MAP 0100: Start of Service Call
MAP 0100: Start of Service Call
This MAP is the starting point for all service calls. It directs you to other MAPs and service information.
Refer to Chapter 2 for locations.
001
Read the following notes before continuing:
Attention: Processor nodes should never be powered-on without the top cover seated in its normal posi-
tion to protect internal components.
Attention: Some 48-volt power cables have in-line circuit breakers. Ensure the in-line circuit breaker
switch is in the Off (O) position before connecting or disconnecting 48-volt power cables from the node.
Attention: Components in the frame are susceptible to damage from static discharge. Always use an ESD
wristband when working inside frame covers.
Attention: The MAPs explain restrictions for performing concurrent maintenance on the system. Notify
customer where indicated before disconnecting or powering-down parts of the system so that critical
information/applications of the customer will not be lost during maintenance/repair.
Attention: Servicing processor nodes with the switch assembly feature (parallel environment) requires that
parallel processing must be stopped before removing the nodes from the active configuration. Powering off
processor nodes without removing them from the active partition will cause all parallel jobs in that partition
to terminate and switch initialization to rerun.
– Problem Determination (found in
IBM Parallel System Support Programs for AIX: Diagnosis Guide
should be performed before using the following MAPs to reduce the likelihood that the problem is due to
a software or user error.
– When you need to power-off a processor node, which has not previously been shut down or flashing
888, check with the Customer to see if the node can be fenced and shut down; otherwise, maintenance
must be deferred.
– Concurrent service operations can be done on the following boundaries without interfering with opera-
tion of other parts of the system:
Bulk power supply or power module (only one can be removed at a time)
Frame supervisor card
Processor node (only when not in the active configuration)
Switch with partitions—for service within a partition, see "Viewing Switch Partitions" in
IBM RS/6000
SP: Maintenance Information, Volume 3, Locations and Service Procedures.
– For additional information, check the error log(s) of the failing node(s) by issuing the following command
from an AIX window open for that node:
errpt -a | pg
See Step 002 on page 1-5 for more information and notes concerning the errpt command.
– When a MAP instructs you to replace a FRU, use "FRU Removals and Replacements" in
IBM RS/6000
SP: Maintenance Information, Volume 3, Locations and Service Procedures and the specific instructions
for removing and replacing that part to ensure that the proper procedure is used. Be sure to observe
danger, caution, and attention notices.
1-4
RS/6000 SP MAPs
Summary of Contents for RS/6000 SP
Page 1: ...RS 6000 SP IBM Maintenance Information Volume 2 Maintenance Analysis Procedures GA22 7376 01...
Page 2: ......
Page 3: ...RS 6000 SP IBM Maintenance Information Volume 2 Maintenance Analysis Procedures GA22 7376 01...
Page 8: ...vi RS 6000 SP MAPs...
Page 10: ...viii RS 6000 SP MAPs...
Page 16: ...xiv RS 6000 SP MAPs...
Page 20: ...xviii RS 6000 SP MAPs...
Page 60: ...MAP 0130 Processor Node Figure 1 7 Thin Node High Level Diagram 1 40 RS 6000 SP MAPs...
Page 383: ......
Page 387: ......