Contents

The information in this Readme contains fix list and other package information about the Hardware Management Console.

• HMC V7 R3.5.0 Upgrade (Recovery Media) package
• Special upgrade notes for HMC V7 R3.5.0
• Enhancements and changes in HMC V7 R3.5.0
• Security fixes in HMC V7 R3.5.0
• Known issues in HMC V7 R3.5.0
• Web browser requirements
• Installation

HMC V7 R3.5.0 Recovery Media.

This package represents the Recovery image that you can use to upgrade the HMC from HMC V6.1.3, HMC V7.3.1.0, HMC V7.3.2.0, HMC V7.3.3.0 or HMC V7.3.4.0 to HMC V7R3.5.0. You can also use this package to install a clean version of HMC V7R3.5.0. You can also reference this package by APAR MB03046.

NOTE: PTF MH01127 (APAR MB02828) is needed to upgrade from HMC V6R1.3 to HMC V7.3.5.0. PTF MH01135 (APAR MB02661) is needed for one to upgrade from HMC V7R3.1.0 to HMC V7.3.5.0. See the Special upgrade instructions for HMC V7 R3.5.0

Upgrade from HMC 6.1.2 to HMC 7.3.5.0

A fix has been released (PTF MH01127, APAR MB02828) for HMC 6.1.3 that enables you to upgrade directly from HMC 6.1.3 to HMC 7.3.5.0.

Follow these basic steps

1. If your HMC is at Release 6.1.2, first update the HMC from 6.1.2 to 6.1.3 by means of the HMC 6.1.3 Update package (PTF MH01012).
2. Apply the following fixes before applying the fix that enables the upgrade to HMC 7.3.5.0: MH01082, MH01110, and MH01128
3. Next, apply PTF MH01127 to HMC 6.1.3 to enable the ability to upgrade to HMC 7.3.5.0
4. Finally, upgrade to HMC 7.3.5.0 by means of the HMC V7R350 Recovery media.

Upgrade from HMC 7.3.1.0 to HMC 7.3.5.0

A fix has been released (PTF MH01135, APAR MB02661) for HMC 7.3.1.0 that enables you to upgrade directly from HMC 7.3.1.0 to HMC 7.3.5.0.

Follow these basic steps

1. Apply PTF MH01135 to HMC 7.3.1.0 to enable the ability to upgrade to HMC 7.3.5.0
2. Upgrade to HMC 7.3.5.0 by means of the HMC V7R350 Recovery media.

Upgrade from HMC 7.3.2.0 to HMC 7.3.5.0

Upgrade from HMC 7.3.3.0 to HMC 7.3.5.0

Follow these basic steps

• Upgrade to HMC 7.3.5.0 by means of the HMC V7R350 Recovery media.
  OR
  
• Upgrade to HMC 7.3.5.0 by means of the HMC V7R350 Update package.

This package provides the following enhancements and fixes:

General fixes

• Redundant devices were being removed incorrectly when modifying the paging VIOS for the AMS pool and seeking to modify paging devices at the same time.
• Fixed a problem where the user will not be able to perform the HOT Node repair, if one of the IBM i partitions owns the IOP from the target node and that LPAR is in Error state.
• Fixed a problem where some of the message detail was lost on a Host Ethernet Adapter dynamic logical partitioning error.
• Fixed a problem where activation of an AMS partition may fail with error HSCLA423 VRM_STREAM_FAIL under certain conditions.
• Fixed a problem where activating an IBM i Active Memory Sharing (AMS) partition may set the partition max memory value to the wrong size.
• Creating a system plan may fail with inventory gathering error when run against POWER6 servers.
.
• Fixed a problem when a partition profile name includes one or more spaces and the "Manage Profile" panel did not resize correctly causing it to not display all available profiles.
• Fixed a problem where HMC dump files were not able to be deleted as a result of the file name being over 40 characters in length.
• Fixed a problem where E3321032 00000002 may be reported. This particular problem only happens in a dual HMC environment.
• Fixed a problem where dynamically changing the shared processor pool of a partition incorrectly sets the uncapped weight of the partition to 0 (zero).
• Fixed the error message text returned when there is an attempt to create a system plan on a POWER6 server that has a non-functional Host Ethernet Adapter (HEA) to identify the failing HEA.
• Fixed a problem where Create System Plan incorrectly reported a failure after running 2 hours, even though it completes successfully
.
• Fixed a problem where activating a VIOS profile with Ethernet adapters having the same trunk priority but different virtual switches fails with HSCL058A and HSCL0292.
• Fixed a problem where an invalid hostname with a numeric first character generated the following message "Error Rendering Task Panel" when attempting to use the "Change Network Settings" panel
• Enhanced performance of the CIM layer by eliminating unnecessary logging.
• Fixed a problem where E332FFFF was logged in large configurations with close to the maximum supported partitions and/or systems.
• Fixed a problem where E212E136 was logged due to the vr directories filling up and contributed to an HMC hang and/or large backup CCD exceeding the 4.7Gb DVD-RAM size limit
• Fixed a problem when a user perform inactive partition migration twice, the uncapped weight value of the last activated profile will be set to 0 even though the original value of that attribute is not 0 (zero).
• Fixed a problem where E212E161 was logged when sockets were exhausted in the framework communication services layer due to misconfigured network settings for call home.
• Fixed a problem where activating a partition results in HSCL11C6 A shared memory pool is not configured.
• Fixed a problem when deploying a sysplan that includes a VIOS at level 2.x. That VIOS level, cannot be selected because it is not available on the list shown by the pulldown menu.

