Troubleshooting Suggestions

• System won't power on:
  • Ensure that system is plugged into a live wall outlet.
  • Ensure that both fans are plugged in and operating properly. A non-operational fan will cause automatic OS shutdown (All OS's) and can prevent power on.
  • Ensure that power connectors are firmly seated in the MLB.
  • Ensure that the CPU module is fully seated in the MLB connector.
  • Ensure that the power switch connector is plugged in.
  • Look for shorts or overcurrent if power won't stay on (bad hard drive, etc).
  • Wait two or more seconds after plugging unit in before powering on.
• System hung at power on:
  • Consult LED code charts and try suggested fixes.
  • 1.44 MB diskette light illuminated indicates possible firmware corruption. Follow instructions for creating firmware update disk, insert disk, power cycle system, update firmware to repair system.
• Memory problems:
  • DIMMs ignored by system, or system unstable - hangs or crashes
  • Ensure that each memory bank has four identical DIMMs installed.
  • DIMMs may be failing memory POST testing, try another set.
  • DIMMs may not have ECC bits:
    All Compaq AlphaStation XP1000 systems need ECC capable PC100 buffered DIMMs.
    Look for 72-bit rather than 64-bit DIMMs.
  • Some third party DIMMs may not be compatible with Compaq AlphaStation XP1000 systems:
    • POST checks 24 SDRAM DIMM performance parameters per DIMM to ensure system reliability.
    • All Compaq AlphaStation XP1000 system qualified DIMMs meet performance requirements.
    • SDRAM DIMMs come in many performance grades which do not all meet the performance requirements of these machines.
  • Suspect system is taking ECC errors - several cases:
    • ECC error is a hard single-bit error on a particular DIMM.
      In a data block -- results in noticeable performance degradation.
      In code stream -- system will appear hung or extremely slow.
    • ECC error is a soft error or only affects a few locations -- results probably won't be noticed except in error logs.
    • Multi-bit ECC error will crash or hang system.
  • On NT systems, try running SRM diagnostics to isolate bank and/or DIMM.
  • Ensure that memory is qualified, and check field blitz data to ensure that DIMM was not from a bad lot.
• PCI parity errors causing OS problems:
  • All PCI devices must correctly generate PCI parity to enable checking feature.
  • Parity errors indicate that a PCI option card is probably at fault.
  • Card may be defective or not fully PCI compliant.
  • If card must be used, disable PCI parity checking in OS firmware.
  • If problem not specific to option cards, replace CPU card or MLB.
• SCSI problems: (Data corruption, problems booting, poor performance)
  • Check SCSI bus termination.
    • Bus is terminated at MLB connector.
    • Bus must be terminated at physical cable end.
    • No terminators in between.
    • Old 50 Pin (narrow) devices must be connected with wide to narrow adapter (SN-PBXKP-BA). Do not cable from connector on card (if used)
    • 50 pin devices on the bus may significantly degrade performance.
  • Any external drives MUST be connected to their own card which has no internal drives connected to it.
    • SCSI has strict bus length requirements.
    • SCSI bus itself can't handle internal plus external cable.
    • Use a separate card for external devices - and terminate properly.
• Windows NT specific problems:
  • IDE / ATAPI hard drives are not supported on XP1000's.
  • Windows NT will not properly boot:
    • Ensure that no boot HW changed (storage, input, display) from last good boot.
    • Ensure that all onboard and installed PCI options and disks are visible in AlphaBIOS display configuration menus.
    • If an option or storage device is suspected bad, try reseating or recabling.
    • Try booting with last known good registry.
    • If that fails, enter BIOS set-up mode, utilities/os selection, edit boot selection.
    • In 'OS Options' type: 'sos' - to force NT to display a verbose boot.
    • See if boot hangs on a specific driver.
  • Boot error suggesting invalid selection, missing boot device, missing or bad files.
    • Can result from firmware update, NVRAM corruption, juggling SCSI IDs.
    • Ensure that HAL is compatible with AlphaBIOS and NT version.
    • View hard disk config, use utilities/os_selection to edit and restore boot selection based on best guess or previously recorded values, validate.
  • Suspected bad video driver
    • In boot entry, add 'BASEVIDEO' to 'OS Options:' .
    • NT will be forced to boot to VGA mode using this selection.
    • If boot is successful, fix video driver or hardware, remove 'BASEVIDEO' option and try again.