4.1 Recovering from a Hardware Failure

This section describes how you can recover each component of the Exalogic Control stack if a compute node running an Exalogic Control vServer crashes.

4.1.1 Recovering the Database vServer from a Hardware Failure

The database repositories for Oracle VM Manager (ovs) and Enterprise Manager Ops Center (emoc, emoc_ro) are deployed to a database vServer. By default, the database vServer is deployed to the first compute node in the first pool. If the database vServer crashes, both Oracle VM manager and Enterprise Manager Ops Center will stop being operational.

After the compute node is restored to its previous state, do the following to start the database vServer:

  1. Log in to the compute node as the root user.

  2. Change directory to /OVS/Repositories/*/VirtualMachines.

    hostname#cd /OVS/Repositories/*/VirtualMachines
    
  3. Find the absolute path to the virtual machine configuration file for the database vServer.

    Run the following grep command to identify the correct configuration file corresponding to the Exalogic Control database vServer:

    hostname# grep -i ExalogicControlDB */vm.cfg
    

    The output is similar to:

    0004fb00000600002c18bee8647fb8f7/vm.cfg:OVM_simple_name = 'ExalogicControlDB'
    
  4. Start the database vServer by using the xm create command.

    xm create absolute_path_to_vm.cfg
    

    Example:

    xm create 0004fb00000600002c18bee8647fb8f7/vm.cfg
    
  5. Verify whether the database vServer and all the database processes are running by logging in to the vServer.

  6. Restart the Oracle VM Manager process:

    1. Log in to the Oracle VM Manager vServer as root.

    2. Run service ovmm stop to stop the Oracle VM Manager process.

    3. Run service ovmm start to start the Oracle VM Manager process.

    4. Verify whether Oracle VM Manager started correctly by logging in to the Oracle VM Manager console.

  7. Restart the Proxy Controller and Enterprise Controller processes:

    1. Log in to each of the Proxy Controller vServers as root.

    2. Run proxyadm stop to stop the Proxy Controller.

    3. Run proxyadm start to start the Proxy Controller.

    4. Log in to the Enterprise Controller vServer as root.

    5. Run satadm stop to stop the Enterprise Controller.

    6. Run satadm start to start the Enterprise Controller.

    7. Verify whether the Proxy Controller and Enterprise Controller vServers restarted successfully by logging in to the Exalogic Control BUI.

4.1.2 Recovering the Oracle VM Manager vServer from a Hardware Failure

The Oracle VM Manager is deployed to the ovmm vServer. By default, it is deployed to the first compute node in the first pool. If the Oracle VM Manager vServer crashes, Enterprise Manager Ops Center functionality will be affected.

After the compute node is restored to its previous state, do the following to start the Oracle VM Manager vServer:

  1. Log in to the compute node as the root user.

  2. Change directory to /OVS/Repositories/*/VirtualMachines.

    cd /OVS/Repositories/*/VirtualMachines
    
  3. Find the absolute path to virtual machine configuration file for the Oracle VM Manager vServer.

    Run the following grep command to identify the correct configuration file:

    Hostname# grep -i ExalogicControlOVMM */vm.cfg
    

    The output is similar to:

    0004fb000006000088afde54f9794d32/vm.cfg:OVM_simple_name = 'ExalogicControlOVMM'
    
  4. Start the Oracle VM Manager vServer using the xm create command.

    xm create absolute_path_to_vm.cfg
    

    Example:

    xm create 0004fb000006000088afde54f9794d32/vm.cfg
    
  5. Verify whether the Oracle VM Manager vServer and all the Oracle VM processes are running by logging in to the vServer.

  6. Verify whether Enterprise Manager Ops Center is fully functional by logging in to the Exalogic Control BUI.

4.1.3 Recovering the Proxy Controller vServer from a Hardware Failure

The Proxy Controller component is deployed as two vServers. By default, the second vServer (pc2) is deployed to the second compute node in the first pool and the first proxy controller (pc1) vServer is deployed to the third compute node in the first pool. Enterprise Manager Ops Center functionality will be affected if either of the Proxy Controller vServers crash.

After the compute nodes are recovered to their previous state, do the following to start the Proxy Controller vServer:

  1. Log in to the compute node as the root user.

  2. Change directory to /OVS/Repositories/*/VirtualMachines.

    cd /OVS/Repositories/*/VirtualMachines
    
  3. Find the absolute path to virtual machine configuration file for the Proxy Controller vServer.

    Run the following grep command to identify the correct configuration file:

    Hostname# grep -i ExalogicControlOpsCenterPC* */vm.cfg
    

    The output is similar to:

    0004fb0000060000821f3e60a6d3502d/vm.cfg:OVM_simple_name = 'ExalogicControlOpsCenterPC2'
     
    0004fb000006000084a183dbe7c3dba0/vm.cfg:OVM_simple_name = 'ExalogicControlOpsCenterPC1'
    
  4. Start the Proxy Controller vServer by using the xm create command.

    xm create absolute_path_to_vm.cfg
    

    Example:

    xm create 0004fb0000060000cf01f02c2fb5adaf/vm.cfg
    
  5. Verify whether Enterprise Manager Ops Center is fully functional by logging in to the Exalogic Control BUI.

4.1.4 Recovering the Enterprise Controller vServer from a Hardware Failure

The Enterprise Controller component of the Enterprise Manager Ops Center is deployed to the enterprise controller vServer. By default, the enterprise controller vServer is deployed to the fourth compute node in the first pool. All provisioning and lifecycle management functionality is affected if the enterprise controller vServer is unavailable. Once the compute node is recovered to its previous state, follow the steps to start the enterprise controller vServer:

  1. Log in to the compute node as the root user.

  2. Change the directory to /OVS/Repositories/*/VirtualMachines:

    cd /OVS/Repositories/*/VirtualMachines
    
  3. Find the absolute path to virtual machine configuration file for the Proxy Controller vServer.

    Run the following grep command to identify the correct configuration file:

    Hostname# grep -i ExalogicControlOpsCenterEC1 */vm.cfg
    

    The output is similar to:

    0004fb0000060000cf01f02c2fb5adaf/vm.cfg:OVM_simple_name = 'ExalogicControlOpsCenterEC1'
    
  4. Manually start the enterprise controller vServer using the xm create command. The syntax for the command is:

    xm create absolute_path_to_vm.cfg
    

    Example:

    xm create 0004fb0000060000cf01f02c2fb5adaf/vm.cfg
    
  5. Validate that Ops Center is fully functional by logging in to the Ops Center BUI.