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

• Fixed a problem with concurrent node DCA repair action taking a non-concurrent path due to a failure to initiate power save mode when the repair should be concurrent.
• Fixed an issue where CEC concurrent maintenance (CCM) takes a non-concurrent path in a multi-CEC configuration when the reporting system is the frame.

Command line changes

• The chsyscfg, chsysstate, and lssyscfg commands have been enhanced to support new managed system power on options. The power_on_lpar_start_policy attribute replaces the power_on_option attribute, which has been deprecated, on the chsyscfg and lssyscfg commands.
• The lshwres -r proc --level sys command has been enhanced to display the maximum number of dedicated or virtual processors that can ever be assigned to a partition. There is a value displayed for each partition type (AIX/Linux, VIOS, and IBM i).
• A new option has been added to the chhwres command to clear a physical I/O bus. This command can be used to clear stale bus information after an I/O drawer has been concurrently removed but not replaced.
• A new option has been added to the migrlpar command to allow the user to specify if an error or only a warning should occur if the HMC is not able to maintain an equivalent multipath I/O configuration of the mobile partition's virtual SCSI and virtual fibre channel adapters on the destination managed system.
• A new option has been added to the migrlpar command to allow the user to specify if an error or only a warning should occur if the HMC is not able to configure each of the mobile partition's virtual ethernet adapters so that it is bridged on the same VLAN to an external network on the destination managed system.
• The virtual_fc_mappings and virtual_scsi_mappings attributes on the migrlpar command have been extended to include the VIOS virtual slot number. The suggested_virtual_fc_mappings and suggested_virtual_scsi_mappings attributes on the lslparmigr -r virtualio command have also been extended to include the VIOS virtual slot number. In addition, the lslparmigr -r virtualio command has been enhanced to display all available VIOS virtual slot numbers.
• New attributes have been added to the migrlpar command to allow the user to specify which IP addresses on the Mover Service Partitions to use for a partition migration operation. In addition, the lslparmigr -r msp command has been enhanced to show all possible IP addresses which can be used on the Mover Service Partitions.
• The topology attribute displayed by the lshwres -r virtualio --rsubtype fc --level lpar -F and lshwres -r virtualio --rsubtype scsi -F commands now displays the end-to-end device mappings for each virtual fibre channel device and each virtual SCSI device for IBM i partitions.
• The backing_devices attribute is no longer displayed when the lshwres -r virtualio --rsubtype scsi command is issued without the -F option. The backing_devices attribute is now only displayed when the -F option is specified. This is due to the fact that to get the backing device information, the HMC has to issue a command via RMC to each VIOS partition that has at least one SCSI adapter. Sometimes, the command takes a long time to return. Now, users of the lshwres -r virtualio --rsubtype scsi command do not have to wait for the backing device information to be retrieved. Note that it is no longer recommended that users use the backing_devices attribute. The topology attribute returned by the lshwres -r virtualio --rsubtype scsi -F command should be used instead. The topology attribute is intended to replace the backing_devices attribute.
• A new attribute, capabilities, has been added to the lssyscfg -r sys -F command to display all of the capabilities of the managed system. A capability will only appear in the output if it is true. The capabilities attribute is intended to replace all of the individual capabilities attributes (*_capable attributes) displayed by the lssyscfg -r sys command. The capabilities attribute is only displayed when the -F option is specified.
• The chsyscfg and lssyscfg commands have been enhanced to support changing and displaying Host Channel Adapter (HCA) bandwidth capability settings.
• The chhmc and lshmc commands have been enhanced to support changing and displaying the jumbo frame setting for a network interface.
• The getdump, lsdump, and startdump commands have been enhanced to support resource dumps (POWER6 servers only).
• The chhmcencr, lshmcencr, and lshmcusr commands have been enhanced to allow users to specify which encryption to use for the passwords of locally authenticated HMC users.
• The chhmcldap and lshmcldap commands have been enhanced to provide full LDAP support on the HMC.
• The chhmcusr, lshmcusr, and mkhmcusr commands have been enhanced to support all of the HMC user properties that are supported on the HMC Web UI.
• A new option has been added to the rmhmcusr command to allow the user to remove all HMC users that have a specified authentication type.

