chungen0126 commented on code in PR #6989: URL: https://github.com/apache/ozone/pull/6989#discussion_r2993177512
########## hadoop-hdds/docs/content/design/storage-policy.md: ########## @@ -19,379 +20,588 @@ status: draft # Terminology -## Terminology +## Definitions - Storage Policy: Defines where key data replicas should be stored in specific storage tiers. -- Storage Type: The types of disks/Container replicas in a Datanode, storage type could include RAM_DISK, SSD, HDD, ARCHIVE, etc. -- Storage Tier: A set of Container replicas in a cluster that satisfy the storage policy. -- Volume: In this document, unless otherwise specified, a volume refers to the volume of a Datanode.. -- prefix: The prefix in this article, unless otherwise specified, refers to the prefix of the storage policy type, not the ACL prefix. The prefix of the storage policy type is used to configure the prefix of the storage policy for the specified prefix. +- Storage Type: The type of each Datanode volume or container replica. Each Datanode volume can be configured with a + storage type, including SSD, DISK, and ARCHIVE. +- Storage Tier: A specific storage tier is composed of all replicas of a container based on their storage type. For + example, a 3-replica SSD tier consists of 3 replicas of SSD type. +- Volume: In this document, unless otherwise specified, a volume refers to the volume of a Datanode. +- Key: In this document, a key refers to an object in Ozone, including entries in both the KeyTable and FileTable. ## Storage Policy vs Storage Type vs Storage Tier  -The relation of Storage Policy, Storage Type and Storage Tier +The relationship between Storage Policy, Storage Type, and Storage Tier: -- The storage policy is the property of key/bucket/ prefix (Managed by OM); -- The storage tier is the property of Pipeline and Container (Managed by SCM); -- The storage type is the property of volume and Container replicas (Managed by DN); -- Only the storage policy can be modified by the user directly via ozone command; +- The storage policy is the property of key/bucket (managed by OM). +- The storage tier is the property of Pipeline and Container (managed by SCM). +- The storage type is the property of volume and container replica (managed by DN). +- Only the storage policy can be modified by the user directly via the ozone command. Example: -For a keyA, its storage policy is Hot, its Container 1 tier is SSD tier, and Container 1 has three replicas, all of which are of the SSD storage type. +For a keyA, its storage policy is Hot, Its Container tier is SSD tier, the Container has three replicas, all of which +are of the SSD storage type. # User Scenarios -- User A needs a bucket that supports high-performance IO, so create a bucket with the storage policy set to Hot. Data written by User A to bucket will automatically be distributed across the SSD disks in the cluster. -- User B needs higher IO performance for the directory/prefix /project/metadata, so set the storage policy for the prefix /project/metadata to Hot. Subsequently, data written to /project/metadata will be automatically distributed across the SSD disks in the cluster. -- User C has already written key1 to the cluster and requires better IO performance. The storage policy for key1 can be set to Hot, and then a migration can be triggered to move key1 to the SSD disks. -- Use D use command `aws s3 cp myfile.txt s3://my-bucket/myfile.txt --storage-class XXX` upload a file the Ozone SSD tier +- User A needs a bucket that supports high-performance IO, so they create a bucket with the storage policy set to Hot. + Data written by User A to the bucket will automatically be distributed across SSD disks in the cluster. +- User B needs higher IO performance for a specific key. They write a key with the storage policy set to Hot. The + key's data will be distributed across SSD disks in the cluster. +- User C uses the command `aws s3 cp myfile.txt s3://my-bucket/myfile.txt --storage-class STANDARD` to upload a file + to the Ozone SSD tier. The key's data will be distributed across SSD disks in the cluster. + +# Goals + +- Storage Policy: Introduce storage policy and related concepts. Define multiple storage policies and support S3 + storage class. +- Storage Policy Writing: Allow writing keys/files to specified storage tiers based on storage policy. Support S3, + API, and shell command interfaces. +- Storage Policy Update: Enable setting and unsetting storage policies for buckets, and setting storage tiers for + containers. +- Storage Policy Display: Support displaying the storage policy attribute of buckets and keys. Support displaying the + storage tier of SCM containers and pipelines. Support displaying Datanode storage type usage information. Support + checking whether the key storage policy is satisfied. +- Container Balancer: Support migrating container replicas between Datanodes to volumes of the matching storage type. + For example, SSD type container replicas will be migrated