Apple Network Server 500 and 700 Troubleshooting

Front panel controls and the LCD (Apple, Service Guide volume V, p. 220)

Fault diagnosis for the Apple Network Server 500 and 700 (Apple Network Server 500, Apple Network Server 700), built on Apple's symptom charts, the LCD start-up messages in the Service Source and the AIX display codes in Apple's AIX manual. Apple lists cures in order of likelihood: try the first, check whether the symptom persists, and refit the original module before trying the next. Where Apple says "replace logic board", it also says to keep the customer's DIMMs.

Tools for diagnosis

The LCD

The four-line LCD on the front shows the ROM version, the power supply status and the power-on self test. During start-up, progress messages appear on line 3. If the server stops with one of them showing, the modules named in Table 1 below are the ones in use at that moment; one of them may have failed, be badly seated or be incompatible. Problem messages appear on line 1. At the end of a good self test the LCD shows something like:

ROM ver.1.1.20.1
0048 MB Parity DRAM
150 MHz 604, 50 MHz Bus
1024 KB Level 2 Cache

During an AIX boot the LCD shows AIX's three-digit progress codes, which Kaiser notes are the same codes as IBM's own servers of the period use.

Diagnostic utility

Apple's Network Server Diagnostic Utility came on a floppy with every Network Server. It reports the hardware configuration and the results of the power-on self test, and tests memory, the Fast/Wide SCSI controllers, the PCI bridges, video and its memory, NVRAM and the LCD. It runs only on the Network Servers. Use it when AIX will not load, and after fitting new modules. Start it from Open Firmware with boot fd:diags. With the front key in the service position the server looks for a bootable CD and then for the file diags on a floppy, the name of the utility's program file.

Open Firmware

To reach the Open Firmware prompt, turn the front key upright, start or reset the server, and hold Command-Option-O-F until the 0 > prompt appears. Enter login and the Open Firmware password, then:

• probe-scsi0 lists devices on the external SCSI bus;
• probe-scsi1 and probe-scsi2 list devices on the first and second internal buses.

Each listing gives the SCSI ID, logical unit, device type, manufacturer, model and firmware version. A device missing from the list needs its termination, ID and cabling checked. Kaiser notes that the server prints further messages on a terminal connected to serial port 2 at 9600 bps during the self test.

Cuda and logic board resets

The Cuda microcontroller switches the power on and off, handles resets, keeps PRAM and the clock, and runs ADB. Press the red Cuda reset button on the logic board to reset it. If that does not help, reset the logic board, which Apple says to do before replacing any module on a unit that will not power up:

1. Unplug the server and slide the rear drawer out.
2. Prise open the latch at the end of the battery holder with a small flat-blade screwdriver, lift off the cover and take out the battery.
3. Disconnect any external SCSI or serial devices.
4. Wait at least 10 minutes, then refit the battery the right way round.
5. Reassemble and test. The reset clears PRAM, so check the date, time and other settings.

The Service Guide version adds pressing the power-on button with the power supply cable disconnected from the logic board before the wait.

LCD messages

Table 1: start-up progress messages (LCD line 3)

Message	Modules in use
DRAM test #1 Begins	DIMMs, logic board
ROM SIMM Test Begins	ROM SIMM, logic board, processor card
DRAM test #2 Begins	DIMMs, logic board, processor card
LONG DRAM test Begins	DIMMs, logic board, processor card
Turning on Caches	Cache DIMM, processor card, logic board
Jumping to RAM Prog	DIMMs, logic board
Testing Parity DIMMs	DIMMs, logic board
MainLBU Enet Setup	Logic board
Sounding Boot Beep	Logic board
Sizing RAM DIMMs	DIMMs
ROM SIMM Data Access	ROM SIMM
Allocating RAM DIMMs	DIMMs, logic board, processor card
MainLBUNVRAM Setup	Processor card, logic board
CPU Card Info Setup	Processor card, logic board
L2 Cache SIMM Setup	L2 cache SIMM, processor card, logic board
Testing L2 Cache SIMM	L2 cache SIMM, processor card, logic board
RAM/ROM/NVRAM: PASSED	(tests complete)

Table 2: problem messages (LCD line 1)

