Home Product Support Operating Systems AIX Upgrade & Migration U2 - Escala Update from AIX 4.3.X to AIX 4.3.3 Latest Technical Level

U2 - Escala Update from AIX 4.3.X to AIX 4.3.3 Latest Technical Level

Release 8.1 (21/06/2004)

Preliminary Remarks

  • This document describes the procedure to follow when updating AIX 4.3.0, AIX 4.3.1 or AIX 4.3.2 systems to AIX 4.3.3 Latest Technical Level,  4.3.3.86 (according to the bos.mp fileset Level number) or Maintenance Level 11.
  • It often refers to the AIX TM 4.3.3& Bull Enhancement SRB (see its reference below), do not forget to get this document.
  • The OSI products are now delivered on several separate CD-ROMs (see U2.0).

Procedure Contents

U2.0 - Media and SRB
U2.1 - Cluster Node Upgrade
U2.2 - Backup the system
U2.3 - DPT adapter specificities
U2.4 - Single User Mode
U2.5 - File systems operations
U2.6 - Filesets to remove
U2.7 - OSI Migration
U2.8 - SSA adapter specificities

WARNING :
If one of the previous operations fail call your support representative.
If you continue unpredictable results may occur and some updates might not be applied

U2.9  - Check the Filesets
U2.10 - Update AIX using "AIX server 4.3.3" CD-ROM'S
U2.11 - Install AIX bundle using "AIX server 4.3.3" CD-ROM'S
U2.12 - Install AIX updates CD-ROMs
U2.13 - Reboot the system
U2.14 - Install Bull Enhancement CD-ROM
U2.15 - Reboot the system
U2.16 - Perform Disks Sub-systems installation
U2.17 - Check the system
U2.18 - Netscape Communicator
U2.19 - Sendmail Configuration

Appendix 1 - How to update a Fiber Channel Adapter Firmware Level


U2.0 - Media and SRB

    • media=CD-ROM 76741108-001 : AIX server 4.3.3 for Escala (4 volumes)
    • media=CD-ROM 76741002-001 : AIX 4.3.3 updates
    • media=CD-ROM 76741370-001 : AIX 4.3.3 Maintenance Level 10 -> 11 (4.3.3.86)

    • To get this media, contact your BULL representative.
    • media=CD-ROM 76741160-001 : Bullenh 433_15
    • SRB 86 A2 98EE Rev 02


    The following media may be useful :

    • Power Sream X25 for AIX 4.3 : 76723980-001 (see U2.14)
    • OSI Media (see U2.7) :
      • CPIC-OSI : 76723992-001
      • OSI Layers 1 to 4 & Netshare - Library : 76723981-001
      • OSI Layers 5 to 7 & TPAD/HPAD - Library : 76723982-001
      • APIs - Library (incl. XTI-API & MAX3) : 76723983-001
      • GCOS Connectivity - Library : 76723984-001
    • Bonus Pack for AIX 4.3 : 76741043-001 (see U2.18)
    • Atf for AIX V4 : 76741022-001 (see U2.16)

U2.1 - Cluster Node Upgrade

    • If you have to upgrade the HACMP software too, please look at the following URL before starting AIX Update :
    • if you just need to upgrade AIX :
      • stop HACMP with takeover option on that node, using the smit clstop command.
      • if you are running the HACMP 4.3.1 software, you must install now the HAS ES 4.3.1 and Bull Cluster Management 4.3.3 Latest Technical Level, please look at : http://epicentre.frec.bull.fr/info/topics/HA/FAQS/faq_ha_level.html (for BULL employees only)
      • if you are running the HACMP 4.2.software, you must upgrade to HACMP 4.4 to cohabit with AIX 4.3.3 and the new Bull Enhancement CD-ROM. See the above referenced URL

U2.2 - Backup the system

 

Perform a bootable backup of your configuration (rootvg) and save other Volume Groups.

