16.Create archive directories for cluster nodes
Log into the virtual service for the package and create archive directories for all cluster nodes.
#cd /var/opt/ignite/recovery/archives
#mkdir
#chown bin:bin *
#cd /
17.Add clients and create recovery archives
Add all cluster nodes as clients and then create recovery archives for each inactive node (not running the
# /opt/ignite/bin/add_new_client \
–s [ha-iux virtual hostname or IP]
From each inactive node, create a recovery archive. These can run in parallel using the virtual server IP. When they all complete, the package can be moved to another node to allow archive creation for the last node.
#/opt/ignite/bin/make_net_recovery
Best practices
Upgrading Ignite-UX
To upgrade the version of
Inactive cluster nodes and swverify
On nodes that are not running the Ignite Serviceguard package, swverify of IGNITE will fail.
This is expected due to /var/opt/ignite, /etc/opt/ignite, and /opt/ignite/boot being set up for exclusive access.
Other depots under shared file systems
If there are other SD depots under shared file systems such as /var/opt/ignite, they will need to be registered on each cluster node individually while the Ignite package is running on that node, or registered by Serviceguard when the Ignite package starts. To manage from Serviceguard, The swreg commands can be added to
of the /var/opt/ignite/depots/recovery_cmds depot. When you swinstall from these depots, be sure to use the virtual hostname or IP for the Ignite Serviceguard package.
19