Create and manage data caches

Parameters

When you create a data cache, you must configure the following parameters:

• Basic properties

  API	Console	Description
  RegionId	Region	The region where the data cache is located.
  Bucket	Cache bucket	The bucket where the data is stored. The default value is default. You can specify a custom bucket to group services and prevent path conflicts. The eci-system bucket is reserved for ECI and cannot be used.
  Path	Cache directory	The storage path for the data.
  Name	Cache name	The name of the data cache.
  Size	Cache size	The size of the data cache. Unit: GiB. The default value is 20 GiB. We recommend that you estimate the required size based on your actual data volume.
  RetentionDays	Retention period (days)	The number of days to retain the data cache. The cache is deleted after this period expires. By default, a data cache is retained indefinitely.

• Data source

  API	Console	Description
  DataSource.Type	Type	The type of the data source.
  DataSource.Options	Parameters	The configuration parameters for the data source, specified as a JSON object. Required parameters vary depending on the data source type.

  The following table describes the data source types and their parameters.

  Important

  When creating a data cache in the OpenAPI Portal, you must prefix each parameter name in DataSource.Options with its character length. For example, you must write url as #3#url, repoId as #6#repoId, and repoSource as #10#repoSource. This is not required when using an SDK.

  Data source type	Source description	Parameters
  URL	URL	url: The download address of the data. accessToken: An access token required to access private data.
  	HuggingFace or ModelScope	repoSource: The source of the model or dataset. Valid values: ModelScope/Model, ModelScope/DataSet, HuggingFace/Model, and HuggingFace/DataSet. repoId: The ID of the model or dataset. revision: The version. Default value: main or master. accessToken: An access token required to access private data.
  NAS	NAS file system	server: The mount point of the Apsara File Storage for NAS (NAS) file system. path: The subdirectory to mount. For Extreme NAS file systems, the path must start with /share, such as /share/path. vers: The NFS protocol version. We recommend that you use v3. Extreme NAS file systems support only v3. options: The mount options. We recommend that you use the options suggested by the NAS service, such as nolock,tcp,noresvport.
  OSS	Object Storage Service (OSS) bucket	bucket: The name of the OSS bucket. url: The endpoint of the OSS bucket. path: The subdirectory to mount. This is the directory relative to the root directory of the bucket. Default value: /. otherOpts: Custom parameters for mounting the OSS bucket. The format is -o *** -o ***. Example: -o max_stat_cache_size=0 -o allow_other. ramRole: The RAM role for authorization. Note You must create a RAM role and grant permissions to it. For more information, see Create a RAM role for a trusted Alibaba Cloud service and Manage RAM role permissions. When you create the role, set Alibaba Cloud Service to Normal Service Role, Elastic Compute Service to Normal Service Role, and Trusted Service to Elastic Compute Service. When you grant permissions, select the AliyunOSSFullAccess permission policy. akId: The AccessKey ID for direct authorization. For more information, see Obtain an AccessKey pair. akSecret: The AccessKey secret for direct authorization. For more information, see Obtain an AccessKey pair.
  SNAPSHOT	Snapshot	snapshotId: The ID of the snapshot.

• Network parameters

  If the specified vSwitch is not configured with an SNAT gateway, associate an Elastic IP address (EIP) to pull data from the internet.

  API	Console	Description
  EipInstanceId	Use existing	Associates an existing EIP.
  Parameters included in EipCreateParam	Auto create	Automatically creates and associates an EIP.

• Other parameters

  In addition to the preceding parameters, you must also specify a Virtual Private Cloud (VPC), a vSwitch, and a security group. The system uses these to create temporary resources during the data cache creation process.

Query data caches

After you create a data cache, you can query its information, such as its status, the bucket to which it belongs, and its storage path within the bucket. A data cache is ready for use when its status is Available.

You can query a data cache in one of the following ways:

• OpenAPI: Call the DescribeDataCaches operation to query information about data caches.

• Console: On the Data Cache page, you can view a list of all data caches in the current region. Click the ID of a data cache to view its details.

Update data caches

You can update the bucket, data source, retention period, and size of a data cache if its state is Available or UpdateFailed.

You can update a data cache in the following way:

• OpenAPI: Call the UpdateDataCache operation to update a data cache.

• Console: This feature is not supported.

Copy data caches

Data caches are region-specific. If your ECI instance is in a different region than the data cache, you must copy the data cache from the source region to the destination region.

You can copy a data cache in one of the following ways:

• OpenAPI: Call the CopyDataCache operation to copy a data cache.

• Console: On the Data Cache page, find the target data cache and click Copy in the Actions column.

Delete data caches

You can delete a data cache in one of the following ways:

• OpenAPI: Call the DeleteDataCache operation to delete a data cache.

• Console: On the Data Cache page, find the target data cache and click Delete in the Actions column.