Preparing Microsoft Failover Cluster for Second Data Tier Installation
With the first data tier installed,
in preparation for
installing the second data tier,
complete
the configuration of the failover cluster to handle all the tasks involved in switching the
two data tier nodes between active and passive
states.
On the active cluster node, verify that all
Unwired Platform database services are started.
Sybase Unwired CacheDB
Sybase Unwired ClusterDB
Sybase Unwired LogDataDB
In Windows Explorer, navigate to the folder in which you installed
the database when you installed the first data tier.
If you did not change the default location, this folder is
SMPData\CDB, where
SMPData is the data folder you added to the cluster share
disk.
Open each *errorlog.txt file and look for "Now accepting
requests" as the last line.
If instead you see error messages, there is a problem with the installation that
you must fix before proceeding.
If you do not check the error logs immediately after installing
the data tier, you will see pairs of lines beginning "Starting checkpoint..." and
"Finished checkpoint...". As long as these are the only lines logged after "Now
accepting requests," the database is operating properly.
On the active cluster node, stop the
Unwired Platform database services.
In Windows Services panel, set all
Unwired Platform database services to be started
manually.
Sybase Unwired CacheDB
Sybase Unwired ClusterDB
Sybase Unwired LogDataDB
These services need to be set to start manually so that Microsoft Failover Cluster
can have full control over starting and stopping them.
Define a managed cluster resource group in the Microsoft Failover Cluster and add
the database services to it, so that they move from the current active node to the
new active node when the active node fails over.
For details, refer to the Microsoft Failover Cluster documentation.
Verify that all the database services stop when you take the resource group offline
and start and when you take the resource group online.
Verify that everything moves to the
new
active node when active and passive nodes are switched.
Fail
over to the
other
node to
make it active.
Verify that all Unwired Platform
database services shut down on the current, passive node.
Go to the desktop on the
other,
active node and
fail
over that node back to the first node.
Go to the desktop on the
new
active node and check the database log files.