HMC V7.3.4.0 Service Pack 2

This package updates your HMC from HMC V7R3.4.0 Service Pack 0 or Service Pack 1 to HMC V7R3.4.0 Service Pack 2. You can also reference this package by APAR MB02891.

Installing MH01162 requires one of the following:

• MH01097, the HMC V7R3.3.0 recovery image, plus MH01109 (the upgrade to HMC V7R3.4.0).
• MH01147, the HMC V7R3.4.0 recovery image.
• MH01097, the HMC V7R3.3.0 recovery image, plus MH01109 (the upgrade to HMC V7R3.4.0) plus MH01163, the HMC V7R3.4.0 SP1 upgrade image.
• MH01147, the HMC V7R3.4.0 recovery image, plus MH01163, the HMC V7R3.4.0 SP1 upgrade image

This update (MH01162) supersedes MH01163.

This package includes the following enhancements and fixes:

• Make Manage Profiles Hyperlinked to Edit profile. This enables the user to quickly edit an lpar's profile using a single click without having to select the profile, then go to Action->Edit Profile.
• Fixed a problem to prevent concurrent active migration from stealing memory or processor resources from other partitions that are still in the middle of a migration.
• Solved a timing issue when calculating system available memory resources.
• Corrected a problem which causes the Network Settings task to fail if the HMC hostname is set to an invalid value.
• Corrected a problem which causes LIC Update error message dialogs to be shown twice.
• Added support to show the migration state for LPARs managed by HMC and IVM.
• Added performance enhancements for HMCs managed by IBM Director.
• Fixed a performance issue where "lshmcusr" CLI was slow.
• Fixed a problem where a user's resource role was replicated incorrectly when the user's resource role was empty. Instead of having the same empty resource role in slave HMC, the user role in slave HMC was assigned an AllSystemResources resource role. An empty resource role implies no access to any managed server resource.
• Corrected a HSCL350B error when a non-hscroot user performs a dynamic logical partitioning IO move using the depreciated syntax of "chhwres -r io --rsubtype slot". This allows scripts created at earlier HMC releases to run successfully after an HMC upgrade without modification.
• Fix for the Collect VPD tasklet (under Systems Management -> Tasks -> Serviceability -> Collect VPD) failing with message "The VPD collection command succeeded, but the resulting output could not be packaged for copying".
• Fixed errors transmitting inventory information from the Transmit Service Information tasklet. Errors include exception parsing XML_INVENTORY and [SQL0803] Duplicate key value specified.
• Fixed a problem where activating a second lpar with a redundant switch configuration may fail to activate with the message: com.ibm.hsc.common.exceptions.PIHscClientException: HSCLA333
• Fixed a problem with the linux rpm command that affects the size of the files made when using the HMC "backup critical console data" command. The fix correctly calculates the set of HMC RPMs that need to be backed up.
• Added additional feature code mappings for 571F scsci disk controller. This allows deploy system plan to successfully deploy on systems containing this card.
• Fixed problem where HMC reports service event E35510C1 0000A346
• Fixed incorrect description of Ethernet adapter feature code 5740.
• Fixed a problem where mksysplan may report error message N3xxxException Occurred in com.ibm.inventory.VPDRegistry:populateSlot1
• Fixed a problem with partition migration where the migration fails if the client partition is using a vlan adapter with a non-zero virtual switch id
• The power on/off unit task has been updated to clearly identify the target of the operation by the unit location code.
• Fixed a problem where E3550046 00001B6B may be reported when using the "Find Managed System" option of the "Add Managed System" task.
• Fixed a problem where service event E355104C 000086E3 may be reported
• Fixed a problem where IBM5250 emulator fails to start from the setup5250 panel
• Fixed error ACT01503 with serviceable events E36C0101 and E3550037 when displaying details of previously scheduled backup critical console application.
• Fixed report of an E212E152 serviceable event (CPU busy), during command line interface (CLI) backups (bkconsdata command) or scheduled backups created at earlier releases. The E212E152 may be accompanied by an E3321011.
• Correct a problem with the Edit profile task where removing an IO adapter from the profile does not correctly clear tagged IO settings that reference the removed adapter.
• Fixed a problem where a system profile with 11 or more partition profiles displays the partitions in the incorrect order however the activation order is correct.
• Fixed a problem where the "lssyscfg -r sys" command output does not include "sp_failover_enabled" and "sp_failover_state".
• Updated help text to specify that LDAP user authentication requires the use of anonymous user binding.
• Fixed problem where service event E355104C 0000B593 may be reported after removing a managed system connection.
• Fixed a problem causing the lpar_netboot command to return the error "Unable to obtain network adapter information."
• Fixed error with host command.
• Fixed a problem that may result in serviceable event E355104D 0000C07
• Fixed a problem where restoring profile data after a POWER5 to POWER6 upgrade may leave the system in recovery state.
• Fixed a problem where dlpar add/remove processor (proc units or dedicated processor) via GUI on 5250-capable POWER5 servers may fail with "Attempted to allocate an amount of 5250 CPW above the maximum setting."
• Fixed problem that caused backups over 4 Gig in size to fail when saved to USB media.
• Fixed a problem where in some cases TCP packets may be dropped when packet sizes exceed the MSS negotiated at startup. The chhmc command is used to turn tso on/off for a network interface. By default, the initial tso setting for the network interfaces will be set to 'off'. Any changes to the network interface tso setting will persist over system restarts.
  