Message	Module
L2 Cache SIMM Failed	Cache DIMM
ROM SIMM Failed	ROM SIMM
MainLBUNVRAM Failed	NVRAM on the logic board
RAM DIMM 1A failed at Address xxxxxxxx	The DIMM named (1A to 4A, 1B to 4B); the address is in hexadecimal
MainLBUVideo ID Bad	Video on the logic board
MainLBU825#1 Failed	SCSI controller 1 on the logic board
MainLBU825#2 Failed	SCSI controller 2 on the logic board
Drive Fan Failed!	The fan on the right, facing the rear of the server
Processor Fan Failed	The fan on the left, facing the server
Temperature Warning!	Cabinet temperature is becoming unsafe
Temperature Too Hot!	Cabinet temperature is unsafe: shut the server down at once
Left Power Fail! / Left Power Hot!	Power supply on the left, facing the front
Right Power Fail! / Right Power Hot!	Power supply on the right, facing the front (700 only)
CudaNotResponding!!!	Logic board
ParityAddrAtAddrFail	Logic board

During the long memory test the server adds a dash to line 4 for each DIMM slot it tests. If it freezes, count the dashes:

Table 3: dash count during the long DRAM test

Dashes	Slot	Dashes	Slot	Dashes	Slot
1-2	Logic board DRAM (reserved)	7	2A	13	3A
3	1A	8	2B	14	3B
4	1B	9	2A	15	4A
5	1A	10	2B	16	4B
6	1B	11	3A	17	4A
		12	3B	18	4B

Kaiser adds one condition that Apple's tables do not cover: a backlit LCD showing only blank or dark blocks means enough of the logic board is working to light the front panel but the processor is not running. The usual cause is a processor card that has worked loose, then a failed card, and less often a badly seated ROM SIMM. On his 700, reseating did not help and a different processor card brought it up.

AIX display codes

Once the firmware hands over to AIX, a three-digit code on the LCD shows how far the boot has got. Apple's AIX manual lists these values for the Network Server:

AIX boot display values (Using AIX, p. 161)

Value	What the system is doing
510	Configuration manager started
511	Configuration manager completed
517	Mounting the / (root) and /usr file systems; changes to 518 if the mount fails during a network boot
585	Configuring the network
591	Configuring the Logical Volume Manager
594	Configuring asynchronous I/O
704	Configuring the 2.2 GB 16-bit SCSI disk drive
723	Configuring an unknown CD-ROM or SCSI device driver
727	Configuring an asynchronous device
731	Configuring an unknown PTY
811	Configuring the processor complex or standard I/O
812	Configuring memory
814	Configuring NVRAM
821	Configuring the standard keyboard adapter
823	Configuring the standard mouse adapter
828	Configuring the standard floppy disk adapter
831	Configuring serial port 2
868	Configuring the integrated SCSI adapter
871	Configuring the graphics subsystem adapter
921	Configuring a keyboard
925	Configuring a three-button mouse
935	Configuring a 3.5-inch diskette drive
9002	Configuring the 825 SCSI card

If the boot stops on one of these, the device named is the one to run diagnostics on. The manual gives fixes for the stop codes:

AIX stop codes and Apple's fixes

Code	Meaning	Fix
522	/etc/inittab damaged or wrongly edited	Boot into maintenance mode and repair the file or copy it from the installation media
551	/dev missing or damaged, bad boot logical volume, or /etc problems	mkdir /dev; bosboot -a; restore files or reinstall; then fsck hdiskn on the root volume group disk
552	Infinite loop	Run fsck, then reformat the log volume with /etc/aix/logform /dev/hd8 and restart
553	/etc/inittab cannot be run or read, usually because /tmp or / is full	Boot into maintenance mode, check space with df, delete files such as /smit.log and /smit.script, check /etc/inittab and the permissions of /etc/environment, /bin/sh, /bin/bsh, /etc/fsck, /etc/profile and /.profile
c31 or c32	Looking for the system console	Check the cables; then from maintenance mode enable the console with chcons -a login=enable /dev/hft/0

Maintenance mode

1. Put the AIX installation CD or a bootable backup tape in the drive and shut down with shutdown -F.
2. Turn the front key to the left (service) position and switch on.
3. Press F1 and Enter to use the local display as the console. An extended keyboard is needed unless a terminal is connected to the modem port.
4. Choose the language, then option 3, "Start maintenance mode for system recovery", then 1, "Access a Root Volume Group".
5. Choose the volume group and file system, and select Mount File System and Start Shell.