U2.3 - DPT adapter specificities

      If the system to update (rootvg) resides on RAID1 DPT disks, and if the adapter firmware level is less than 07M8 (ex: 07L4), the cache must not be validated before updating. To do that, you have to configure the 2 RAID1 disks in "write-through" mode instead of "write-back" mode (See DPT documentation)
      To know if a DPT is present, run : lsdev -Ccadapter
      a DPT adapter looks like :
          sra0   Available    04-07     BULL PCI RAID adapter
      To know if rootvg is on a DPT lun , run :  lspv
      which returns (for example):
          hdisk3         0000002060b7bd8c   rootvg
      and run then :     lsdev -Ccdisk
      which returns, in case of DPT lun :
          hdisk3 Available 04-07-01-0,0 BULL PCI RAID 1

U2.4 - Single User Mode

    • Remove smw fileset:
    installp -u -g smw*
      Then check to be sure they are not present :
      lslpp -l |grep -i smw
    • Check if lft or tty0 is the correct console with a view to reboot the machine on it :

    • lscons (to check)
      smit chcons (to assign)
    • the system must be in single user mode :
      • reboot the system in single mode user after changing this /etc/inittab line
        • "init:2:initdefault:" by "init:s:initdefault:"
      • Check with the "ps -eaf" command to be sure all applications have stopped running.
      • If you use lft display perform the following command:
      export TERM=lft

    WARNING : If applications are still running, unpredictable results may occur and some updates
    might not be applied.

U2.5 - File systems operations

    • Increase the /usr PP (Physical Partition) number with the command:

    • chlv -x 512 hd2
    • I t is recommended that your /tmp file system has at least 40 MB of free space; if necessary, free up in /tmp or extend the file system with SMIT or the chfs command (example here to increase /tmp with 20,000 512 blocks = 10 MB) :
    chfs -a size='+20000' /tmp
      WARNING : If the tmp logical volume name is different from /dev/hd3 the update will fail.
    • Be aware that if the disk is completely full, you may not be able to increase the size of the

    • /tmp or /etc if they are too small.
      If this is the case, make a backup (smit mksysb), and restore with the shrink option.
    • T he minimum free disk space required is 200 MB. Check the rootvg place and PP size with the command:

    • lsvg rootvg
    • Verify the fsize parameter in the /etc/security/limits file. Within AIX 4.3.2, fsize = 0 means that the file size is unlimited, but within AIX 4.3.3, you must either give a size (default is 2097151 x 512 Bytes), either put the "-1" value that means unlimited.
    • It is better to extend the hd5 partition in the following cases :
      • you have a system disk less than 4GB
      • the PP size is 4MB and the hd5 PP number is not greater than 1

      • To check the PP number , run :
          lspv -p hdiskXX| grep -i boot (hdiskXX is rootvg)
        In the case of only 1 PP, you get the follwing result : 1-1    used    outer edge   hd5                 boot      N/A
      If you do not extend hd5, you may have the following problem:
      >0301-161 bosboot: dd failed to copy boot image
      >0301-165 bosboot: WARNING! System not bootable
      To extend the hd5 partition:
      • i f the PP number is 1 extend it with the command:
          • extendlv hd5 1
          WARNING : the two hd5 PP have to be contiguous.
            Two PP not contiguous : 1-1     used   outer edge   hd5                 boot     N/A
            13-13   used   outer edge    hd5                 boot    N/A
            If it is not, you have to remove hd5 and create it again with the commands:
                rmlv -f hd5
                mklv -y hd5 -t boot -u2 rootvg 2 hdiskXX
                lspv -p hdiskXX | grep -i boot (for checking)
            ( Two PP contiguous : 162-163   used   outer middle hd5              boot      N/A)
      • as you modify the hd5 partition, you must run then :
        • bosboot -a -d /dev/ipldevice

U2.6 - Filesets to remove

    • If the following filesets exist : ifor_ls.base.cli 4.3.0.0 and bos.net.ncs 4.3.0.0 , remove them :
          • installp -ug ifor_ls.base.cli
            installp -u bos.net.ncs
      If there are requisite filesets, you can remove them with "-g" option.
    • only when upgrading from AIX 4.3.0 or 4.3.1 remove the bos.perf.pmr fileset with the command :
          • installp -u bos.perf.pmr
    • If you do not need the bullx25 filesets (if you do not use X25 on MCA type machines as DPX/20, EPC800, ESCALA M or D) de-install them because they can introduce problems (see chapter Known Problems of the migration guide)