to SSD type volumes, and will not be migrated to DISK + type volumes. +- ReplicationManager: Support managing the storage type of container replicas to ensure that container replicas on + Datanodes reside on the correct volumes. Ensure that the storage types of container replicas forming a storage + tier are correct. For example, a 3-replica SSD storage tier container in SCM should consist of 3 SSD type container + replicas, and each container replica should reside on an SSD type volume. +- DiskBalancerService: Support migrating container replicas within a Datanode to volumes of the matching storage type. + For example, SSD type container replicas will be migrated to SSD type volumes, and will not be migrated to DISK + type volumes. + +# Design + +## Supported Storage Policies + +- Supported storage policies: Hot / Warm / Cold +- Supported storage tiers: SSD / DISK / ARCHIVE / EMPTY +- Supported storage types: SSD / DISK / ARCHIVE +- Supported bucket layouts: FILE_SYSTEM_OPTIMIZED, OBJECT_STORE, LEGACY +- S3 storage classes: STANDARD / STANDARD_IA / GLACIER + +### Storage Policy Map to Storage Tier -# Current Status +| Storage Policy | Storage Tier for Write | Fallback Tier for Write | +|----------------|------------------------|-------------------------| +| Hot | SSD | DISK | +| Warm | DISK | EMPTY | +| Cold | ARCHIVE | EMPTY | -- Ozone currently has some support for tiered storage such as storage type, and some parts of this article may already be implemented. -- Currently, in Ozone, when a key is created, the key's Block can appear on any volume of a Datanode. When a key is created, SCM first needs to allocate a Block for the key through Pipelines. The Client then writes the Block to the corresponding Datanode based on the Pipeline information. In this process, the smallest element managed by the SCM Pipeline is the Datanode, and when the Datanode creates a Container, the Container may appear on any volume with enough remaining space. Under the current architecture, Ozone does not support writing data to specific disks +- Storage Tier for Write: The primary storage tier where data is written when a storage policy is specified. +- Fallback Tier for Write: If the specified storage policy cannot be satisfied with the primary storage tier, SCM + will attempt to use this fallback tier to meet the policy requirements. EMPTY means no fallback is available. -# Goal Requirements Specification +### Storage Tier Map to Storage Type -### **Support for Storage Policy Writing and Management** +| Tier | Storage Type of Pipeline | One Replica Container Storage Type | Three Replica Container Storage Type | EC Container Replicas Storage Type | +|---------|--------------------------|-------------------------------------|--------------------------------------|-------------------------------------| +| SSD | SSD | SSD | 3 SSD | n SSD | +| DISK | DISK | DISK | 3 DISK | n DISK | +| ARCHIVE | ARCHIVE | ARCHIVE | 3 ARCHIVE | n ARCHIVE | +| EMPTY | - | - | - | - | -- **Writing keys**: Allow keys to be written to specified storage tiers based on storage policies. -- **Policy Management**: Enable setting, unsetting, and inheriting storage policies for keys, prefixes, and buckets. Inherit policies based on the longest matching prefix or bucket if no specific policy is set. +### Fallback Storage Type for Container Replica Replication/Migration -### **Support for Data Migration Across Different Storage Policies** +| Container Replica Storage Type | Fallback Storage Types (ordered) [1] | +|--------------------------------|--------------------------------------| +| SSD | DISK, ARCHIVE | +| DISK | ARCHIVE | +| ARCHIVE | none | -- **Data Migration**: Support data migration across different storage policies via manual triggers, ensuring data is moved to the appropriate storage tiers. +- Fallback Storage Type: During the container replica replication or migration process, if SCM cannot find a suitable Review Comment: During the container replica replication or migration process, I think we should prioritize migrating those fallbacked data back to their storage type once there is enough space. ########## hadoop-hdds/docs/content/design/storage-policy.md: ########## @@ -19,379 +20,588 @@ status: draft # Terminology -## Terminology +## Definitions - Storage Policy: Defines where key data replicas should be stored in specific storage tiers. -- Storage Type: The types of disks/Container replicas in a Datanode, storage type could include RAM_DISK, SSD, HDD, ARCHIVE, etc. -- Storage Tier: A set of Container replicas in a cluster that satisfy the storage policy. -- Volume: In this document, unless otherwise specified, a volume refers to the volume of a Datanode.. -- prefix: The prefix in this article, unless otherwise specified, refers to the prefix of the storage policy type, not the ACL prefix. The prefix of the storage policy type is used to configure the prefix of the storage policy for the specified prefix. +- Storage Type: The type of