For use with Version 9 Release 2 Maintenance 950 (V9 R2 M950)

Updated: 04 March 2021
(C) Copyright International Business Machines Corp., 2021 All rights reserved.

Contents

These instructions contain the following information specific to the HMC V9 R2 M950 code level

• Contents
• Terminology
• HMC V9 R2 M950 Recovery Media - for 7063 Machine Type
• Upgrade notes
• Enhancements and new function
• General and Security fixes
• Command line changes
• Performance and Scale
• Known issues and limitations
  
• Best Practices
• Web browser requirements
• Clearing the Browser Cache
• Other Web Browser Considerations
• BMC Browser Requirements
• Installing HMC in the OpenPower Environment, 7063 Machine type
• Virtual HMC Installation for x86 hypervisors
• Virtual HMC for PowerVM LPARS
• General Virtual HMC notes
• National Language Support (Supported languages)
• Copyright and Trademark Information

• The minimum HMC levels required for the upgrade to HMC V9 R2 M950 are:
  
  • HMC V9 R1
  • Mandatory fix for vHMC on x86: MH01857
    
  • Mandatory fix for vHMC on PowerVM and 7063 Machine type: MH01858
  • The mandatory fix requirement:
    
  • Ensures that once the HMC is upgraded to the 950 version, all CECs and BPCs connect to the HMC. Without the fix, all  CECs and BPCs will go to "Failed Authentication" state.
  • Includes fix for PCM database incompatibility issues while upgrading to 950.
  • User MUST run Save Upgrade Data after installing mandatory fix and prior to upgrading. Refer to https://www.ibm.com/support/pages/node/6409986
    
    
• HMC V9 R M950 supports POWER7 servers with the following minimum  firmware levels:
  
  • FW730_146 (730.A0)
  • FW740_156 (740.90)
  • FW760_087 (760.50)
  • FW770_098 (770.50)
  • FW780_068 (780.21)
  
  
• HMC V9 R M950 continues to provide support for POWER8 and POWER9 processor-based systems.
  
  
• Management of POWER6 and older servers is not supported by HMC V9. Attempts to manage POWER6 and older servers will result in a connection state of Version Mismatch and connection_error_code “Connection not allowed 0009-0008-00000000”.
  
  
• Transport protocol TLSv1.1 support is removed from all HMC communication channels
• Verify that all remote browsers, scripts and programs that interface with HMC REST APIs or web browser (including  HA programs, AIX Live Kernel Update, etc) are enabled for TLS 1.2.
  
  
• HMC V9 R2 M950 supports the following HMC models:
  
  • x86 : KVM, XEN, ESXi (6.0 / 6.5)
  • ppc64le : 7063-CR1, vHMC on PowerVM (Power 8 & Power 9 systems)
    
  
  
• HMC V9 R2 M950 does not support 7042 models.

• If two HMCs manage the same server, both HMCs must be at the same version,  release and maintenance level.
  
  
• It is recommended to have all connected VIOS servers on the latest level i.e., Virtual I/O Server (VIOS) level 3.1.x. To have optimal performance from the GUI, ensure all Virtual I/O Servers are at level 3.1.1.x level, or later.
  
  
• All running partitions (AIX, Linux, and VIOS) should be at Reliable Scalable Cluster Technology (RSCT) version 3.2.0.0, or later. It is recommended to have AIX 7.1 TL4 or higher and AIX7.2 TL0 or higher; these levels contain RSCT 3.2.1.0.
  
  
• V9 R2 does not support HMC user profile names that are all numeric  (i.e. user "123456").   Any existing hmc user profiles with numeric ids are deleted on upgrade.   This issue is under investigation.
  
  
• HMC V9 network installation images and installation instructions can be found here:  https://www.ibm.com/support/pages/node/687979
  
  
  

• Limit new managed resource role and task role names to 47 characters.
  
• Added additional validation for the -h parameter on the chhmc -c syslog -s add command to return an error if a comma-separated list of host names is specified.
  
