|
|
Configuring Failover and Failback in a Domains Environment
Two types of failover can be performed in a Domains environment: link-level failover and Domains-level failover. This section provides instructions for both:
If you want failover and failback functionality in your domain, you must configure your Domains configuration file to support it.
For details about the Domains configuration file, see the DMCONFIG(5) in the File Formats, Data Descriptions, MIBs, and System Processes Reference.
How to Configure Domains to Support Link-level Failover
Link-level failover is a mechanism that ensures that an alternate network link becomes active when a primary link fails. To use link-level failover, the primary and alternate gateways must reside on different remote domains (that is, gateway mirroring must be used). Currently, link-level failover does not support multiple alternate links to the same gateway.
To implement link-level failover, specify it in the DM_TDOMAINS section of the Domains configuration file (DMCONFIG) as follows:
*DM_TDOMAINS
RDOM1 NWADDR=//addr1:0
RDOM1 NWADDR=//addr2:0
The first entry refers to the primary network link for remote domain RDOM1; the second entry refers to the alternate link.
Link-level failback is a manual procedure. When the primary link is restored, the administrator must bring down the alternate link manually. This operation may cause requests that are in progress to fail, and new traffic to be resumed over the primary link.
Note: For more detailed information on gateway mirroring, see DMCONFIG(5) in the File Formats, Data Descriptions, MIBs, and System Processes Reference.
Configuring Domains-level Failover and Failback
Domains-level failover is a mechanism that transfers requests to alternate remote domains when a failure is detected with a primary remote domain. It also provides failback to the primary remote domain when that domain is restored.
This level of failover/failback depends on Dynamic Status. The domain must be configured with a CONNECTION_POLICY of ON_STARTUP or INCOMING_ONLY to enable Domains-level failover/failback.
Domains-level failover/failback defines a remote domain as available when a network connection to the remote domain exists, and unavailable when a network connection to the remote domain does not exist.
Prerequisite to Using Domains-level Failover and Failback
To use Domains-level failback, you must specify ON_STARTUP or INCOMING_ONLY as the value of the CONNECTION_POLICY parameter.
A connection policy of ON_DEMAND is unsuitable for Domains-level failback as it operates on the assumption that the remote domain is always available. If you do not specify ON_STARTUP or INCOMING_ONLY as your connection policy, your servers cannot fail over to the alternate remote domains that you have specified with the RDOM parameter.
Note: A remote domain is available if a network connection to it exists; a remote domain is unavailable if a network connection to it does not exist.
How to Configure Domains to Support Failover
To support failover, you must specify a list of the remote domains responsible for executing a particular service in your Domains configuration file. Specifically, you must specify such a list as the value of the RDOM parameter in the DM_REMOTE_SERVICES section. You can also specify alternate domains, as follows:
RDOM=identifier_1, identifier_2, identifier_3
Example
Suppose the TOUPPER and TOUPPER2 services are available from three remote domains: R1 (the primary remote domain), R2, and R3. Include the following entry in your Domains configuration file:
*DM_REMOTE_SERVICES
DEFAULT: RDOM=R1, R2, R3
TOUPPER
TOUPPER2
How to Configure Domains to Support Failback
Failback occurs when a network connection to the primary remote domain is reestablished for any of the following reasons:
Note: For automatic retries, connection retry must be turned on (that is, MA54ETRY>0).
|
Copyright © 2001 BEA Systems, Inc. All rights reserved.
|