    • Before de-installing the bullx25 filesets, run the following command :  /usr/lib/methods/udefifxd
    • EXABYTE, BX11 and ipx.rte
      • Perform the following de-installation commands :
          • rmdev -dl rmtX
      installp -u -g Bulltape
            installp -u EXB*
            installp -u BX11*
            installp -u -g ipx.base.rte (if you don't use netware)
      • Then check to be sure they are not present :
          • lslpp -l |grep -i exb or bx11 or ipx.base.rte
    • cda software

    • Check the cda software Release using :  lslpp -l cda (the cda software is used by the EMC disks Sub-system)
      If the Release is less than 2.1.9 :
      • if you do not use EMC Sub-systems, de-install the package now :
          • installp -u  cda
      • if you do use EMC Sub-systems, follow the instructions given in the EMC section.
    • Disks Sub-systems :

    • You have to remove the hdisks attached to these Sub-systems
      • DAS :
        • Take note of the names of the volumes group and the corresponding PV-ID.
        • Unmount all the files system.
              • umount <filesystem*>
        • Vary off the Volume Groups :
              • varyoffvg <vg>
        • DAS without ATF , remove the hdisks :
              • rmdev -dl hdisk<x>
        • DAS with ATF

        • WARNING: If you do not de-install DAS & ATF the machine may crash during the boot after the migration process.
          • Remove atf and sp devices running :
                • rmdev -dRl atf*
                  rmdev -dl sp*
          • Take note of the Atf software Release using : lslpp -l atf.obj
          • Remove the ATF lpp :
                • installp -u -g atf.obj
          • If your current Atf Release is less than 2.1.8, you have to remove the following ODM objects :
                • odmdelete -q"uniquetype=array/scsi/CLARiiON" -o PdAt
                  odmdelete -q"uniquetype=array/fcscsi/CLARiiON" -o PdAt
                  odmdelete -q"uniquetype=array/scsi/CLARiiON" -o PdDv
                  odmdelete -q"uniquetype=array/fcscsi/CLARiiON" -o PdDv
        • Remove the DiskArray and ArrayGUIde lpp with smit or with the following commands :
                • installp -u -g DiskArray
                  installp -u -g ArrayGUIde
      • EMC :
        • Take note of the names of the volumes group and the corresponding PV-ID.
        • Unmount all the files system.
              • umount <filesystem*>
        • Vary off the Volume Groups :
              • varyoffvg <vg>
        • Remove the hdisks :
              • rmdev -dl hdisk<x>
        • If your current cda Release is less than 2.1.9, remove now the already installed cda fileset :
              • installp -u cda
      • Navisphere :
        • WARNING: r emove the hdisks references in /etc/Navisphere/agent.config file (very important) , they will have to be replaced by the "sp" devices after the new Bull Enhancement installation (see Managing & Configuring a DAS : 86 A2 20PN 11)

U2.7 - OSI Migration

    • If you use OSI products it is recommended to update them now, using the OSI CD-ROMs you need. See the description in the U2.0 section
    • Insert the CD-ROMs and type the command:
          • smit update_all
    • to get more information about the OSI CD-ROMs, see Chapter OSI Products in the AIX 4.3.3 & Bull Enhancement 433_15 SRB.

U2.8 - SSA adapter specificities

 

Verify if the devices.mca.8f97.com fileset is installed using the command :

 

 

lslpp -L devices.mca.8f97.com

If it is installed, now run the ssa_pre_i.efix script ., as root.

U2.9  - Check the Filesets

 

Check the Filesets using the following command :
lppchk -v -m3
if you do not use the fileset displayed, remove them using the following command:
installp -u "FilesetName"
otherwise call your support representative.

U2.10 - Update AIX using "AIX server 4.3.3" CD-ROM'S

    • as you are in single_user mode, run :  /usr/sbin/srcmstr & (otherwise, some packages will not be installed)
    • If you are updating a Version 4.3.0 or Version 4.3.1 system, you must update and commit the bos.rte.install fileset prior to updating the rest of the system.
      • Insert the installation media and issue this command as root :
        • installp -acgXd <DEVNAME> bos.rte.install 4.3.3.75
      where <DEVNAME> represents the device name of your installation device, for example, /dev/cd0 (CD-ROM device), /dev/rmt0.1 (tape device), or a directory where the bos.rte.install 4.3.3.75 is available.
      • Check to be sure "bos.rte.install 4.3.3" is installed with the command:
          • lslpp -l |grep -i bos.rte.install
      • Enter then the following command and accept the default menu values:
          • smit update_all
        WARNING : smit will ask you several times to insert a different CD-ROM volume
    • If you are updating a Version 4.3.2 system on ESCALA T430, T450, EPC430 and EPC450 models, you must now install PTF Z002928 (devices.pci.isa_BULL_chrp.rte 4.3.2.1) available on the Bullenh CD-ROM in the /PTF directory (to install the PTF, first mount the CD-ROM, then copy the PTF in the /usr/sys/inst.images directory and install it from that location)

