This topic identifies example network configurations and then describes two sample IP configuration exercises. The first example illustrates a typical case of a database application dependent upon a single IP resource and configured on a pre-existing subnet. The second example illustrates an active/active scenario where multiple IP resources are configured.
Network Configuration
The first two configuration examples assume the network configuration diagrammed in the following figure.
The network configuration has these components:
- Servers. The configuration has two servers, Server 1 and Server 2, each with the appropriate LifeKeeper and application software installed.
- Interfaces. Each server has two Ethernet interfaces, eth0 and eth1, configured as follows:
eth0 eth1 |
Server1 25.0.3.6 Server11 25.9.1.8 |
Server2 25.0.3.7 Server21 25.0.1.7 |
- Network. The network consists of three subnetworks:
º Low traffic backbone (25.0.3) primarily for servers
º High traffic backbone (25.0.1) with both servers and clients
º High traffic client network (25.0.2.)
A gateway provides interconnection routing between all LANs. A Domain Name Server (not shown) is used for address resolution.
- Heartbeat. TCP heartbeat communication paths would be configured using either or both of the server subnetworks.
Typical Configuration Example
Server 1 and Server 2 have access to an application called mydatabase that resides on a shared disk. To ensure that the application mydatabase and the IP resources used to access it are switched together, the system administrator creates a mydatabase application resource and adds the IP resource to the application hierarchy as a dependency.
These are the configuration issues:
- Application hierarchy. The application hierarchy must exist before the administrator names it as a parent of the IP resource. For the purposes of this example, Server 1 is the primary server. The application resource tags are mydatabase-on-server1 and mydatabase-on-server2.
- IP resource name. The administrator adds the name and address of the IP resource to the /etc/hosts file on both Server 1 and Server 2 and to the DNS database. In this example, the IP resource name is databaseip and its network address is 25.0.1.2. If no name-to-IP address association is necessary, then this is not required.
- Routers, gateways, and users. Because databaseip is an address on an existing subnet, no additional configuration is necessary. The IP resource is on the 25.0.1 subnet. All users connect to databaseip via the route they currently use to get to the 25.0.1 subnet. For example, users on 25.0.2 go through the gateway and users on 25.0.1 connect directly.
- IP instance definition. When the administrator enters databaseip as the IP resource on the Resource Hierarchy Create screen, the software performs several tests. It verifies that Server 1 can determine the address that goes with databaseip (it is in the hosts file and/or can be retrieved from the DNS). It also verifies that the address retrieved, address 25.0.1.2, is not already in use. Since the IP resource is on the 25.0.1 subnet, the IP Recovery software will ensure that it is configured on the eth1 interface. If the IP resource is acceptable, the software fills in the remainder of the wizard dialog boxes with default values, as shown in the table below Figure 3. If you selected all the default values, an independent IP resource hierarchy called ip-databaseip would be created.
Note: The tables associated with each configuration illustration provide examples of the appropriate information that would be entered in the Create Resource Hierarchy wizard for the primary server (Server 1) and Extend Resource Hierarchy wizard for the backup server (Server 2). For additional details on what information should be entered into the wizards, refer to the LifeKeeper Configuration Tasks section later in this section. These tables can be a helpful reference when configuring your recovery kit.
Figure 3. Typical Configuration Example of IP Resource Creation
Configuration Notes:
- The application resource is mydatabase-on-server1.
- The IP resource is databaseip with a tag name of ip-databaseip.
- If mydatabase-on-server1 fails, LifeKeeper switches it to Server 2; (ip-databaseip is only switched if a dependency exists).
- If Server 1 fails, both resources are brought in-service on Server 2.
- During a switchover, databaseip users would be disconnected. When they log back in, they can access any applications on Server 2.
- During a manual switchover, users connected to Server 1 via connections other than databaseip remain connected to Server 1.
Creating an IP resource hierarchy on Server 1:
Server: | Server1 |
IP Resource: | databaseip |
Netmask: | 255.255.252.0 |
Network Interface: | eth1 |
IP Resource Tag: | ip-databaseip |
Note: See the topic Guidelines for Creating an IP Dependency before extending an IP resource to a backup server.
