Superdome 2 Release notes 2018.06 (4.2.54)

TITLE: HPE Integrity Superdome 2 Server, I/O and HDD Firmware Bundle - for Smart Update Manager

VERSION: 
Smart Update Firmware Bundle
Bundle Version: 2018.06
Complex and nPartition Firmware Version: 4.2.54

DESCRIPTION:
This bundle contains the firmware for HPE Integrity Superdome 2 Servers with Smart Update Manager (SUM) application - which includes Complex and nPartition firmware as well as Superdome 2 I/O and HDD firmware.

WARNING:  HPE Integrity Superdome 2 Global Instant Capacity (GiCAP) users must upgrade GiCAP software to v10.11.00.01 or above to manage OA members with firmware version 4.2.54 or above. Refer to “GiCAP does not control SD2 members” in known problems section of HP Instant Capacity (iCAP) Support release notes for more details.

 

WARNING: To ensure that the OA GUI continues to work after December 31, 2016, after upgrading to version 4.1.34 or later, the OA SHA1 self-signed certificate will be removed and replaced with SHA256 self-signed certificate.

• To prevent security warnings, user is encouraged to re-generate the self-signed certificate with the common name (CN) matching exactly the OA hostname as known by the web browser. See the Certificate Administration section in the OA user guide and this advisory c05315790 for more information.
• Also GiCAP users must update the OA certificates stored on the Active Group Manager (AGM) and the Standby Group Manager (SGM) so they can continue to borrow or loan core usage rights in their GiCAP group or perform TiCAP prefetch. This is because AGM and SGM will keep the stale OA certificate for the GiCAP member. More information is available in this advisory c05242542.

 

WARNING:

• Due to the use of a stronger signing key starting with version 4.1.34, you need to contact HPE support in order to downgrade to a version prior to 4.1.34.

I/O smart components versions:
Firmware Flash for Add-in Smart Array Adapters version 2013.09.25
    P411: ROM 6.42
    P812: ROM 6.42

Firmware Flash for Qlogic Fibre Channel Host Bus Adapters version 2018.06.15:
        1-port/2-port 8 Gb FC AH400A, AH401A: ROM 5.06.06, EFI 2.27 (i2 and i4 blades only)
        1-port/2-port 16 Gb FC SN1000Q: ROM 8.05.43, EFI 6.39

Firmware Flash for InfiniBand Adapters version 2012.11.30:
   HPE 4X QDR Infiniband (AT083A): 2.9.1000

Firmware Flash for Emulex Adapters version 2018.05:
     1-port/2-port 4 Gb FC + 1000BT AD221A, AD222A, AD393A: FC ROM:  2.82x4, FC EFI:  4.01a5; LAN 4.3.02 (i2 and i4 blades only)
     1-port/2-port 8 Gb FC AH402A, AH403A: ROM:  2.02x2, EFI: 4.01a5  (i2 and i4 blades only)
     2-port CN1000E AT111A: LAN & FC ROM:  4.9.416.17, LAN EFI:  49.00A12, FC EFI: 5.21.A13
     8 Gb FC/10 Gb LAN AT094A: LAN ROM: 4.9.416.17, LAN EFI:  49.00A12 , FC ROM:  2.02x2 , FC EFI:  4.01.a5 (Universal Boot Code: 5.03a15)

Firmware Flash for Intel and Broadcom Adapters version 2014.10.27 - installs at UEFI:
    Warning: For Broadcom adapters, use this smart component only if IEXGBE-DRV version is lower than B.11.31.1303 (otherwise, use the Firmware for Broadcom NIC version 2014.2.18)
    Intel 2-port/4-port 1000B-T/SX AD337A, AD338A, AD339A: EFI 4.3.02 (i2 and i4 blades only)
    Intel 2-port 1000B-SX + 2-port 4Gb FC AD393A: no longer supported
    Intel 1-port/2-port 4 Gb FC + 1000BT AD221A, AD222A: FC ROM:  2.82x4, FC EFI:  4.01a5; LAN 4.3.02 (i2 and i4 blades only)
    Broadcom Superdome 2 embedded LOM: EFI 5.2.60, ROM 6.2.28   
    Broadcom 2-port 10 GbE AM225A, AM232A, AM233A: EFI 5.2.60, ROM 6.2.28 

 Firmware Flash for Broadcom NIC HP-UX version 2014.2.18 - installs from HP-UX:
    Warning: Use this smart component only if Broadcom IEXGBE-DRV version is B.11.31.1303 or greater (otherwise, use the firmware for Intel and Broadcom adapters, version 2014.10.27)
       Broadcom 2-port 10 GbE AM225A, AM232A, AM233A: EFI 5.2.60, ROM 6.2.28

Firmware Flash Component for Superdome 2 iexgbe - HP-UX version 5.2.60 - installs from HP-UX
    Warning: Use this smart component only if Broadcom IEXGBE-DRV version is B.11.31.1303 or greater (otherwise, use the firmware for Intel and Broadcom adapters, version 2014.10.27)
           Broadcom Superdome 2 embedded LOM: EFI 5.2.60, ROM 6.2.28

