Configuring IBM PowerVM LPAR guest for disaster recovery
Prev Configuring for disaster recovery in virtualized environments Next

Configuring IBM PowerVM LPAR guest for disaster recovery

The IBM PowerVM is configured for disaster recovery by replicating the boot disk by using the replication methods like Hitachi TrueCopy, EMC SRDF, IBM duplicating, cloning rootvg technology, and so on. The network configuration for the LPAR on the primary site may not be effective on the secondary site, if the two sites are in the different IP subnets. To apply the different network configurations on the different sites, you will need to make additional configuration changes to the LPAR resource.

Figure: Across site disaster recovery of the managed LPAR using VCS in the management LPARs

Across site disaster recovery of the managed LPAR using VCS in the management LPARs

To configure LPAR for disaster recovery, you need to configure VCS on both the sites in the management LPARs with the GCO option. See the Cluster Server Administrator's Guide for more information about the global clusters.

Perform the following steps to set up the LPAR guest (managed LPAR) for disaster recovery:

  1. On the primary and the secondary site, create the PowerVM LPAR guest using the Hardware Management Console (HMC) with the ethernet and the client Fibre Channel (FC) virtual adapter's configuration.

    Note:

    The installed OS in the LPAR guest is replicated using the IBM rootvg cloning technology or the DR strategy N_Port ID Virtualization (NPIV).
  2. On the LPAR guest, copy and install the VRTSvcsnr fileset from the VCS installation media. This fileset installs the vcs-reconfig service in the LPAR guest. This service ensures that the site-specific-network parameters are applied when the LPAR boots. You can install the VRTSvcsnr fileset by performing the following steps: 
    # mkdir /<temp_dir>
# cp <media>/pkgs/VRTSvcsnr.bff /<tmp_dir>
# cd /<temp_dir>
# installp -a -d VRTSvcsnr.bff VRTSvcsnr
  3. Create a VCS service group and add a VCS LPAR resource for the LPAR guest. Configure the DROpts attribute of the LPAR resource with the site-specific values for each of the following: IPAddress, Netmask, Gateway, DNSServers (nameserver), DNSSearchPath, Device, Domain, and HostName.

    Set the value of the ConfigureNetwork attribute to 1 from the DROpts attribute to make the changes effective. The LPAR agent does not apply to the DROpts attributes for the guest LPAR, if the value of the ConfigureNetwork attribute is 0. For more info about DROpts attribute see the Cluster Server Bundled Agents Reference Guide.

  4. [ This step is optional:] To perform the rootvg replication using NPIV, the boot disk LUN is mapped directly to the guest LPARs via NPIV, and the source production rootvg LUN is replicated using the hardware technologies like Hitachi TrueCopy, EMC SRDF, and so on for the DR Site. Subsequently, add the appropriate VCS replication resource to the LPAR DR service group. Examples of hardware replication agents are SRDF for EMC SRDF, HTC for Hitachi TrueCopy, MirrorView for EMC MirrorView, and so on. VCS LPAR resource depends on the replication resource.

    See Figure: Sample resource dependency diagram for NPIV base rootvg replication using the hardware replication technology

    For more information about the appropriate VCS replication agent that is used to configure the replication resource, you can visit our website at the following URL: https://sort.veritas.com/agents

    The replication resource ensures that when the resource is online in a site, the underlying replicated devices are in the primary mode, and the remote devices are in the secondary mode. Thus, when the LPAR resource is online, the underlying storage is always in the read-write mode.

  5. Repeat step 1 through step 4 on the secondary site.

Figure: Sample resource dependency diagram for NPIV base rootvg replication using the hardware replication technology

Sample resource dependency diagram for NPIV base rootvg replication using the hardware replication technology

When the LPAR is online, the LPAR agent creates a private VLAN (with VLAN ID 123) between the management LPAR and the managed LPAR. The VLAN is used to pass the network parameters specified in the DROpts attribute to the managed LPAR. When the managed LPAR boots, it starts the vcs-reconfig service that requests for the network configuration from the management LPAR. As a result, the network configuration is resent, as a part of the response through the same VLAN. The vcs-reconfig service subsequently applies this configuration when the appropriate commands are run.

Prev Up Next
About disaster recovery configurations in virtualized environments Home Section VI. Reference