HP StoreAll OS 6.5.1 Upgrade Guide

Manual upgrade
Check the following:
• After upgrading from StoreAll version 5.5 and running the restore script, the /etc/multipath
file might contain "hp_scsi_id". As a result, the device mapper will not work. To resolve
this issue, change "hp_scsi_id" to "scsi_id" by entering the following command:
sed -i -e 's/hp_scsi_id/scsi_id/g' /etc/multipath.conf
• If the restore script fails, check /usr/local/ibrix/setup/logs/restore.log for
details.
• If configuration restore fails, look at /usr/local/ibrix/autocfg/logs/appliance.log
to determine which feature restore failed. Look at the specific feature log file under /usr/
local/ibrix/setup/logs/ for more detailed information.
To retry the copy of the configuration, do one of the following, depending on the version from
which you are upgrading:
◦ If you are upgrading from version 6.1.5, use the following command:
/usr/local/ibrix/autocfg/bin/ibrixapp upgrade -f -s
◦ If you are upgrading from version 5.6, use the command appropriate for your server:
File serving node: /usr/local/ibrix/autocfg/bin/ibrixapp upgrade -s
Agile node (file serving node hosting the agile management console):
/usr/local/ibrix/autocfg/bin/ibrixapp upgrade -f -s
6.1.5 upgrade issues
Offline upgrade fails because iLO firmware is out of date
If the iLO2 firmware is out of date on a node, the auto_ibrixupgrade script will fail. The /usr/
local/ibrix/setup/logs/auto_ibrixupgrade.log reports the failure and describes how
to update the firmware.
After updating the firmware, run the following command on the node to complete the StoreAll
software upgrade:
/root/ibrix/ibrix/ibrixupgrade -f
Moving the Fusion Manager VIF to bond1
(X9720 Storage platforms) When the X9720 system is installed, the cluster network is moved to
bond1. The 6.1.5 release requires that the Fusion Manager VIF (Agile_Cluster_VIF) also be moved
to bond1 to enable access to ports 1234 and 9009. To move the Agile_Cluster_VIF to bond1,
complete these steps:
1. From the active Fusion Manager, list all passive management consoles, move them to
maintenance mode, and then unregister them from the agile configuration:
# ibrix_fm -f# ibrix_fm -m maintenance -A# ibrix_fm -u
<management_console_name>
2. Define a new Agile_Cluster_VIF_DEV and the associated Agile_Cluster_VIF_IP.
Issues with cascading upgrades 47