HDD firmware:
Online ROM Flash Component for HP-UX - MB2000FBUCL and MB3000FBUCN: version HPDA
Online ROM Flash Component for HP-UX - MB1000FCWPP, MB2000FCVBV, MB3000FCVCA, MB4000FCVCB: version HPD4
Supplemental Update / Online ROM Flash Component for HPUX - EG0900FBVFQ, EG0600FBVFP, EG0450FBVFM, EG0300FBVFL: version HPDE
Online ROM Flash Component for HP-UX - MB4000FCWDK, MB3000FCWDH, MB2000FCWDF, MB1000FCWDE: version HPDD
Online ROM Flash Component for HP-UX - EG0900FCVBL, EG0600FCVBK, EG0450FCVBH, EG0300FCVBF: version HPD9
Supplemental Update / Online ROM Flash Component for HPUX - EG1200FDJYT, EG0900FDJYR: version HPD4
Online ROM Flash Component for HP-UX - EG0300FCHHR, EG0450FCHHT, EG0600FCHHU, EG0900FCHHV: version HPD8
Online ROM Flash Component for HP-UX - EG0300FBLSE, EG0450FBLSF, EG0600FBLSH, EG0900FBLSK: version HPD8
Online ROM Flash Component for HP-UX - EH0300JDYTH,EH0450JDYTK,EH0600JDYTL: version HPD6
Supplemental Update / Online ROM Flash Component for HPUX - MB6000FEDAU: version HPD7
Online ROM Flash Component for HP-UX - EG1200FCVBQ,EG1200FDNJT: version HPD8
Supplemental Update / Online ROM Flash Component for HPUX -MB2000FCZGH, MB3000FCZGK, MB4000FCZGL: version HPD7
Online ROM Flash Component for HP-UX - MB3000FBNWV and MB2000FCQPF: version HPD9
Online ROM Flash Component for HP-UX - EH0300JEDHC, EH0450JEDHD and EH0600JEDHE: version HPD4
Supplemental Update / Online ROM Flash Component for HPUX - MB4000JEQNL,MB6000JEQNN: version HPD7
Online ROM Flash Component for HP-UX - MM1000JEFRB, MM2000JEFRC: version HPD7
Supplemental Update / Online ROM Flash Component for HPUX -EG0300JEHLV,EG0600JEHMA, EG0900JEHMB,EG1200JEHMC: version HPD5
Supplemental Update / Online ROM Flash Component for HPUX - EG1800JEHMD: version HPD6
Supplemental Update / Online ROM Flash Component for HPUX - EH0600JDYTN: version HPD7
Online ROM Flash Component for HP-UX - EG1200JEMDA,EG0900JFCKB,EG0600JEMCV,EG0300JFCKA: version HPD6
Supplemental Update / Online ROM Flash Component for HPUX -MM1000FECVH: version HPD2
Online ROM Flash Component for HP-UX - EH0146FAWJB,EH0072FAWJA,DH0146FAQRE,DH0072FAQRD: version HPDK
Online ROM Flash Component for HP-UX - EH0146FBQDC,EH0300FBQDD: version HPD5
Online ROM Flash Component for HP-UX -MM0500FBFVQ,MM1000FBFVR:: version HPD9
Supplemental Update / Online ROM Flash Component for HPUX -DG0300FAMWN, DG0146FAMWL, EG0300FAWHV,EG0146FAWHU: version HPDG
Supplemental Update / Online ROM Flash Component for HPUX -DG0300BALVP, DG0146BALVN, DG0072BALVL,DG0300BAHZQ,DG0146BAHZP: version HPD5
Supplemental Update / Online ROM Flash Component for HPUX - MB6000JEFND, MB4000JEFNC: version HPD5
Online ROM Flash Component for HP-UX - EH0300JDXBA, EH0450JDXBB, EH0600JDXBC: version HPD5
Supplemental Update / Online ROM Flash Component for HPUX - EG1800JEMDB: version HPD3
Online ROM Flash Component for HP-UX - MB4000JFDSN,MB2000JFDSL: version HPD4
Supplemental Update / Online ROM Flash Component for HPUX - MM1000JFJTH: version HPD2
Online ROM Flash Component for HP-UX - EF0600FARNA, EF0450FARMV, EF0300FARMU: version HPD9
Online ROM Flash Component for HP-UX - EG1200JETKC, EG0900JETKB, EG0600JETKA: version HPD7
Online ROM Flash Component for HP-UX - EH0146FCBVB, EH0300FCBVC: version HPD9
Supplemental Update / Online ROM Flash Component for HPUX - MB4000JFEPB, MB2000JFEPA: version HPD5
Supplemental Update / Online ROM Flash Component for HPUX - MB6000JEQUV, MB8000JEQVA: version HPDB
Online ROM Flash Component for HP-UX - EG000300JWBHR: version HPD4
Online ROM Flash Component for HP-UX - EG0900FCSPN,EG0600FCSPL,EG0450FCSPK,EG0300FCSPH: version HPD2
Online ROM Flash Component for HP-UX - MB6000JEXYB, MB4000JEXYA: version HPD8
Online ROM Flash Component for HPUX- MB4000JVYZQ, MB3000JVYZP, MB2000JVYZN, MB1000JVYZL: version HPD2
Online ROM Flash Component for HP-UX - MB6000JVYYV: version HPD2
Online ROM Flash Component for HPUX- EH000900JWCPN, EH000600JWCPL, EH000300JWCPK: version HPD3
Online ROM Flash Component for HP-UX - EH000900JWCPH, EH000600JWCPF: version HPD3
Supplemental Update / Online ROM Flash Component for HPUX - EG000600JWEBH, EG000300JWEBF: version HPD3
Online ROM Flash Component for HP-UX - EG1800JFHMH: version HPD5
Supplemental Update / Online ROM Flash Component for HPUX - MB6000JEFND, MB4000JEFNC: version HPD9
Online ROM Flash Component for HP-UX - EG001800JWFVC: version HPD6

Smart Update Manager (SUM)
SUM Version: 7.6.6

Notes:

• Complex firmware version must be equal to or higher than the nPartitions firmware version at all times.
• Minimum OS and Firmware requirements per blade type:
• Partitions containing CB900s i6 blades (with Intel Itanium 9700 series processors) must run HP-UX 11i v3 Update 2017 or later and both the Complex and nPartitions must run firmware version 4.2.36 or later.
• Partitions containing CB900s i4 blades (with Intel Itanium 9500 series processors) must run HP-UX 11i v3 September 2012 or later and both the Complex and nPartitions must run firmware version 3.6.48 or later.
• Partitions containing CB900s i2 blades (with Intel Itanium 9300 series processors) must run HP-UX 11i v3 September 2011 update or later, the Complex must run firmware version 3.6.48 or later and nPartitions must run firmware version 2.53.0 or later.
• Rolling upgrades:
• Servers with complex firmware version 1.3.0 or 1.3.1 require bringing down the entire complex to perform the firmware update.
• Servers with complex firmware version 2.2.27 or later and running HP-UX 11i v3 September 2011 update or later can be updated online without the need to bring down the complex. A partition reboot must be scheduled when convenient to activate a new nPartition firmware.
• You may update the complex and all the nPartitions firmware at the same time.
• If you plan to update the nPartition firmware later, make sure that your intermediate combination of complex and nPartitions firmware is supported.
• For details on supported combinations of Complex and nPartition/OS firmware versions and features supported in each combination, refer to the table under the Compatibility section.
• Firmware update process notes:
• If LLF (Link Loss Failover setting) is enabled and a loss of link occurs on the active OA during a firmware update from complex version 2.53.0 or earlier, blades can be left in an unusable state. Workaround: Disable LLF prior to a firmware update using the OA CLI command "disable LLF" and re-enable LLF upon completion using the OA CLI command "enable LLF”
• For gWLM users, it is recommended to stop gWLM while performing an online update.
• If running Business iCAP software, you must update to Managed iCAP. See dependencies under Pre-Requisites. For more details on the iCAP upgrade, please contact HPE support.
• When updating active IOX enclosures from a version 2.53.0 or earlier, the update will be partial until each IOX can be updated after shutting down affected partitions.  It is supported to run indefinitely in the partially updated (Mixed) state.  IOX enclosures in this state will not have the increased thermal tolerance introduced in version 3.6.48. To complete the update, you must shutdown all partitions using an IOX and then reapply the firmware update to the complex only (not to the nPartition). This update is quick as it only targets the IOX device(s).
• When updating to this release from 2.2.27 or earlier, Socket Level Memory (SLM) values for nPars that have not been explicitly set i.e. still have default factory values, will be set to 0. In order to easily recover these values, before the update, document your Interleaved Memory (ILM)/SLM settings using the parstatus command. After the update has completed, reset the values using the parmodify command.
• Firmware update can be performed using SUM or from the OA CLI or GUI.
• User access to commands for managing the partition configuration and the partitions themselves can be controlled through the Parcon_Admin access right and partition access assignments using the ASSIGN PARCON_ADMIN and ASSIGN PARTITION commands. This was enforced starting with version 3.6.48 and OA user with operator capabilities can’t use “parstatus, parcreate, etc”.  

Refer to the SUM User Guide and Release notes for more information.

UPDATE RECOMMENDATION: Recommended
HPE recommends systems used in Mission Critical environments update to this revision of firmware during the next system maintenance period in order to improve system reliability.
Note: This is the initial release for support of CB900s i6 blades (Intel® Itanium® processor 9700 series).
[  ] Critical 
    [ ] Panic, [ ] Hang, [ ] Abort, [ ] Corruption, [ ] Memory Leak,
    [ ] Security
    [ ] Hardware Enablement, [ ] Software Enablement
[X ] Recommended
[ ] Optional
    [ ] Hardware Enablement, [ ] Software Enablement, [ ] non-critical
[ ] Initial Customer Release

SUPERSEDES:  
Complex Firmware Bundle Version: 2017.12 (4.2.48)
Smart Update Manager (SUM) version: 7.6.5

PRODUCT MODEL(S):

• Superdome 2 – 8s
• Superdome 2 – 16s
• Superdome 2 – 32s

OPERATING SYSTEMS:

• HP-UX 11i v3 September 2011 update (or later) is required to support this firmware with CB900s i2 blades only
• HP-UX 11i v3 March 2012 update
• HP-UX 11i v3 September 2012 update or later is required to support this firmware with CB900s i4 blades

• HP-UX 11i v3 Update 2017 or later is required to support this firmware with CB900s i6 blades

LANGUAGES:

• International English

 

ENHANCEMENTS:

 

Enhancements in Complex Firmware:

·       Enabled "Content-Security-Policy" (CSP) header in HTTP responses from Onboard Administrator.

·       Enhanced OA on Superdome 2 to offer “Advance Security Settings” in strong encryption mode. This feature allows Administrator user to enable/disable SSL (Secure Sockets Layer)/TLS (Transport Layer Security) protocols and ciphers on monarch Onboard Administrator.

