SCSI Adapters - Problem determination

10.310 bytes

Service Hints & Tips

Document ID: VLAR-3WGNH9

SCSI Adapters - Problem determination

Applicable to: World-Wide

The SCSI (Small Computer Systems Interface) is not a fixed-disk controller. It is a channel attachment that meets industry standards and it is capable of controlling any device that meets the SCSI signal protocol.

In troubleshooting problems, when the SCSI device(s) will not configure or 'come ready', there are a number of possible causes. The following tip will help provide a basic understanding of SCSI operation and will also be a useful aid in problem determination.

Configuration
For a SCSI device to 'come ready' and be configured, the following events must occur:

1.

Upon Power-on Reset, SCSI Post issues an Inquiry command to each SCSI address - to poll all devices present.

2.

All SCSI devices attached to the SCSI Bus must answer with their inquiry data.

3.

A Test Unit Ready command is issued to all devices that responded to the Inquiry command.

4.

If a hard disk drive answers 'Not Ready', a Start Unit command is then issued to sequentially start the drive (Motor Start). The devices must then run 'Self Test' and report their status.

5.

Once the drive is ready, the Read Device Capacity command is issued - to determine the device size for example 320 MB, 1GB, and so on).

6.

All responding devices are compared with the configuration data.

7.

PS/2 Post beeps twice and displays errors if there are any new, missing, or defective SCSI devices.

8.

Following a successful Post (one beep), the system will attempt to load the operating system.

Procedure Following Configuration Errors
If SCSI Post or configuration errors occur, the following items should be checked:

1.

Run diagnostics - to test the equipment. If a failure is indicated, use the error code extension to isolate the cause. Complete SCSI device error codes are 12 characters long.

Note: The first eight characters of the SCSI error code are for device information - not error information. The error information is in the next four characters (error extension bytes). Failure to use the entire error code may result in incorrect FRU parts replacement.

2.

Each device must be set for a unique SCSI ID via jumpers, switches, or remotely with a cable-connected switch. Run 'Set and View SCSI Configuration' from the Configuration menu - to configure new or powered-off devices.

3.

Check that all devices are receiving power:

a.

Is the 'Power Good' (green) light illuminated on the system unit?

'Power Good' is more complex than a 'Power On' light. It indicates that the power supply is delivering all the correct dc output voltages within specifications (that is, +5 Vdc for logic circuits and +12 Vdc for motor circuits, though some SCSI fixed disks may also use +5 Vdc). Any missing or incorrect voltage will cause the system to shut down (self-protect).

b.

Is the power cable connected correctly (that is, not inverted)?

c.

Check whether any contacts have been inadvertently pushed back into the power cable connector. Examine all power cable connections and check whether any female contacts are distorted and thereby making intermittent contact. If so, reform them using the tip of a small screwdriver (or similar tool).

4.

Check that all other connections and cabling are correct.

a.

Is the SCSI adapter fully seated? Tension on the external SCSI interface cable may cause the SCSI adapter to partially unplug (unseat) from the system board. (Systems jammed against a wall are subject to this.)

b.

Are the cables fully seated?

5.

It is critical that the SCSI Bus is terminated properly.

a.

The physical ends of the internal and external SCSI Bus must be terminated at the last device. Devices connected to the middle of the Bus MUST NOT BE TERMINATED.

b.

Terminator types include the following:
-Socketed DIP (Dual In-line Package) on the adapter.
-Connector plug on the adapter's internal or external connector or cable.
-Two or more socketed SIPs (Single In-line Packages) on a device.
-In-line terminator attached between the device and cable.
-Enabled via a jumper.

c.

The polarity of terminators must be correct:
-Pin 1 on DIPs and SIP are usually indicated by a dot or arrow.
-Pin 1 on sockets are usually indicated by a 1, dot, or square solder pad.

6.

Read errors may be caused by a bad or interrupted write operation if the drive was bumped or power failed during the write operation (sector incompletely written).

a.

If only the data is corrupted, rewriting the data will correct the error.

b.

If the error occurs in the directory or the operating system's file structure, it may be necessary to perform a high-level format (Format C), or even to re-create the partition with FDISK.

c.

If the sector ID is corrupted (ID or Sector Not Found errors), an advanced diagnostic format should be attempted prior to replacement of the fixed disk. The advanced diagnostic format will re-write all sector addresses and data fields. When the format completes, the diagnostics should be re-run. If the diagnostics then run error-free, it is unlikely that a hardware failure has occurred.

7.

If failures return after a particular software application has been loaded, it could mean that the software is corrupting the fixed disk. Assistance should be sought from the appropriate software source.

Note:
Failures may occur during system operation, yet diagnostics might run error-free. This could be due to improper termination. The terminator is considered to be part of the 'Cabling Being Intact' assumptions (described above).

