Configuring Multi-Target Geographic Mirroring (7.5)

1 Background
2 Before you begin
3 Procedure
- 3.1 Create the Cluster Resource Group
  - 3.1.1 Automatic Failover
    - 3.1.1.1 Disabling Automatic Failover
    - 3.1.1.2 Enhanced Automatic Failover with Cluster Monitor
  - 3.1.2 Optional CRG Exit Program
- 3.2 Configure Geographic Mirroring
  - 3.2.1 Configuring the First Target Node
  - 3.2.2 Configuring Additional Target Nodes
4 Results

Background

Geographic Mirroring supports replication between up to six nodes across two sites. Replication within a site is synchronous; replication across sites can be either synchronous or asynchronous.

Before you begin

This scenario assumes that the following tasks have been completed prior to these steps:

A cluster between the nodes has been created, and the cluster nodes have a status of Active.
The nodes are all in the same device domain. When creating a cluster or adding a node to a cluster with default options, PowerHA automatically adds nodes to a single device domain.
An IASP has been created on the production node.

As part of this process, when configuring Geographic Mirroring between the first target system, the production copy of the IASP will need to be varied off. Configuring and deconfiguring Geographic Mirroring between additional nodes does not require the production IASP to be varied off.

Procedure

Create the Cluster Resource Group

A cluster resource group (CRG) defines the following:

A recovery domain is a list of nodes that are potential hosts for the IASP along with:
- The current role the node has (primary or backup)
- The data port IP address(es) to use for replication.
A list of configuration objects, such as IASPs or IP addresses, that are switched between the nodes.

A CRG can be created using any of the following methods:

Type the CRTCRG command and press F4 to go to the Create Cluster Resource Group (CRTCRG) command screen.
Using the Work with Cluster (WRKCLU) command menu. Select option 9, Work with cluster resource groups, and use option 1, Create, and enter a name for the CRG to go to the Create Cluster Resource Group (CRTCRG) command screen.
From the command line, enter the CRTCRG command and your parameters.

For this example, a CRG is created using the following command:

CRTCRG CRG(GEOCRG) 
       CRGTYPE(*DEV) 
       RCYDMN((MN01 *PRIMARY *LAST MN ('192.0.2.100' '192.0.2.101'))
              (MN02 *BACKUP  1     MN ('192.0.2.110' '192.0.2.111'))
              (WI01 *BACKUP  2     WI ('192.0.2.200' '192.0.2.201'))
              (WI02 *BACKUP  3     WI ('192.0.2.202' '192.0.2.203'))) 
       CFGOBJ((MYIASP *DEVD *ONLINE))

This command creates a device CRG called GEOCRG. The CRG defines four nodes: two in a site named MN and two in a site named WI. Each node has two data port IP addresses defined for replication. The CRG has a single IASP device specified that is switched with the CRG called MYIASP.

Data Port IP Addresses

It is possible to use anywhere from 1 to 4 dataport IP addresses per node. Dataport IP addresses are used for Geographic Mirroring data replication. While only a single data port IP address is required, it is generally recommended to have more than one per node for improved performance and throughput. The greatest performance benefit is typically noticed in going from one data port IP address to two.

Automatic Failover

By default, a cluster resource group has automatic failover enabled. When a failure is detected, automatic failover will automatically switch the IASP to a backup node.

Disabling Automatic Failover

Automatic failover can be disabled using the QCST_CRG_CANCEL_FAILOVER PowerHA policy. The following example disables all automatic failover for the CRG named GEOCRG:

ADDHAPCY PCY(QCST_CRG_CANCEL_FAILOVER) PCYDMN(GEOCRG) QUAL('SCOPE(*CRG)') VALUE('EVENT(*ALL)')

For additional information on customizing automatic failover, see the QCST_CRG_CANCEL_FAILOVER policy.

Enhanced Automatic Failover with Cluster Monitor

In environments where automatic failover is desired, it is possible to enhance automatic failover to monitor for additional events via a Hardware Management Console (HMC) using cluster monitors. See Advanced Node Failure Detection for additional information.