    • Once this PTF is installed,  remove the CD-ROM and insert the "AIX server 4.3.3 CD-ROM Vol.1" , then run the following script :
          /usr/local/bin/updt_433_BULL_chrp | tee /tmp/update_433
      this will cause the  update_all command to run automatically, look at the result of the installation in the /tmp/update_433 file .
    • If you are updating a Version 4.3.2 system on all other ESCALA models, de-install now the fileset devices.pci.isa_bull_rspc.rte, running the command :
        • installp -u devices.pci.isa_Bull_rspc.rte
      Do not worry if you get the answer " nothing to de-install "
      Then, enter the following on the command line and accept the default menu values:
              smit update_all
      WARNING : smit will ask you several times to insert a different CD-ROM volume

      WARNING : if , at the end of this update, you get a "Bosboot failed" message, re-run the bosboot command as follow :

          bosboot  -a
      if you get then "missing proto file /usr/lib/drivers/isa/isa_busBull", perform the following command :
          ln -sf /usr/lib/drivers/isa/isa_busdd /usr/lib/drivers/isa/isa_busBull
      and then re-run the bosboot -a command.
      For any other message, contact your Bull representative.

U2.11 - Install AIX bundle using "AIX server 4.3.3" CD-ROM'S

    • Using the volume 1 CD-ROM, install the fileset bundle_aix433 withthe following command :
      • installp -acgXd  /dev/cd0 bundle_aix433
    • then install AIX_4.3.3 bundle with " smit install_bundle "

    • If the result of installation is "Failed" because of bos.perf.perfstat fileset, you will have to re-install it after rebooting (see U2.13)

      WARNING : in case of DPT disks use,  you have to re-install now the devices.bullraid.rte fileset

U2.12 - Install AIX updates CD-ROMs

 

  • Insert the "AIX 4.3.3 updates"CD-ROM and type the command :

 

    smit update_all

  • Insert the "AIX 4.3.3 Maintenance Level 10 -> 11 (4.3.3.86)" CD-ROM and run the command :
  • smit update_all

    If the fileset devices.pci.14109f00.rte is installed on your system, you will get a Requisite Failure applying the "AIX 4.3.3 Maintenance Level 10 -> 11 (4.3.3.86)" CD-ROM.
    If your customer does not use the IBM PCI 4658 CRYPTOGRAPHIC COPROCESSOR ADAPTER SOFTWARE (this adapter is not sold by BULL), the filesets devices.pci.14109f00.* can be removed. If the customer uses such an adapter, he has to download the U480740 PTF which will be installed using smit install_latest.

U2.13 - Reboot the system

    • Before rebooting, :
      • remove the devices.pci.00101010.rte (no more correct for AIX 4.3.3) fileset, running :
          • installp -u  devices.pci.00101010.rte
        In case of AIX mirroring on the LVD additional controller (EPC430 & ESCALA T 430 models), de-installation of this fileset may fail, see TSB 500.00007.00.tsb
      • remove the devices.pci.df1000f7* filesets (Fiber channel drivers not used), running :
          • installp -u  devices.pci.df1000f7*
    • Run then :     shutdown -Fr
    • Install bos.perf.perfstat fileset from the AIX server 4.3.3 CD-ROM Vol.2, running :
        • installp -acgXd  /dev/cd0 bos.perf.perfstat