Search Keywords

Hint Category

Error Messages, Performance

Date Created

17-07-98

Last Updated

27-01-99

Revision Date

26-01-2000

Brand

IBM Options

Product Family

SCSI Adapters

Machine Type

ISA, MCA, PCI, PCMCIA

Model

All

TypeModel

Retain Tip (if applicable)

Reverse Doclinks
and Admin Purposes

1.		Upon Power-on Reset, SCSI Post issues an Inquiry command to each SCSI address - to poll all devices present.
2.		All SCSI devices attached to the SCSI Bus must answer with their inquiry data.
3.		A Test Unit Ready command is issued to all devices that responded to the Inquiry command.
4.		If a hard disk drive answers 'Not Ready', a Start Unit command is then issued to sequentially start the drive (Motor Start). The devices must then run 'Self Test' and report their status.
5.		Once the drive is ready, the Read Device Capacity command is issued - to determine the device size for example 320 MB, 1GB, and so on).
6.		All responding devices are compared with the configuration data.
7.		PS/2 Post beeps twice and displays errors if there are any new, missing, or defective SCSI devices.
8.		Following a successful Post (one beep), the system will attempt to load the operating system.

1.		Run diagnostics - to test the equipment. If a failure is indicated, use the error code extension to isolate the cause. Complete SCSI device error codes are 12 characters long. Note: The first eight characters of the SCSI error code are for device information - not error information. The error information is in the next four characters (error extension bytes). Failure to use the entire error code may result in incorrect FRU parts replacement.
2.		Each device must be set for a unique SCSI ID via jumpers, switches, or remotely with a cable-connected switch. Run 'Set and View SCSI Configuration' from the Configuration menu - to configure new or powered-off devices.
3.		Check that all devices are receiving power:
	a.	Is the 'Power Good' (green) light illuminated on the system unit? 'Power Good' is more complex than a 'Power On' light. It indicates that the power supply is delivering all the correct dc output voltages within specifications (that is, +5 Vdc for logic circuits and +12 Vdc for motor circuits, though some SCSI fixed disks may also use +5 Vdc). Any missing or incorrect voltage will cause the system to shut down (self-protect).
	b.	Is the power cable connected correctly (that is, not inverted)?
	c.	Check whether any contacts have been inadvertently pushed back into the power cable connector. Examine all power cable connections and check whether any female contacts are distorted and thereby making intermittent contact. If so, reform them using the tip of a small screwdriver (or similar tool).
4.		Check that all other connections and cabling are correct.
	a.	Is the SCSI adapter fully seated? Tension on the external SCSI interface cable may cause the SCSI adapter to partially unplug (unseat) from the system board. (Systems jammed against a wall are subject to this.)
	b.	Are the cables fully seated?
5.		It is critical that the SCSI Bus is terminated properly.
	a.	The physical ends of the internal and external SCSI Bus must be terminated at the last device. Devices connected to the middle of the Bus MUST NOT BE TERMINATED.
	b.	Terminator types include the following: -Socketed DIP (Dual In-line Package) on the adapter. -Connector plug on the adapter's internal or external connector or cable. -Two or more socketed SIPs (Single In-line Packages) on a device. -In-line terminator attached between the device and cable. -Enabled via a jumper.
	c.	The polarity of terminators must be correct: -Pin 1 on DIPs and SIP are usually indicated by a dot or arrow. -Pin 1 on sockets are usually indicated by a 1, dot, or square solder pad.
6.		Read errors may be caused by a bad or interrupted write operation if the drive was bumped or power failed during the write operation (sector incompletely written).
	a.	If only the data is corrupted, rewriting the data will correct the error.
	b.	If the error occurs in the directory or the operating system's file structure, it may be necessary to perform a high-level format (Format C), or even to re-create the partition with FDISK.
	c.	If the sector ID is corrupted (ID or Sector Not Found errors), an advanced diagnostic format should be attempted prior to replacement of the fixed disk. The advanced diagnostic format will re-write all sector addresses and data fields. When the format completes, the diagnostics should be re-run. If the diagnostics then run error-free, it is unlikely that a hardware failure has occurred.
7.		If failures return after a particular software application has been loaded, it could mean that the software is corrupting the fixed disk. Assistance should be sought from the appropriate software source.

	Search Keywords
	Hint Category	Error Messages, Performance
	Date Created	17-07-98
	Last Updated	27-01-99
	Revision Date	26-01-2000
	Brand	IBM Options
	Product Family	SCSI Adapters
	Machine Type	ISA, MCA, PCI, PCMCIA
	Model	All
	TypeModel
	Retain Tip (if applicable)
	Reverse Doclinks and Admin Purposes