FIXES:

 

Fixes in Complex Firmware:

 

·       Fixed unexpected error message issue on B32 Door Display

·       Fixed an issue where icapmanage -Q execution would cause a failure with the message "could not contact, active manager unknown". Then either Active or Standby GM would no longer contact the Superdome 2 GiCAP members with firmware version 4.2.36 or 4.2.48.

 

 

Note: For I/O components’ enhancements and fixes, please refer to the I/O components’ release notes. Links to the I/O component firmware packages can be found in the HPE Release set for Superdome 2 servers.

PREREQUISITES:
WARNING: If you are updating from a version prior to 4.0.4, you must first install to 4.0.4 (or 4.1.8 as this version also qualifies as a bridge) and then upgrade to 4.1.34 or 4.2.36 (bridge release for digital signing) prior to updating to this version.

WARNING: If you are updating from version 4.0.4 or 4.1.8, you must install 4.1.34 or 4.2.36 (bridge release) prior to updating to this version.

WARNING: iCAP users of version 10.10.00.12 or earlier should follow the steps below prior to upgrading to version 4.2.54 to prevent issues with handling of GICAP CA certificates.

1. Check the signature hash algorithm for GiCAP_OA_*.pem:

 

# openssl x509 -in /etc/opt/iCAP/GiCAP_OA_CA_CERT.pem -text -noout | grep Signature

 

  You may see either sha1WithRSAEncryption or sha256WithRSAEncryption.

 

2. Store the current GiCAP_OA_CA_CERT.pem file on both GMs before updating iCAP SW to 10.11.00.01 or later:

# cp /etc/opt/iCAP/GiCAP_OA_CA_CERT.pem /tmp

 

3. Update iCAP SW to 10.11.00.01 or later on both GMs

 

4. Skip the step here if the above #1 steps had indicated sha1WithRAEncryption.

Run the following command to re-create GiCAP_OA_USER_CERT.pem if the above #1 step had indicated sha256WithRSAEncryption:

 

# rm –f /etc/opt/iCAP/GiCAP_OA_USER_CERT.pem

# /opt/icod/bin/GiCAP_OA_keygen -u

 

  NOTE: The above assumes that you were using iCAP SW 10.10.00.07, 10.10.00.10, or 10.10.00.12 before updating iCAP SW to 10.11.00.01 or later.

 

5. Execute updateGiCAPCert from Active GM to transfer GiCAPcert.pem to Standby GM and upload CA & User certificates to each GiCAP member:

 

# /opt/icod/bin/updateGiCAPCert

 

  NOTE: You need to have the root password for Standby GM OS in handy.

  NOTE: You should defer the choice of restarting OA until the next step.

 

6. Execute the above step from Standby GM to transfer GiCAPcert.pem to Active GM and upload CA & User certificates to each GiCAP member:

 

# /opt/icod/bin/updateGiCAPCert

 

  NOTE: You need to have the root password for Active GM OS in handy.

  NOTE: If Category 1 is applicable to your systems, you should restart the member at this time. Category 2 would not need to restart members.

 

7. Wait a while until each member OA starts up.

 

8. Execute the following command from both GMs to check the connectivity to each member OA:

 

# icapmanage -sv

 

You may consult “HP Instant Capacity (iCAP) Version 10.x User Guide” at https://support.hpe.com/hpsc/doc/public/display?docId=emr_na-c05054382 if you see any error.

If there is no error, you can now upgrade the Superdome 2 firmware to version 4.2.54 or later.

 

  NOTE: Review the Superdome 2 firmware release notes for stepping requirements if upgrading from a version prior to 4.0.4 or 4.1.8.

 

You need to re-add the members to the existing GiCAP group:

 

# icapmanage -a -m member1:member1-oa1.corp.com -g prod

# icapmanage -a -m member2:member2-oa1.corp.com -g prod

...

 

Finally, execute the following command from both GMs to check the connectivity to each member OA:

 

# icapmanage –sv

HPE strongly recommends implementing the following security best practices to help reduce both known and future security vulnerability risks:

1.       Isolate the management network by keeping it separate from the production network and not putting it on the open internet without additional access authentication.

2.       Patch and Maintain LDAP and web servers.

3.       Run the up-to-date viruses and malware scanners in your network environment

4.       Apply HPE Firmware updates as recommended.

 

Minimum HP-UX OS and patch requirements for supporting the CB900s i4 blades are documented in the HP-UX Errata for HP Integrity Superdome 2 (Superdome 2 CB900s i2 or i4 Server Blade model) Additional HP-UX 11i v3 (B.11.31) drivers, products, and patches.
Note: No specific patches are required when using HP-UX 11i v3 March 2013 update or later.

Note: As part of fixing a FPL log issue in Superdome 2 complex firmware, FPL logs will be cleared when updating from version 3.7.14 (or lower) to version 3.7.60 (or higher). SEL logs will remain intact.  HPE recommends capturing the FPL logs prior to the firmware update by running the SHOW FPL command and selecting “F’ (first record) then ‘D’ (dump).

Note: Before updating firmware from version 1.3.x to 2.2.27 or later, the IO slots 6 & 7 (USB and ILO) should be removed from the vPars that currently own these resources as you won’t be able to delete them from the vPar after the update.

For information on supported I/O firmware versions, please refer to this Matrix: HPE Release set for Superdome 2 servers.

INSTALLATION INSTRUCTIONS:
Please review all instructions and the "Hewlett Packard Enterprise Support Tool License Terms" or your Hewlett Packard Enterprise support terms and conditions for precautions, scope of license, restrictions, and limitation of liability and warranties, before installing this package. It is important that you read and understand these instructions completely before you begin. This can determine your success in completing the firmware update.

Important Note: This version of SUM uses an Internet browser as the graphical user interface.  Currently supported browsers are:

	Win2k8	Win2012	Win2016	SLES11	SLES12	RHEL6	RHEL7
IE 9, 10 or 11	x	x	x
Edge			x
FireFox 17 ESR				x	x	x	x
FireFox 18 or later	x	x	x	x	x	x	x
Chrome 24 or later	x	x	x	x	x	x	x

Important Note: SUM uses port 63001 by default for pulling FW files from the server running SUM. 

A. Extracting and Running SUM

1. Download the firmware bundle binary <filename>.exe or <filename>.tar.gz in a new folder on your system.
2. Extract the bundle:
• Windows: double click <filename>.exe then click “Extract”
• Linux: gunzip <filename>.tar.gz; tar -xvf <filename>.tar
3. The above steps will extract the contents of the bundle in the new folder where the bundle file is located. SUM and firmware components will reside in the same folder.
4. Execute hpsum.bat (Windows) or hpsum.sh (Linux) located in the extracted folder, by logging as a user with Administrator or root privilege.  Executing hpsum will bring up your default browser to show the user interface of the Smart Update Manager (SUM).

NOTE: SUM also provides a Command Line Interface (CLI) that allows user to perform scripted installations in silent mode. For more information please refer to the SUM User Guide located at http://www.hpe.com/info/sum-docs  or the "CLIHelp.txt" file in the extracted folder.

B. Installing the Firmware Components using SUM - Graphical User Mode

• Welcome Screen:

• Once started, a page displays with SUM main content on it.

• For local updates on a Windows or Linux node, use Localhost Guided Update:

• Select Localhost Guided Update from the main content window
• Choose Interactive or Automatic mode
• Continue to step through the process of completing the inventory of the baseline, inventory of the node and deployment of updates to the localhost node.
• Note: You must have components in the same directory as SUM in order for the baseline to be added during Localhost Guided Update.

