IBM RS/6000 SP, Problem Determination Manual

Page: 194 / 316

This soft copy for use by IBM employees only.
In the following example, we will remove the service collections created using
the table ssp.tab:
splm -a service -t /spdata/sys1/logtables/ssp.tab -r
Gather service collections
The fast path to invoke for the Gather Service Collections menu is:
smit spgather_collect
The following example will show how the compressed tar files from the nodes in
the table ssp.tab are gathered to the local node and written to the tape device
/dev/rmt0. The
-s
option provides a method for gathering all the compressed tar
files to tape or disk or mailing them without requiring disk space on the local
node.
splm -a gather -k service -t /spdata/sys1/logtables/ssp.tab -d /tmp -s \
-o /dev/rmt0
6.2.7 SP Error Log
It may be helpful in system problem determination to look at all error logs at
once in parallel. It is not a good idea to copy the /var/adm/ras/errlog files from
all the nodes to one central place and then run
errpt
against this central file, for
the following reasons:
1. The time needed for copying is added to the sequential processing time of
all nodes.
2. The total time required will be longer than that required for viewing the logs
in parallel, and longer than error log processing requires for node
information from the ODM database on each node.
Therefore it is better to use the
dsh
command in combination with the
errpt
command and options to view the error log. To establish this, you can perform
the following:
1. View the s u m m a r y information for all nodes to determine which one must be
examined more closely:
dsh -a errpt -s 0810050096 | pg
2. In the previous example, all error entries are listed that occurred after
August 10, 1996 at 5 a.m. for every node defined in the SDR.
3. Select the nodes that have error entries that need m o r e examination.
4. View the selected nodes, as in the following example:
dsh -w sp21n1,sp21n4,sp21n10 errpt -a -s 0810050096 > /tmp/810errors
When the hardware supervisors indicate a warning or shutdown condition, the
SP System Monitor writes a message using the BSD syslog facility and the AIX
Error Log facility. For example, when the hardware supervisors determine that a
fan has failed, the SP System Monitor writes a precise message into the log file
that includes the time, node, type of error, variable name and, in some cases,
associated values.
The installation process creates the default syslog file
/var/adm/SPlogs/SPdaemon.log on the Control Workstation. You may want to
configure your system to send the system log information to other locations. For
example, you may want to send the SPdaemon.log messages to another
174
SP PD Guide