  chhmc -c network -s modify -i eth0 --tso on
• Fixed a problem when performing an HMC backup to a USB Flash Memory device hangs when the HMC backup file exceeds the maximum file size limit for the device.
• The HMC Manage User Profiles and Access was modified to allow for the User Properties to be modified for users hscroot, hscpe, and root. The Inactivity Values remain unchangeable to prevent accidentally disabling these users.
• Option added to lshwres to display virtual device (V SCSI, VEth, and V FC) topology - end to end mapping of virtual device. Customers can use the new option to have a global view of their virtual devices. This release only supports RPA partitions (AIX, Linux). The partition needs to be updated to latest OS for this support.

Repair and Verify fixes for Power 570, 575 and 595 servers (POWER6-based technology)

• Updated Repair and Verify (R&V) procedures for POWER6 575 servers to include additional instructions for handling of the Node Shipping Bracket node to the frame in the operating position. Instructions added to RV service procedures requiring placing the node into service position.
• Addition of Repair & Verify support for new 5803, 5873, 5802 and 5877 IO units.
• Update to Repair & Verify service procedures for POWER 575 where the Modular Water Unit (MWU) is drained/filled. Updates include the MWU exchange, the Initial System Fill, and the RDHX Fill. The procedures to deactivate and activate MWU were also updated to ensure MWU is in active cooling mode following the service procedure.
• Updated Repair & Verify Procedure for nonconcurrent service of 5791 IO unit. Updates include the removal of the infocenter panels/instructions for unconfiguring IO resources and power off of the enclosure. This information is no longer needed since R&V actively guides the user through the procedure using information from the Hypervisor.
• Fixed incorrectly displaying non-concurrent panels during FSP concurrent maintenance on multi-CEC POWER6 MMA Repair and Verify Exchange. Fixed Repair and Verify problem on POWER6 595 where the HMC may not display the partition detail for Node Evacuation error messages.
• Fixed problem where Hardware maintenance procedures may fail with a HSCS0245 message, "An unrecoverable HMC internal error has occurred.", when executed on a Power 570 or Power 595 server at firmware level 330 or lower. The failing tasks included Add FRU, Exchange FRU and Remove FRU.

