IBM Books

Managing Shared Disks

Command and SMIT interfaces for the IBM Recoverable Virtual Shared Disk software

This section tells how to maintain twin-tailed Logical Volume Manager (LVM) components shared by two nodes in an SP system running the IBM Recoverable Virtual Shared Disk software. It includes procedures for volume groups, logical volumes, and physical volumes.

Defining twin-tailed LVM components in the Recoverable Virtual Shared Disk environment

Some of the definitions you need to understand are:

twin-tailed volume group
A volume group consisting of supported disks that are all twin-tailed to the same nodes.

twin-tailed physical volume
A disk in a twin-tailed volume group.

twin-tailed logical volume
A logical volume in a twin-tailed volume group.

Coordinating the Recoverable Virtual Shared Disk environment

The twin-tailed components must have the same definition on both twin-tailed nodes. Any change to an LVM component must be reflected in the Object Data Manager (ODM) definitions on both nodes.

Follow the procedures in this chapter carefully in the order given to keep the LVM ODM definitions on both nodes synchronized.

Note:
Do not reboot or issue ha_vsd reset until you have completed all the steps in a task.

Overview of shared volume groups administrative tasks

The overall procedure for modifying a twin-tailed LVM component is the same for all tasks. In general, you export the volume group from a secondary node, make the change on the primary node, and then reimport the volume group on the secondary node. Specific operations, however, have unique steps. They are described in the following sections.

The table below summarizes the steps you must complete on the primary node and the secondary node to change a twin-tailed LVM component in a recoverable virtual shared disk system. Perform all the steps in the correct order so the data does not become corrupted.

Table 11. General Procedure for Changing a Twin-tailed LVM Component

Step Number Primary Node Secondary Node
Step 1 Complete prerequisite tasks Complete prerequisite tasks
Step 2   Export a volume group
Step 3 Vary on a volume group  
Step 4 Make changes to the twin-tailed LVM component  
Step 5 Vary off the volume group  
Step 6   Import a volume group
Step 7   Change a volume group to remain dormant at start up
Step 8   Vary off the volume group
Step 9 Complete follow-up tasks Complete follow-up tasks

Prerequisite tasks

The prerequisite tasks, while not directly involved in modifying LVM components, must be completed before you begin to make the change.

Note:
You must suspend, stop, and unconfigure all virtual shared disks on all nodes in the volume group to be changed before changing a twin-tailed LVM component.

The prerequisite tasks can vary for the different operations. The descriptions for each operation on the following pages have a list of specific prerequisite tasks.

Maintaining twin-tailed volume groups

Maintaining twin-tailed volume groups requires the following administrative tasks:

Creating a twin-tailed volume group

The table below summarizes the steps you must complete on the primary node and the secondary node to create a twin-tailed volume group. Perform all the steps in the correct order so the data does not become corrupted.

Note:
Twin-tailed disks that have SCSI IDs must have a different ID on the secondary than the one on the primary.

Table 12. Steps to Create a Twin-tailed Volume Group

Step Number Primary Node Secondary Node
Step 1 Complete prerequisite tasks Complete prerequisite tasks
Step 2 Create a twin-tailed volume group  
Step 3 Create logical volumes  
Step 4 Vary off the volume group  
Step 5   Import a volume group
Step 6   Change a volume group to remain dormant at startup
Step 7   Vary off the volume group
Step 8 Complete follow-up tasks Complete follow-up tasks

Complete the following steps to create a twin-tailed volume group.

Step 1: Complete prerequisite tasks

  1. The physical volumes (hdisks) should be installed, configured, named, and available.
  2. Make sure the primary and the secondary nodes are active.

Step 2: Create a twin-tailed volume group on the primary node

Use the smit mkvg fastpath to create a volume group.

  1. As the root user at the primary node, enter: 
    smit mkvg

    SMIT returns a screen similar to the following:



    View figure.

  2. Enter field values as follows:

    VOLUME GROUP name
    Enter the name of the twin-tailed volume group.

    Physical partition size in megabytes
    Accept the default value, unless your site has another specific partitioning requirement.

