HP XC System Software Installation Guide Version 3.2.1

upgradesys output logged to /var/log/upgradesys/upgradesys.log
CAUTION: Do not proceed to the next step in the upgrade process if the output from the
upgradesys script indicates failures. If you cannot determine how to resolve these errors,
contact your local HP support center.
2. Review the /opt/hptc/systemimager/etc/base_exclude_file to determine if you
want to exclude files from the golden image beyond what is already excluded. The
base_exclude file file is the file that is read when the golden image is re-created as
part of the upgrade process. The HP XC System Software Administration Guide describes how
to add exclusions to this file.
3. If you are using the InfiniBand interconnect, verify that the appropriate configuration files
were successfully changed so that the XC system is using the same type of InfiniBand software
stack as the HP SFS system. The sfsconfig should have changed the following files:
/etc/modprob.conf
/etc/modprobe.conf.lustre
/etc/modprobe.conf.lustre.*
a. If you are using the OFED InfiniBand software stack, verify that the lnet option is
added in /etc/modprobe.conf.lustre and it looks similar to the following line:
options lnet networks=o2ib0
b. Verify the /etc/sfstab.proto file to ensure that the appropriate lnet entries were
created and the InfiniBand interconnect interface used is o2ib0 (and not vib0). If this
is not the case, contact HP SFS Support for more information on how to manually change
the /etc/sfstab.proto file.
c. If necessary, when all of the above mentioned files are correct (and you are able to
automatically mount the HP SFS file system on the head node), verify that the
corresponding files under /var/lib/systemimager/images/base_image/etc
are also correct and replace the files.
4. Verify that the newly created HP XC Golden Image is a reasonable size (typically 2 to 3 GB)
otherwise the imaging process might fail.
# du –sk /var/lib/systemimager/images/base_image
5. Decide on the cluster_config option to use to configure the upgraded system. Table 5-6
describes two options to the cluster_config utility that you can use to reconfigure a
system after a software upgrade. The option you choose depends upon how you want the
upgrade to proceed.
Table 5-6 Upgrade Options for the cluster_config Utility
--init Option--migrate Option
Initializes (resets) your existing node role assignments
and configures the system with the default node role
assignments in this release.
This option does not preserve role-to-node assignments
from the previous release.
The default roles and assignments have been optimized
for performance and, you might decide that this
configuration is better suited for your environment.
See Appendix F (page 199) for a description of roles
and the services provided by them, as well as the
default node role assignments.
Brings existing, recognized roles in the configuration
and management database into alignment with the
new roles introduced in this release.
This option retains information about role assignments
and preserves role-to-node assignments.
Using this option does not guarantee the correct
migration process for unrecognized (user-created) roles
and services in the configuration and management
database.
Before you decide to use this option, view the
/opt/hptc/etc/sysconfig/upgrade/role_migration.ini
file to see how the previous role assignments compare
to the roles provided in the new release.
5.9 Task 8: Configure the System and Propagate the Golden Image 131