IBM SG24-5131-00 User Manual

Cluster Management and Administration 

  • If the Cluster Manager is active on the local node, synchronization triggers 

a cluster-wide, dynamic reconfiguration event. In dynamic reconfiguration, 
the configuration data stored in the DCD is updated on each cluster node, 
and, in addition, the new ODM data replaces the ODM data stored in the 
ACD (Active Configuration Directory) on each cluster node. The cluster 
daemons are refreshed and the new configuration becomes the active 
configuration. In the HACMP for AIX log file, reconfig_resource_release, 
reconfig_resource_acquire, and reconfig_resource_complete events mark 
the progress of the dynamic reconfiguration.

  • If the Cluster Manager is active on some cluster nodes but not on the local 

node, the synchronization is aborted.

The HACMP for AIX software provides a Dynamic Reconfiguration (DARE) 
Resource Migration utility that allows for improved cluster management by 
allowing a system administrator to alter the placement of resource groups 
(along with their resources—IP addresses, applications, and disks) to specific 
cluster nodes using the 

cldare

 command. The command lets you move the 

ownership of a series of resource groups to a specific node in that resource 
group’s node list, as long as the requested arrangement is not incompatible 
with the current resource group configuration. It also lets you disable 
resource groups, preventing them from being acquired during a failover or 
reintegration.

Dynamic resource group movement essentially lets a system administrator 
better use hardware resources within the cluster, forcing resource traffic onto 
one or more high-powered or better-connected nodes without having to shut 
down HACMP on the node from which the resource group is moved. Dynamic 
resource group movement also lets you perform selective maintenance 
without rebooting the cluster or disturbing operational nodes.

Using the DARE Resource Migration utility does not affect other resource 
groups that might currently be owned by that node. The node that currently 
owns the resource group will release it as it would during a “graceful 
shutdown with takeover”, and the node to which the resource group is being 
moved will acquire the resource group as it would during a node failover.

The following section covers the types and location keywords used in DARE 
resource migrations, and also how to use the 

cldare 

command and the -M 

flag to perform the migration.