after a system migration has been completed (a migration that has used the VMSES/E MIGRATE command):

     ...
     WN:DMSUPD174W Sequence error introduced in output file: 00032000 to 00032000
     SV:VMFMLB1965E The command, UPDATE, failed with return code 8 while
     SV:            processing object IRAQVS. It was operating upon file
     SV:            IRAQVS MACRO on 493 (K)
     SV:VMFBLD1851E (1 of 1) VMFBDMLB completed with return code 100. Some
     SV:            objects were not built
     SV:VMFSUI1965E The command, VMFBLD, failed with return code 8 when issued
     SV:            with the argument(s):
     ...
    
    Check for the presence of an IRAQVS MACRO file on the MAINT 493 minidisk. This file should not exist on this disk. If it is found at this location, delete the file from the MAINT 493 disk, and restart the subject PTF application.

    This type of error might also occur for any new part that has been added via a PTF.