Return the front key to the upright position before restarting normally. Kaiser warns that the Apple version of AIX must be repaired or installed from Apple's own CDs: installing base packages from a standard IBM AIX CD is likely to corrupt the installation.

Power

Power symptoms (Apple's cures, in order)

Symptom	Cures
Does not power up: screen blank, fan not running, power LED not lit	Check the power cord, any power strip and the wall supply (plug the monitor straight into the wall to test it); check the rear drawer key is horizontal (locked); check the power supply is fitted and seated; check the power LED; check the rear drawer is fitted; check the toggle bolts on the rear of the logic board are fully tightened; check that the frame is not bent so that the micro switch at the rear does not make contact (it sits near the lower left edge of the logic board face plate, visible through the grille, where a plastic protrusion fits into a slot in the chassis); check the metal ground strip on the logic board; reseat the processor card, other PCI cards and cache DIMM; test the DIMMs one at a time; try the keyboard power key (if the server starts, try the front switch to turn it off, and if it will not power down replace the NMI reset switch cable; if it does not start, replace the power supply); reset the Cuda chip; reseat the logic board so its interconnect is fully engaged; reset the logic board; then replace the power cord, rear micro switch, power supply, on/off switch, front keyswitch, powerplane interconnect board and logic board
Shuts down intermittently	Check the power cord; check the fans; check every front drive tray is fully in (a loose tray disrupts air flow); clear the vents (the thermal protection circuit should let it restart after 30 to 40 minutes); run the diagnostic utility; check the battery; reset the Cuda chip; reset the logic board; then replace the power cord, power supply, powerplane interconnect board, logic board and processor card
Clicking, chirping or thumping	Check the fans are not loose (they can be replaced with the server running); check the power supply is seated; check the front drive trays are fully in; check the hard drives, DAT drive and floppy drive; then the logic board and processor card. Apple notes that a noisy fan may be more annoying than a cause for concern.

The clock battery should read 3.2 V or more. When the rear panel has been off, check that the fan-to-SCSI-backplane cable went back on the right connector: the rear drives' power connector is similar, and with the fan cable on it the server will not power on.

Start-up and system

Start-up and system symptoms (Apple's cures, in order)

Symptom	Cures
Cannot boot from the hard drive	Check AIX is installed on the drive and has booted from it before; check the Open Firmware boot path names the right drive; read any LCD problem message or three-digit code (above); run the diagnostic utility; for a boot drive in the 700's rear bracket, check its connections and termination, then replace the rear drives' SCSI cable, SCSI ID cable and power cable, the power backplane-to-SCSI backplane cable, the SCSI backplane, the power backplane and the drive; for a front drive, move it to another bay (the boot path may need resetting), then replace the 68-pin SCSI cable, SCSI ID cable, power cable, SCSI backplane and drive; then the logic board and processor card
Long DRAM test never completes	Check every DIMM is properly seated
Will not boot; LCD lines 1 and 2 report a memory failure at address ECCBEBAD	Check the DIMM specification: ECC DIMMs with non-quad CAS logic are not supported
No beep at start-up, or when the AIX "beep" command is used	Disable the ROM checksum in the power-on self test (Apple Technical Information Library article 19814, "Network Server 500 or 700 No Sound at Startup or in AIX")
Crashes or hangs intermittently	Check the power cord, power supply, rear drawer and front drive trays are seated; check AIX is 4.1.4 or later; run the diagnostic utility; check the DIMMs are fast-page mode, 60 ns or faster; reseat the processor card and cache DIMM; test the DIMMs one at a time; remove all PCI cards and refit them one at a time; then the logic board and processor card
System inactive	Check the power LED; look for a three-digit code on the LCD; check the cables; adjust the monitor brightness; press Control-D or Control-C to cancel a stalled process; check the rear drawer key is horizontal; restart; run the diagnostic utility

Kaiser's 500 kernel-panicked with a Sonnet level-2 cache upgrade and was stable with a 1 MB cache DIMM taken from a 700. Apple also notes that some cards had residue on their gold edge connectors that causes intermittent faults; clean the pins gently with a pencil eraser.