U2.14 - Install Bull Enhancement CD-ROM

      WARNING :
      • It is very important to install the last release of Bull Enhancement. If you do not perform this installation, you can introduce incompatibility problems.
      • The psx25 product is no more delivered with the Bull Enhancement CD-ROM. If you use psx25, Version 2.0.11 must be installed on AIX 4.3.3 systems, or as soon as fileset bos.rte.tty has to be updated to 4.3.3 level.
        You can install it from the Power Sream X25 for AIX 4.3 CD (Ref. 76723980-001).
    • as you are in single_user mode, run :  /usr/sbin/srcmstr &
            • and  /usr/sbin/cron &
      (otherwise, some packages will not be installed)
    • check if your configuration includes an IBM PCI 8-Port Asynchronous Adapter, running the command :
        • lsdev -Ccadapter|grep ^sa
      If it does, you will get IBM 8-Port EIA-232/RS-422A (PCI) Adapter for one of the "sa<x>" adapters. In that case, take note of the bullasync.pci.rte fileset release using the command:
          lslpp -l bullasync.pci.rte
    • If your configuration includes FC DAS 5300, 4500, 470x and (or) NDAS CXxxx and if the failover mode software is PowerPath :
      • insert the Bull Enhancement CD-ROM and run :
        • smit install_latest
      • select : INPUT device/directory for software :  [/dev/cd0]

      • and install all or only the filesets you need.
    • If your configuration includes FC DAS 5700, 5300, 4500, 470x and (or) SCSI DAS and if the failover mode software is ATF

    • or
      if your configuration does not include any DAS at all :
      • Insert the Bull Enhancement CD-ROM and run then :
        • mount /cdrom
          cd /cdrom/FC_ATF
          smit install_latest
      • select : INPUT device/directory for software :  .

      • and install all or only the filesets you need.
    • If you need the bullx25 filesets (if you use X25 on MCA type machines as DPX/20, EPC800, ESCALA M or D), you have to install them from the /MCA directory of the Bullenh CD-ROM but have a look to the Known Problems chapter of the Migration Guide (installing these filesets can introduce problems).
    • the ArrayGUIde Manager is no longer supported within this Bull Enhancement Release, use the Navisphere Manager.
    • if you introduce a DAS 4500 or a DAS 4700 in your configuration, you must use Navisphere Manager for Windows
      • in case of other DAS, if you still want to use Navisphere Manager for AIX :
        • first de-install the following filesets :
          • installp -u -g Navisphere.Agent
            installp -u -g Navisphere.CLI
            installp -u -g DiskArray
        • then install all the filesets located in the /Mgnt_AIX of the Bullenh CD-ROM
    • for all that concerns the Navisphere product, it is strongly recommended to read :
      • AIX 4.3.3 & Bull Enhancement SRB , PART 2, "DAS Subsystems-Navisphere"
      • Configuring and Managing a DAS : 86 A2 20PN 11
    • the devices.pci.48110040 package present on the Bull Enhancement CD-ROM is dedicated to PCI FDDI adapters with the label B5-3, B5-4 or B5-5. Consequently, if you have PCI FDDI adapters without such a label, you have first to de-install the devices.pci.48110040 package , then to re-install it from the FDDI media delivered with the adapter (see AIX 4.3.3 & Bull Enhancement  SRB , PART 1, "Install Bull Enhancement Packages").
    • if your configuration includes an IBM PCI 8-Port Asynchronous Adapter and if the previous bullasync.pci.rte fileset release you took note before installing the new Bull Enhancement CD-ROM is greater or equal to 1.8.2, you have nothing to do.

    • If it is less than 1.8.2, you have to re-install the devices.pci.4f111100 package from the AIX server 4.3.3 CD-ROM.
    • Warning : if your configuration includes LP6000 adapters with firmware level = 2.20 , you have to update the firmware level to 2.21 before rebooting your system, otherwise your system will not reboot.

    • To verify the type and level of your fiber channel adapter, run the following command : lscfg -vl fchan<x>
      You will get :
      -----
      Configuration which prevents booting :
      ----------

      fchan1            04-01             PCI Fibre Channel Adapter

      Adresse de réseau...........0000C9205704
      Niveau de modif. technique..BIU 2
      Référence ..................76730052
      Message affichable..........Fibre Channel Host Adapter
      Niveau microcode télécharg..2202000
      Niveau diagnostics..........2322000
      Fabricant...................EMULEX
      Spécification unité (Z1)....ENDEC 21acb057
      Spécification unité (Z2)....SM 11acc057

      -------
      Right configuration got after firmware updating :
      ------------
      fchan2            04-02             PCI Fibre Channel Adapter