each Datanode volume or container replica. Each Datanode volume can be configured with a + storage type, including SSD, DISK, and ARCHIVE. +- Storage Tier: A specific storage tier is composed of all replicas of a container based on their storage type. For + example, a 3-replica SSD tier consists of 3 replicas of SSD type. +- Volume: In this document, unless otherwise specified, a volume refers to the volume of a Datanode. +- Key: In this document, a key refers to an object in Ozone, including entries in both the KeyTable and FileTable. ## Storage Policy vs Storage Type vs Storage Tier  -The relation of Storage Policy, Storage Type and Storage Tier +The relationship between Storage Policy, Storage Type, and Storage Tier: -- The storage policy is the property of key/bucket/ prefix (Managed by OM); -- The storage tier is the property of Pipeline and Container (Managed by SCM); -- The storage type is the property of volume and Container replicas (Managed by DN); -- Only the storage policy can be modified by the user directly via ozone command; +- The storage policy is the property of key/bucket (managed by OM). +- The storage tier is the property of Pipeline and Container (managed by SCM). +- The storage type is the property of volume and container replica (managed by DN). +- Only the storage policy can be modified by the user directly via the ozone command. Example: -For a keyA, its storage policy is Hot, its Container 1 tier is SSD tier, and Container 1 has three replicas, all of which are of the SSD storage type. +For a keyA, its storage policy is Hot, Its Container tier is SSD tier, the Container has three replicas, all of which +are of the SSD storage type. # User Scenarios -- User A needs a bucket that supports high-performance IO, so create a bucket with the storage policy set to Hot. Data written by User A to bucket will automatically be distributed across the SSD disks in the cluster. -- User B needs higher IO performance for the directory/prefix /project/metadata, so set the storage policy for the prefix /project/metadata to Hot. Subsequently, data written to /project/metadata will be automatically distributed across the SSD disks in the cluster. -- User C has already written key1 to the cluster and requires better IO performance. The storage policy for key1 can be set to Hot, and then a migration can be triggered to move key1 to the SSD disks. -- Use D use command `aws s3 cp myfile.txt s3://my-bucket/myfile.txt --storage-class XXX` upload a file the Ozone SSD tier +- User A needs a bucket that supports high-performance IO, so they create a bucket with the storage policy set to Hot. + Data written by User A to the bucket will automatically be distributed across SSD disks in the cluster. +- User B needs higher IO performance for a specific key. They write a key with the storage policy set to Hot. The + key's data will be distributed across SSD disks in the cluster. +- User C uses the command `aws s3 cp myfile.txt s3://my-bucket/myfile.txt --storage-class STANDARD` to upload a file + to the Ozone SSD tier. The key's data will be distributed across SSD disks in the cluster. + +# Goals + +- Storage Policy: Introduce storage policy and related concepts. Define multiple storage policies and support S3 + storage class. +- Storage Policy Writing: Allow writing keys/files to specified storage tiers based on storage policy. Support S3, + API, and shell command interfaces. +- Storage Policy Update: Enable setting and unsetting storage policies for buckets, and setting storage tiers for + containers. +- Storage Policy Display: Support displaying the storage policy attribute of buckets and keys. Support displaying the + storage tier of SCM containers and pipelines. Support displaying Datanode storage type usage information. Support + checking whether the key storage policy is satisfied. +- Container Balancer: Support migrating container replicas between Datanodes to volumes of the matching storage type. + For example, SSD type container replicas will be migrated to SSD type volumes, and will not be migrated to DISK + type volumes. +- ReplicationManager: Support managing the storage type of container replicas to ensure that container replicas on + Datanodes reside on the correct volumes. Ensure that the storage types of container replicas forming a storage + tier are correct. For example, a 3-replica SSD storage tier container in SCM should consist of 3 SSD type container + replicas, and each container replica should reside on an SSD type volume. +- DiskBalancerService: Support migrating container replicas within a Datanode to volumes of the matching storage type. + For example, SSD type container replicas will be migrated to SSD type volumes, and will not be migrated to DISK + type volumes. + +# Design + +## Supported Storage Policies + +- Supported storage policies: Hot / Warm / Cold +- Supported storage tiers: SSD / DISK / ARCHIVE / EMPTY +- Supported storage types: SSD / DISK / ARCHIVE +- Supported bucket