Video

Video symptoms (Apple's cures, in order)

Symptom	Cures
Screen black; drive operates and fan runs	Use an SVGA or multiple-scan monitor, and an extension cable rated for multiple-scan monitors; adjust the brightness; check the monitor cable; check whether the server is set to use a serial terminal as its console; run the diagnostic utility; reset the Cuda chip; reset the logic board; try a known-good monitor; then the monitor cable, logic board and processor card
Screen lights but shows nothing	SVGA or multiple-scan monitor and rated extension cable; reset the Cuda chip; reset the logic board; known-good monitor; run the diagnostic utility; then the monitor cable, logic board and processor card
Objects too large or distorted	Adjust the display resolution in AIX

The built-in video supports multisync displays only, at 640 × 480, 800 × 600 and 1024 × 768.

Drives

Drive symptoms (Apple's cures, in order)

Symptom	Cures
One front-bay hard drive does not work or spin	Check the tray is fully in; check the drive appears with probe-scsi or the diagnostic utility; check the SCSI devices and PCI cards are within the power allowance; run the diagnostic utility; move the drive to another bay; check its SCSI, SCSI ID, LED and power cables; then replace those cables, the drive and the SCSI backplane interconnect board
No front-bay hard drive works	Check every tray is fully in; check the drives appear on the bus; check the power allowance; run the diagnostic utility; check the power backplane-to-SCSI backplane cable and the 10-inch and 20-inch mezzanine-to-SCSI backplane cables; then replace those cables, the SCSI backplane interconnect board, the powerplane interconnect board, the logic board and the processor card
Drive or drives in the rear bracket do not work (700 only)	Check for duplicate SCSI IDs; check the drives appear on the bus; check the power allowance; check the rear drives' SCSI, SCSI ID and power cables; check that the last drive on the cable is terminated and that the bottom drive has both the SCSI ID cable and the last connector (a single drive goes on the last connector); check the backplane cables; then replace the cables, the SCSI backplane interconnect board, the drive, the powerplane interconnect board, the logic board and the processor card
Floppy drive does not work	Try a known-good disk; check the mezzanine-to-floppy cable and that the rear drawer is fitted; then the floppy drive, the cable, the mezzanine interconnect board and the logic board
Floppy disk will not eject	Push an opened paper clip into the hole beneath the drive; reseat the bezel and drive so the slot lines up; check the rear drawer; then the floppy drive, cable, mezzanine board and logic board
CD-ROM drive does not work	Try a known-good disc; reseat the drive or move it to another bay; check it appears on the bus; check the power allowance; run the diagnostic utility; check its cables; then the drive, the SCSI backplane board and the logic board
DAT drive does not work	Use DDS-2 tapes; try a known-good tape; move the drive to another bay; check it appears on the bus; check the power allowance; run the diagnostic utility; check its cables; then the drive, SCSI backplane board and logic board
DAT self-test fails, right light flashing amber	Clean the tape heads
Amber light flashes when a tape is loaded	The tape is worn out: copy the data to a new tape
DAT self-test fails, right light solid amber	Move the drive to another bay; then the drive, SCSI backplane board and logic board

Apple's floppy charts include checking that the rear drawer is fully installed; the floppy cable runs from the mezzanine board that the drawer plugs into. Kaiser found that with the SCSI bus cable to backplane connector J11 left off, the drives on that bus did not work and the diagnostic utility reported the drive fan as off.

Keyboard, mouse and sound

Peripheral and sound symptoms (Apple's cures, in order)

Symptom	Cures
Cursor does not move	Check the mouse connection; clean the mouse; plug the mouse into the server's ADB port instead of the keyboard (if it then works, replace the keyboard); replace the ADB cable; replace the mouse; then the logic board and processor card
No response to keys	Check whether the server is set to use a serial terminal; use an ADB extended keyboard; check the connection; then the keyboard cable, keyboard and logic board
No sound from the speaker	Check the LCD/speaker/keylock cable; then the speaker, mezzanine interconnect board and logic board
No sound from known-good external speakers	Check the volume and connections; test the speakers on a working Macintosh; then the logic board

Related pages

• Apple Network Server 500 and 700 General Maintenance
• Apple Network Server 500
• Apple Network Server 700
• Apple Network Server 500 and 700 Documentation
• Hard Drive Maintenance and Repair
• Floppy Drive Repair