SGI Origin 300 Troubleshooting Guide
Fault-finding on the SGI Origin 300, from SGI's SGI Origin 300 User's Guide (007-4365-002) and SGI Origin 300 System with NUMAlink Module User's Guide (007-4498-002), with owners' reports from the IRIX Network forum.[1][2] The main diagnostic tool is the L1 system controller: its display on the front panel, and its command prompt on the L1 console port at the rear.[3]
Front panel LEDs
[edit | edit source]- Power button LED, green: the internal components are powered.
- Service-required LED, yellow: an item such as a fan has failed and the server is still running.
- Failure LED, red: a failure has occurred and the server is down.
- Drive LEDs, green: disk activity.
[4] The four heartbeat LEDs on the rear panel, one per processor (A to D), show processor activity.[3]
Connecting to the L1 controller
[edit | edit source]Connect a PC or terminal to the DB-9 L1 console port with a null-modem cable. Set 38400 baud, 8 data bits, no parity, 1 stop bit and RTS/CTS hardware flow control. The L1 runs whenever the server has mains power, and its display then shows L1 running; if it does not, check the power connection and the circuit breaker.[4] At the L1 prompt (for example 001c01-L1>), Ctrl+D switches to the system console and Ctrl+T returns to the L1.[4]
| Command | What it shows or does |
|---|---|
env |
Each monitored voltage, fan and temperature with its warning and fault limits and current value |
fan |
Whether the fans are on, and each fan's speed |
power |
The state and measured voltage of each supply |
power up, power down |
Switch the server on or off; prefix with * for joined servers
|
log, log reset |
Show or clear the L1 event log |
leds |
The status codes behind each processor's LEDs |
env reset |
Clear current warnings and faults |
reset, nmi |
Reset the system; send a nonmaskable interrupt |
version, flash status |
L1 firmware version and the two stored firmware images |
L1 messages
[edit | edit source]SGI counts a voltage more than 10% from nominal as a warning and more than 20% as a fault.[6]
| Message | Meaning |
|---|---|
ATTN: x.xV high (or low) fault limit reached @ x.xxV |
A voltage is 20% out; the module starts a 30-second power-off |
ATTN: x.xV high (or low) warning limit reached @ x.xxV |
A voltage is 10% out |
ATTN: x.xV level stabilized @ x.xV |
The voltage is back within limits |
ATTN: FAN # x fault limit reached @ xx RPM |
A fan has reached its maximum speed; the intake air may be too hot, or another fan has failed |
ATTN: FAN # x warning limit reached @ xx RPM |
A fan has speeded up; check the room temperature and whether the fan settles |
ATTN: TEMP # advisory / critical / fault temperature reached |
Intake air above 30, 35 or 40 °C (27, 31 or 35 °C at the high-altitude setting) |
Auto power down in xx seconds |
The L1 has registered a fault and is shutting the module down |
Base module appears to have been powered down |
The L1 has shut the module down after a fault |
The env output on a working server, as printed in SGI's guide, gives the monitored rails as 12V IO, 12V DIG, 5V, 3.3V, 5V aux, 3.3V aux, 2.5V, the CPU core and 1.5V, five fans (left, centre, right and two in the supply) with a 2160 rpm warning threshold, and two node temperature sensors.[5]
Will not power on
[edit | edit source]From SGI's troubleshooting chart and power-on notes:[7][6]
- For a rack system, check that the PDU power cord is seated and its circuit breaker is on.
- Check the server's own power cord. The Origin 300 base module has no power switch: its L1 controller runs whenever the module has power.[7]
- Read the L1 display and look up any message.
- If the L1 controller is not running, check the module's power connection; beyond that SGI refers the module to its engineers.