Optional CRG Exit Program

CRG Exit programs provide a way to define a program to call whenever a cluster event occurs and enable enhanced automation. See the Cluster Resource Group Exit Program topic in the IBM i Documentation for more information

Configure Geographic Mirroring

Configuring the First Target Node

As part of this process, when configuring Geographic Mirroring between the first target system, the production copy of the IASP will need to be varied off. Configuring and deconfiguring Geographic Mirroring between additional nodes does not require the production IASP to be varied off.

Vary off the IASP on the production system if it currently has a status of Available by using the VRYCFG command. For example: VRYCFG CFGOBJ(MYIASP) CFGTYPE(*DEV) STATUS(*OFF)
Use the Configure Geographic Mirroring (CFGGEOMIR) command to configure Geographic Mirroring replication to the first node. For this example, the following command is used:
CFGGEOMIR ASPDEV(MYIASP) ACTION(*CREATE) DELIVERY(*SYNC) COMPRESS(*RESYNC) TGTNODE(MN02)
This command configures Geographic Mirroring for the IASP named MYIASP to the node MN02.
Since MN02 is within the same site as the production system, delivery must be *SYNC.
A value of *RESYNC is specified for transmission compression. Transmission compression can reduce the time required for initial synchronization and resynchronization and uses hardware-accelerated compression for Power10 and later systems. On systems prior to Power10, software compression is used. Compression can have higher CPU overhead on the production system.
Since *SELECT was specified for the disk units, a panel showing all non-configured units is displayed on the target node, as shown below. Use option 1 to select the disk units you want from the list.
PowerHA Select Non-Configured Disk Units ASP device . . . . . . : MYIASP Mirror copy node . . . : MN02 Prod copy capacity . . : 28634 Mirror copy capacity . : 0 Prod disk units . . . : 3 Mirror disk units . . : 0 Type options, press Enter. 1=Select Resource Opt Name Serial Number Type Model Capacity Rank Eligible 1 DMP476 YF5258000079 2145 0050 9544 002 Yes 1 DMP481 YF525800007A 2145 0050 9544 002 Yes 1 DMP485 YF525800007B 2145 0050 9544 002 Yes More... F1=Help F9=Calculate Selection F11=View 2 F12=Cancel
After confirming the selection, PowerHA will initialize and format the disk units on the target system, configure Geographic Mirroring, and create copy descriptions and a session to manage the replication. The default session name if *GEN is used is GEOMIRSSN.
Since the default value for the start indicator parameter was left as *YES in the command, PowerHA will also start the cluster resource group, vary on the IASP on the production system, and start replication.

Configuring Additional Target Nodes

End the cluster resource group to configure replication to the next node. For example: ENDCRG CRG(GEOCRG)
Use the Configure Geographic Mirroring (CFGGEOMIR) command to configure Geographic Mirroring replication to the additional node. For this example, the following command is used:
This command configures Geographic Mirroring for the IASP named MYIASP to the node WI01.
Since WI01 is in a different site from the production system and is the lowest numbered node at that site, it is possible to specify asynchronous delivery, enabling further distances between the nodes.
Since *SELECT was specified for the disk units, a panel showing all non-configured units is displayed on the target node. Use option 1 to select the disk units you want from the list.
After confirming the selection, PowerHA will initialize and format the disk units on the target system and configure Geographic Mirroring.
Since the default value for the start indicator parameter was left as *YES in the command, PowerHA will also start the cluster resource group, vary on the IASP on the production system if it was not already varied on, and start replication.
This example would repeat this process starting at step 1 for the final node WI02. With WI02, a delivery value of *SYNC must be specified since synchronous replication will be used between the systems within the IA site containing nodes WI01 and WI02.

Results

Geographic Mirroring is now configured between the nodes in the CRG. To verify and monitor the status of replication, use the Display ASP Session (DSPASPSSN) command. For example, DSPASPSSN SSN(GEOMIRSSN).