New function or enhancement

• Updated screen for Power On and Server IPL Parameters.
• Performance enhancements have been made for external management servers such as IBM Director.
• Two kinds of new indications are supported by CIM provider. One is monitor device driver changes, and the other is to monitor memory and processor changes for dynamic logical partitioning and profile changes.

The following security fixes are included in the HMC V7.3.5.0 release.

• Call Home message of SRC E35510A2 can be ignored if it is generated when the HMC at 7.3.5 levels is rebooted. A fix is planned for a later HMC 7.3.5 service pack.
• Call-Home Incompatibility. J Older versions of the HMC will no longer be able to act as a call-home server for HMC V7.350.0 and higher. When HMC V7.350.0 calls home though an incompatible HMC version, SRC E35F0012 will be generated on the call-home server. This compatibility issue is the result of changes in the call-home format, which is not supported by earlier HMCs. HMC V7.350.0 and higher can act as a call-home server and they can accept call-home requests from earlier versions. To prevent call-home issues, use HMC V7.350.0 as the call-home server or select only HMC V7.350.0 and higher in Manage Outbound Connectivity screen.
  
  If a user doesn't specify a Backup Filename in the "Set up a Scheduled Operation"panel from the "Systems Management->Servers->Schedule Operations->Backup Profile Data" panel, then the default filename will be profData_bak.YYYYMMDDHHmmSS where YYYY=year, MM=month, DD=day, HH=hour, mm=minute and SS=seconds.
  Workaround: Specify a Backup filename on the last tab of the dialog.
• When a partition is activated with virtual adapters one time, then changed and activated as an "all resource" partition, the partition may report errors against the virtual adapters that were assigned earlier.
  Workaround: Activate the LPAR with normal profile but with no virtual adapter in it and then activate the LPAR with full system profile.
• Chinese text (zh_SG and zh_HK) not translated.
  Workaround: If you desire zh_SG, use zh_TW. If you desire zh_HK, use zh_CN.
• Japanese text in quick filter window for system plans turns into English.
  Workaround: none
• Any tasklet using a table with numbers is missing the "second" number box when doing the "between" filter.
  Workaround: none
• HMC web UI appears to hang when doing repeated rapid operations such as delete a number of partitions.
  Workaround: Force a refresh by navigating to a different task and then back.
• Scheduled operations sometimes fail to execute, and no indication is made that the operation was attempted nor that the operation failed or succeeded.
  Workaround: none
• One cannot make a custom group of custom groups.
  Workaround: Use a single custom group of all the elements you wish to group.
• E212E161 00000385 call home: Problem can occur in an environment where network settings are configured incorrectly. The problem can occur in either of these specific instances:
  • Invalid login information for a call-home proxy is provided.
  • Call-home is configured to to "dial" through another HMC, but fails to allow FCS (framework communication services - the hmc-to-hmc communication protocol, running on 9920/TCP 9900/UDP) through the firewall.
  Workaround: Reboot the HMC
• Misleading shortcut keys are displayed in Japanese window or pull-down menus. Shortcut keys are displayed in various window or pull-down menus when the HMC is configured for a locale that uses double-byte languages, but these short cut keys do not work. There are two problems, related to same root cause:
  • HMCs configured for double-byte locales: In some action menus English shortcut key identifiers appear before localized menu item text.
  • HMCs configured for single-byte locales: Shortcut key identifiers do not appear for "submenus," even though they are specified in the properties file.
  Workaround: None
• After removing the last task in the Modify Task Role "Current Tasks" object window, the rest of the tree is not removed.
  Workaround: You can manually delete the rest of the tree.
• On the view utilization data panel, if one puts in an invalid time or date, then one gets the proper error message. Then when one clicks on "Close Message", only the invalid date will close the message and return to same panel. The invalid time message will not close and one has to cancel to get back to that panel.
  Workaround: Cancel out and go back to view events panel.
• E212E161 00000388 too many open files in process. The l2tpd daemon has a file descriptor leak. This leak is exposed if IBM i partitions are configured to use the HMC modem for call home.
  Workaround: Configure IBM i partitions to use another call home method.
• E355104C 0000B5CE. This error code is due to a web UI refresh timing issue that rarely occurs. The table updates the model (usually when removing the selected server) while the tasks pad is also being updated.
  Workaround: Deselect the server in the work area table before removing the server connection.
• "Error Rendering Task Panel" error occurs in the update wizard during an update of the HMC level. This problem is very rare and only triggered when a user switches repeatedly between media and FTP installs.
  Workaround: Cancel UI Update and retry operation, selecting desired media
• The HMC may run out of memory when displaying a very large amount of utilization data either using the Web UI, or using the HMC command line (the lslparutil command). When the HMC runs out of memory, the user will get an error on the Web UI indicating that connectivity to the console has been lost, and the lslparutil command will fail with the error "HSCL8016 An unknown error occurred while trying to perform this command. Retry the command. If the error persists, contact your software support representative." The HMC may need to be restarted.
  Workaround: On the Web GUI, specify a short time interval between the beginning time/date and ending time/date for the events, and/or specify a small maximum number of events. To work around this problem on the lslparutil command, use the -h option to specify a small number of hours, and/or use the -n option to specify a small number of events.