layouts: FILE_SYSTEM_OPTIMIZED, OBJECT_STORE, LEGACY +- S3 storage classes: STANDARD / STANDARD_IA / GLACIER + +### Storage Policy Map to Storage Tier -# Current Status +| Storage Policy | Storage Tier for Write | Fallback Tier for Write | +|----------------|------------------------|-------------------------| +| Hot | SSD | DISK | +| Warm | DISK | EMPTY | +| Cold | ARCHIVE | EMPTY | -- Ozone currently has some support for tiered storage such as storage type, and some parts of this article may already be implemented. -- Currently, in Ozone, when a key is created, the key's Block can appear on any volume of a Datanode. When a key is created, SCM first needs to allocate a Block for the key through Pipelines. The Client then writes the Block to the corresponding Datanode based on the Pipeline information. In this process, the smallest element managed by the SCM Pipeline is the Datanode, and when the Datanode creates a Container, the Container may appear on any volume with enough remaining space. Under the current architecture, Ozone does not support writing data to specific disks +- Storage Tier for Write: The primary storage tier where data is written when a storage policy is specified. +- Fallback Tier for Write: If the specified storage policy cannot be satisfied with the primary storage tier, SCM + will attempt to use this fallback tier to meet the policy requirements. EMPTY means no fallback is available. -# Goal Requirements Specification +### Storage Tier Map to Storage Type -### **Support for Storage Policy Writing and Management** +| Tier | Storage Type of Pipeline | One Replica Container Storage Type | Three Replica Container Storage Type | EC Container Replicas Storage Type | +|---------|--------------------------|-------------------------------------|--------------------------------------|-------------------------------------| +| SSD | SSD | SSD | 3 SSD | n SSD | +| DISK | DISK | DISK | 3 DISK | n DISK | +| ARCHIVE | ARCHIVE | ARCHIVE | 3 ARCHIVE | n ARCHIVE | +| EMPTY | - | - | - | - | -- **Writing keys**: Allow keys to be written to specified storage tiers based on storage policies. -- **Policy Management**: Enable setting, unsetting, and inheriting storage policies for keys, prefixes, and buckets. Inherit policies based on the longest matching prefix or bucket if no specific policy is set. +### Fallback Storage Type for Container Replica Replication/Migration -### **Support for Data Migration Across Different Storage Policies** +| Container Replica Storage Type | Fallback Storage Types (ordered) [1] | +|--------------------------------|--------------------------------------| +| SSD | DISK, ARCHIVE | +| DISK | ARCHIVE | +| ARCHIVE | none | -- **Data Migration**: Support data migration across different storage policies via manual triggers, ensuring data is moved to the appropriate storage tiers. +- Fallback Storage Type: During the container replica replication or migration process, if SCM cannot find a suitable + volume type that matches the original container replica's storage type, it will attempt to use the fallback storage + types in order. -### **Adaptation of AWS S3 StorageClass** +[1] A container replica does not know the storage policy of the key or the storage tier of the SCM container it belongs +to. The container replica only knows its own expected storage type, which is why the column name is "Fallback Storage +Types" rather than "Fallback Storage Tier". -- **S3 StorageClass Mapping**: Map AWS S3 storage classes to Ozone storage policies, supporting related API operations (PutObject, CopyObject, Multipart Upload, GetObject, HeadObject, ListObjects). +### AWS S3 StorageClass -### **Management and Monitoring Tools** +| AWS S3 StorageClass | Ozone Storage Policy | +|---------------------|----------------------| +| STANDARD [1] | Hot | +| STANDARD_IA | Warm | +| GLACIER | Cold | +| DEEP_ARCHIVE | Warm | -- **Storage Policy Commands**: Provide tools to view storage policies of containers, datanode usage, and pipeline information. -- **Metrics and Monitoring**: Enable visibility into storage policy compliance, container storage types, and space information across different storage policies. +> AWS StorageClass Valid Values: STANDARD | REDUCED_REDUNDANCY | STANDARD_IA | ONEZONE_IA | INTELLIGENT_TIERING | +> GLACIER | DEEP_ARCHIVE | OUTPOSTS | GLACIER_IR | SNOW | EXPRESS_ONEZONE +> According to AWS S3 documentation, STANDARD is the highest performance S3 StorageClass, but its name is STANDARD, +> which is not straightforward to map to the Ozone SSD tier. -### **Future Enhancements** +[1] The field names here reuse the AWS S3 field names, but the actual semantics differ from AWS S3. For example, in +Ozone, STANDARD represents the Hot storage policy, while in AWS S3, STANDARD has different semantics. -- **Intelligent Storage Policies**: Plan to support automatic data migration based on access frequency, similar to S3 Intelligent-Tiering. -- **Bucket StorageClass Lifecycle Rules: Support setting storage policies Lifecycle Rules at the bucket level.