Flash movie: Disabling a node
Flash movie: Resetting a node
Flash movie: GPFS snapshots
Flash movie: Health monitoring
Flash movie: File notify
Flash movie: Software Upgrade
Resetting a node
This movies demonstrates a simulated hardware failure of a node while
a WindowsXP box is copying files on the node. We use
echo b > /proc/sysrq-trigger
to simulate the hardware failure. That tells the Linux kernel to do an
immediate reset of the CPUs, without any sync or clean shutdown. The
file copy continues after a few second pause.GPFS Snapshots
This movie demonstrates accessing GPFS snapshots from a WindowsXP
explorer window. Snapshots appear under the "previous versions" tab in
explorer, giving users a very easy way to access old versions of
their files.Health Monitoring
This movie demonstrates how CTDB monitors the health of the cluster,
and automatically disables nodes when one of the monitored services
becomes unavailable. In this example the NFS service is disabled, and
ctdb notices the failure of the NFS service, triggering the node to
lose its public IP address and a failover of file services to another
node.File Notify
This movie demonstrates coherent file notify support across the
cluster. A file is added from a windows box connected to one node of
the cluster, and the file immediately appears in the explorer window
of a windows box attached to a different node.Software Upgrade
This movie demonstrates software upgrade of a cluster node, taking
advantage of the all-active nature of the cluster to ensure that
service is not interrupted during the upgrade. The software upgrade
happens while a NFS client is actively copying files on the server.