• For remote OA and iLO nodes

• Select Baselines from the main content menu
• Verify that the current directory baseline completes inventory
• Select Nodes from the main content menu
• Select 'Add Node' and enter IP address and credentials for the Onboard Administrator, select the Baseline to be assigned from the Baseline drop down, and click 'Add'
• After initial verification of credentials is done, click Inventory from the Actions menu
• After Inventory is done on the OA, iLO nodes will be added automatically and Inventory started on them.
• If a Virtual Connect node was found, click on that node and enter credentials which will start inventory
• Once inventory is done on all OA, VC and iLO nodes, click on the OA node and start Review/Deploy
• On the Review page, click to select the associated nodes to update along with the OA.  SUM will sequence the updates as needed.
• Start the Deploy
• After Deploy is done, the logs for the updates can be reviewed on each node page

• For remote Windows, Linux, HP-UX and VMware nodes

• Select Nodes from the main content
• Verify that the current directory baseline completes inventory
• Select Nodes from the main content menu
• Select 'Add Node' and enter IP address and credentials for the remote node, select the Baseline to be assigned from the Baseline drop down, and click 'Add'
• After initial verification of credentials is done, click Inventory from the Actions menu
• After Inventory is done on the node, click Review/Deploy from the Actions menu
• Review the updates to be applied and make changes to installation options as needed
• Start Deploy
• After Deploy is done, the logs for the updates can be viewed on the node page

• Refer to SUM User Guide for more information.
• Refer to the SUM Release Notes for a complete list of known SUM installation issues.

COMPATIBILITY/INTEROPERABILITY:

• WARNING: Do not disconnect AC power to HPE Integrity Servers or remove HPE Server Blades from the enclosure during the firmware upgrade process. Doing so may corrupt the programming of the firmware, rendering the Server inoperable.

• The table below describes the supported firmware/OS combinations.
  Notes for using the table:
  - Complex firmware can be updated and activated while the complex is online as long as both the “before” and “after” combinations of firmware/OS are supported (see table). From a visual viewpoint, complex firmware can be updated online as long as the two configuration boxes (“before” and “after”) in the same column indicate "Yes".
  - Partition firmware can be installed while the partition is online.  However, a partition reboot is required to activate the new partition firmware.

·         Note 1: No PCIe OL* support (requires 3.6.48 or later partition firmware)

·         Note 2: Supports PCI OL*, but won’t have the following functionality (only enabled with September 2012 or later):

• Improved I/O performance in virtual environments
• Added fault isolation assistance to the proper FRU for some catastrophic errors
• Support for Cisco B22HP Fabric Extender and B22HP Fabric Extender with 16 FET (standalone only)
• Note 3: One Stop Fault Management requires HP-UX 11i v3 March 2012 update or later

• The following recommendations are in place for optimizing PCI OLRAD performance:
• Customer Requirements to enable support for PCI OL*
• Customers will be required to run Superdome 2 system firmware release 3.6.48 or later
• Customers are required to install the following patches (or their current replacements) to fully support PCI OL* with HP-UX 11i v3 September 2011 update with system firmware release 3.6.48 or later:
• PHKL_42230    1.0            kernel access infrastructure cumulative patch
• PHKL_42314    1.0            TLB,MCA Recovery,HPVM,NUMA,P-State,IPMI,OsName,GPE
• PHKL_42331    1.0            EFI Cumulative patch
• PHKL_42482    1.0            ACPI CA cumulative patch
• HPE strongly recommends for optimal performance for OLRAD functionality, customers also install the following patch: (required for 11i v3 September 2011, March 2012 and September 2012 updates)
• PHCO_42997                1.0            olrad(1M) cumulative patch
• For HP-UX 11i v3 March 2012 update and September 2012 update, no additional patches required
• While not directly related to OLRAD functionality, HPE recommends all customers install or ensure the following list of additional patches are on their HP-UX 11i v3 September 2011 update and March 2012 update systems to ensure optimal performance and stability for OLRAD
• PHKL_41942 (enable CPU OL*) : included with 11i v3 September 2011 update cold install
• PHCO_42306 (enhancement for machinfo)
• Note: PCI OL* is supported in vPars v6 and HPVM environments with very limited capabilities. See customer advisory c03757066 for details.

DETERMINING CURRENT VERSION:

• Establish a telnet session with OA and login to the get the command prompt
• Type the OA command "UPDATE SHOW FIRMWARE" to display the version of complex bundle installed.  Type the OA command “UPDATE SHOW NPARTITION ALL” to display the version of nPartition firmware installed on each partition.

KNOWN ISSUES & WORKAROUNDS: 

·       OA web interface stops working when LDAP user modifies the “Advanced Security Settings” (SSL/TLS ciphers, protocols). HPE recommends not to modify “Advance Security Settings” when logged in as LDAP user.    

Workaround: Login to OA CLI as local administrator user and modify the “Advanced Security Setting”. 

·       Door-Display firmware update fails, if SSL/TLS protocol version-TLSv1.0 is disabled on Onboard Administrator.

        Workaround: Enable SSL/TLS protocol version-TLSv1.0 using “Advanced Security Settings”

• Remote serial console applet may not launch with Java Runtime Environment (JRE) version 6. Workaround: It is recommended to update to the latest JRE version on the client system.
• The user might see a warning or notice as follows while launching the Remote Serial Console (RSC) applet from the OA GUI: “Applet or Pericom version mismatch between the local applet the browser already has, and what server has to send to the browser. Close all browser instance and then start the browser again.”
  Workaround: It is recommended to delete the previously cached TeemWorld.jar file from java cache using Java Control Panel.
• The Firefox browser versions 52 or later does not support following set of functionalities in OA GUI – Launch Virtual Media (VMedia), Remote Serial Console (RSC), Complex/Partition Log viewer and OA CLI Java applets.
  Workaround: Internet Explorer or Firefox version 51 (or earlier) can be used to access these services.
• Updating the Superdome 2 firmware via FTP may fail if the password contains some special characters, for instance, §, $, & or space (but not !). This is due to interpreting the special character as part of the command instead of the password.
  Workaround: Use quotes around the URI (ex. update firmware ‘ftp://user:passwd@15.1.1.75/firmware/hpsdx-8.7.84 -fw.bundle’ all).
• iLO may hang occasionally when running with complex firmware version 3.7.60 or higher. 

 Workaround: A new "reset ilo <bay #>" CLI command was introduced in firmware 4.0.4 to recover from the hang.

• No error is reported for loss of power supply redundancy if all but one of the upper power supplies are removed or failed.
• If SSL certificates are renewed with GiCAP_OA_keygen on GMs after some members are removed, you must restart all the OAs specified for a member before re-adding the member or the addition will fail.

• iLO is affected by the OpenSSL Poodle (Padding Oracle on Downgraded Legacy Encryption) CVE-2014-3566 vulnerability. More information can be found in this advisory c04496345 and on the HPE support center at this link (Enter “POODLE CVE-2014-3566” in the Search box).