    PHYSICAL VOLUME names
    Enter the names of all the disks to be used in the volume group. You can click on List to display all available physical volumes. . Click on OK when you have selected all the entries you want. The physical volumes you selected are automatically entered in the PHYSICAL VOLUME names field.

    Activate volume group AUTOMATICALLY at system restart?
    Set the field to no so that the volume group can be activated as appropriate by the IBM Recoverable Virtual Shared Disk recovery scripts.

    ACTIVATE volume group after it is created?
    Set this field to yes.

    Volume Group MAJOR NUMBER
    Use the default, the next available number in the valid range.

    Create VG Concurrent Capable?
    Accept the default of no. Do not specify yes.

    Auto-varyon in Concurrent Mode?
    Accept the default of no. Do not specify yes.
  3. Click on OK. The system asks if you are sure. Check and correct your entries, if needed.
  4. Create logical volumes. Refer to the LVM reference pages.
  5. After the command completes, press F12 to exit SMIT and return to the command line.

Step 3: Create logical volumes

Use the smit mklv fastpath to create logical volumes. Create the logical volumes you want for this virtual shared disk.

Step 4: Vary off the volume group on the primary node

Use the varyoffvg command to quiesce the affected volume group.

To vary off the volume group so that it can be activated as appropriate by the IBM Recoverable Virtual Shared Disk recovery scripts, enter: 

varyoffvg  volume_group_name

Step 5: Import volume group information onto the secondary node

You now return to the secondary node. First, import the volume group onto the secondary node. Importing the volume group onto the secondary node synchronizes the ODM definition of the volume group on both nodes.

  1. To use the smit importvg fastpath to import the volume group, enter: 
    smit importvg

    SMIT returns a screen similar to the following:



    View figure.

  2. Enter field values as follows:

    VOLUME GROUP name
    Enter the name of the volume group that you are importing. Make sure the volume group name is the same name that you used on the primary node.

    PHYSICAL VOLUME name
    Enter the name of one of the physical volumes that resides in the volume group. Note that a disk can have a different physical name on different nodes. Make sure that you use the disk name as it is defined on the secondary node. Every hdisk has an ID. If you know that the name of a disk on the primary node is hdisk01, use the name that matches that ID on the secondary node. You can use the lspv command to determine the name.

    ACTIVATE volume group after it is imported?
    Set the field to yes.

    Volume Group MAJOR NUMBER
    Use the default, the next available number in the valid range.

    Make this VG Concurrent Capable?
    Accept the default of no. Do not specify yes.

    Make default varyon of VG Concurrent?
    Accept the default of no. Do not specify yes.
  3. Click on OK.
  4. Press F12 to exit SMIT and return to the command line.

Step 6: Change a volume group to remain dormant at start up

By default, a volume group that was just imported is configured to automatically become active at system restart. However, a recoverable virtual shared disk volume group should be varied on as appropriate by the IBM Recoverable Virtual Shared Disk recovery scripts. Therefore, after importing a volume group, use the SMIT Change a Volume Group screen to reconfigure the volume group to remain dormant at start up.

  1. To use the smit chvg fastpath to change the characteristics of a volume group, enter: 
    smit chvg

    SMIT prompts you to select the volume group.

  2. Enter the name of the volume group you just imported, or click on List and select the name from the display. Click on OK.

    SMIT returns a screen similar to the following: The first field contains the volume group name you specified.



    View figure.

  3. Enter remaining field values as follows:

    ACTIVATE volume group automatically at system restart?
    Set this field to no.

    A QUORUM of disks required to keep the volume group on-line?
    Accept the default, yes.

    Convert this VG to Concurrent Capable?
    Accept the default no. Do not specify yes.

    Autovaryon VG in Concurrent Mode?
    Accept the default no. Do not specify yes.
  4. Click on OK.
  5. Press F12 to exit SMIT and return to the command line.

Step 7: Vary off the volume group on the secondary node

Use the varyoffvg command to quiesce the affected volume group after making the change.

To vary off the volume group so that it can be activated as appropriate by the IBM Recoverable Virtual Shared Disk recovery scripts, enter: 

varyoffvg volume_group_name

Step 8: Complete follow-up tasks

Once you have created the volume group, do the following tasks:

Adding physical volumes to a twin-tailed volume group

The following table summarizes the steps you do on both the primary and secondary nodes to extend (add one or more physical volumes to) a twin-tailed volume group. Perform all the steps in the correct order so the data does not become corrupted.

