Managing Storage Failover
Prev Storage Tasks Next

Managing Storage Failover

This section describes the tasks you can perform to manage storage failover.

Displaying Information about Storage-Failover Configurations

To display information about storage-failover configurations, use the storage failover show command. By default, the command displays the following information:

  • Node name.

  • Partner node name.

  • Whether storage failover is enabled.

  • Whether storage failover is possible.

  • Whether the storage-failover interconnection is available.

  • The current state of storage failover. Possible values include the following:

    • CONNECTED

    • TAKEOVER SCHEDULED

    • TAKEOVER STARTED

    • TAKEOVER

    • TAKEN OVER

    • SENDHOME STARTED

    • ERROR

    • WAITING

Other information is displayed in a detailed view and if other parameters are specified. See the reference page for the command for details.

The following example displays information about all storage-failover configurations.

node::> storage failover show
Node       Partner     Enabled      Reason    InterConn     State
                                     Not       Up 
                                     Enabled 
node0a     node0b     True                   False         TAKEOVER
                                                            SCHEDULED
node0b     node0a     True                   False         CONNECTED
node1a     node1b     False        NOT_INIT  False         CONNECTED
node1b     node1a     False        NOT_INIT  False         CONNECTED
node2a     node2b     True                   False         TAKEN OVER
node2b     node2a     True                   False         WAITING
node3a     node3b     True                   False         TAKEOVER
                                                            STARTED
node3b     node3a     True                   False         WAITING

Modifying a Storage-Failover Configuration

To modify a storage-failover configuration for a node, use the storage failover modify command. You can modify the following attributes of a storage-failover configuration:

  • Whether storage failover is enabled. The default is true.

  • Whether automatic send-home operations are enabled. An automatic send-home operation is invoked when one file of a failover pair is in takeover mode and the failed node is repaired and restarts. When the repaired node restarts, the node in takeover mode detects this and initiates a send-home operation. The default is false.

  • Whether long-running jobs such as NDMP operations and volume operations are terminated when an automatic send-home operation is initiated. If this option is set to false (the default), the automatic send-home operation is deferred until the long-running jobs have completed.

  • Whether the node checks its partner's readiness before initiating a send-home operation that was invoked by using the storage failover send-home command. The default is true, which reduces downtime caused by a send-home operation.

  • The number of seconds that the node remains unresponsive before its partner initiates a takeover operation. The default is 15 seconds; possible values range from 10 to 180 seconds.

  • Whether the node automatically takes over for its partner if the partner fails. The default is true.

  • Whether the node automatically takes over for its partner if the partner panics. The default is true. Setting this option on one node automatically makes the same setting on its partner node.

  • Whether the node takes over for its partner if the partner fails within a default 60 seconds of starting up. The default is true. Setting this option requires the advanced privilege level or higher.

  • The number of seconds, used by the previous option, before the node takes over for its partner if the partner fails within the specified number of seconds after starting up. The default is 60 seconds. Setting this option requires the advanced privilege level or higher.

  • The number of times the node attempts an automatic send-home operation within a default 10 minutes. The default value is three attempts. Setting this option requires the advanced privilege level or higher.

  • The number of minutes, used by the previous option, during which the node attempts automatic send-home operations. The default value is 10 minutes. Setting this option requires the advanced privilege level or higher.

  • Whether storage-failover status is propagated via mailbox disks. The default is true. Setting this option requires the advanced privilege level or higher.

  • How frequently the node reads its partner's status from the mailbox disks. The default is five seconds. Setting this option requires the advanced privilege level or higher.

  • How frequently the node writes its status to the mailbox disks. The default is five seconds. Setting this option requires the advanced privilege level or higher.

The following example enables the storage-failover service, automatic send-home operations, and partner-readiness checks on a node named node1b.

node::> storage failover modify -node node1b -enabled true
-auto-sendhome true -check-partner true

Taking Over the Storage of a Node's Partner

The storage failover takeover command initiates a takeover of the partner node's storage. When you run the command, specify the name of the node that is taking over the partner's storage, not the name of the partner. You can also specify the style of takeover operation; possible styles include the following:

  • normal—Specifies a normal takeover operation; that is, the partner is given the time to close its storage resources gracefully before the takeover operation proceeds. This is the default value.

  • immediate—Specifies an immediate takeover. In an immediate takeover, the takeover operation is initiated before the partner is given the time to close its storage resources gracefully.

  • allow-version-mismatch—If this value is specified, the takeover operation is initiated even if the partner is running a version of Data ONTAP 10.0 that is incompatible with the version running on the node. In this case, the partner must be cleanly halted. Use this value as part of a nondisruptive upgrade procedure.

  • force—If this value is specified, the takeover operation is initiated even if the node detects an error that normally prevents a takeover operation from occurring. Setting this option requires the advanced privilege level or higher.

The following example causes a node named node1b to initiate an immediate takeover of its partner's storage.

node::> storage failover takeover -bynode node1b -option immediate

Sending Storage Back to its Home Node

By using the storage failover send-home command, you can return storage that has failed over to a node's partner back to the home node. The command fails if other resource-intensive operations (for example, system dumps) are running, thereby making the send-home operation potentially dangerous or disruptive. When you run the command, specify the name of the node that currently holds the storage, not the node to which the storage is being returned. You can also specify the following options:

  • Whether the storage is sent home regardless of whether the partner is available to take back the storage. The default is true. If this option is set to true and the partner is not available to take back its storage, the storage failover send-home command fails.

  • Whether to override partner veto votes during a send-home operation. The default is false. Setting this option requires the advanced privilege level or higher.

The following example sends home storage that is currently held by a node named node1b. The partner must be available for the send-home operation to occur.

node::> storage failover send-home -fromnode node1b
-require-partner-waiting true

Bringing Storage Back to its Home Node

If the node to which storage has failed over is unavailable but the storage's home node is available, you can use the storage failover bring-home command to return the storage to its home node. This command is available only in maintenance mode. NetApp recommends that you use this command only when all other attempts to return storage to its home filer have failed. When you run the command, specify the node to which the storage is to be returned.

The following example returns storage to its home node, named node0b.

node::maint::> storage failover bring-home -tonode node0b

Displaying Information about Mailbox Disks

Storage failover uses mailbox disks to communicate status between the nodes in a storage-failover pair. You can use the storage failover mailbox-disk show to display information about mailbox disks. This information is not useful for the daily administration of a Data ONTAP 10.0 system; use it only with the advice of support personnel. The command requires the advanced privilege level or higher.

Displaying Information about Storage-Failover Interconnections

Storage failover operations occur over an Infiniband interconnection between the nodes of a storage-failover pair. The following commands are available at the advanced privilege level and higher to display the status of the interconnection. Use these commands only with the advice of support personnel. For detailed information about these commands, see the reference pages.

  • storage failover interconnect link—Displays information about the interconnection link

  • storage failover interconnect queue—Displays information about the interconnection's send and receive queues

  • storage failover interconnect status—Displays the state of the interconnection and any active logical links

  • storage failover interconnect statistics performance basic—Displays basic performance statistics for the interconnection

  • storage failover interconnect statistics performance per-interface—Displays performance statistics for the interconnection on a per-interface basis

Prev Up Next
Managing Disks Home Chapter 7. Managing Virtual Servers