• Several iCAP commands may fail when using GiCAP on an AGM/SGM system running a version of OpenSSL prior to 0.9.8za. See the Pre-Requisites section for error examples.
  Workaround: Install OpenSSL 0.9.8za or later on the AGM and if one exists on the SGM (see Pre-Requisites section for details).
• Update show nPartition can fail to read the configured firmware version
  Workaround: Since this issue is intermittent, repeating the command may return the correct information. The SHOW BLADE INFO command will also return the nPartition firmware version for each blade in the partition.
• The connection to the system console logs is lost after hitting CTRL-c.
• Java Runtime Environment (JRE) must be installed on the client system to view the remote serial console screens (see advisory c04675356).
• When adding privileges to a user or group account via the OA GUI, only the first enclosure’s Server and IO Bays are displayed for selection.  This issue only impacts systems with multiple enclosures (32-socket systems).  Since Superdome privileges are assigned on a partition basis, not a Server or IO Bay basis, there is no effect on system operation.
• When a Fabric configuration change needs to be propagated across partitions and blades, and one of the nPars is a single blade nPar where the blade is missing, FAB_DAEMON_ERROR is generated.  There is no impact to performance or functionality.
• Rapid navigation through the partition pages can result in incorrect content being displayed.  Please wait for pages to load completely before navigating away to another tab.
• When a user navigates from Create nPartition to Modify nPartition in the GUI, fields in the Modify nPartition page may end up displaying empty fields.  Select the newly created partition from the left navigation menu and click on the Modify button on the Status tab to update the blank fields.
• vPar state handling gets stuck in INIT when a TOC is issued on the vPar while it is activating. 
  Workaround: Force a reboot or shut down the vPar to recover the state.
• An nPar may get stuck in the reboot state if a TOC is issued on the nPar and a reboot is issued before the TOC completes. 
  Workaround:  Force a power off on the nPar.
• If a blade is left in a bad state after an online firmware update, the GUI does not provide enough information to diagnose the problem. 
  Workaround: The OA CLI command “show update firmware” can provide additional information.
• Under conditions of very high, unnatural stress involving reboot cycles and core license and assignment migration, some core movement requests may fail.  An example scenario would be a user running an OL* operation via a script in a stressful way, issuing both icapmodify and vparmodify commands on the same vPar. 
  Workaround: The problem is resolved by retrying the failed request.
• If a vPar does not have SLM (Socket Local Memory) assigned to it, the SLM field under “Memory Details” from the vparstatus -v command is blank.
• Upon the rapid issue of multiple parallel vparcreate commands on the same nPar, one of the commands may fail.  This failure is typically encountered when a script is created to issue multiple vparcreate commands in rapid succession. 
  Workaround: It is advisable to insert a delay of 1-2 seconds between successive vparcreate commands in scripts to avoid facing this issue.
• De-configuring a CPU socket causes the memory associated with the socket to be lost to the partition but this change is not reflected in the nPar "Status" page.
• The OA GUI shows the status of CPUs and DIMMs on an indicted/de-configured blade as OK instead of indicted or de-configured.
  Workaround: Check the status using the CLI interface.
• A user without partition access privileges may select "Update All Firmware" in the OA GUI Firmware Update page but the command will fail with the error "The action did not complete successfully."
  Workaround: If this message is displayed, verify that the user has sufficient permissions.
• If the SLM (socket local memory) for a blade is provided in absolute value (GB) during nPartition creation and/or modification, then “Total SLM” field in parstatus output (parstatus –c <blade_path> –V) and “SLM at Next Boot” field in parstatus output (parstatus –Z <parspec_name>) will display --.-% instead of the percentage of SLM memory configured. The values however are properly configured internally.
  Workaround: To view the values configured at the nPartition level, you may run the command “parstatus –p <partition> -V”.
• In vPars mode, "Virtual Devices" tab, which allows access to internal DVD and other virtual devices, is not available to OA users configured with "Operator" level privilege.
  Workaround: Use the privilege "Administrator" for this task.
• After a FW update and after power-on, the nPartition ILM/SLM setting may rarely be altered due to a timing issue.
  Workaround: Powering off and powering on the partition will restore the expected values.
• Some BPS may rarely show in dynamic power saving mode although dynamic power saving is turned off for the complex.
  Workaround: Turn on dynamic power saving for the complex and turn it off again.
• For powered off partitions, after OA failover or OA restart, 'show xfm status all’ and ‘show IOX status all’ may show J-LINKs as “OK” instead of “Dormant”.
• If an active J-LINK cable gets disconnected, ‘show XFM status all’ and ‘show IOX status all’ may report that LINK as “Degraded” instead of “Failed” and as “OK” instead of “Dormant.
  Workaround: Restart OA to see the proper status.

• Superdome 2 OA/CAMNet uses 169.254.0.x and 10.254.x.x addresses for internal management. Those addresses are visible to the outside LAN via “SNMPwalk”, which may cause incorrect reports of IP address conflicts. However since the internal management network traffic is contained within the complex, this will not affect the customer’s management network.
  Note: The OA EBIPA cannot use addresses 169.254.x.x or 10.254.x.x.

• After a CAMNET link failure, “Show complex status” continues to report CAMNET status as “Degraded” even after recovery of the CAMNET cable and acquitting the IOX indictment.
  Workaround: Restart OA to see proper status.
• The system does not detect or report when there is a problem with the standby CAMNET.
• After an OA restart, a partition state may become Unknown/Detached. Refer to the HPE Superdome 2 Partitioning Administrator Guide HP-UX 11i v3 for details.

Known IO Issue with 10 GbE Superdome 2 LOM and AM225A/AM232A/AM233A cards (EFI 5.2.60/ROM 6.2.28):

• A general exception fault may rarely occur when using the Broadcom 10 GbE adapter on a Superdome 2 server connected to a network switch with STP disabled. This failure only occurs in the UEFI environment, in presence of abundant broadcast traffic, and prevents from booting the partition to the OS. As a workaround, enable STP on the switch.

Known issues and workarounds related to PCI OL*:

• When performing an OLAR operation for network cards that use the iocxgbe driver, the system may crash when using a driver version prior to B.11.31.1303.01.
• PCI OL* is supported in vPars v6 and HPVM environments with very limited capabilities. See customer advisory c03757066 for details.
• For Superdome 2 customers using Infiniband, to use PCI OL*, customers must install the IB4X-00 product from the HP-UX 11i v3 March 2012 update or later release media.
• Because of PCI OL* enablement issues found with several of the HP-UX IO drivers, customers must install:
• HP-UX 11i v3 September 2012 update IO drivers.
  Note:  HP-UX 11i v3 September 2012 update or later drivers can be installed on HP-UX 11i v3 September 2011 update and March 2012 update, without having to install anything else from HP-UX 11i v3 September 2012 update or later.  Customers should visit software.hp.com for the latest driver release packages which may have important updates.
• Web Release 1209 SFM and IO Providers

KNOWN INSTALLATION ISSUES: 

• The door display device may not work as expected after updating the complex firmware to 4.2.54 or above. Contact HPE support team for further help.
• The ‘show update firmware’ OA command may report a FRU(s) firmware mismatch after updating from system firmware version 1.3.1 to version 2.51.102 or later. Reapplying the update will solve the mismatch (the second update will be much faster as it will only update the metadata).

      Refer to the SUM Release Notes for a complete list of known SUM installation issues.

DISCLAIMER:
The information in this document is subject to change without notice.

Hewlett Packard Enterprise makes no warranty of any kind with regard to this material, including, but not limited to, the implied warranties of merchantability and fitness for a particular purpose. Hewlett Packard Enterprise shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material.

This document contains proprietary information that is protected by copyright. All rights are reserved. No part of this document may be reproduced, photocopied, or translated to another language without the prior written consent of Hewlett Packard Enterprise.

(C) Copyright 1994-2018 Hewlett Packard Enterprise Development L.P.

FEEDBACK
As we are continuing to improve the firmware management process we welcome your feedback on this document and on the firmware update process: TEAM-FWupdateFeedback@groups.ext.hpe.com

SUPERSEDES HISTORY:

Bundle Version 2017.12 (4.2.48):
ENHANCEMENTS:
Enhancements in nPartition Firmware:

• Optimized the memory settings to improve the performance of the memory system, when every DIMM in the nPar supports unrestricted MAC (Maximum Active Count).

Enhancements in Complex Firmware:

• Added support for OA Battery Diagnostic status.
• Clarified the action to be taken in case of CAE event 9666 generated due to an XFM failure 

FIXES:
Fixes in nPartition Firmware:

·   Fixed a vPar environment boot error condition.

Fixes in Complex Firmware:

• Upgraded SNMP package to avoid memory leaks.
• Upgraded NTP to address CVE-2016-7434 vulnerability.
• Addressed an instance of DOM (Document Object Model) based Cross-Site Scripting (XSS) vulnerability.
• Addressed an instance of Stored Cross-Site Scripting (XSS) vulnerability.
• Enabled HTTP Strict Transport Security (HSTS) headers in HTTPS response from Onboard Administrator.
• When HEAP_SPACE_CORRUPT is reported in the FPL/SEL, it is now also reported in the Core Analysis Engine
• Due to a race condition on the available rights of the complex, one partition would only get partial authorization despite of having enough available rights in the group that could have been used. This is now fixed.

