This section covers the additional tasks you must perform if you want to
set up InterChange Server Windows services in a high-availability
environment. The examples in this section assume you have installed
InterChange Server software to the C:\ directory. The
following topics are covered:
For InterChange Server (ICS) to work in a high-availability (HA)
environment, you must provide the Mfc42u.dll DLL file on
each machine in the cluster. This DLL resides in the bin subdirectory
of the product directory. For the HA environment, you must copy this
DLL to the C:\WINNT\system32 directory on your Windows
system.
To perform this copy, follow these steps:
- Navigate to the bin subdirectory of the
product directory.
- Copy the Mfc42u.dll file.
- Navigate to the C:\WINNT\system32 directory.
- Make sure that a file named Mfc42u.dll does not already
exist here:
- If Mfc42u.dll already exists, check to see if it is
newer than the file you are copying:
- If it is newer, do not copy the older version of the file; instead,
keep the new file in the C:\WINNT\system32 directory.
- If it is older, rename the older file, then paste the version you copied
from the ProductDir\bin directory.
- If Mfc42u.dll does not exist, paste the version you
copied from the ProductDir\bin directory.
- Change ownership of the non-cluster master to the cluster master, then
repeat steps 1 through 4 on the other machine in the cluster (which is now the
non-cluster master).
This section describes how to move the management of InterChange Server
from Windows services to the MSCS (Microsoft Cluster Server) Administrator by
creating an InterChange Server resource instance.
- Note:
- Create an InterChange Server instance on only one of the machines in the
cluster.
- From the Cluster Administrator, click New
> Resource from the Group.
- In the New Resource screen, enter the following information, then click
Next:
Name: ha_interchange
Description: HA InterChange Server
Resource type: ICSinstanceName
Group: Cluster Group
- In the Possible Owners screen, add each machine in the cluster, then click
Next.
- In the Dependencies screen, add the WebSphere MQ queue manager, the shared
drive, and the Persistent Naming Server. Click Next to continue.
- In the InterChange Server Parameters screen, type the server's name
preceded by CW, for example,
CWserver_name. Click Finish.
- Use the Cluster Administrator to bring InterChange Server online and
offline. Be sure to test it on each machine in the cluster.
This section describes how to move the management of an adapter from
Windows services to the
MSCS (Microsoft Server Cluster) Administrator by creating an adapter
instance.
- Note:
- Create the adapter instance on only one of the machines in the
cluster.
- From the Cluster Administrator, click New
> Resource from the Group.
- In the New Resource screen, enter the following information, then click
Next:
Name: adapterName
Description: adapterDescription
Resource type: Generic Service
Group: Cluster Group
- In the Possible Owners screen, add each machine in the cluster, then click
Next.
- In the Dependencies screen, add the WebSphere MQ queue manager, the shared
drive, and the InterChange Server resource, then click Next.
- In the Service Parameters screen, type
CWadapter_nameConnector.
Leave the Startup Parameters field blank, then click Next.
- Click Finish in the Registry Replication screen without adding any
registry keys.
- Use the Cluster Administrator to bring the
adapter online and offline. Be sure to test it on each machine in the
cluster.
- Repeat steps 1 through 7 for additional adapters you want to manage
under MSCS.
When setting up InterChange Server components as Windows services, you can
configure the restart and failover properties of the components by using the
MSCS GUI tools. Although each environment is unique, the following
guidelines may help as you configure the components:
- Setting connectors to failover after a certain number of attempted
restarts might create the situation in which the system continuously fails
over between the two nodes. To avoid this situation, set some (or all)
adapters not to fail over after a failed restart.
- Always configure InterChange Server to fail over to the other node after a
failed restart.