** -- **Recon Support**: Enhance Recon to display relevant storage tier information. +## Component Changes -# Detailed Requirements Specification +### Datanode Container Replica -## Storage Policy and Storage Types +A storage type field is added to container replicas on Datanodes, which is persisted in the container's metadata YAML +file. -### Supported Storage Types +### Bucket, Key -- Specify the Storage Type for each volume through configuration. If no Storage Type is specified, the default value will be DISK. -- Support Storage Type:SSD / DISK / ARCHIVE / RAM_DISK +A storage policy attribute is added to buckets and keys on OM. -### Supported Storage Policies +### SCM Container, Pipeline -Support storage policy: Hot , Warm, Cold +A storage tier attribute is added to containers and pipelines on SCM. A pipeline can support multiple storage tiers. +For example, if all Datanodes in a pipeline have both SSD and DISK type volumes, the pipeline's supported storage tier +attributes will include both SSD and DISK. -### Storage Policies Map To Storage Tiers +## Datanode Volume Storage Type -| Storage Policy | Storage Tier for Write | Fallback Tier for Write | -| --- | --- | --- | -| Hot | SSD | DISK | -| Warm | DISK | none | -| Cold | ARCHIVE | none | -- **Storage Tier For Write**: The priority storage tier where data is written when storage policy is specified. -- **Fallback Tier for Write**: If the specified storage policy cannot be satisfied with the priority storage tier, the SCM will attempt to use this fallback tier to meet the policy requirements. +- Referencing HDFS [1], use configuration to define the storage type of each volume. If no storage type information is + configured for a volume, the storage type will be DISK. + - For example: + ```xml + <property> + <name>hdds.datanode.dir</name> + <value>[SSD]/mnt/disk/0/ozone,[DISK]/mnt/disk/1/ozone</value> + </property> + ``` + Volume /mnt/disk/0/ozone will be SSD storage type, and Volume /mnt/disk/1/ozone will be DISK storage type. -### Storage Tier Map To Storage Type +[1] Refer to https://hadoop.apache.org/docs/stable/hadoop-project-dist/hadoop-hdfs/ArchivalStorage.html section +Configuration. -| Tier | StorageType of Pipeline | One Replication -Container Replicas Storage Type | Three replication -Container Replicas Storage Type | EC -Container Replicas Storage Type | -| --- | --- | --- | --- | --- | -| SSD | SSD | SSD | 3 SSD | n SSD | -| DISK | DISK | DISK | 3 DISK | n DISK | -| ARCHIVE | ARCHIVE | ARCHIVE | 3 ARCHIVE | n ARCHIVE | +## Writing Keys with Storage Policy -### Fallback Storage Type For Container replicas Replication/Migration +### Ozone Filesystem -| Container Replicas Type | Container Replicas Fallback Storage Type [1] | -| --- | --- | -| SSD | DISK | -| DISK | none | -| ARCHIVE | none | -- Container Replicas Fallback Storage Type: During the Container replicas replication or migration process, if the SCM cannot find a suitable volume type that matches the original Container replica's storage type, it will attempt to use this fallback storage tier. +- Support specifying a storage policy when writing a key. + - If a storage policy is specified when writing a key, the key storage policy is the specified storage policy. + - If a storage policy is not specified when writing a key, the storage policy of the key will depend on the bucket. + If the bucket does not have a storage policy set, the default storage policy is determined by the configuration + `ozone.default.storagepolicy`, the default value is WARM. -[1] For a Container replicas, it will not know the Storage Policy of the Container’s key or the tier of the SCM Container located, the Container replicas just know its own expected storage type, So column name is “Fallback Storage Type” +- Fallback strategy: + - When writing a key, if data cannot be written to the specified storage tier, it can be written to the fallback + storage tier, provided the fallback storage tier is not EMPTY and fallback is allowed. + - Fallback control: Fallback can be controlled at the bucket granularity. If allowed, the key will attempt to write + to the fallback storage tier. The default is to allow fallback. -## Support for Ozone Storage Policy Writing and Management +### S3 -### Support storage policy writing +- If StorageClass is specified in the request, the storage policy is derived from the StorageClass in the request. +- If StorageClass is not specified in the request, the storage policy of the bucket is used. +- If the storage policy of the bucket is null, the storage policy is determined by the configuration + `ozone.s3.default.storagepolicy`, the default value is STANDARD_IA, which maps to Ozone storage policy WARM. -- Support specifying a storage policy when writing a key. - - If a storage policy is specified when writing a key, the key storage policy is the specified storage policy. - - If no storage policy is specified, the default behavior