Bundle Version 2017.06 (4.2.36):
ENHANCEMENTS:

·         Added support for CB900s i6 blades (Intel® Itanium® processor 9700 series)

 

FIXES:

Fixes in nPartition Firmware:

·         Fixed an MCA loop condition on a vPar environment.

·         Fixed an issue where a memory controller issue on a blade could cause the partition to fail power-on with the FPL log showing the events MEM_SMBUS_WRITE_FAILED and/or MEM_MC_INIT_FAIL. The blade will now power-on with the blade indicted.

Fixes in Complex Firmware:

·         Virtual Media applet can now be used with Java 7 or above without requiring the iLO URL to be added to Java Exception Site list.

·         Added missing syslog messages for IOX Power Supply Insert and IOX Power Supply Remove.

·         CAE event 2015 is no longer erroneously reported after encountering some hardware issue.

·         Fixed a file lock issue causing the 'show partition' command to fail when run by user with operator permissions (error was: /etc/small_parconf_update_lock: No such file or directory).

·         Fixed the issue wherein ‘remove snmp trapreceiver <IP address>’ command would incorrectly remove SNMP trap receivers having the input <IP address> as its substring.

Bundle Version 2016.09 (4.1.34):
FIXES: 
Fixes in Complex Firmware:

• iLO watchdog timer has been enabled which can reset and recover iLO in certain cases when the iLO crashes or gets into a hung state.
• After upgrading to version 4.1.34, iLO’s self-signed certificate will be removed and replaced with SHA256 self-signed certificate.
• After OA reboot, stale vpar information was sometimes displayed in vparstatus. This is now fixed.
• Parstatus would rarely show the blade product name as "[Unknown]". This is now fixed.
• Parstatus would occasionally report CPU speed as 0MHz on new monarch OA after an OA failover causing GiCAP to mark the Group as non-compliant. This is now fixed.
• Complex Firmware would rarely show as [Unknown] on new monarch after an OA failover, causing iCAP to reject the complex from joining the GICAP group. This is now fixed.
• As a non-supported feature, Blade PowerDelay is now always disabled and can no longer be changed from the CLI or GUI.
• SNMP no longer incorrectly sends trap 22041 whenever the active OA status changes. 
• The OA no longer incorrectly returns an error to snmpwalk searching for OID 1.3.6.1.4.1.232.22.2.3.2.2 of the cpqRackLastEAEEvent.
• An LDAP user logged into the OA GUI of the monarch enclosure of a 32 socket Superdome 2 system will now see the second enclosure in the GUI page.
• Fixed an issue where, if connected to the OA via telnet, the message “fatal: Cannot bind any address” could be seen in the OA syslog after running the command “disable ntp”, “disable https” or “disable snmp”.
• The HTTPoxy vulnerability CVE-2016-5387 has been addressed.
• Addressed iCAP memory leaks in the 2 situations below, that could potentially lead to an OA reboot in GiCAP environment:

• When Group Manager requests Member information.
• Upon icapmodify command failure, when member communication with Group Manager fails due to certificate errors; error reported:
  “Error:Unable to connect to AGM.
  Response from AGM : HTTP/1.1 401 Unauthorized"

Bundle Version 2015.09 (4.1.8):

ENHANCEMENTS:

• Release notes updated to clarify that the exception fault with Broadcom is a known issue in the IO card firmware and not the Superdome 2 firmware.

FIXES:

• Fixed a corner case where a partition could unexpectedly shut down in case of USB access failure during OA failover.
• Incorrect handling of recovery from fabric link failures on a running partition could rarely cause a partition to MCA. This is now fixed.

Bundle Version 2015.05 (4.0.4):

ENHANCEMENTS:

Enhancements in Superdome 2 Complex Firmware bundle 2015.05 (Version 4.0.4): 

• Added a reset iLO command capability from the CLI (Command Line Interface) in case the iLO is failed or hung.

FIXES: 
Issues fixed in Superdome 2 nPartition Firmware bundle 2015.05 (Version 4.0.4): 

• Fixed a virtual mapping issue that could rarely cause an HP-UX partition (data TLB) panic.

 

Issues fixed in Superdome 2 Complex Firmware bundle 2015.05 (Version 4.0.4): 

• Addressed the OpenSSL Poodle (Padding Oracle on Downgraded Legacy Encryption) CVE-2014-3566 vulnerability in Superdome 2 OA (Onboard Administrator).
• NTP (Network Time Protocol) was upgraded in this version to address the CVE-2014-9295 vulnerability.
• Addressed the GHOST (GetHostByName) CVE-2015-0235 vulnerability.
• After an OA failover, partitions could very rarely be powered down by the system if the OA was unable to get partition data from the internal USB flash drive (robust store). This is now fixed.
• Standby OA now correctly updates certificate replacements initiated by GiCAP (Global instant Capacity).
• OA was updated to support new time zone changes in Russia.
• The device bays of fan zones 3 and 4 are now correctly reported in the active OA GUI (Fan Zones TAB under Thermal Subsystem Screen).
• In the OA GUI, after selecting "Local User" and then "User information", Interconnect Bay Access is now shown for all bays, including Bay 5 to Bay 8.
• The OA GUI now properly detects and displays a valid Firmware bundle present on a USB stick in the external USB drive.
• Addressed the openSSL CVE-2014-3571, CVE-2015-0209, CVE-2015-0286, CVE-2015-0287,  CVE-2015-0288, vulnerabilities.

Bundle Version 2014.10 (3.8.0):

FIXES:

• The bash “Shellshock” vulnerability has been fixed in the OA version included in this firmware. Firmware versions prior to 3.8.0 could be exploited remotely to allow execution of code. For details on this issue, please refer to CVE-2014-6271, CVE-2014-6277, CVE-2014-6278, CVE-2014-7169, CVE-2014-7186, CVE-2014-7187.

Bundle Version 2014.07:

ENHANCEMENTS:

• Implemented internal USB bus auto-recovery and monitoring to improve USB bus robustness

FIXES:

The following issues have been fixed in Complex firmware version 2014.07 (3.7.98): 

• SSL/TLS MITM vulnerability (CVE-2014-0224) was fixed by implementing OpenSSL version 0.9.8za.
• When attempting to borrow core usage rights on a GiCAP member system, the icapmodify command sometimes reported the error: "Parse error at line 1: unknown encoding".
• If an SD2 complex, member of a GiCAP group, was removed from the GiCAP group and its hostname changed, adding back the complex into the GiCAP group would fail with the error "icapmanage command failed. Could not install member certificate".
• If the vparmodify command was run while the communication to the active group manager was lost, it was possible to activate a TiCAP core despite having the TiCAP pre-authorization turned off for the complex.
• A user logged in the OA as Administrator with administrator privileges and trying to load configuration scripts from the OA GUI would get the error "USER_REQUEST 8 uploadConfig Insufficient privileges for operation".
• In rare cases, the error 9708 "Failed to access a FPGA component" would occur due to a blade link failover causing a temporary communication loss to the blade across the management network. 
• After booting an nPar with a link stopped, the OA command "show complex status" sometimes reported XFABRIC as DEGRADED, although there were no indictments or degraded links.
• If a partition was booted with a de-configured J-link, a subsequent 'start link' command would sometimes leave the partition with non-redundant routes until the next partition reboot.
• After a BIOX chassis power loss, partitions with IO bays configured in that BIOX would sometimes MCA and fail to boot back up. With this version, both IO bays will now get de-configured, allowing the partitions to boot back up.
• A system crash due to an unrecoverable error sometimes triggered multiple unnecessary indictments and de-configurations. These indictments were side effects of treating certain CPU cache errors (primary cause) and will no longer happen.
• When using the trusted host feature introduced in version 3.6.48, any telnet/ssh/ftp connection to the OA would be disabled after updating the trusted host list.
• After running the ‘force takeover’ command or after an OA restart from the OA CLI, XFM and IOX links (J-link cables) sometimes showed as unknown in ‘show xfm status’ or ‘show IOX status’ output.
• On very rare occasions, the OA would lose contact with the upstairs fans due to excessive internal (CAMNet) network traffic, causing the cabinet to power off. Secondary events would also sometimes be generated, such as OA_MULTI_ENCSYNC_LOST, HR_FMP_OFF_NETWORK, CFW_FPGA_READ_ERROR,  SD_PDHC_DOWN, FAN_STATUS_FAILED  and COOLING_STATUS_DEGRADED.
• After an OA restart, when remote syslog logging was enabled and the remote syslog server was specified on the OA as a hostname (as opposed to an IP address), the OA would stop logging to the remote syslog server.