      Adresse de réseau...........0000C9204986
      Niveau de modif. technique..BIU 2
      Référence ..................76730052
      Message affichable..........Fibre Channel Host Adapter
      Niveau microcode télécharg..2212000
      Niveau diagnostics..........2332000
      Fabricant...................EMULEX
      Spécification unité (Z1)....ENDEC 21acb057
      Spécification unité (Z2)....SM 11acc057

      --------------

      See : How to update a Fiber Channel Adapter firmware level

U2.15 -Reboot the system

    • Before rebooting, restore the multi-user mode in /etc/inittab file, replacing :
        • "init:s:initdefault:" by "init:2:initdefault:"
    • If you do not have to perform any DAS or EMC installation, reboot the system now, running :
        • shutdown -Fr
    WARNING : if your configuration includes LP7000 adapters in FC Loop environment, they may not be configured after microcode update (fileset.devices.pci.df1001f7.ucode 1.3.3.7 or later).

    Run the following command :

          lsdev -Cc adapter | grep fchan
    If the fchan are in the defined state, this may be due to connection problem (MIA, cable, hubs, ...) which were tolerated by microcode level 3.20 but no more accepted by microcode levels 3.22 or 3.30.
    Solution is to fix connection problem or, as a workaround, go back to microcode level 3.20.
        To identify LP7000 (BIU 4) adapter and microcode (322xxxx), run :
          lscfg -vl fchan<x>
        fchan1            30-70             PCI Fibre Channel Adapter

        Network Address.............0000C920C77A
        EC Level....................BIU 4
        Part Number.................76730259
        Displayable Message.........Fibre Channel Host Adapter
        Loadable Microcode Level....322xxxx (or 330xxxx)

        To downgrade microcode level:

          cd /usr/lpp/devices.pci.df1001f7
        • rename the current new microcode :

        • cp df1001f7.awc  df1001f7.awc.NEW
          cp df1001f7.mcode  df1001f7.mcode.NEW
        • copy microcode 320  to current location :

        • cp df1001f7.awc.320 df1001f7.awc
          cp df1001f7.mcode.320 df1001f7.mcode
        See then : How to update a Fiber Channel Adapter Microcode level


    Don't install further PTFs related to fileset devices.pci.df1001f7.ucode
    as long as connection problem is not fixed ...

U2.16 - Perform Disks Sub-systems installation

    • only in case of DAS :
      • since some important problems have been described in TSBs, please have a look at the suggested URL in the following cases :
      • the DiskArray and fcDiskArray packages have been re-installed with the Bull Enhancement CD, verify the flare Firmware release of your DAS, and re-install it if it has changed (Refer to AIX 4.3.3 & Bull Enhancement SRB - PART 2 - DAS Subsystems).

      • Note :
          After the DAS flare upgrading to :
          • 5.11.14 for DAS 5700 , 5.11.64 for 5720
          • 5.24.05 for DAS 5300
          • 5.32.13 for DAS 4500
          the disks LEDS blink continuously. Do not worry, this is a normal behaviour and will remain forever.
      • if you have ordered it, now install the Atf package from the Atf CD-ROM (see the Atf SRB : 86 A2 70WF).

      • WARNING : compatibility between Atf release and other software packages as DiskArray is important. It is described in the Atf SRB and at the following URL : http://epicentre.frec.bull.fr:80/info/topics/AIX/compatibility/Atf.html
      • do not reboot the system at this moment
    • if you have upgraded a Cluster Node :
      • as other Nodes have taken the upgraded Node resources during the Upgrading process, Volume Groups will be varied on on these Nodes.

      • So to be able to recover the Volume Groups  on the upgraded Node:
        • on "split_bus" configurations (one SCSI bus by Node to a shared DAS), you have to stop HACMP on the other Node ( smit clstop ) and then run :
        • /usr/lpp/DiskArray/das_available  hdisk<x>

          hdisk<x> is the hdisk where the Volume Group has to be re-imported
        • on other configurations (DAS with Atf, EMC...), it is mandatory to break the disk reservation running the following command on the other Nodes (where the VGs are varied on):
          • varyonvg -b -u <VGname>
      • If you have not already done it (because of Atf installation), run then, on the upgraded Node, for each hdisk :
    rmdev -dl hdisk<x>
      • reboot the system
      • verify that all the Volume Group have been recovered
      • you can now restart HACMP running the smit clstart command and verify that the upgraded Node recovers its resources
    • otherwise :
      • reboot the system
      • import the Volume Groups that  you took note before the migration
      • run the varyonvg command for each Volume Group
      • mount the filesystems