refers to the "Inheritance of storage policy" section. - - If a key neither inheriting any storage policy nor specified a storage policy when writing a key, then the key storage policy will be default storage policy (can refers to the "default storage policy" section) - - If the priority storage policy is not satisfied, support writing to the fallback tier if the fallbackStrategy is “allow” +### Supported APIs -### Support fallback strategy configuration +#### Ozone Filesystem -- fallbackStrategy - - Allow (default): In this case, the behavior is similar to HDFS, with automatic fallback, and it does not trigger errors or additional alerts; - - Prohibit: Prohibit fallback; if a tier that satisfies the storage policy cannot be found, the write operation fails directly. +- createKey +- createStreamKey +- createFile +- createStreamFile +- initiateMultipartUpload -### Inheritance of storage policies +#### S3 Request -- If no storage policy is specified (undefined storage policy) when writing a key, the key's storage policy inherits the longest matching prefix. If there is no matching prefix, it inherits the storage policy of the bucket. If the bucket has no effective storage policy [1], the key's storage policy will be the default storage policy . -- If a key is created with an effective storage policy, the storage policy of the key will not change with the storage policy changing of the bucket or prefix. +Refer to +[Using Amazon S3 storage classes - Amazon Simple Storage Service](https://docs.aws.amazon.com/AmazonS3/latest/userguide/storage-class-intro.html) -[1] Effective storage policy means a non-empty storage policy. +- PutObject: + - Support specifying the StorageClass parameter in the PutObject request to determine the storage policy for the + object. +- CopyObject: + - Support specifying a new storage policy (StorageClass) in the CopyObject request and applying the new storage + policy when copying the object from the source location to the target location. If no new storage policy is + specified, inherit the storage policy of the source object. +- Multipart Upload: + - Support specifying the StorageClass parameter in the CreateMultipartUpload request and following the StorageClass + parameter of CreateMultipartUpload in UploadPart. +- GetObject: + - Return the current storage policy of the object in the GetObject response. +- HeadObject: + - Return the metadata of the object, including its storage policy (StorageClass). +- ListObjects: + - Include the storage policy (StorageClass) information of each object in the ListObjects response. -### undefined storage policy +### Write Key Process -- If the user does not specify any storage policy when creating a key, the user's storage policy is undefined storage policy. -- Even if the user's key inherits the storage policy of the bucket/prefix, the user's storage policy is still undefined storage policy. -- Undefined storage policy does not mean the key no storage policy. if the key inherits a storage policy, the key actual storage policy is the inherited storage policy. -- The undefined storage policy will change as the changing of the prefix/bucket storage policy, including when the key is renamed to a prefix with a different storage policy. +#### Normal Write (No Fallback) -### default storage policy + -- If a key neither inheriting any storage policy nor specified a storage policy when writing a key, then the key storage policy will be default storage policy -- The default storage policy is the storage policy for existing keys before the storage policy feature is launched. That is, all keys have at least the default storage policy, even if the key was created before the storage policy feature was launched. -- If the user has not configured a default storage policy, the default storage policy should be Warm. -- The default storage policy can be configured. +- Client sends CreateKey request to OM with storage policy Hot. +- OM requests SCM to allocate a container with storage policy Hot. +- SCM allocates a container and pipeline with storage tier SSD according to storage policy Hot. +- Client determines the storage type of the chunk to be SSD according to the storage tier SSD. +- DN creates the container or writes the chunk on the specified storage medium according to the storage type. -## Storage policy management +#### Fallback Write -### key storage policy + -- Support setting and unsetting the storage policy for keys. - - After unsetting the storage policy, the actual storage policy of the key refers to the "Inheritance of storage policy" section, or is the default storage policy if key do not inherit any storage policy. -- Support displaying the storage policy in the key list/info results(Include whether the storage policy is the default storage policy). +- Client sends CreateKey request to OM with storage policy Hot. +- OM requests SCM to allocate a container with storage policy Hot. +- SCM attempts to allocate a container and pipeline with storage