• If the Mover Service Partitions (MSPs) used in partition mobility operations have multiple IP addresses, then you may experience a significant performance degradation for partition migration operations, partition migration validation operations, and the lslparmigr -r msp command. This is due to the additional connectivity tests performed between each pair of source and destination MSP IP addresses.
  Workaround: turn off the MSP attribute on any VIOS partition that is not needed as an MSP for partition mobility operations.
• The rsthwres -r io -p <partition name> and rsthwres -r io --id <partition ID> commands return a good return code, however no action is performed.
  Workaround: issue the rsthwres -r io command without specifying a partition name or ID, and all recoverable I/O adapters on the managed system will be restored.
• The rsthwres -r hea command returns a good return code, however no action is performed.
  Workaround: Issue the chhwres command to remove each HEA logical port to be restored.
• The rmlparutil command does not remove utilization data snapshot files from the HMC.
  Workaround: None
• Error message "VIOSERR2" is displayed after removing a partition, then starting Virtual Storage Management - If customers configure a VIOS with server virtual SCSI adapters (SVSA) that point to a client partition then delete the client partition without removing or remapping the SVSA 's, they will encounter an error when they attempt to open Virtual Storage Management.
  Workaround: Customers should remove the SVSA's that point to the deleted client LPAR and open Virtual Storage Management again.
• The name "i5/OS" is used instead of "IBM i" in the Problem Reporter code. This only affects data that is headed to the notelog and to the backend and will not be displayed on the HMC.
  Workaround: None
• In the "Service Management - Manage Remote Connections" function of the HMC, some descriptions of remote connections are translated and some are not. The problem arises on the remote HMC, which creates the support request using its locale. The locale of the user context on that machine determines the locale for the short description. The support request is then sent over the web to the local HMC. The local HMC may have a different locale. There is no way to coerce the locale of the messages from the remote machine in the code at present. Also, if the remote machine does not have a logged-in user, the user context is null and the default context will be used to determine the locale of the remote short description. Often, this is English.
  Workaround: None
• When SSH is used to access an HMC that uses Japanese, the banner file is not displayed in the correct language. The Japanese locale in the HMC is EUC-JP only and the banner function only supports Unicode.
  Workaround: None
• "Interaction with History" is displayed in the language which is the same as the language when HMC booted last time. Here is the message that is being displayed in Japanese, shown in English here for convenience:
  "Interaction with History Page
  You have apparently interacted with a page from the browser history. This is not supported. Your browser will now be redirected to the the current panel for the running task."
  Workaround: None
• Prefecture text area is not selectable under Japanese environment. The error message reads: 'Translation error i--"State or province" shouldn't use for "Prefecture" in Japanese'. Only the United States and Canada have a drop-down list for the state or province. This field is not selectable for all other countries, that is, it is disabled.
  Workaround: None
• Wrong name on title when create a system plan. The limitation in the HMC is that since the system plan code has allowed the single selection, the code takes the selected object to be the target of the task and uses the name of that target on the progress panel. There is no functional break here, just a display issue, and that only occurs when a single system plan is selected.
  Workaround: None
• There is no "Description" column on for the system plan table even when one configures columns for the Custom Groups table
  Workaround: None
• Filter displayed in wrong language at table of "System Management ? Servers" on Japanese HMC. The problem arises when the locale used to define the filters is different than the locale used to view the filters. This may occur when accessing an HMC defined in a locale different than the locale of the remote web browser.
  Workaround: None
• In the HMC, when viewing the Processor tab of a Managed Profile, the read-only text labels are shown in the correct locale, but the dynamic values and the values in the text input fields are not.
  Workaround: None
• In the HMC, when using the "View HCM Events" dialog, and using a selected different time range, the code displays up to 1000 events after the initial time, instead of up to 1000 events before the desired time.
  Workaround: None
• In the HMC UI, the width of column changes after execute filter function. In general, the resizing behavior of the HMC UI is limited in its flexibility.
  Workaround: None
• In the HMC User Profile, the user properties are modified individually, not all or none. The code has been designed to minimize validation problems as much as possible, but some errors are detected only after the dialog is confirmed, making it difficult to roll back the changes if the user cancels the dialog.
  Workaround: None
• If one cancels a User Profile dialog in which some changes have been made, go back to the dialog to confirm that the changed properties are correct.
  Workaround: None
• The copy user function does not copy the user ID nor the user description. These two fields must be entered for each user.
  Workaround: None
• When using the command 'chhmcencr' to remove one or more cipher algortithms, after reboot the HMC will still show all the cipher algorithms enabled for the HMC on web server port 443.
  Workaround: None
• SSL is based on negotiation of cyphers between the two endpoints. Customers who wish to block a particular algorithm may restrict the set of cipher algorithms at the other endpoint of the SSL link.
  Workaround: None
• The user password expiration day for newly created users and saved/restored users are not displayed properly on the UI. The CLI lshmcusr can be used to check users' password expiration date.
  Workaround: None

