Solving undetermined problems

If the diagnostic tests did not diagnose the failure or if the server is inoperative, use the information in this section.

If you suspect that a software problem is causing failures (continuous or intermittent), see Software problems.

Damaged data in CMOS memory or damaged server firmware can cause undetermined problems. To reset the CMOS data, use the CMOS jumper to clear the CMOS memory and override the power-on password; see System-board switches and jumpers. If you suspect that the server firmware is damaged, see Recovering the server firmware.

If the power supplies are working correctly, complete the following steps:

  1. Turn off the server.
  2. Make sure that the server is cabled correctly.
  3. Remove or disconnect the following devices, one at a time, until you find the failure. Turn on the server and reconfigure it each time.
    • Any external devices.
    • Surge-suppressor device (on the server).
    • Printer, mouse, and non-IBM devices.
    • Each adapter.
    • Hard disk drives.
    • Memory modules. The minimum configuration requirement is 2 GB DIMM in slot 3.
    The following minimum configuration is required to start the server:
    • One microprocessor in socket 1.
    • One 2 GB DIMM for each microprocessor installed (one 2 GB DIMM in slot 3 if only one microprocessor installed).
    • One power supply.
    • Power cords.
    • Five cooling fans (fan 1, 2, 3, 4, and 5).
    • One PCI riser-card assembly in PCI connector 2.
    • One ServeRAID SAS/SATA controller.
  4. Turn on the server.

If the problem is solved when you remove an adapter from the server but the problem recurs when you reinstall the same adapter, suspect the adapter; if the problem recurs when you replace the adapter with a different one, suspect the riser card.

If you suspect a networking problem and the server passes all the system tests, suspect a network cabling problem that is external to the server.