U2.17 - Check the system

    • Perform :   lppchk -v -m3
    • Run :  oslevel

    • if you do not get 4.3.3.0 , run oslevel -l 4.3.3.0 to get the filesets which cause the AIX level to be wrong. If you don't use these filesets, remove them (installp -u <Fileset>), otherwise call your support representative.
    • Check if all the devices are available (tty, cd-rom, tape, hdisk ....)
    • If a problem occurs, take a snap , trace your system with the command cfgmgr -v (forbidden if ATF installed) and send them to your Support Representative.
    • Check if all your applications work.
    • Create a dump partition, or increase it so that it can contain at least the estimated dump size (run sysdumpdev -e to know it), then configure your dump devices.
    • Check that remote maintenance is operational (using RSF).

U2.18 - Netscape Communicator

 

WARNING : Netscape Communicator 4.76 at least is mandatory in an AIX 4.3.3.81 environment.
It is recommended to install Netscape Communicator 4.79, delivered in the following CD-ROM :
- Bonus Pack for AIX 4.3 : 76741005-001

U2.19 - Sendmail configuration

 

When updating from the previous version of AIX 4.3, the version 8.9.3 sendmail binary will be installed, but the version 8.9.3 /etc/sendmail.cf file will not. The original /etc/sendmail.cf is kept. However, you can generate an 8.9.3 sendmail.cf if desired.

Sendmail version 8.7 (supported on AIX 4.3.0),  sendmail version 8.8.8 (supported on AIX 4.3.2) and sendmail version 8.9.3 are somewhat compatible. The version 8.7 /etc/sendmail.cf will work with the the version 8.8.8 and version 8.9.3 sendmail binaries. The version 8.8.8 /etc/sendmail.cf will work with the version 8.9.3 sendmail binary. However, each new version of /etc/sendmail.cf contains new rewrite rules, macros and options that are particular to the updated sendmail binary. So, if you choose to run version 8.7 /etc/sendmail.cf with the version 8.9.3 sendmail binary, you will lose some new features and functionality particular to version 8.9.3. You may wish to merge in the new options, macros, and rewrite rules. The new sendmail.cf should have been saved in /usr/lpp/save.config/etc/sendmail.cf.new. You will have to merge this manually as there are no migration scripts to do this.
NOTE :
When the version 8.9.3 sendmail binary is running with a previous sendmail.cf file, the following warning will appear :

Warning: .cf file is out of date: sendmail AIX4.3/8.9.3 supports version 8, .cf

This is just a warning. Change the version to 8 in the V macro in the /etc/sendmail.cf to get rid of this warning.


Appendix 1 - How to update a Fiber Channel Adapter Firmware Level
      Updating FW on a FC board (whatever the model is) implies loading a specific maintenance and diagnostics driver in AIX for performing the operation. To do that, the adapter must be free of  attached devices (i.e rmdev/mkdev of adapter should be possible). So two solutions:

      First solution :

       

    • stop applications and umount/varyoff/export VG and FS
    • stop other applis that could have open access to disks (for example navisphere)
    • run the diag command (it will automatically perform a rmdev and mkdev of the adapter, that should succeed if the adapter and the disks are no more used).
    • Download the firmware using :
    • Task Selection -> Download Microcode

      (follow plug/unplug instruction when needed)

       

    • If this still fails, manually try a rmdev -Rl fchan<x> to check what object is still used (hdisk, fcp, sp ,atf, etc...?). also check IP on FC ( ifconfig fe<x> down or detach )
    • Once all applis are really stopped and fchan objects are in define state, run diag/Download Microcode again: it should succeed.
    • run cfgmgr to put back objects in available state, if needed.
    • re-start admin tools such as navisphere.
    • importVGs and mount FileSystems
    • restart applications.

    • Second solution :

      Consider upgrading FW on an adapter is a maintenance operation...So:

       

    • boot your system in service mode
    • use diag command to download the firmware
    • reboot in normal mode.

    • For LP7000 and LP8000, an automatic download feature has been added so that, early in the system 's life, and surely before any application is started, the FW version is checked on the FC board, and replaced by a newer one if available. The adapter is free at this time, so the download microcode operation is always possible.


 


Assistance request
Create and track
Bull Search