Table 13. Procedure for Extending a Twin-tailed LVM Component

Step Number Primary Node Secondary Node
Step 1 Complete prerequisite tasks Complete prerequisite tasks
Step 2   Export a volume group
Step 3 Vary on a volume group  
Step 4 Extend a twin-tailed volume group  
Step 5 Vary off the volume group  
Step 6   Import a volume group
Step 7   Change a volume group's characteristics
Step 8   Vary off the volume group
Step 9 Complete follow-up tasks Complete follow-up tasks

Step 1: Complete prerequisite tasks

  1. The physical volumes (hdisks) should be installed, configured, named, and available.
  2. Suspend, stop, and unconfigure all virtual shared disks in the volume group involved in the change on all nodes.

Step 2: Export volume group information from the secondary node

Before making any changes to the LVM elements on the primary node, you must export the appropriate volume group from the secondary node. Exporting the volume group deletes the information about this volume group from the ODM.

To use the exportvg command to export the volume group containing the component you are going to change on the secondary node, enter: 

exportvg volume_group_name

Step 3: Vary on a volume group on the primary node

To use the varyonvg command to activate the affected volume group after making the change, enter: 

varyonvg  volume_group_name

Step 4: Extend the twin-tailed volume group on the primary node

Use the smit extendvg fastpath to extend a volume group.

  1. As the root user at the primary node, enter: 
    smit extendvg

    SMIT returns a screen similar to the following:



    View figure.

  2. Enter field values as follows:

    VOLUME GROUP name
    Enter the name of the volume group that you are extending.

    PHYSICAL VOLUME names
    Enter the names of the physical volumes you are adding to the volume group.
  3. Click on OK to extend the volume group.
  4. After the command completes, press F12 to exit SMIT and return to the command line.
  5. Create new logical volumes or extend existing ones.

Step 5: Vary off the volume group on the primary node

Use the varyoffvg command to quiesce the affected volume group after making the change.

To vary off the volume group so that it can be activated as appropriate by the IBM Recoverable Virtual Shared Disk recovery scripts, enter: 

varyoffvg volume_group_name

Step 6: Import volume group information onto the secondary node

You now return to the secondary node. First, import the volume group onto the secondary node. Importing the volume group onto the secondary node synchronizes the ODM definition of the volume group with the primary node.

  1. To use the smit importvg fastpath to import the volume group, enter: 
    smit importvg

    SMIT returns a screen similar to the following:



    View figure.

  2. Enter field values as follows:

    VOLUME GROUP name
    Enter the name of the volume group that you are importing. Make sure the volume group name is the same name that you used on the primary node.

    PHYSICAL VOLUME name
    Enter the name of one of the physical volumes that resides in the volume group. Note that a disk can have a different physical name on different nodes. Make sure that you use the disk name as it is defined on the secondary node. Every hdisk has an ID. If you know that the name of a disk on the primary node is hdisk01, use the name that matches that ID on the secondary node. You can use the lspv command to determine the name.

    Volume Group MAJOR NUMBER
    Use the default, the next available number in the valid range.

    Make this VG Concurrent Capable?
    Accept the default of no. Do not specify yes.

    Make default varyon of VG Concurrent?
    Accept the default of no. Do not specify yes.
  3. Click on OK.
  4. Press F12 to exit SMIT and return to the command line.

Step 7: Change volume group characteristics on the secondary node

By default, a volume group that was just imported is configured to automatically become active at system restart. However, a recoverable virtual shared disk volume group should be varied on as appropriate by the IBM Recoverable Virtual Shared Disk recovery scripts. Therefore, after importing a volume group, use the SMIT Change a Volume Group screen to reconfigure the volume group to remain dormant at start up.

  1. To use the smit chvg fastpath to change the characteristics of a volume group, enter: 
    smit chvg

    SMIT prompts you to select the volume group.

  2. Enter the name of the volume group you just imported, or click on List and select the name from the display. Click on OK.

