Managing provisioned SSD read cache - FSx for Lustre
Considerations when updating SSD read cacheUpdating a provisioned SSD read cacheMonitoring SSD read cache updates

Managing provisioned SSD read cache

When you create a file system with the Intelligent-Tiering storage class, you have the option to also provision an SSD-based read cache which provides SSD latencies for reads of your frequently-accessed data, up to 3 IOPS per GiB.

You can configure your SSD read cache for frequently-accessed data with one of these sizing mode options:

  • Automatic (proportional to throughput capacity). With Automatic, Amazon FSx for Lustre automatically selects an SSD data read cache size based on provisioned throughput capacity.

  • Custom (user-provisioned). With Custom, you can customize the size of your SSD read cache and scale it up or down at any time based on your workload's needs.

  • Choose No Cache if you do not want to use an SSD data read cache with your file system.

In Automatic (proportional to throughput capacity) mode, Amazon FSx automatically provisions the following default read cache size based on the throughput capacity of your file system.

Provisioned throughput capacity (MBps) SSD read cache in Automatic (proportional to throughput capacity) mode (GiB) Supported SSD read cache size

minimum (GiB)

maximum (GiB)
Every 4000 20000 32 131072

After your file system is created, you can modify your read cache's sizing mode and storage capacity at any time.

Considerations when updating SSD read cache

Here are a few important considerations when modifying your SSD data read cache:

  • Any time you modify the SSD read cache, all of its contents will be erased. This means that you may see a decrease in performance levels until the SSD read cache is populated again.

  • You can increase or decrease the capacity size of an SSD read cache. However, you can only do this once every six hours. There is no time restriction when adding or removing an SSD read cache from your file system.

  • You must increase or decrease the size of an SSD read cache by at least 10% every time you modify it.

Updating a provisioned SSD read cache

You can update your SSD data read cache using the Amazon FSx console, the AWS CLI, or the Amazon FSx API.

  1. Open the Amazon FSx console at https://console.aws.amazon.com/fsx/.

  2. In the left navigation pane, choose File systems. In the File systems list, choose the FSx for Lustre file system that you want to update the SSD read cache for.

  3. SSD On the Summary panel, choose Update next to the file system's SSD read cache value.

    The Update SSD read cache dialog box appears.

  4. Select the new sizing mode that you would like for your data read cache, as follows:

    • Choose Automatic (proportional to throughput capacity) to have your data read cache automatically sized based on your throughput capacity.

    • Choose Custom (user-provisioned) if you know the approximate size of your dataset and would like to customize your data read cache. If you select Custom, you will also need to specify the Desired read cache capacity in GiB.

    • Choose None if you do not want to use an SSD data read cache with your Intelligent-Tiering file system.

  5. Choose Update.

To update the SSD data read cache for an Intelligent-Tiering file system, use the AWS CLI command update-file-system or the equivalent UpdateFileSystem API action. Set the following parameters:

  • Set --file-system-id to the ID of the file system that you are updating.

  • To modify your SSD read cache, use the --lustre-configuration DataReadCacheConfiguration property. This property has two parameters, SizeGiB and SizingMode:

    • SizeGiB ‐ Sets the size of your SSD read cache in GiB when using USER_PROVISIONED mode.

    • SizingMode ‐ Sets the sizing mode of your SSD read cache.

      • Set to NO_CACHE if you do not want to use an SSD read cache with your Intelligent-Tiering file system.

      • Set to USER_PROVISIONED to specify the exact size of your SSD read cache.

      • Set to PROPORTIONAL_TO_THROUGHPUT_CAPACITY to have your SSD data read cache automatically sized based on your throughput capacity.

The following example updates the SSD read cache to USER_PROVISIONED mode and sets the size to 524288 GiB.

aws fsx update-file-system \
   --file-system-id fs-0123456789abcdef0 \
   --lustre-configuration 'DataReadCacheConfiguration={SizeGiB=524288,SizingMode=USER_PROVISIONED}'

To monitor the progress of the update, use the describe-file-systems AWS CLI command. Look for the AdministrativeActions section in the output.

For more information, see AdministrativeAction in the Amazon FSx API Reference.

Monitoring SSD read cache updates

You can monitor the progress of an SSD read cache update by using the Amazon FSx console, the API, or the AWS CLI.

Monitoring updates in the console

You can monitor file system updates in the Updates tab on the File system details page.

For SSD read cache updates, you can view the following information:

Update type

Supported types are SSD read cache sizing mode and SSD read cache size.

Target value

The updated value for the file system's SSD read cache sizing mode or SSD read cache size.

Status

The current status of the update. The possible values are as follows:

  • Pending – Amazon FSx has received the update request, but has not started processing it.

  • In progress – Amazon FSx is processing the update request.

  • Completed – The update finished successfully.

  • Failed – The update request failed. Choose the question mark (?) to see details on why the request failed.

Request time

The time that Amazon FSx received the update action request.

Monitoring SSD read cache updates with the AWS CLI and API

You can view and monitor file system SSD read cache update requests using the describe-file-systems AWS CLI command and the DescribeFileSystems API operation. The AdministrativeActions array lists the 10 most recent update actions for each administrative action type. When you update a file system's SSD read cache, a FILE_SYSTEM_UPDATE AdministrativeActions is generated.

The following example shows an excerpt of the response of a describe-file-systems CLI command. The file system has a pending administrative action to change the SSD read cache sizing mode to USER_PROVISIONED and the SSD read cache size to 524288.

"AdministrativeActions": [
    {
        "AdministrativeActionType": "FILE_SYSTEM_UPDATE",
        "RequestTime": 1586797629.095,
        "Status": "PENDING",
        "TargetFileSystemValues": {
            "LustreConfiguration": {
                "DataReadCacheConfiguration": {
                     "SizingMode": "USER_PROVISIONED"
                     "SizeGiB": 524288,
                }
            }
        }
    }
]

When the new SSD read cache configuration is available to the file system, the FILE_SYSTEM_UPDATE status changes to COMPLETED. If the SSD read cache update request fails, the status of the FILE_SYSTEM_UPDATE action changes to FAILED.