The following browsers have been tested.

Microsoft Internet Explorer 6.0 or Internet Explorer 7.0 is supported.

• Internet Explorer 7.0 is required if you are using IPv6 addresses on the network interface over which the browser is connected.
• Internet Explorer 8.0 is not supported; however, the Internet Explorer compatibility mode seems to alleviate most problems.
• If your browser is configured to use an internet proxy, then local internet addresses should be included in the exception list, consult your network administrator for more information. If you still need to use the proxy to get to the Hardware Management Console, enable Use HTTP 1.1 through proxy connections under the Advanced tab in your Internet Options window.

Firefox 1.5.0.7

Note: For Firefox 2.0 make sure the JavaScript options to raise or lower windows and move or resize existing windows are enabled. To enable these options, go to the Content tab in the browser's Options dialog, click Advanced? next to the Enable JavaScript option, then select Raise or lower windows option (a check mark appears) and Move or resize existing windows options (a check mark appears). These features allow you to switch easily between HMC tasks.

Other web browser considerations

Session cookies need to be enabled in order for ASMI to work when connected to HMC remotely. The asm proxy code saves session information and uses it.

Internet Explorer

1. Select Tools -> Internet Options
2. Select Privacy tab and select 'Advanced'
3. Check if 'Always allow session cookies'
4. If not checked, check 'Override automatic cookie handling' and check 'Always allow session cookies'
5. The handle First-party Cookies and Third-party Cookies, you can choose either block or prompt or accept. Pprompt is preferred, in which case you will be prompted every time a site tries to write cookies. It may be a little annoying, but it is the safe thing to do. Some sites need to be allowed to write cookies.

Firefox

1. Tools -> Options
2. Select Cookies Tab
3. Select check box Allow sites to set cookies.
4. If you want to allow only specific sites then select 'Exceptions' and then you can just add this HMC to allow.

Installation instructions for HMC V7 R.3.5.0 updates and fixes can be found here:

Update corrective services instructions for HMC Version 7

Recovery media upgrade and installation instructions for HMC V7 R.3.5.0 can be found here:

Recovery media upgrade and installation