Extending an IP resource hierarchy to Server 2:
Template Server: | Server1 |
Tag to Extend: | databaseip |
Target Server: | Server2 |
Target Priority: | 10 |
**IP Resource: | 25.0.1.2 |
Netmask: | 255.255.252.0 |
Network Interface: | eth1 |
IP Resource Tag | ip-databaseip |
Note: The actual IP address associated with the DNS name is displayed in the Extend Wizard as the IP resource.
Test Your IP Resource
To verify the successful creation of the IP resource, the administrator should perform the following tasks:
- From the LifeKeeper GUI, observe whether ip-databaseip is in-service (ISP) on Server 1.
- From a remote server, connect to address databaseip using ping or telnet.
- Test manual switchover by selecting the in_service option on Server 2 and selecting ip-databaseip. Verify that the IP address migrates to Server 2.
Active/Active Configuration Example
The second example, using the same network configuration, describes two IP resources, one active on each server.
Resource Addresses
For this example, the IP resources are server1ip (address 25.0.6.20) and server2ip (address 25.0.6.21). Entries for these resources must be in the /etc/hosts files on each server and in the DNS database.
Router Configuration
Because the selected addresses are on a new (logical) subnet, they can be configured for either eth0 or eth1. However, both must go on the same interface.
For this example, choosing eth0 means that all users would have to go through the gateway. Choosing eth1 would allow the users on the 25.0.1 subnet to access the resources directly (assuming that the new subnet had been added to their internal routing tables). Users on subnet 25.0.2 would still require the gateway. For the purposes of this example, the selected interface is eth1.
Regardless of which physical network is chosen to support the new subnet, the network administrator would have to add routing information to the gateway system before creating the IP resources.
First IP Resource Definition
The administrator creates the first IP resource on Server 1. eth0 is the first available interface on each server and would appear as the default. To define eth1 as the interface, the administrator selects it from the list of available interfaces.
Creating an IP resource hierarchy on Server 1:
Server: | Server1 |
IP Resource: | server1ip |
Netmask: | 255.255.252.0 |
Network Interface: | eth1 |
IP Resource Tag: | ip-server1ip |
Note: See the topic Guidelines for Creating an IP Dependency before extending an IP resource to a backup server.
Extending an IP resource hierarchy to Server 2:
Template Server: Tag to Extend: Target Server: Target Priority: **IP Resouce: Netmask: Network Interface: IP Resource Tag: |
Server1 server1ip Server2 10 25.0.6.20 255.255.252.0 eth1 ip-server1ip |
Note The actual IP address associated with the DNS name is displayed in the Extend Wizard as the IP resource.
Second IP resource definition
The administrator creates the second IP resource on Server 2. eth0 is the first available interface on each server and would appear as the default. To define eth1 as the interface, the administrator selects it from the list of available interfaces.
Creating an IP resource hierarchy on Server 2:
Server: IP Resource: Netmask: Network Interface: P Resource Tag: |
Server2 server2ip 255.255.252.0 eth1 ip-server2ip |
Note: See the topic Guidelines for Creating an IP Dependency before extending an IP resource to a backup server.
Extending an IP resource hierarchy to Server 1:
Template Server: Tag to Extend: Target Server: Target Priority: **IP Resouce: Netmask: Network Interface: IP Resource Tag: |
Server2 server2ip Server1 10 25.0.6.21 255.255.252.0 eth1 ip-server2ip |
Note: The actual IP address associated with the DNS name is displayed in the Extend Wizard as the IP resource.
Testing IP Resources
The administrator should verify that the new resources are functioning on both servers by performing the following tests:
- With each resource on its primary server, verify that each is accessible by using either ping or telnet. The administrator may also want to test connectivity from all user sites.
- Test switchover by manually bringing ip-server1ip into service on Server 2. Verify both resources are functional on Server 2.
- Bring both resources into service on Server 1. Verify both resources are functional on Server 1.
- Bring ip-server2ip back into service on its primary server, Server 2.
Post your comment on this topic.