Recovering from a Failed System ROM Upgrade



Recovering from a Failed System ROM Upgrade

During the ROM upgrade process, the whole process must complete for the ROM upgrade to be successful.

 

|[pic] |

|[pic] |WARNING: If an error such as a power failure, operating system lockup, or some other |

| |catastrophic event occurs during the system ROM upgrade process, the ROM upgrade will|

| |not complete and the system ROM image on the target server will be corrupted. Should |

| |this occur, the server will not operate correctly and ROM disaster recovery |

| |procedures must be implemented to regain server functionality. |

|[pic] |

|[pic] |

|IMPORTANT:  Supported Servers |

|While ROMPaq disaster recovery is supported on all ProLiant server models, each server has unique|

|and specfic directions for exicuting the Disaster Recovery Procedure. Older models require |

|powering down, opening the chassis, and locating the system maintenance switch block on the |

|server’s system board. Specfic switches have to be set and very from server model to server |

|model. Newer models do not require setting any system board switches. For the server specific |

|information, warnings and procedures, please refer to the server’s Setup and Installation user |

|guide for details.  |

|[pic] |

Follow these steps to perform ROMPaq disaster recovery on any Compaq server that has a corrupted ROM image:

1. Obtain the latest ROMPaq version for the target server.

The ROMPaq files can be obtained in several ways. See Question 5 below in the "Questions and Answers" section for sources where the ROMPaq files can be obtained.

 

2. Create the ROMPaq diskette that will be used to restore the server’s system ROM.

|[pic] |

|IMPORTANT: If System ROM is Corrupted |

|If the system ROM is corrupted on the target server because the ROMPaq ROM restoration procedure |

|was interrupted, the initial ROMPaq attempt may have affected the contents of the ROMPaq |

|diskette. It is recommended that a new ROMPaq diskette be created.  |

|[pic] |

|[pic] |

|NOTE: Compaq SmartStart |

|The Compaq SmartStart and Support Software CD 4.80 or later includes Diskette Builder software |

|that facilitates the creation of ROMPaq diskettes.  |

|[pic] |

Unlike the System ROMPaq Utility, the ROMpaq Disaster Recovery procedure does not allow for saving the original System ROM image by writing to diskette. Write operations to the diskette are not performed by ROMpaq Disaster Recovery procedure.

Top of page

Older ProLiant System Models

• Power down the server that contains the system ROM that needs to be restored.

• Open the server chassis and record the default positions of the switches on the server's Configuration Maintenance Switch Pack. If possible, obtain the revision level of the System ROM firmware originally loaded. This information can be obtained from an INSPECT or Survey Report.

• Set configuration switches 1, 4, 5, and 6 on systems with 6 switches, or switches 1,2,3 and 4 on systems with 4 switches on the system maintenance switch block to the ON position. Setting these switches to the ON position enables Disaster Recovery Mode on older ProLiant servers.

|[pic] |

|NOTE: Configuration Switches |

|Refer to the server’s documentation and hood labels for the location of the configuration |

|switches.  |

|[pic] |

6. Insert the ROMPaq diskette into the diskette drive.

7. Power up the target server. The keyboard, mouse, and monitor will all be inactive.

8. When the server has powered up, the server will make two long beeps to indicate that it is in Disaster Recovery Mode. For systems with the Integrated Management Display, check the IMD display for:  "Insert ROMPAQ diskette".

9. Insert the System ROMPaq diskette in the server's diskette drive. Check the IMD display for: "ROMPAQ flashing ROM". Then, the server will read the ROMPaq diskette for the latest ROM image and will proceed to flash the system ROM on the target server. Flashing the System ROM should take a few minutes to complete. The procedure will take longer to complete on servers with Redundant ROMs since it is flashing two images; the primary and redundant.

10. Listen for a sequence of ascending (increasing) tone beeps after flashing is complete. Check for a message in the IMD that will indicate a successful operation is completed. The diskette drive may appear to still be in use, but the procedure is complete per the IMD messages and ascending tone beeps.

If the flashing fails, check for a message in the IMD display indicating the failure, and listen for a series of descending beeps will be generated. A second attempt should be made just in case the procedure should complete on that iteration. A quick series of beeps indicates successful completion of the system ROM recovery process.

11. When the disaster recovery process has successfully completed:

a. Remove the ROMPaq diskette from the drive and power down the server.

b. Return configuration switches 1, 4, 5, and 6 to their original positions.

c. Reassemble the server chassis.

d. Power up the server as usual.

11. If the Disaster Recovery procedure can not be completed and before dispatching the System or Peripheral Board depending on the server, components like the first bank of memory, the processors and associated hardware such as PPMs, Processor Power Modules, and Terminator Boards, and the power supplies should be eliminated as much as possible to causing or masking the problem. Remember to remove unnecessary and redundant hardware from the server that is not required to complete POST. Some of these components include multiple banks of memory, multiple processors, any PCI device, and redundant power supplies and PPMs. These extra components become available as on-site spares.

Top of page

Switch Settings

The following table provides a summary of the locations of the Switch Packs on older models of ProLiant ML and DL servers and what switches need to be set to the ON position in order to invoke the System ROMpaq Disaster Recovery Procedure. The above ROM upgrade information about this procedure is found in the "Using Compaq ROM Flash Technologies to Upgrade System and Array Controller ROM Firmware User Guide".

The original System ROMpaq Disaster Recovery Procedure was written for the ProLiant 8000 and ProLiant 8500 model servers. The ML and DL series servers that have been announced since then also support the ROMpaq Disaster Recover procedure. Some of these servers require only a single switch to be set to the ON position. Later models with redundant ROMs do not require any system board switch settings.

 

|Server Model |

|ProLiant ML330 |System Board |SW2 |3 |Configuration Page |

|ProLiant ML350 |System Board |SW1 |4 |Configuration Page |

|ProLiant ML370 |System Board |SW2 |1, 4, 5, 6 |Configuration Page |

|ProLiant ML530 |System Board |Only Switch Pack on |1, 4, 5, 6 |Configuration Page |

| | |Board | | |

|ProLiant ML570 |Peripheral Board |SW1 |1, 4, 5, 6 |Configuration Page |

|ProLiant DL Models |

|ProLiant DL320 |System Board |SW2 |3 |Configuration Page |

|ProLiant DL360 |System Board |System Maintenance |1, 4, 5, 6 |Configuration Page |

|ProLiant DL380 |System Board |SW2 |1, 4, 5, 6 |Configuration Page |

|ProLiant DL580 |Peripheral Board |SW2 |1, 4, 5, 6 |Configuration Page |

Top of page

Recovery Summary

|[pic] |

|IMPORTANT: Switches 1, 4, 5, and 6 Must Be Set ON |

|If switches 1, 4, 5, and 6, are not set to the ON position and other switches are inadvertently |

|set instead, the System ROM recognizes each individual switch setting as valid operations as |

|described above. It is only when these specific switches are set does the System ROM invoke the |

|ROMpaq Disaster Recovery procedure. It is not likely that this specific switch configuration will|

|ever be set during normal operations. Without the ROMpaq Disaster Recovery correctly configured |

|from the servers' Configuration Switch Pack, the System ROM will proceed as it would during a |

|normal boot sequence honoring each of the functions of the switches that are set.  |

|[pic] |

After power up, the BootBlock determines whether the switches on the Configuration Switch Pack are correctly set. Once the BootBlock reads the correct switches are configured, it will generate two long beeps indicating the procedure has been invoked. The second step occurs when the system verifies that a valid System ROMPaq diskette has been inserted and if so will begin flashing the System ROM. Note that the BootBlock is never modified by the System ROMPaq or ROMpaq Disaster Recovery procedure. It should not need to be modified in the Field.

 

|[pic] |

|IMPORTANT: Some Servers Don't Have Protected Boot Blocks |

|Certain models of servers (for example, the ML 330 and ML 350) do not have protected boot blocks.|

|If the boot block becomes corrupted, then it is possible that the  procedure will not recover the|

|server from a catastrophic event. Therefore, the physical ROM part(s) would need to be replaced. |

|[pic] |

Note that on the ProLiant 8500, the Configuration Switch Pack is located on the I/O Board to the left of the fan assemblies. On the ProLiant 8000, this switch pack is located on the Processor Board below the Processors. Check for the location of the switch pack since it may vary from server to server. For example, on the ProLiant 6000, 6500, and 7000 Xeon models, it is located on the Standard Peripheral Board. It is also important to verify what PCB the System ROM is located on.

On servers with Integrated Management Displays or IMD, appropriate messages will be displayed indicating the progress of the procedure and if user intervention is required. In servers like the ProLiant 8000 and 8500, the IMD will prompt the user with the message, "Please insert ROMPAQ diskette". When the server begins flashing the ROM, the IMD will display, "ROMPAQ flashing System ROM". If the procedure is successful, the IMD will display, "ROMPAQ flash successful, power cycle the System." There will also be a series of audio tone beeps generated during the procedure. In models without the IMD, these audio signals will help indicate the procedure's progress.

If the diskette is bad or the System ROMPaq diskette is corrupted, a failure message will be displayed on the IMD and the system will continue to beep. These beeps will continue until a valid System ROMPaq diskette is detected. Once flashing is complete, a message will be displayed on the IMD indicating a successful operation and an ascending series of tone beeps will be generated. If the flashing operation is unsuccessful, a descending series of tone beeps will be heard.

Top of page

New Redundant System ROM Models

In servers with Redundant System ROMs like the ProLiant 8000 and 8500 and later ML and DL models, the ROM is divided into two logical parts. When the system boots, the primary ROM side is executed and used in server operation. During a ROM flash, the backup image is flashed. Once the flash is fully completed, the backup image becomes the primary ROM image. If under rare conditions your flash does not complete safely, potentially due to interruption of power, the backup ROM image will be available to boot the system. A "POST 105 - Current System ROM is corrupt ... now booting redundant System ROM" message is displayed if this occurs. Therefore a corrupt primary System ROM image is automatically corrected by the loading of a redundant image during boot time. In the situation where both images are valid, the user can select which image to use at boot time in RBSU.

Another feature of the Redundant ROMs is the Remote-Flash Redundant ROM. This management feature lets you upgrade your ROM remotely with confidence. When your ROM is flashed from a remote site, a backup ROM is always stored on the server. When a new version is loaded, it replaces the older of the two versions; if the ROM flash fails, the server can re-start using the previous copy of the ROM.

If the bootblock determines that both the primary and redundant images of the ROM are corrupt, then the ROMpaq Disaster Recovery procedure needs to be applied only once since the primary and redundant images are both flashed during the procedure.

 

|[pic] |

|IMPORTANT: When To Perform ROMpaq Disaster Recovery Procedure |

|The ROMpaq Disaster Recovery Procedure should not be performed on these servers unless both |

|primary and redundant images are suspected as being corrupt.  |

|[pic] |

A corrupted system ROM requires that you recreate the ROM BIOS informatin by a process called flashing the ROM. This operation can be accomplished only when the system is in disaster recovery (emergency repair boot) mode. If both sides of the redundant ROM are corrupted, the system automatically enters the disaster recovery mode. The server emits three extended beeps. There will be only one sequence of tone beeps even though two flashing operations are involved. This is unlike the System ROMPaq Utility that needs to be applied twice to recover from a corrupted primary and redundant image.

To perform ROMPaq disaster recovery:

1. Power on the server.

2. After hearing the three beeps, insert a ROMPaq diskette with the latest system ROM from SmartStart CD.

3. Wait until the server emits three rising beeps, indicating the completion of the ROM flash process.

4. Cycle the system power.

|[pic] |

|IMPORTANT: Flashing both System ROMs |

|The ROMPaq flashes both sides of the redundant ROM, and this process may take up to 10 minutes. |

|[pic] |

|Server Model |Redundant ROM |Recovery Procedure |

|ProLiant ML Models |

|ProLiant ML330 G2 |Yes |Automatic: No system board switches need to be set. |

| | |If both sides of the redundant ROM are corrupted, the|

| | |system automatically enters the disaster recovery |

| | |mode. The server emits three extended beeps. Use |

| | |steps above to reload new ROM image. |

|ProLiant ML330 G3 |Yes | |

|ProLiant ML350 G2 |Yes | |

|ProLiant ML370 G2 |Yes | |

|ProLiant ML370 G3 |Yes | |

|ProLiant ML530 G2 |Yes | |

|ProLiant DL Models |

|ProLiant DL360 G2 |Yes |Automatic: No system board switches need to be set. |

| | |If both sides of the redundant ROM are corrupted, the|

| | |system automatically enters the disaster recovery |

| | |mode. The server emits three extended beeps. Use |

| | |steps above to reload new ROM image. |

|ProLiant DL360 G3 |Yes | |

|ProLiant DL380 G2 |Yes | |

|ProLiant DL380 G3 |Yes | |

|ProLiant DL760 |Yes | |

Questions and Answers

Q1. What is the System ROMpaq Disaster Recovery Procedure?

Q2. When should the System ROMPaq Disaster Recovery Procedure be Invoked?

Q3. What are some possible causes of System ROM corruption?

Q4. What are some possible symptoms of System ROM corruption?

Q5. Where do I obtain the required System ROMPaq diskettes?

|Q1. What is the System ROMpaq Disaster Recovery Procedure? |

|A1. The System ROMpaq Disaster Recovery is a procedure that can be invoked in an attempt to clear suspected corruption in the |

|server's System ROM. |

|  |

|[pic] |

|IMPORTANT: Problems Caused By Hardware Faults |

|This procedure will not bypass or correct problems caused by hardware faults with the System ROM and supporting components.  |

|For older system models the procedure requires a special system configuration that is set up from the switches on the server's |

|Configuration Switch Pack; 1,4, 5, and 6 switches set to the ON position. The server's System ROMPaq diskette must be inserted in|

|the server's diskette drive when prompted during the procedure. A hard reboot (cycling power to the server) is the final step to |

|the procedure. It is important to use a known good or recently created diskette from the server's System ROMPaq Utility. There |

|are many System ROMPaqs for the servers in Compaq's product line. Be sure to use the one for the server being serviced. |

|[pic] |

| |

|  |

|Q2. When should the System ROMPaq Disaster Recovery Procedure be Invoked? |

|A2. There are two conditions when the procedure should be invoked. |

|The first is when power is lost to the server when applying the System ROMPaq. If the procedure does not resolve problems after |

|power is lost, then the System Board or PCB, Printed Circuit Board, that the System ROM is located on will need to be replaced.  |

|The second is when intermittent and inconsistent error symptoms occur during POST. The System ROM manages operations such as |

|initializing video for output to the console monitor, counting memory, initializing processors, and preparing the server for |

|loading the Operating System. If the System ROM is corrupted, then POST may likely fail. The System ROMPaq Disaster Recovery |

|Procedure can be invoked in an attempt to fix these problems. |

|  |

|Q3. What are some possible causes of System ROM corruption? |

|A3. The most common cause of System ROM corruption occurs when power is lost to the server during a System ROMPaq upgrade or |

|flashing operation. Flashing is a special transfer of data from one medium such as a floppy diskette to an integrated device such|

|as memory. The reference to Read Only Memory is used loosely in this context since there are special areas of memory associated |

|with the System ROM that are writeable.  |

|  |

|[pic] |

|IMPORTANT: System ROMpaq Upgrade Diskette May Be Corrupt |

|The System ROMPaq upgrade diskette should also be considered corrupt after an event like a power loss during the System ROMPaq |

|upgrade or the ROMpaq Disaster Recovery procedure. A new upgrade diskette should be created since it will be required by the |

|procedure. A second attempt at flashing the System ROM should be attempted after a power loss, but the chances of successfully |

|completing are not good.  |

|Other possible causes of corruption are hardware intermittent or solid failures in the System ROM itself and supporting chipsets,|

|a corrupt System ROMPaq upgrade diskette, a virus infection, or a manufacturing defect in the System ROMPaq Utility.  |

|[pic] |

| |

|  |

|Q4. What are some possible symptoms of System ROM corruption? |

|A4. One of the most likely symptoms of System ROM corruption are intermittent or inconsistent symptoms occurring at POST. These |

|tests may unexpectedly hang, abort, or error messages may be inconsistent. These symptoms could also occur when other components |

|necessary to complete POST such as Memory, the Processor Power Modules, NVRAM, and Power Supplies are defective.  |

|Note that the application of the System ROMPaq Utility is a test in itself since it uses much of the same hardware used by the |

|System ROMpaq Disaster Recovery Procedure. If a clean pass of the System ROMPaq upgrade can not be completed, then the ROMpaq |

|Disaster Recovery Procedure should be invoked.  |

|One aspect about firmware or software corruption is that it is very difficult to determine where problems will occur without |

|knowing exactly where and how much corruption has occurred. To make matters more complicated, there may already be hardware |

|failures occurring in the server that created the service call.  |

|In the event that corruption has occurred in areas of the firmware accessed by the Operating System, it is possible that |

|erroneous status and error conditions could be reported by the OS.  |

|Once the F9 or F10 prompt is displayed on the console monitor, the System ROM's control of the system is nearly complete.  If F9 |

|is selected, then system control is still managed by the System ROM. If F10 is selected, control is passed to the System |

|Partition / System Configuration Utility. If the approximately 5 second time out window to select the F9 or F10 options during |

|the later phase of POST is allowed to expire, then system control is passed to the Operating system. |

|Q5. Where do I obtain the required System ROMPaq diskettes? |

|A5. The only software that is required for flashing the system ROM firmware is the individual System ROMPaq Flash executable |

|file, specific to your model of server and ROM family. However, if remote ROM upgrading capabilities are desired, Smart |

|Components for Remote ROM Flash are designed to integrate seamlessly with the Compaq Remote Deployment Console Utility for |

|Windows NT 4.0 and Windows 2000.  |

|  |

|[pic] |

|NOTE: Compaq Remote Deployment Console Utility |

|The Compaq Remote Deployment Console Utility is part of the Compaq Support Paq for Windows NT 4.0 or Windows 2000. Smart |

|Components for Remote ROM Flash and the Compaq Remote Deployment Console Utility can be obtained from the following sources:  |

|Compaq Website |

|The latest deployment utility for Windows NT 4.0 and Windows 2000 and individual ROM flash components are always available on the|

|Compaq website.  |

|Compaq ActiveUpdate |

|The latest deployment utility and individual ROM flash components for Windows NT 4.0 and Windows 2000 are also available from |

|Compaq ActiveUpdate. Compaq ActiveUpdate is a Web-based client application that eliminates time wasted searching the Web for the |

|latest server updates by proactively delivering updates to a software repository defined by the Compaq ActiveUpdate user.  |

|[pic] |

| |

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download