If the L1 accepts commands but power up only spins the fans, read log and env. In one owner's server env reported the 5V aux rail at 3.22 V and 3.3V aux at 1.7 V; swapping the power supply, VRMs and I/O-8 assembly from a working server changed nothing. Collectors on the forum pointed out that the server could not run at all with rails that low, and suspected the Dallas DS1780 monitoring chips that take the readings.[8] Measure the suspect rail with a meter before replacing parts. The same owner first saw ERROR: no command response from 001c24, TIMEOUT; a forum member pointed to the L1 console-selection section of SGI's L1 and L2 guide.[8]
Fan warnings
[edit | edit source]An owner whose left fan rattled and stopped saw these entries in the L1 log after power up:[9]
FAN 0 warning limit reached @ 0 RPM. Environmental redundancy lost.
weblacky, who repairs these boards, gives three possible causes for a fan warning: the fan has failed; the fan runs but a failed environmental monitoring chip cannot see it; or the MOSFET beside the fan connector that switches the fan's 12 V has failed, so that a new fan will not run either. He links the last case to replacement fans plugged in with the wrong pinout. SGI's fans were commercial parts, but their connector pinouts were often non-standard, so check the wire order against the markings on the old fan's circuit board before plugging in a new one.[9] The owner swapped the plug of the stopped fan with a working fan's; the working fan ran from the suspect connector, which ruled out the third case.[9] The L1 command env off disables environmental monitoring altogether and env temp off only the temperature checks, so that the server can be powered up past a false reading, but either also removes protection against real faults.[9]
Booting
[edit | edit source]- After
power up, Ctrl+D shows the power-on diagnostics and then the System Maintenance Menu.[4] - The reset button resets the processors and ASICs and loses memory contents. The NMI button resets them without losing memory and saves register and memory data to
/var/adm/crash.[3] - When two servers are joined, give the slave a higher system ID with
brick slotbefore cabling them. If the slave is not found at boot, use the L1resetcommand. For server-to-server communication the network mode must be set to 422 with the L1networkcommand.[4]
PCI and disks
[edit | edit source]- A PCI slot with its PWR LED off: reseat the card. A slot with its fault LED on: reseat the card, then replace it if the LED stays on.[6]
- An amber LED on a disk drive: replace the drive.[6]
- 5 V-only PCI cards are not supported.[10]
Related pages
[edit | edit source]- SGI Origin 300
- SGI Origin 300 General Maintenance
- SGI Origin 300 Capacitor Replacement Guide
- SGI Origin 300 Documentation
- Capacitor Failure Symptoms
References
[edit | edit source]- ↑ Silicon Graphics, SGI Origin 300 User's Guide, 007-4365-002, April 2002 (File:SGI Origin 300 User's Guide (007-4365-002).pdf).
- ↑ Silicon Graphics, SGI Origin 300 System with NUMAlink Module User's Guide, 007-4498-002, 2002 (File:SGI Origin 300 System with NUMAlink Module User's Guide (007-4498-002).pdf).
- ↑ 3.0 3.1 3.2 SGI, 007-4365-002, chapter 1, pp. 6-10. Front panel, NMI and reset, heartbeat LEDs, L1 controller.
- ↑ 4.0 4.1 4.2 4.3 4.4 SGI, 007-4365-002, chapter 3, pp. 34-41. Console settings, powering on and off, connecting two servers, status LEDs, L1 modes.
- ↑ 5.0 5.1 SGI, 007-4498-002, appendix C, pp. 177-205. L1 command set, including
env(Example C-12),fan,leds,log,power,flash. - ↑ 6.0 6.1 6.2 6.3 6.4 SGI, 007-4498-002, chapter 5, pp. 106-109, Tables 5-1 and 5-2. Troubleshooting chart and L1 controller messages.
- ↑ 7.0 7.1 SGI, 007-4498-002, chapter 3, p. 81. The Origin 300 base module has no power switch; its L1 controller is on whenever the PDU circuit breaker is on.
- ↑ 8.0 8.1 "Origin 300 is not starting", IRIX Network forum, posts by blacksmith, weblacky and jan-jaap, 31 March to 11 April 2022, forums.irixnet.org/thread-3466.html. Owner's report and community diagnosis.
- ↑ 9.0 9.1 9.2 9.3 "How to Incorporate an Origin 300 In Homelab?", IRIX Network forum, posts by jander31 and weblacky, 6 May 2025, forums.irixnet.org/thread-4523.html. Fan failure log, causes of fan warnings, fan pinouts, the
env offandenv temp offcommands. - ↑ SGI, 007-4365-002, chapter 4, p. 45. PCI card voltages.