The following issues have been fixed in nPartition firmware version 2014.07 (3.7.98): 

• In a single-core vPar configuration, shutting down and restarting a booted vPar would rarely leave the CPU in a low power state, causing performance to be degraded.

SD2 version 2013.08 (3.7.60):

• Optimizes the memory settings to improve the reliability of the memory system.
• Added support for new DIMMs (AT127A, AT128A, AT129A, AT130A) for use on CB900s i4 blades only 
• Support of these new DIMMs requires Complex and nPartition firmware version 2013.08 (3.7.60) or later
• DIMMs are enabled in the system in groups of 4 (quads) and DIMM types cannot be mixed in a quad.
• Mixing different DIMM sizes is not supported on Superdome 2 blades (CB900s i2 or i4).
• The total memory in a system fully populated with 16 GB DIMMs may reach 8 TB; however HP-UX supports only up to 4 TB of memory  per HP-UX instance

Bundle Version 2013.04 (3.7.14)

ENHANCEMENTS:

SD2 Version 2013.04 (3.7.14): 

• Improved error resiliency on fabric links 

UEFI Driver 5.2.59 (from 5.2.58) for embedded LOM, AM225A, AM232A, AM233A:

•  Added Dynamic Network detection support

FIXES: 

SD2 Version 2013.04 (3.7.14): 

• A failed fabric link now results in the complex-level fabric status showing as degraded instead of remaining as OK while the “Xfabric link status” field of the “show xfm status” or “show iox status” command was showing the link as Failed.
• If an XFM module cannot be powered on as part of a complex initialization, "show complex status" now shows as DEGRADED instead of OK and the powered off XFM module will be de-configured until the next complex initialization.
• Upon a cell replacement or upgrade, warranty and entitlement information is no longer migrated from the original blade to the replacement blade.
• The "show enclosure power_summary" command no longer shows 549W more power than is actually allocated in "Power Allocated" on a per blade basis.
• Fixed a firmware issue which could result in the OA reporting erroneous failures like “unexplained firmware mismatch”, or grayed-out display on the “Insight Display” page, or a firmware update failing with the error "Firmware update cancelled by system”, and where the user had to restart the OA to fix the issue.
• Vparstatus no longer shows more cores than expected if vparmodify is run with the option “-m cpu::<num>” or “-m core::<num>” while the vpar is down and if that vpar already has some SLP (Socket Local Processors) or cores added by path.
• On a Superdome 2-32s system with blades in different slots in each enclosure, the ParGUI partition status page no longer displays some blade product names as "[Unknown]".

SAS 6.22 (from 5.78):

• HP Integrity Smart Array P411 and P812 Controllers, running version 5.78 or lower and connected to Superdome 2 servers may, under extremely rare conditions, silently return stale or out-of-order data. For more details, please refer to c03689769.

• With HP P411 or P812 controllers installed in a Superdome 2 server and external RAID connected, no post message is printed if “reconnect --r” is used to list cards at EFI shell. However during reboot, when selecting “s” at boot menu in EFI shell, post messages are printed. This is a generic issue and not specific to any server.
• HP-UX boot may fail with READ Error (elf_load) if boot volume is built on SAS Logical Volume larger than 2TB

Boot Code 6.2.25 (from 6.2.21) for embedded LOM, AM225A, AM232A, AM233A:

• In some cases, only 2 PCI functions are exposed in multi-function mode.

AT094A Lan FW  4.2.401.20 (from 4.1.450.14) fixed issues:

• Boot failure might be seen after running commands "reconnect -r", "map -r" commands at EFI environment.
• Boot from SAN intermittently fails during EFI bootstrap of a FCoE device that is busy.
• FCOE EFI version truncated in VPD data displayed by ODE-FCFUPDATE tool at EFI.
• On HP-UX, the fcoc port driver state is displayed as AUTHENTICATION/DISCOVERY when fcocutil command is run to check the port information.

AT094A LAN UEFI Driver 2.22A0 (from 2.21AC) fixed issue:

• BL8x0c i4 LOM Product Name is truncated when displayed from NIC interface using FCFUPDATE.

HBA FC FW version 2.02x2 (from 2.00a4) for AT094A, AH402A, AH403A:

• I/O “illegal frames received” errors may be reported intermittently by FC disks linked through an 8Gb Hp/Emulex FC card if a disk reset is performed during heavy I/O load.

Bundle Version 2012.12

ENHANCEMENTS:

• Superdome 2 servers now support CB900s i4 blades containing Intel Itanium 9500 Series processors  in addition to the already supported CB900s i2 blades (with Intel Itanium 9300 Series)
• Support for 8GB Low Voltage DIMMs on CB900s i4 blades. – Up to 4TB per OS image
• PCIe OL* support for CB900s i2 and CB900s i4 blade partitions – Add, Replace or Delete IO cards without partition downtime
• New One Stop Fault Management solution for Superdome 2 platforms, allowing single point of access for all hardware failure indications
• Additional Error Analysis capabilities

New Onboard Administrator Management features

• Per-Partition User Authorization – User rights may be restricted to access per individual partition
• Enclosure IP Reduction – Reduces the demand for external IP addresses for blades
• Archive Storage – Internal storage capacity for firmware images, configuration scripts, and other data
• Partition Configuration Backups – Enables disaster recovery protection for the complex partition configuration

IOX enclosure thermal thresholds and actions are now consistent with the compute enclosure limits

FIXES: 

• Interleaved or Socket Local Memory (ILM or SLM) added to a vPar are no longer lost if iCAP authorization fails on vPar boot.
• Blades are no longer left in an unusable state if Link Loss Failover (LLF) is set and a loss of link occurs on the Active OA during a firmware update.
• ILM (Interleaved Memory) and SLM (Socket Local Memory) values are no longer automatically reassigned when switching between nPars and vPars mode.
• During an OA reboot, if an ILO/VGA port is reported but not present, the OA will go into a continuous reboot loop. This behavior is now fixed.
• Some PCIe errors might cause a local MCA which might in turn cause a global MCA. This behavior is now fixed.
• vPar no longer gets into a hung state if a CPU migration is cancelled by an unknown entity.
• If an OA restart is done while some nPars are active, parstatus -p -V now displays the firmware revision for the active partitions.
• Fixed an erroneous amber LCD display following an upgrade from Superdome 2-16s to Superdome 2-32s.
• One large nPar exhibiting a blade failure will no longer cause a resource issue, preventing all vPars configured in the partition from booting.
• Corrected indictment errors erroneously reported for blade 255.
• It is now possible to TOC and power off a vPar that is hung due to a defective PCI card driver.
• If the OA is restarted while iLO is unresponsive, the system no longer erroneously reports Power Redundancy Lost.
• Fixed an issue where iLO SSH may hang when OA uses SSH public keys for the auto login.

Bundle Version 2012.09

ENHANCEMENTS:

None

FIXES:

• Fixed an issue where a memory update operation may not complete properly when an L2 cache line encounters a second single bit data error. This may result in the possibility of undefined system behavior. See Intel Specification Update for more information. If Specification Update link does not work, HP recommend to go to www.intel.com  and search for the latest version of the Intel® Itanium® Processor 9300 Specification Update.
• Fixed an issue where PAL returned no information on bus log overflow
• Fixed an error where an uncorrected MCA with PSP that does not indicate an error type when PAL is catastrophic
• Fixed an issue on 8s system following OA reboot or takeover or PDHC restart which can result in Detached partitions.