tier SSD according to storage policy Hot, but finds + that there are no available containers or pipelines with SSD tier. SCM then selects the fallback tier DISK. +- Client determines the storage type of the chunk to be DISK according to the fallback storage tier DISK. +- DN creates the container or writes the chunk on the specified storage medium according to the storage type. -### Bucket storage policy management +### Datanode Container Replica Creation -- Support setting and unsetting storage policies for buckets. - - After unsetting the storage policy, the storage policy is the default storage policy. -- Support specifying a storage policy when creating a bucket. - - If no storage policy is specified, the storage policy is the default storage policy. -- Support displaying the storage policy in the bucket list/info results (Include whether the storage policy is the default storage policy). +- When the Ozone client writes data to a container replica on a Datanode, the storage type is included in the request. +- When the Datanode receives a write request for a container replica and the container does not exist, the Datanode + creates the container by selecting an appropriate volume (VolumeChoosingPolicy) based on the storage type in the + request. +- If the storage type is not specified when creating a container replica on the Datanode, the container replica will be + created on a volume of the storage type specified by the configuration `hdds.datanode.default.storagetype` (default + is DISK). This typically only occurs when a client from an older version that does not support the storage policy + feature writes data to the Datanode. +- The Datanode does not perform fallback when creating container replicas. If a volume of the storage type specified in + the client's write request cannot be found, an exception will be thrown. Since SCM selects appropriate Datanodes to + create pipelines and allocate containers based on Datanode reports, this situation should not occur during normal + operation. -### Prefix management +### SCM Pipeline Creation and Selection -- Support creating, deleting, setting, getting, and listing prefixes. -- The display of the prefix storage policy should display whether the storage policy is the default storage policy -- A prefix can only have one type of policy. -- Prefixes do not support unsetting storage policies; deleting a prefix is equivalent to unsetting the storage policy. +- Datanodes report their storage space information to SCM through StorageReportProto, including the space usage and + storage type of all volumes (the default storage type for a Datanode volume is DISK). +- SCM calculates the supported volume types and capacities for each Datanode based on StorageReportProto. +- When creating a container, SCM selects appropriate Datanodes to create a pipeline based on the storage tier + corresponding to the storage policy in the request. + - For example, the Hot storage policy corresponds to the SSD storage tier. For a 3-replica SSD storage tier, the + pipeline will consist of 3 Datanodes that have SSD type volumes. + - For the Hot storage policy with EC 6+3 replication, the pipeline will consist of 9 Datanodes that have SSD + type volumes. +- If SCM cannot find enough suitable Datanodes to create a pipeline, it will try to create a pipeline using the + fallback tier of the storage policy. If the fallback tier is EMPTY, or the fallback tier also cannot find enough + suitable Datanodes, the creation fails. +- SCM does not mix storage types within a storage tier. An SSD storage tier pipeline will only consist of Datanodes + with SSD type volumes, and will not include Datanodes with only DISK type volumes. +- The storage tier is an attribute of the pipeline, indicating the storage tiers supported by that pipeline. A pipeline + can support multiple storage tiers. For example, if Datanodes all have both SSD and DISK type volumes, the pipeline + created on these Datanodes will have supported storage tier attributes of both SSD and DISK. +- When allocating a container through an existing pipeline, SCM filters matching pipelines based on their supported + storage tiers. -### Support for persistent storage of storage policy changes: -- Use ozone admin storagepolicies satisfyStoragePolicy to trigger the migration of corresponding changes, and mark the corresponding storage policy changes as completed. +- BackgroundPipelineCreator will attempt to create pipelines for all storage tiers. If the Datanodes in the cluster + can support a certain storage tier, BackgroundPipelineCreator will automatically create pipelines for that tier. +- The pipeline count limit is calculated independently for each storage tier. -### FSO type buckets: +## Storage Policy Update -- Use prefixes to implement directory-level storage policy management, not directly support setting storage policies for directories. -- Do not support setting storage policies for directory-type keys. +Setting the bucket storage