Known issues

• A problem has been identified with HMC V7.3.4 where unchecking the “Managed System and Power LIC” check box during the “Change Licensed Internal Code Wizard” will not stop the managed system and power code from getting installed when the user is electing to install only I/O microcode. If the repository selected contains system firmware that can be installed concurrently, the system will remain operational but if the system firmware is a disruptive update, it will be installed disruptively. This problem will be corrected in the release of Service Pack 3 for the V7R3.4.HMC code.
  Workaround: Utilize the operating system capability of installing I/O microcode.
• A problem may occur when creating an lpar, if the user selects dedicated processor mode, clicks Next, then clicks Back and changes to Shared processor mode. This may place the lpar creation wizard in an invalid state that prevents the wizard from completing properly.
  Workaround: Cancel the wizard and restart the create lpar operation.
• Use of a script to remove a physical I/O adapter from a partition and immediately add it to another partition may fail because the HMC may not have a chance to update its repository after the remove and before the add. That causes the HMC to think the physical I/O adapter hasn't been released for add.
  Workaround: It is advised to insert a delay after the remove command in the script.
• B181EF88 may occur after a system dump if there are two HMCs managing the server or a user invalidates a dump
  Workaround: None
• An HTTP 500 error can occur when launching tasks on the HMC if the "tasks pad" is disabled in the user interface settings.
  Workaround: Enable the task pad using HMC Management, Change User Interface Settings.
• The "Edit Sort" button on both "Manage Serviceable Events" and "Reference Code History" does not respond to user interaction.
  Workaround: Click up/down triangle next to each column headers to sort.
• On a HMC set to a non-US English Language/Locale, the Add install resources task does not allow a customer to exit if an incorrect DVD is used. The customer is constantly prompted to put in the correct DVD with no option to cancel.
  Workaround: The user can always terminate the task by going to HMC Management -> Manage Users and Tasks, selecting the running tasks and clicking Terminate.
• Attempting to change the utilization data sampling rate fails with a ClassCastException
  Workaround: None
• When using the Guided Setup Wizard to configure the HMC Firewall, the "Allow Incoming" setting is not adding the correct entry for IPv6 addresses.
  Workaround: Set the firewall setting using "Change Network Settings"
• Whenever a VIOS partition has a FixPack installed on it, any system plan created for that managed system will lack fix information.
  Workaround: None
• The Save Upgrade Data wizard has a panel that incorrectly lists "DVD-RW media". Customers can only use DVD-RAM media.
  Workaround: Use DVD-RAM media
• Unable to create pattern match custom group under Japanese language environment.
  Workaround: Change HMC language to English to create pattern match groups.
• Table columns for viewing virtual adapter slots are not in the correct order after a sort action.
  Workaround: None
• Unable to sort "View Console Events" when HMC language is Japanese
  Workaround: None
• When starting an ssh session, the proper banner text is displayed. However, when listing the banner text with the HMC command line (e.g. lsusrtc -t b), the banner text is blank (e.g. "text=").
  Workaround: None
• When a partition profile name includes one or more spaces, the "Manage Profile" panel may not resize correctly causing it to not display all available profiles.
  Workaround: None
• In an environment with tens of systems and hundreds of LPARs managed by a single HMC, the performance of SAP or Director Management of HMC will downgrade.
  Workaround: None
• Virtual Storage Management - After assigning a physical disk to a Server Virtual SCSI adapter that accepts connections from "Any" partition, the user is unable to reassign the physical disk to a different partition if there is not a client virtual scsi adapter owned by a running lpar attached to the server virtual scsi adapter.
  Workaround: Users should reassign the physical disk to "None" (remove the assignment) manually before attempting to assign the disk directly to another partition.
• The DLPAR GUI Panel for a Active Memory Sharing (AMS) Partition displays values only in multiples of 16 and will increment and decrement values only in multiples of 16 for the I/O Entitled memory (MB) field. It displays an invalid warning message when No changes are made to values in the panel, and "Auto" Mode is already set. If a Multiple of 16 is not entered, then the DLPAR GUI Panel will set the I/O Entitled Memory(MB) value to the nearest multiple of 16.
  Workaround: Use CLI and the chhwres command to specify values that are not multiples of 16.
• Partition migration is not supported between a server managed by a V7.342 HMC and a server managed by a V7.3.4.0 or V7.3.4.1 HMC
  Workaround: None
• Active Memory Sharing. Please note the following limitation when migrating a shared memory partition: When a shared memory partition is migrated to a target system that has a shared memory pool configured with redundant paging VIOS partitions, you must select this option: "Paging VIOS redundancy required"
  Workaround: Ensure that there is an available redundancy-capable paging space device on the target system.
  