• Fixed an issue where a UI DLPAR memory operation times out at 30 minutes even when the command timeout is greater than 30 and the command is still running.   The UI task reports message REST0255 An internal error occurred.
  
• Fixed an issue that ignored the -w option of the chhwres memory remove command for IBM i partitions.
  
• Fixed an issue on Memory tab of IBM i partitions where the Options panel is not enabled.
  
• Enabled hover text to display the complete name for VIOS names longer than 13 characters in the VIOS And Adapters tab of the Add Virtual Network wizard.
  
• Fix to prevent call home of SRC E2FF1801.
  
• Fix to prevent E212E13A against /hmcdump from reporting when getupgfiles is executed due to increased disk1.img size
  
• Fixed an issue that caused Save Current Configuration to generate a saved profile named profile_0 where the number incremented each time rather than the specified profile name.  This issue only impacts partitions that have Remote Restart enabled but have a configuration error that causes Local Storage Update Failed.
  
• Fixed an issue with profiles of migration incapable partitions where the Specify MAC Address Restrictions option is neither reflected nor able to be set while editing SR-IOV logical ports.
  
• Fixed a CEC periodically going to Incomplete state with NovaLink co-management
  
• Fixed the  recovery behavior for CPU DLPAR remove errors to rollback to the original values for virtual processors.
  
• Fix to long running pedbg execution collecting the sosreport.
  
• Fixed a UI issue where renaming a vSwitch  did not take effect when there was only a change to the case (upper case to lower or vice-versa) of the characters
  
• Fixed an issue  with installios not using the user specified image name while installing VIOS from NFS server when multiple files existed in the directory.
  
• Fixed an issue where additional vlan remains in the last valid configuration of VIOS even though a virtual network that is an additional vlan on existing trunk adapter is deleted.  This issue also impacts the last
  
• activated profile when automatically save configuration to profile is enabled.
  
• Fix to prevent call home of SRCs E23D040A and E2FF4500.
  
• Fixed gaps in the PCM data not showing up in the dashboard when run for a few days in a scaleup environment and improved response time  of the Aggregated metrics API
  
• Fix to prevent the secondary HMC to take over as primary during daily 'primary arbitration' scheduled task if a repair operation is running.
  
• Fixed an intermittent issue where the  partition storage diagram cannot be viewed due to some stale entry in the Storage Information
  
• Fixed an issue with the REST call "/rest/api/uom/ManagedSystem?hwinventory=true" that does not return the results of the PowerSupplies and FANs XML on the first attempt.
  
• Fixed a NullPointerException while creating a sysplan resulting in error Exception Occurred in com.ibm.inventory.VSPPComponent:setPartitionProfiles
  
• Fix to prevent power off of the managed system when a system dump is currently being performed so the dump won’t be invalidated.
  
• Fixed an issue where BMC events would get re-reported by the HMC.
  
• Fix to the actions menu options based on selection in VNIC Tab of Edit Partition Template.
  
• Fix to display an error message for VIOS RMC state busy while viewing the license during Deploy System Template.
  
• Fix to block the user from modifying the backing device type while the partition template is opened in view mode.
  
• Fix for a failed partition migration operation without error message resulting in LPM in the other direction error of HSCL8012 The partition named was not found
  
• Fixed an issue that caused a Power enterprise pool to show an incorrect number of unreturned processors or unreturned memory for the pool after a server in the pool that was unreachable from the master HMC came back up.  This issue only occurs if the server had unreturned resources before it came back up, and those unreturned resources were able to be automatically reclaimed when the server came back up.
  
• Fix to discover a newly added interface in vhmc on PowerVM without requiring a reboot.
  
• Large Send attribute issue in network bridges using Deploy System Template has been fixed.
  
• Persistence issue of link aggregation configuration in network configuration tab of Deploy System Template has been fixed.
  
• Fix for accurate values  of a system's stealable proc/memory resources.
  
• Fix lock timeout problem for cross HMC concurrent bi-directional partition migration operations.
  