policy is supported. Existing buckets can be updated with a new storage policy. -## Adaptation of AWS S3 +- `ozone sh bucket update --storagepolicy <storagePolicyStr>` + Allowed values: HOT, WARM, COLD, or null to unset. -### Adaptation of AWS S3 StorageClass +Setting the SCM container storage tier attribute is supported. This is mainly used to update existing containers in the +cluster so that they have the specified storage tier attribute. -Not all the StorageClass will be support by the Ozone +- `ozone admin container setstoragetier --storage-tier=<storageTierStr>` + Allowed values: SSD, DISK, ARCHIVE, or null to unset. -A possible solution +## Storage Policy Display -| AWS S3 StorageClass | Ozone StoragePolicy | -| --- | --- | -| STANDARD | Hot | -| STANDARD_IA | Warm | -| GLACIER | COLD | +Display the bucket storage policy attribute: -> According to AWS S3 documentation, STANDARD is the highest performance S3 StorageClass, but its name is STANDARD, which is not easy to convert it to OZONE SSD +- `ozone sh bucket info` -> AWS StorageClass Valid Values: STANDARD | REDUCED_REDUNDANCY | STANDARD_IA | ONEZONE_IA | INTELLIGENT_TIERING | GLACIER | DEEP_ARCHIVE | OUTPOSTS | GLACIER_IR | SNOW | EXPRESS_ONEZONE +Display the key storage policy attribute: -### Adaptation of AWS S3 Related API +- `ozone sh key info` -refer to -[Using Amazon S3 storage classes - Amazon Simple Storage Service](https://docs.aws.amazon.com/AmazonS3/latest/userguide/storage-class-intro.html) +Display the container storage tier, the storage type of its replicas, and the storage type of the volumes where the +replicas reside: -- PutObject: - - Support specifying the StorageClass parameter in the PutObject request to determine the storage policy for the object. -- CopyObject: - - Support specifying a new storage policy (StorageClass) in the CopyObject request and applying the new storage policy when migrating the object from the source location to the target location. If no new storage policy is specified, inherit the storage policy of the source object. -- Multipart Upload - - Support specifying the StorageClass parameter in the CreateMultipartUpload request and following the StorageClass parameter of CreateMultipartUpload in UploadPart. -- GetObject operation: - - Return the current storage policy of the object in the GetObject response. -- HeadObject: - - Support the HeadObject request to return the metadata of the object, including its storage policy (StorageClass). -- ListObjects: - - Include the storage policy (StorageClass) information of each object in the ListObjects response. -- Bucket StorageClass Lifecycle Rules: - - Support setting storage policies Lifecycle Rules at the bucket level and automatically managing the storage policy conversion of objects through policies and lifecycle rules. For example, automatically transfer objects from SSD to HDD or from HDD to NVMe SSD based on the object's age or access frequency. -- ~~RestoreObject operation: [Not Supported]~~ +- `ozone admin container info` -## Support for Storage Policy Management Commands/Metrics +List containers filtered by storage tier: -Lists commands that need to be added/adapted to storage policies, but may not be all commands +- `ozone admin container list --storage-tier=<storageTierStr>` -- Storage policy management - - `ozone admin storagepolicies list`, lists all supported policies in the current cluster. - - `ozone admin storagepolicies get ${key}`, get the key storage policy (include the inherited storage policy) - - `ozone admin storagepolicies check ${key}`, check whether the key storage policy is satisfied - - `ozone admin storagepolicies satisfyStoragePolicy -bucket / -prefix / -key`, triggers migrations to satisfy the corresponding data's storage policy. - - `ozone admin storagepolicies satisfyStoragePolicy status -bucket / -prefix / -key`, checks the migration status of the storage policy. - - `ozone admin storagepolicies checkSatisfyStoragePolicy -bucket / -prefix / -key`, checks whether the specified resource's storage policy is satisfied. - - Containers may be migrated to a fallback tier. At this time, the storage policy of the corresponding resource is not satisfied. -- `ozone admin container info`, supports displaying the storage policy of Containers. -- `ozone admin datanode usageinfo`, supports displaying space information according to different storage policies. -- `ozone admin pipeline list`, supports displaying the storage tier supported by Pipelines. -- `ozone admin pipeline create`, supports creating Pipelines that support specified storage tier. -- Datanode volume related Metrics support displaying the storage policies of volumes. -- SCM supports displaying the data space information of different storage policies in the cluster. +Display the storage tiers supported by pipelines associated with Datanodes: -## Storage Policy Satisfier Service +- `ozone admin datanode list` -- Support data migration across different storage policies. - - