15.5. Troubleshooting a Failed Installation

Not all upgrades go as planned. For example, unforeseen software incompatibilities and user error may cause an unsuccessful installation. It is important to gather all relevant system information and follow all instructions when planning and installing additional hardware.

Other reasons why upgrades fail include the following:

• Hardware not supported

• Hardware problem (power, cable, and so forth)

• Software not supported

The first step to recovering from the failure is to review the system information and the installation instructions. Questions to ask include the following:

• Was the right part ordered for the computer?

• Were the hardware installation instructions followed correctly?

• Were all cables or accessories connected properly?

• Did the hardware activate at system power-on?

• Did the computer display an error during the POST process?

• Were the software installation instructions followed correctly?

• Were any known incompatibilities addressed?

• Did the operating system or software display an error?

You can use logic to eliminate possibilities. For instance, if all procedures were followed properly, but the device remains unresponsive or unacknowledged, it is likely that the part is defective. The operating system or software may or may not display an error if a part is faulty. Always check the documentation included with the part for help or information on any known issues or last-minute errata.

Replacing the newly installed but suspect part with another one is the easiest way to determine or verify where the fault lies. This is also true in most cases of an installed part that fails, such as a fan in a power supply. However, replacement is not always practical because of inventory costs or time constraints. In such cases, eliminating as many failure points as possible can expose the true cause of the problem.

A repeated failure in one area of the system can be a symptom of a problem in another area. For example, if PPMs repeatedly fail, it could be that there is nothing wrong with the PPM, but actually a problem with the processor board, power supply, or even line voltage.

15.5.1 Recovering from a Failed SmartStart Installation

Often, sudden power failures can be disruptive. Unless the system has power protection, some work may be lost or corrupted. On an existing system, this is an inconvenience. When it happens during a SmartStart installation, it can be a much bigger problem. A power failure is one way the SmartStart process can fail.

Recovering from a failed SmartStart installation depends on the following:

• The installation path chosen

• The operating system being installed

• The point in the process at which the interruption occurs

An interrupted manual installation of Linux, for example, must be restarted from the beginning. A complete restart is necessary because in manual installations there is no SmartStart involvement after the hardware is configured until after the operating system installation is complete.

If the interruption occurs during the early phase of an attended installation, where SmartStart is copying files to a temporary partition on the hard disk, the whole process also must be restarted, beginning with a system erase, because not all the drivers or operating system setup files have been copied. This is true for both Windows and NetWare at this stage.

If the interruption occurs after SmartStart has finished the file copying process and the server has restarted, the operating system might be able to pick up from where it left off, especially if it is a Windows product. If recovery is not possible, the whole process must be restarted, beginning with a system erase.