 

Bundle Version 2012.02

ENHANCEMENTS:

None

FIXES:

• Fixed an issue to eliminate deconfiguration of healthy CPUs due to false cache errors (event 100819).

Bundle Version 2012.01

ENHANCEMENTS:

• Support for vPars v5 dynamic cores
• Dynamic core migration with Global Workload Manager(gWLM) and HP Instant Capacity (iCAP)
• Dynamic Processor Resiliency (DPR)
• Global Instant Capacity (GiCAP) grouping
• Locality Optimized Resource Alignment (LORA) for cores
• Support for rolling upgrades

FIXES:

• After an A/C power cycle or a complex “force” restart, the “SHOW TOPOLOGY” command or the GUI now correctly shows the second enclosure as part of the complex.
• After removing and putting back a blade into a slot, the DVD health LED no longer erroneously flashes yellow if the DVD is healthy.
• The command “show iox status” no longer displays the 'populated' and 'powered' fields which were invalid for IO.
• Syslog now shows proper error message at boot (CFW_EXCEED_LIMIT instead of “invalid output”) if the number of vPars to be activated exceeds the maximum number configured.
• The command “show oa info all” may no longer report the standby OA Firmware Bundle Revision as Unknown.
• vMedia applet now returns an error if “Launch Virtual media” is selected from the GUI while an existing applet window is already opened.
• When a firmware update is initiated from the OA via the UPDATE FIRMWARE command, existing OA GUI sessions are now disconnected or logged out.
• On a Superdome 2-8S server, the OA no longer reboots after issuing the command "parmodify -a" for adding a fifth blade to the nPartition.
• The ILM/SLM ratio is no longer changed to 100% (resulting in performance degradation) when the OA is rebooted after creating an nPartition in nPars mode without specifying an ILM/SLM value.
• An nPartition in MCA state will now auto power after A/C cycle or restart “force” complex.
• In nPars mode, the OA no longer restarts upon running a “reboot partition ” command on an nPartition
• The parstatus command no longer displays a status and run state “Active/MCA” after recovering from an LMCA (Local Machine Check Abort).
• Instant Capacity rc script startup failures are no longer reported during HP-UX boot as the Instant Capacity (iCAP) feature is now operational.
• The checkrevs command no longer reports mismatch warnings after OA switchover
• After running a “forced takeover” command from the OA CLI, 'show blade status' now shows a valid CPU type and populates the other CPU fields.
• Fixed an issue where some DIMM ECC errors were not reported
• Fixed an issue where deconfigured resources back to healthy state would remain unavailable (HR_DECONFIG_STATE_MISMATCH IPMI event logged)
• Reduced the possibility of the loss of OA network connectivity that can be caused by excessive OA network traffic
• Added a threshold for FMP reboots so that one time occurrences do not cause the assembly or assemblies to be falsely indicted

Complex Firmware Bundle Version 2.2.27

ENHANCEMENTS:

• Support for 32 socket complex (128 cores)
• Support for 8 IOXs for Superdome 2-16s and Superdome 2-32s
• Enablement of online firmware update and activation of complex firmware for future firmware releases.
• HP-SUM support for Customer Engineers
• 4TB memory
• Upgrades: Superdome 2-8s to Superdome 2-16s, Superdome 2-16s to Superdome 2-32s
• More I/O options: SAS I/O & Storage, Infiniband I/O cards, FCoE I/O cards
• Enablement of online firmware update of partition firmware (reboot of nPartition is require to activate).

FIXES:

• Fixed an issue where a good processor could be deactivated and de-configured due to excessive correctable L1 Instructions cache error (100821 event)
• Fixed an issue where a good processor could be deactivated and de-configured due to no spare directory cache lines (7010 event)
• SEGFAULT error reported when launching a CLI shell from the OA GUI using a user login other than the one that originally launched the CLI
• A partition can now successfully power on, even if one of its blades experiences a power-on failure
• Updated OA firmware to include the latest OA tray firmware
• Fixed an issue where the system might crash and reboot while attempting to update FPGA bits.
• Enhanced the “show xfm status all” command output to include full diagnostic status information
• The Insight Display “Health Summary" output now displays correctly after executing an OA force takeover
• OA USB connectivity status is now verified and cable status is reflected in the OA status output.
• Fixed an issue where a partition found in a transition state could be forced off upon an OA reboot
• Fixed an issue where DIMM address and parity errors were not correctly detected, causing the partition to crash, and error analysis showed a bad CPU or blade, when there was no actual hardware issue.
• Setting the Date time in the OA GUI using an NTP server now works.
• After recovering from a CAMNET connection failure, Parstatus no longer shows BLADE UNINTEGRATED/DEGRADED and 'show blade status' shows PDHC status OK.
• Fixed an issue where the DVD content could be seen even after the DVD drive was disconnected, or where the DVD content could not be seen after the DVD drive was reconnected.
• An XFM power fault no longer prevents the remaining XFMs from powering on, nor causes OA reboots or DL daemon death.
• Upon a nPar reboot in a server with 8 blades, SHOW SYSLOG OA no longer shows extra entries such as "Server Blade in bay 8 has been powered down" and "Server blade in bay 8 has been powered on".
• Improper error messages are shown for partition pages when a non-privileged user logs into the OA GUI.
• "show cae -L" output listing sequence is now in order even if the system time is changed
• Fixed an issue where some events could be dropped from the OA CLI DUMP command output.
• Prevent any user name that begins with a Superdome 2 reserved name from being added to the OA
• A partition may now boot successfully when an IOX is indicted or de-configured
• A LDAP user is now able to login into the OA CLI or GUI
• After a IOX indictment is cleared, the front panel health LED now goes back to green when health is restored to OK
• Fixed an issue where partitions may fail to power on when AC redundancy is enabled and no power is available from one power rail
• Fixed an issue where blade iLOs may become unresponsive.
• Fixed an issue where server would intermittently show the fan events FAN_STATUS_DEGRADED followed by FAN_STATUS_OK and an amber fan LED after an AC power cycle or a GPSM failover
• Show IOX status output no longer displays temperature for IOX Outlet Zone (not monitored)

Complex Firmware Bundle Version 1.3.0/1.3.1

ENHANCEMENTS:
None

FIXES:

• A firmware issue has been identified on HP Integrity Superdome 2 servers installed with firmware version 1.0.35 (initial release) or 1.2.67 (released 15 November 2010). Due to a polling timeout issue in the firmware, Superdome 2 servers running 99.4 days since the last firmware update or ‘AC power on’ can fail unexpectedly resulting in an unplanned outage of the entire system or a loss of partition management capability. This issue is resolved in firmware version 1.3.0.

Complex Firmware Bundle Version 1.2.67

ENHANCEMENTS:

• CPE Fixes and "Smart" offline Firmware Update -- improves firmware update when replacing spares with earlier revisions of firmware (outside of Manufacturing mode)

FIXES:

• Erroneous slot type information may be displayed in parstatus command verbose output
• Unable to assign some blade I/O to vPar with nPar powered on
• No alert is generated when a bulk power supply BPS_REMOVED or BPS_REDUNDANT event occurs.
• Boot options changes in EFI may be lost when followed by a partition reset and an AC power cycle occurred during the partition reset
• A system may crash after issuing a vPar memory addition command resulting in unaligned memory.
• Some vPars may hang during boot after the creation of new vPars if too many vPars have already been created.
• When the GUI is used to modify an active partition, clicking on the “Pending changes” TAB displays a page load error.
• When partition controller is down, partition related GUI pages may fail to load (page hangs with a loading screen).
• Momentary press button should be grayed out in the GUI for an nPar booted using EFI
• Partition power off/on button behavior is confusing while the partition is being powered off or on.
• During a virtual partition boot, the OA GUI does not display CPU information.
• Unable to use the vMedia applet after disconnecting a previously connected media.
• CFW error messages such as CFW_LOAD_ERROR_VPAR_DEF or CFW_EFI_VARIABLE_LOAD_FAILED may be generated during high activity.

INITIAL RELEASE:

Complex Firmware Bundle Version: 1.0.35