Switch Hardware Problems - Sun Datacenter InfiniBand Switch 648 Topic Set

Skip Navigation Links
Exit Print View
	Sun Datacenter InfiniBand Switch 648 Topic Set

Oracle Technology Network

Document Information

Related Documentation

Documentation, Support, and Training

Third-Party Web Sites

Sun Welcomes Your Comments

Administering the Switch

Understanding Administrative Commands

Monitoring the Switch

Monitoring the InfiniBand Fabric

Controlling the Switch

Controlling the InfiniBand Fabric

Diagnosing a Problem

Identifying the Problem

Checking Logs and Error Messages

Switch Event Message Overview

Display the Switch Message Log for Power Supplies

Display the Switch Message Log for Fabric Cards

Display the Switch Message Log for Line Cards

Display the System Event Log

Checking Status LEDs

Check Power Supply LEDs

Check CMC Status LEDs

Check the NET MGT LEDs

Check Fabric Card LEDs

Check Line Card LEDs

Check Link LEDs

Troubleshooting the Switch

Switch Hardware Problems

InfiniBand Fabric Problems

Understanding Signal Routing Through the Switch

CXP Connectors and LEDs to Line Card Switch Chip and Port Routes

Line Card Switch Chip to Line Card XBOW Connector Routes

Fabric Card XBOW Connector to Fabric Card Switch Chip Routes

Fabric Card Switch Chip to Fabric Card XBOW Connector Routes

Line Card XBOW Connector to Line Card Switch Chip Routes

Line Card Switch Chip and Port to CXP Connector and LED Routes

Signal Route Through the Switch

Understanding Switch Startup and Component Addition

Switch Startup Process

Fabric Card Addition Process

Line Card Addition Process

Switch GUIDs Overview

Remote Management

Understanding ILOM on the Switch

ILOM Documentation

Supported Features

Understanding ILOM Targets

Installing the ILOM Firmware

Firmware Layout

rupgrade_tool Application

Download the ILOM Firmware

Install the ILOM Firmware

Administering ILOM (CLI)

Access ILOM From the CLI

Switching Between the ILOM Shell and the Linux Shell

Monitoring ILOM Targets (CLI)

Controlling ILOM Targets (CLI)

Upgrading the Switch Firmware Through ILOM (CLI)

Administering ILOM (Web Interface)

Web Interface Overview

Access ILOM From the Web Interface

Monitoring ILOM Targets (Web Interface)

Controlling ILOM Targets (Web Interface)

Upgrading the Switch Firmware Through ILOM (Web Interface)

Understanding ILOM Commands

create Command

delete Command

version Command

Preparing the Site

Routing Service Cables

Understanding InfiniBand Cabling

Understanding Switch Specifications

Installing the Switch

Understanding the Installation

Installing the InfiniBand Software Stack

Unpacking the Chassis

Preparing the Chassis

Inspecting the Components

Installing the Components

Powering On the CMCs

Powering On the Fabric Cards

Powering On the Line Cards

Connecting InfiniBand Cables

Verifying the InfiniBand Fabric

Servicing the Switch

Understanding Service Procedures

Servicing the Power Supplies

Servicing the CMCs

Servicing the Fans

Servicing the Fabric Cards and Fabric Card Fillers

Servicing the InfiniBand Cables

Servicing the Line Cards

Doing Supportive Software Tasks

Command Reference

Understanding Switch-Specific Commands

Understanding CLIA Commands

Understanding InfiniBand Software Commands

Switch Hardware Problems

The following table lists situations that might occur with switch hardware and corrective steps that can be taken to resolve the problem.

Situation	Corrective Steps
The Attention LED on a power supply is lit or the power supply seems dysfunctional.	Check that the power supply is present. See Display Power Supplies Present. Check the power supply status. See Check the Status of a Power Supply. Unplug the respective power cord, wait 15 minutes, then reattach the power cord. If the previous steps do not rectify the situation, replace the power supply. See Switch Service, servicing a power supply.
The Attention LED on a CMC is lit or the CMC seems dysfunctional.	If you are able to access the CMC, reboot the CMC. See Restart a CMC. If the previous step does not rectify the situation, perform a hot-swap of the CMC. See Switch Service, removing a CMC, installing a CMC. If you are unable to access the CMC, or the previous step does not rectify the situation, replace the CMC. See Switch Service, servicing a CMC.
The Attention LED on a fan is lit or the fan seems dysfunctional.	Check that the respective fabric card is present. See Display Fabric Cards Present. Check the fan speed. See Check Fan Speed and Status. Check the logs for any indication of fan failure. See Display the Switch Message Log for Fabric Cards. If the previous steps do not rectify the situation, replace the fan. See the Switch Service, replacing a fan. If replacing the suspect fan with a known good fan does not rectify the situation, replace the fabric card. See Switch Service, servicing a fabric card.
The Attention LED on a fabric card is lit or the fabric card seems dysfunctional.	Check that the fabric card is present. See Display Fabric Cards Present. Ensure that the enviromental conditions are within limits. See Switch Installation, environmental requirements. Check overall switch health. See Display the General Health of the Switch Check that there are no fabric card power faults. See Check Fabric Card Power Faults. Check the logs for any indication of fabric card failure. See Display the Switch Message Log for Fabric Cards. Restart the fabric card. See Restart a Fabric Card or Filler. If the previous steps do not rectify the situation, replace the fabric card. See Switch Service, servicing a fabric card.
The Attention LED on a line card is lit or the line card seems dysfunctional.	Check that the line card is present. See Display Line Cards Present. Ensure that the environmental conditions are within limits. See Switch Installation, environmental requirements. Check overall switch health. See Display the General Health of the Switch Check that there are no line card power faults. See Check Line Card Power Faults. Check the logs for any indication of line card failure. See Display the Switch Message Log for Line Cards. Restart the line card. See Restart a Line Card. If the previous steps do not rectify the situation, replace the line card. See Switch Service, servicing a line card.
After installation, not all internal links are operational.	Determine if the non-operational links are localized to one fabric card or line card. See Check Fabric Card Link Status (Simple) and Check Line Card Link Status (Simple). If the non-operational links are localized to one fabric card, reset the links for that card. See Enable Downed Fabric Card Links. If the non-operational links are localized to one line card, verify that each IB cable connection has appropriate Link LEDs lit. See Check Link LEDs. If the non-operational links are localized to one line card, reset the links for that card. See Enable Downed Line Card Links. If both fabric cards and line cards have non-operational links, verify if there is commonality to the links. See Understanding Signal Routing Through the Switch If there is commonality through an XBOW connector: Unscrew the retainer bolts of the affected fabric cards or line cards or both by 3/4 turn. Alternate between retainer bolts, turning each 1/4 turn counter-clockwise each time. If the previous step did not rectify the situation, reseat the fabric card and line card. See Switch Service, servicing a fabric card, servicing a line card. If the previous steps do not rectify the situation, remove the fabric card or line card and inspect the midplane XBOW connectors and the fabric card or line card XBOW connectors. See Switch Service, removing a fabric card, removing a line card. See Switch Installation, inspecting the midplane connectors, inspecting the XBOW connectors. If the XBOW connectors are in good condition, exchange the slot positions of the suspect fabric card with a known good fabric card. See Switch Service, servicing a fabric card. If the problem has moved to the other fabric card slot, replace the suspect fabric card with a new one. If the problem remains at the original fabric card slot, replace the suspect line card with a new one. See Switch Service, replacing a fabric card, replacing a line card.

Related Information

InfiniBand Fabric Problems

Copyright © 2009 , Oracle and/or its affiliates. All rights reserved. Legal Notices

Previous

Next