  To address this issue, PTF MH01181 will be available for download on May 27, 2009.
• When using the web browser interface of the HMC to monitor the managed system, changes in status of the managed servers and LPARs may hang the User Interface (UI) and all the other connected web UIs as well.
  Workaround: HMC reboot
• User may experience error "HSCL1585 There was an unknown error while querying the object manager database" while performing a Hot Node Repair operation with insufficient available memory.
  Workaround: Perform DLPAR remove memory operation or shutdown some LPARs to release memory and retry the operation.
• When performing concurrent active partition migrations which require taking memory or processor resources from inactive partitions on the destination managed system, a timing issue may occur which will cause one or more migration operation to fail with the following error message: "Partition migration failed with the following errors: HSCLA27B. The partition cannot be migrated because the processor and/or memory resources it requires exceeds the available resources on the destination managed system. If possible, free up resources from running partitions on the destination managed system, and try the operation again."
  Workaround: Perform a manual migration recovery operation, and then try the migration again.
• Copying a user does not copy the password expiration date; password expiration cannot be set for that user.
  Workaround: Reboot HMC
• i5OS lpar create wizard does not allow the user to select tagged IO for full system resources partition.
  Workaround: edit the profile after creation.
• chhmcusr ?t passwd no longer prompts for password when run without ?v
  Workaround: use command with ?v flag to specify new password
• When data replication is configured, Passwords are not replicated automatically if they are changed using the command line interface (chhmcusr)
  Workaround: Sync HMCs manually, or change passwords through the GUI.
• The partition profile menu does not update the available options when a partition is activated or shut down
  Workaround: Reopen the lpar profile list.
• The partition "Active profile" field is not updated on a redundant HMC for a newly created partition that has never been activated from the redundant HMC .
  Workaround: Use primary HMC or activate once from secondary HMC.
• Tagged I/O on HEA does not display which physical port will be used.
  Workaround: Use the information on the profile property LHEA tab to map the logical port to the physical port.
• When performing concurrent active partition migrations and there are not enough free memory or processor resources available (for example, resources that are not owned by inactive partitions) on the destination managed system, some migration operations may fail with the following message:
  Partition migration failed with the following errors: HSCLA27B The partition cannot be migrated because the processor and/or memoryresources it requires exceeds the available resources on the destination managed system. If possible, free up resources from running partitions on the destination managed system, and try the operation again.
  Workaround: You must perform a manual migration recovery operation. If there are enough resources available on the destination managed system to support all the partitions involved in the concurrent migration operations (for example, if resources currently owned by inactive partitions do not need to be used), you will not encounter this problem. If resources owned by inactive partitions do need to be used, however, it is strongly recommended that you perform the migration operations sequentially.
• A shared memory partition that was first activated using single paging VIOS configuration and then activated using redundant paging VIOS configuration may run into a situation that, when the primary paging VIOS is down, the shared memory partition may not function because no paging service is provided.
  Workaround:
  1. Go to Virtual Resources->Shared Memory Pool Management->Paging Space Device(s) panel.
  2. Locate the paging device associated with shared memory partition to be reactivated.
  3. If the paging space device is redundant and shows a Device Status of Active/Inactive (instead of Active/Active or Inactive/Inactive), remove the paging space device (while the shared memory partition is still shutdown) and re-add it to the pool.
• Modifying one of the paging space partition(s) for an active memory sharing memory pool will remove redundant paging devices from the pool.
  Workaround: Re-add the paging devices after the modification is complete.

Command Line

• The following commands have been enhanced to support Active Memory Sharing, which includes support for a shared memory pool and partitions that use shared memory (POWER6 servers only): chsyscfg, lssyscfg, mksyscfg, chhwres, lshwres, chsysstate, lslparutil, lslparmigr, and migrlpar. In addition, a new command, lsmemdev, has been added to list block storage devices that can be used as paging space devices for partitions that use shared memory.
• The chhmc and lshmc commands have been enhanced to support changing and displaying the TCP segmentation offload (TSO) setting for a network interface.
• A new attribute, topology, has been added to the lshwres -r virtualio --rsubtype fc --level lpar -F command to display the end-to-end device mappings for each virtual fibre channel device. The topology attribute has also been added to the lshwres -r virtualio --rsubtype scsi -F command to display the end-to-end device mappings for each virtual SCSI device. The topology attribute is only displayed when the -F option is specified. For virtual SCSI adapters, the existing backing_devices attribute is displayed with the value null when the topology attribute is displayed.
• A new attribute, device_name, has been added to the lshwres -r virtualio --rsubtype eth --level lpar -F command to display the device name of each virtual ethernet device. The device_name attribute is only displayed when the -F option is specified.
• The defsysplanres, lssysplanres, and rmsysplanres commands have been deprecated. You should use the Install Resources wizard on the HMC Web UI for managing install resources.