    SMIT returns a screen similar to the following. The first field contains the volume group name you specified.



    View figure.

  3. Enter remaining field values as follows:

    ACTIVATE volume group automatically at system restart?
    Set this field to no.

    A QUORUM of disks required to keep the volume group on-line?
    Accept the default, yes.

    Convert this VG to Concurrent Capable?
    Accept the default no. Do not specify yes.

    Autovaryon VG in Concurrent Mode?
    Accept the default no. Do not specify yes.
  4. Click on OK.
  5. Press F12 to exit SMIT and return to the command line.

Step 8: Vary off the volume group on the secondary node

Use the varyoffvg command to quiesce the affected volume group after making the change.

To vary off the volume group so that it can be activated as appropriate by the IBM Recoverable Virtual Shared Disk recovery scripts, enter: 

varyoffvg volume_group_name

Step 9: Complete follow-up tasks

Once the you have extended the volume group, do the following tasks:

Adding or removing logical volumes to or from a twin-tailed volume group

The following table summarizes the steps you do on both the primary and secondary nodes to add or remove logical volumes to or from a twin-tailed volume group. Perform all the steps in the correct order so the data does not become corrupted.

Table 14. Procedure for Extending a Twin-tailed LVM Component

Step Number Primary Node Secondary Node
Step 1 Complete prerequisite tasks Complete prerequisite tasks
Step 2   Export a volume group
Step 3 Create or remove logical volumes  
Step 4 Vary off the volume group  
Step 5   Import a volume group
Step 6   Change a volume group's characteristics
Step 7   Vary off the volume group
Step 8 Vary on the volume group  
Step 9 Complete follow-up tasks Complete follow-up tasks

Step 1: Complete prerequisite tasks

  1. Shutdown your applications using the volume group to ensure data consistency and ODM integrity.
  2. Suspend, stop, and unconfigure all virtual shared disks on all the virtual shared disk nodes.
  3. Undefine virtual shared disks on any logical volumes being removed. Ensure all data is moved from the logical volumes before removing.

Step 2: Export volume group information from the secondary node

Before making any changes to the LVM elements on the primary node, you must export the appropriate volume group from the secondary node. Exporting the volume group deletes the information about this volume group from the ODM.

To use the exportvg command to export the volume group containing the component you are going to change on the secondary node, enter: 

exportvg volume_group_name

Step 3: Create or remove required logical volumes on the primary node

To create the logical volumes, use the fastpath invocation to the SMIT panel: 

smit mklv

Alternatively, you can use the command line interface to create the logical volumes: 

mklv

To remove the logical volumes, use the fastpath invocation to the SMIT panel: 

smit rmlv

Alternatively, you can use the command line interface to remove the logical volumes: 

rmlv

Step 4: Vary off the volume group on the primary node

Use the varyoffvg command to quiesce the affected volume group after making the change.

To vary off the volume group so that it can be activated as appropriate by the IBM Recoverable Virtual Shared Disk recovery scripts, enter: 

varyoffvg volume_group_name

Step 5: Import volume group information onto the secondary node

You now return to the secondary node. First, import the volume group onto the secondary node. Importing the volume group onto the secondary node synchronizes the ODM definition of the volume group with the primary node.

  1. To use the smit importvg fastpath to import the volume group, enter: 
    smit importvg

Step 6: Change volume group characteristics on the secondary node

By default, a volume group that was just imported is configured to automatically become active at system restart. However, a recoverable virtual shared disk volume group should be varied on as appropriate by the IBM Recoverable Virtual Shared Disk recovery scripts. Therefore, after importing a volume group, use the SMIT Change a Volume Group screen to reconfigure the volume group to remain dormant at start up.

To use the smit chvg fastpath to change the characteristics of a volume group, enter: 

smit chvg

For more information, refer to Step 7: Change volume group characteristics on the secondary node.

Alternatively, to use the command line interface to change the volume group to remain dormant at startup, enter: 

chvg -a `n' -Q`y' volume_group_name

Step 7: Vary off the volume group on the secondary node

Use the varyoffvg command to quiesce the affected volume group after making the change.

