Performing storage node maintenance recovery

Verify that the status of the storage node to be recovered is "MaintenanceBlockage", "PersistentBlockage", "RemovalFailedAndTemporaryBlockage", "RemovalFailedAndMaintenanceBlockage" or "RemovalFailedAndPersistentBlockage", and then go to the next step.

However, if the storage node auto-recovery function is disabled, go to the next step even if the status of the storage node to be recovered is "TemporaryBlockage."

• (Bare metal) Power on the storage node to be recovered. For details about how to turn on the power, see the hardware document for your storage node.

  If the storage node you want to recover was active at the time of starting this procedure, shut down the storage node by OS shutdown operation. After confirming that the storage node is stopped, power on the storage node again.

  If the OS does not stop after approximately 5 minutes of the shutdown operation, stop the storage node by performing forced shutdown, and then power on the storage node again. For details about how to perform forced shutdown, see the hardware documentation.

  After powering on the storage node and verifying that it has started, proceed to the next step.

  Note:

  • If you are using Hitachi Advanced Server HA800 series, HA800 G2 series, HA800 G3 series, or HA800 G6 series, to shut down the OS, click the Graceful Shutdown (or Momentary Press) button in the Power Controls (or Power & Thermal) section in iLO. To shut down the OS forcibly, click the Force Power Off (or Press and Hold) button in the Power Controls (or Power & Thermal) section in iLO. For detailed operations, see the hardware documentation. When you are using a physical server other than Hitachi Advanced Server HA800 series, HA800 G2 series, HA800 G3 series, and HA800 G6 series, for details about specific procedures for shutting down or stopping the OS forcibly, see the respective hardware documentation.

  • If the storage node cannot be started and you have replaced parts on the target storage node, the replacement parts might be the cause.

    For the action to be taken for each replacement part, see Cause and action taken in storage node maintenance in the VSP One SDS Block and SDS Cloud System Administration or the hardware documentation. If the storage node does not become active after taking action according to the preceding documentation, recover the storage node according to Replacing storage nodes (Bare metal) in the VSP One SDS Block and SDS Cloud System Administration.

• (Cloud for AWS) Use the AWS Management Console to start the storage node VM (EC2 instance) to be recovered.

  If the storage node VM (EC2 instance) you want to recover was active at the time of starting this procedure, stop the EC2 instance by the EC2 instance shutdown operation. Then start the EC2 instance again.

  If the EC2 instance does not stop after approximately 5 minutes of the shutdown operation, stop the EC2 instance by performing forced shutdown, and then start the EC2 instance again.

  From the AWS Management Console, verify whether the storage node VM (EC2 instance) you want to recover has started.

  If start of the storage node VM is verified, proceed to the next step.

  Note:

  • For details about how to start and stop an EC2 instance, see the AWS documentation.
  • Stopping an EC2 instance forcibly can be accomplished by stopping the EC2 instance again while it is stuck in the "stopping" state. For details about operations, see the AWS documentation.

• (Cloud for Google Cloud) Use the Google Cloud console to start the storage node VM (Compute Engine instance) to be recovered.

  If the storage node VM (Compute Engine instance) you want to recover was active at the time of starting this procedure, stop the Compute Engine instance by the Compute Engine instance shutdown operation. Then start the Compute Engine instance again.

  If the Compute Engine instance does not stop after you perform shutdown operation, Google Cloud forcibly stops the Compute Engine instance in approximately 90 seconds. After that, start the Compute Engine instance again.

  From the Google Cloud console, verify whether the storage node VM (Compute Engine instance) you want to recover has started.

  If start of the storage node VM is verified, proceed to the next step.

  Note:

  For details about how to start and stop Compute Engine instances, see the Google Cloud documentation.

• (Cloud for Microsoft Azure) Use the Azure portal to start the storage node VM (Azure VM instance) to be recovered.

  If the storage node VM (Azure VM instance) you want to recover was active at the time of starting this procedure, stop the Azure VM instance by the Azure VM instance shutdown operation. Then start the Azure VM instance again.

  If the Azure VM instance does not stop after you perform shutdown operation, Microsoft Azure forcibly stops the Azure VM instance in approximately 10 minutes.

  After that, start the Azure VM instance again. From the Azure portal, verify whether the storage node VM (Azure VM instance) you want to recover has started.

  If start of the storage node VM is verified, proceed to the next step.

  Note:

  For details about how to start and stop Azure VM instances, see the Microsoft Azure documentation.

• In the Storage Nodes list tab in the Storage Nodes window, select the storage node (one node) to be recovered, click the preceding icon to the right of the Monitor link icon, and then select Maintenance Recovery.

• In the Storage Nodes list tab in the Storage Nodes window (displayed as inventory view), click the preceding icon displayed for the storage node to be recovered, and then select Maintenance Recovery.

• Click the preceding icon in the Storage node detailed information window for the storage node to be edited, and then select Maintenance Recovery.

The storage node maintenance recovery dialog is displayed.

The storage node maintenance recovery dialog is closed and storage node maintenance recovery is performed.

• Successfully maintenance recovered for storage node.

In some cases, storage node maintenance recovery takes a long time, which can cause logout.

In the Jobs window, verify a job whose PATH is "/v1/objects/storage-nodes/<id>/actions/recover/invoke"*. If the STATUS of the job is "Completed", storage node maintenance recovery is successfully completed.

* <id> represents the ID of the target storage node.

For details about the procedure, contact customer support.

Note that, if you continue operations with other procedures, you must back up the configuration information after you have completed all operations.