Testing Hardware

There are a few things you can do to test your hardware.

Testing RAM

A rather thorough test is possible with Memtest86. For ISG administrators this test is available via the net-boot menu.

Testing Harddisks

First of all you have to find out what type of harddisk you are using. When running Linux (e.g., Knoppix if you don't have a running Linux installation on the box) you can look it up in the /proc filesystem. E.g., for first IDE harddisk (hda, master on the first IDE port) look at the file /proc/ide/hda/model . You will find something like IC35L040AVVN07-0. Enter this model name in a Google search to find some data sheets and testing software made by the manufacturer (in the example mentioned, a IBM/Hitachi Deskstar 120GXP with 40GB).

Some often used test programs for harddisks:

  • Drive Fitness Test by IBM/Hitachi Global Storage Technologies. ISG administrators find it under ~isg/tests/Hardware-Tests or in the net-boot menu
  • SeaTools Diagnostic Suite by Seagate Technology. ISG administrators find it under ~isg/Hardware-Tests or in the net-boot menu
  • independent of the manufacturer: try to read the entire hard disk once (or a couple of times). E.g., start with Knoppix (local mirror) and use dd with a command similar to this: dd if=/dev/hda of=/dev/null bs=1024
  • some drive infos: smartctl -a -d ata /dev/discs/disc0/disc (ide=ata, scsi, 3ware)
  • You can also use the tool badblocks(8) to do either read-only or read-write checks.

Testing CPUs

  • cpuburn is a tool to stress the CPU. If lockups occur it may indicate problems with cooling or just sub-standard CPUs. See the cpuburn package for details.
  • WARNING: improperly cooled CPUs may be destroyed by these tests. Watch carefully and know what you are doing.
  • Note: observe the temperature (e.g. in /proc/acpi/thermal_zone/THRM/temperature). It should not go beyond 70 degrees Celsius
  • According to our Hardware supplier Dalco the CPU is OK when it runs k7burn for one hour.