To vary off the volume group so that it can be activated as appropriate by the IBM Recoverable Virtual Shared Disk recovery scripts, enter: 

varyoffvg volume_group_name

Step 8: Vary on the volume group on the primary node

Use the varyonvg command to quiesce the affected volume group after making the change.

To vary on the volume group so that it can be activated as appropriate by the IBM Recoverable Virtual Shared Disk recovery scripts, enter: 

varyonvg volume_group_name

Step 9: Complete follow-up tasks

Once the you have extended the volume group, do the following tasks:

Reducing a twin-tailed volume group

The following table summarizes the steps you do on both the primary and secondary nodes to reduce (remove one or more physical volumes from) a twin-tailed volume group:

Table 15. Procedure for Reducing a Twin-Tailed Volume Group

Step Number Primary Node Secondary Node
Step 1 Complete prerequisite tasks Complete prerequisite tasks
Step 2   Export a volume group
Step 3 Vary on a volume group  
Step 4 Remove data from the physical volume  
Step 5 Reduce a twin-tailed volume group  
Step 6 Vary off the volume group  
Step 7   Import volume group
Step 8   Change volume group characteristics
Step 9   Vary off the volume group
Step 10 Complete follow-up tasks Complete follow-up tasks

Complete the following tasks to reduce a twin-tailed volume group.

Step 1: Complete prerequisite tasks

  1. Both the primary and secondary nodes must be up and in the active group.
  2. Suspend, stop, and unconfigure virtual shared disks in the affected volume group on all nodes.

Step 2: Export volume group information from the secondary node

Before making any changes to the LVM elements on the primary node, you must export the appropriate volume group from the secondary node. Exporting the volume group deletes the information about this volume group from the ODM.

To use the exportvg command to export the volume group containing the component you are going to change on the secondary node, enter: 

exportvg volume_group_name

Step 3: Vary on a volume group on the primary node

To use the varyonvg command to verify that the affected volume group is varied on after making the change, enter: 

varyonvg volume_group_name

Step 4: Move data from the volume to be removed

Use the smit migratepv fastpath to move data on the physical volume being removed from the volume group to a different physical volume. If you do not, data will be lost.

  1. As the root user on the primary node, enter: 
    smit migratepv

    SMIT returns a screen similar to the following:



    View figure.

  2. Enter the names of the physical volumes being moved in the SOURCE physical volume name field or click on List and select the names from the display. Click on OK.

    SMIT returns a screen similar to the following. The physical volumes being migrated are entered in the SOURCE physical volume names field.



    View figure.

  3. Enter field values as follows:

    DESTINATION physical volumes
    Enter the names of the physical volumes to which you want to move the data.

    Move only data belonging to this LOGICAL VOLUME?
    Set this field to no.
  4. Click on OK to migrate the physical volumes.
  5. After the command completes, press F12 to exit SMIT and return to the command line.

Step 5: Reduce the shared volume group on the primary node

Use the smit reducevg fastpath to reduce a volume group.

  1. As the root user at the primary node, enter: 
    smit reducevg

    SMIT returns a screen similar to the following:



    View figure.

  2. Select the Remove a Physical Volume from a Volume Group option and click on OK.

    SMIT returns a screen similar to the following:



    View figure.

  3. Enter the name of the volume group from which you are removing the physical volumes in the VOLUME GROUP name field. Click on OK.

    SMIT returns a screen similar to the following. The name of the volume group you specified in the preceding screen is entered.



    View figure.

  4. Enter field values as follows:

    PHYSICAL VOLUME names
    Enter the name of the physical volumes to remove from the volume group.

    FORCE deallocation of all partitions on this physical volume?
    Set this field to no.
  5. Click on OK to remove the designated physical volumes from the volume group.
  6. After the command completes, press F12 to exit SMIT and return to the command line.

Step 6: Vary off the volume group on the primary node

Use the varyoffvg command to quiesce the affected volume group after making the change.

To vary off the volume group so that it can be activated as appropriate by the IBM Recoverable Virtual Shared Disk recovery scripts, enter: 

varyoffvg volume_group_name

Step 7: Import volume group information onto the secondary node

You now return to the secondary node. First, import the volume group on the secondary node. Importing the volume group onto the secondary node synchronizes the ODM definition of the volume group on the node.

