IBM RS/6000 SP, Руководство по определению проблем

Страница: 166 / 316

This soft copy for use by IBM employees only.
The
startsrc
uses the
-g
flag in order to start the multiple daemons. It runs the
following script to start up the daemons and ensure that multiple instances of the
same daemon do not get created:
/usr/lpp/ssp/bin/sdr
This script passes the IP address associated with the partitions to the respective
daemons.
If there are SDR problems on a particular system, always check that all the
respective SDR daemons are running. If not, then run:
stopsrc -g sdr
startsrc -g sdr
Also ensure that the daemons remain up after restarting them. If they are dying,
then check in the following directory for log files to get an indication of why this
is happening:
/var/adm/SPlogs/sdr
These logs are covered in more detail later in the chapter.
The SP_NAME variable is not set by default. The user needs to explicilty set and
export this variable; otherwise the following file is referenced:
/etc/SDR_dest_info
In the example used, this file contains the following:
default: 9.180.6.199
primary: 9.180.6.198
This file was actually taken from one of the affected nodes because the IP
address of the primary is the IP alias given to the second partition created (that
is, not the default).
Following is the /etc/SDR_dest_info file from a node in the
default partition:
default: 9.180.6.199
primary: 9.180.6.199
Both the default and the primary are set to the IP address associated with the
hostname of the Control Workstation. The node knows how to communicate with
the SDR.
The
default entry in the file serves two purposes. The first is that whenever the
node is rebooted, the rc.sp script that runs from the inittab does the following as
part of its function:
•
Gets the default IP address from the /etc/SDR_dest_info file on that node.
• Contacts the SDR and checks the global object class Syspar_map to find out
whether the IP address entry in the /etc/SDR_dest_info file for the primary is
correct.
•
If it is not correct, then it will update the SDR_dest_info with the correct
entry.
This is a critical step in getting partitioning to work and is one of the reasons
why the affected nodes should be shutdown. The reboot ensures that the
/etc/SDR_dest_info file is updated with the new IP alias address.
The second reason for having a default IP address is that if the partition gets
deleted by restoring an archive of the SDR, or by applying a different partition,
146
SP PD Guide