RAID Configuration Validation and Troubleshooting

Before installing ARTESCA, you must validate your RAID configuration on storage controllers to ensure that your system meets the recommended settings.

This pre-check step prevents potential performance issues or installation failures.

Drive Type

Configuration Requirement

HDD data drives

RAID 0 with No Read Ahead/Write Back/Direct IO

SSD/NVMe drives

JBOD mode when available, otherwise RAID 0 with No Read Ahead/Write Through/Direct IO

Note

Newer RAID cards requiring perccli2 and storcli2 utilities have not yet been fully validated. In the meantime, the tests requiring these utilities are being skipped without impacting the installation process.

Use the validate_raid_configuration function to check the RAID configuration of all nodes on your system.

Find below the potential errors reported by the function, and the recommended resolutions for each error.

HDD configuration on volume {logical_volume_id} on node {node_name} should use Write Back mode

Description

A logical volume on the specified node, backed by an HDD drive, is not configured with the “Write Back” cache policy.

Root Cause

The HDD drive is not configured with the expected “wb” write cache policy.

Resolution

  1. Review the cache settings for the logical volume identified by its ID on the specified node.

  2. Ensure that the write cache policy for this volume is configured to “Write Back”. Adjust the settings using your system’s RAID controller management tools if needed.

HDD drive {physical_drive_id} on node {node_name} should be part of RAID0

Description

A physical HDD data drive on the specified node is not part of a RAID 0 configuration.

Root Cause

The HDD data drive is expected to be configured as part of a RAID 0 configuration.

Resolution

  1. Review the RAID configuration for the physical drive identified by its ID on the specified node.

  2. Ensure that the physical drive is part of a RAID 0 configuration.

HDD/SSD configuration on volume {logical_volume_id} on node {node_name} should use Direct IO mode

Description

A logical volume on the specified node, backed by an HDD drive, is not configured with the “Direct IO” policy.

Root Cause

The HDD drive I/O policy is not configured with the expected “direct” policy.

Resolution

  1. Review the I/O settings for the logical volume identified by its ID on the specified node.

  2. Ensure that the I/O policy for this volume is configured to “Direct IO”. Adjust the settings using your system’s RAID controller management tools if needed.

HDD/SSD configuration on volume {logical_volume_id} on node {node_name} should use No Read Ahead mode

Description

A logical volume on the specified node, backed by an HDD drive, is not configured with the “No Read Ahead” cache policy.

Root Cause

The HDD drive is not configured with the expected “nora” read cache policy.

Resolution

  1. Review the cache settings for the logical volume identified by its ID on the specified node.

  2. Ensure that the read cache policy for this volume is configured to “No Read Ahead”. Adjust the settings using your system’s RAID controller management tools if needed.

RAID Level must be RAID0 to be used on node {node_name}. Found volume {device_path} with ID {logical_volume_id}: {raid_level}

Description

A logical volume is configured with a RAID level other than RAID 0 on the specified node.

Root Cause

The HDD data drives are not configured as RAID 0, which is the expected configuration level.

Resolution

  1. Review the selected data drives to make sure you selected the right ones.

  2. Ensure that the RAID level for this logical volume is set to RAID 0. You will need to re-create the RAID array following the Data Disks Preparation procedure.

RAID Level RAID0 must have exactly one drive on node {node_name}. Found {device_path} with ID {logical_volume_id}: {count} drives found

Description

A logical volume is configured with RAID 0 on the specified node, but it contains more than one drive.

Root Cause

RAID 0 is expected to have a one-to-one mapping with a physical drive, but this configuration has a different number of drives.

Resolution

  1. Check the physical drives associated with the RAID 0 logical volume identified by its ID on the specified node.

  2. Adjust the RAID configuration to ensure that a single drive is present.

SSD configuration on volume {logical_volume_id} on node {node_name} should use Write Through mode

Description

A logical volume on the specified node, backed by an SSD or or NVMe drive, is not configured with the “Write Through” cache policy.

Root Cause

The SSD/NVMe drive is not configured with the expected “wt” write cache policy.

Resolution

  1. Review the cache settings for the logical volume identified by its ID on the specified node.

  2. Ensure that the write cache policy for this volume is configured to “Write Through”. Adjust the settings using your system’s RAID controller management tools if needed.

SSD/NVMe drive ({physical_drive_id}) on node {node_name} should use JBOD mode when supported, but is not configured as JBOD

Description

An SSD or NVMe drive on the specified node is part of a RAID configuration instead of using JBOD mode when supported.

Root Cause

The SSD/NVMe drive is not configured as JBOD, even though the RAID controller supports and has JBOD mode enabled.

Resolution

  1. Check if the RAID controller on the specified node supports and has JBOD mode enabled.

  2. Reconfigure the SSD or NVMe drive with the specified ID to operate in JBOD mode instead of being part of a RAID array, following the Data Disks Preparation procedure.

Tip

If you encounter an error message not listed here, contact Scality Support.