  1. To use the smit importvg fastpath to import the volume group, enter: 
    smit importvg

    SMIT returns a screen similar to the following:



    View figure.

  2. Enter field values as follows:

    VOLUME GROUP name
    Enter the name of the volume group that you are importing. Make sure the volume group name is the same name that you used on the primary node.

    PHYSICAL VOLUME name
    Enter the name of one of the physical volumes that resides in the volume group. Note that a disk can have a different physical name on different nodes. Make sure that you use the disk name as it is defined on the secondary node. Every hdisk has an ID. If you know that the name of a disk on the primary node is hdisk01, use the name that matches that ID on the secondary node. You can use the lspv command to determine the name.

    Volume Group MAJOR NUMBER
    Use the default, the next available number in the valid range.

    Make this VG Concurrent Capable?
    Accept the default of no. Do not specify yes.

    Make default varyon of VG Concurrent?
    Accept the default of no. Do not specify yes.
  3. Click on OK.
  4. Press F12 to exit SMIT and return to the command line.

Step 8: Change volume group characteristics on the secondary node

By default, a volume group that was just imported is configured to automatically become active at system restart. However, a recoverable virtual shared disk volume group should be varied on as appropriate by the IBM Recoverable Virtual Shared Disk recovery scripts. Therefore, after importing a volume group, use the SMIT Change a Volume Group screen to reconfigure the volume group to remain dormant at startup.

  1. To use the smit chvg fastpath to change the characteristics of a volume group, enter: 
    smit chvg

    SMIT prompts you to select the volume group.

  2. To select the volume group you just imported, click on OK.

    A screen similar to the following appears. The first field contains the volume group name you specified.



    View figure.

  3. Enter remaining field values as follows:

    ACTIVATE volume group automatically at system restart?
    Set this field to no.

    A QUORUM of disks required to keep the volume group on-line?
    Accept the default, yes.

    Convert this VG to Concurrent Capable?
    Accept the default no. Do not specify yes.

    Autovaryon VG in Concurrent Mode?
    Accept the default no. Do not specify yes.
  4. Click on OK.
  5. Press F12 to exit SMIT and return to the command line.

Step 9: Vary off the volume group on the secondary node

Use the varyoffvg command to quiesce the affected volume group after making the change.

You vary off the volume group so that it can be activated as appropriate by the IBM Recoverable Virtual Shared Disk recovery scripts. Enter: 

varyoffvg volume_group_name

Step 10: Complete follow-up tasks

Once the you have reduced the volume group, complete the following tasks:

Removing a twin-tailed volume group

See the following table for an overview of the steps needed to remove a twin-tailed volume group.

Table 16. Steps to Remove a Shared Volume Group

Step Number Primary Node Secondary Node
Step 1 Complete prerequisite tasks Complete prerequisite tasks
Step 2   Export volume group information
Step 3 Vary on a volume group  
Step 4   Delete a twin-tailed volume group

Complete the following steps to remove a twin-tailed volume group.

Step 1: Complete prerequisite tasks

  1. Suspend, stop, and unconfigure the active virtual shared disks on this volume group on all nodes.
  2. Undefine the volume group with vsdelvg at the control workstation.

Step 2: Export volume group information from the secondary node

Before making any changes to the LVM elements on the primary node, you must export the appropriate volume group from the secondary node. Exporting the volume group deletes the information about this volume group from the ODM. To use the exportvg command to export the volume group containing the component you are going to change on the secondary nodes, enter: 

exportvg volume_group_name

Step 3: Vary on a volume group on the primary node

To use the varyonvg command to activate the affected volume group after making the change, enter: 

varyonvg volume_group_name

Step 4: Delete a twin-tailed volume group on the primary node

  1. As the root user at the primary node, enter: 
    smit reducevg
    SMIT returns a screen similar to the following:



    View figure.

  2. Select the Remove a Volume Group option and click on OK.

    SMIT returns a screen similar to the following:



    View figure.

  3. Enter the name of the volume group to be removed. Click on OK.
  4. When the command completes, press 10 to leave SMIT and return to the command line.

[ Top of Page | Previous Page | Next Page | Table of Contents | Index ]