• Fixed a problem of failed migration operations hitting the error message "HSCLA319 The partition's virtual Fibre Channel client adapter xx cannot be hosted by the existing Virtual I/O Servers (VIOSs) on the destination managed system.  Set up the necessary VIOS host on the destination managed system, then try this operation again."
  
• User is not allowed to performing Deploy System Template on a CEC with "Power off on last LPAR Shutdown" setting enabled.
  
• Fix for handling code update when one FSP times out during a firmware update operation while the other FSP completes. 
  
• Fixed an issue where the last character of a 9 character SRN event was stripped off.
  
• Fixed an issue while trying to view the Manage Inbound Connectivity panel when a system was connected using IPV6.
• Fix to prevent network traffic from the HMC attempting to contact IBM servers even when the HMC is not set as a callhome server.
  
  

Security fixes

• Addressed tomcat vulnerability: CVE-2020-9484
  
• Addressed a problem in the authentication flows to avoid thread hung/idle state during access server calls.

Command line changes

• The following new commands have been added to support partition virtual serial numbers: lsvsn, rmvsn, xfervsn.
  
• The following existing commands have been enhanced to support partition virtual serial numbers: chsyscfg, lscod, lssyscfg, mksyscfg.
  
• The following commands have been enhanced to support partition keystores:  chsyscfg, lssyscfg, mksyscfg, rstprofdata.
  
• A new option has been added to the lslparmigr command to display memory requirements for a partition to be migrated as well as available memory on a destination managed system.
  
• Enhanced the lslparmigr -r msp and lslparmigr –r lpar commands to display MSP thread information.
  
• The chsysstate command to power off a managed system has been changed to fail if the managed system is currently offloading a dump to the HMC.  To override this new behavior and power off the managed system anyway, specify the --force option on the command.
  
• New options have been added to the chled and lsled commands to support identify LEDs.
  
• New options have been added to the chled command to run an LED lamp test.
  
• The chhmc command no longer allows the SSH protocol version to be set to 1, Kerberos weak encryption to be set to on, or the legacy anonymous cipher mode for the HMC chatlet interface (port 9920) to be enabled.
  
• A new option has been added to the chsvc command to automatically add the required certificates to the HMC's truststore when the cloud connector encounters a non-trusted certificate chain when connecting to cloud endpoints.
  
• The chhmccert command has been enhanced to support adding and removing certificates to and from the HMC's truststore, and the lshmccert command has been enhanced to list the trusted certificates added to the HMC's truststore.
  
• The chhmcfs and lshmcfs commands have been enhanced to support the /data file system and to support the /var/hsc/log file system as the separate file system that it now is (/var/hsc/log is no longer in the /var file system).  Prior to this enhancement, these commands included the /var/hsc/log file system with the /var file system. 
  
• New options have been added to the chsacfg and lssacfg commands to configure and display connection monitoring settings.
  
• New options have been added to the chsvcevent command to add a comment to a serviceable event when closing it.
  
• The lssvceventfile command has been added to list the contents of a file associated with a serviceable event.
  
  
  

• Navigating the enhanced UI when managing systems and partitions can result in message:
  Proxy Error
  The proxy server received an invalid response from an upstream server.
  The proxy server could not handle the request GET /ui/sfp/.     
  Reason: Error reading from remote server
  An upcoming fix is planned to be delivered soon.  In the interim, refer to the proxy error techdoc for more information.
  
• Any user installed certificate must have key size of 2048 bit or higher. If the certificate of key size 1024 bits was installed the HMC web server and 5250 console will fail to start after the upgrade.
  
• HMC V9 R2 does not support HMC user profile names that are all numeric (i.e. user "123456").  Any existing HMC user profiles with numeric ids are deleted on upgrade.
• The System Virtual NICs task shows no vnics configured even when vnics exist
  
• When user runs getupgfiles, an SRC E212E13A might be  observed due to increased size of the installation image files
  
• Occasional Out of Memory Exception during Virtual I/O Server Installation when user logs out or the session expires
  
• Proxy time out issue when multiple mappings are created or deleted as part of single POST operation for the same RR partition
  
• Accessibility: In some panels, navigating to different input fields by TAB key doesn't work
  
• "Internal error" message is observed, on removing stand alone vSCSI client adapter from inactive AIX partition
  
• During LPM, the state of the LPAR is shown as "Migrating - Running" or "Migrating - Not Activated", whereas PCM shows the state as just "Running" or "Not Activated".
  
• Customer may observe multiple heapdumps and javacores if there is any OutOfMemory for a service. It can result into triggering of SRC "E332FFFF" multiple times.
  
• Occasionally, during lpar activation, command execution becomes very slow affecting HMC performance
  
• A failure occurs during a pool (CPU/Memory) update operation that prevents adding the resource to the server
  
• REST API error messages are not translated to the language of the browser or HMC.
  
• Error launching task message is thrown in Partition virtual storage topology diagram page
  
• Occasionally, user shall see that System level Hardware Virtualized IO tab takes more time to launch
  
• In SRIOV Logical port advanced settings, Promiscuous Mode field is missing in add VIOS wizard
  
• During partition deployment, user will not see the configured client slot id for SCSI adapters when Remote Slot ID is configured to next available ID.
  
• Launching UI task might show errors for a server (General Settings, Processor, Memory,I/O, Virtual Networks, Virtual NICs) in some specific scenarios with running Linux Partitions post LPM operation. Remove the system connection and add it back as a workaround.
  
• The validation fails for IPV6 address in the NetBoot wizard, when the user enters the prefix along with IPV6 address. The user can provide an IPV6 address in Unicast Addressing excluding the prefix as a workaround.
  
• Automatic installation of HMC using Activation Engine fails to process the configuration file. If the installation using activation engine presents you with a license to accept, then there is a possibility that the Activation Engine ISO image deployed cannot be mounted. Try to install again by rebooting the partition.
  
• Remote Restart operation fails when tried with REST Job.  Circumvention is to use HMC CLI for Remote restart.  This issue will impact older PowerVC versions which do not support V9R2 HMC.  
  
• On older POWER7 server models that have HEA but do not support SRIOV, the user is unable to launch the Hardware Virtualized IO tab for a partition.   Circumvention:  Use CLI or profile edits to view or modify the partition HEA configuration.   
  
• HMC is not accessible remotely after user deletes one of the network bonds. User can select one interface name for the default gateway instead of selecting 'any' .  Circumvention is to use local console to when deleting a bond interface that is used for remote access.
  
• Throttling may not be started/adjusted/stopped on a server in an EP 2.0 pool when that server is managed by 2 HMCs and the state of the server on one of the HMCs is not in Standby or Operating state. The workaround is to fix the bad server state on the HMC.

Google Chrome

Safari

Microsoft Internet Explorer

Mozilla Firefox

Safari

Clearing the Browser Cache

1. Open the Browser
2. Select Tools
3. Select Internet Options
4. Under Browser History, select Delete
5. Check "Temporary Internet files and websites files" and "Cookies and website data"
6. Click on Delete

In Mozilla Firefox

In Google Chrome

Other web browser considerations

Internet Explorer

1. Click Tools > Internet Options.
2. Click the Privacy tab and select Advanced.
3. Determine whether Always allow session cookies is checked.
4. If not checked, select Override automatic cookie handling and Always allow session cookies.
5. For the First-party Cookies and Third-party Cookies, choose block, prompt, or accept. Prompt is preferred, in which case you are prompted every time a site tries to write cookies. Some sites need to be allowed to write cookies.

Mozilla Firefox

1. Click Tools > Options.
2. Click the Cookies Tab.
3. Select Allow sites to set cookies.
   If you want to allow only specific sites, select Exceptions, and add the HMC to allow access.

BMC Browser requirements

• Google Chrome Version 60
• Firefox Version 52.1.2
• Internet Explorer Version 11
• Safari Version 10.1.2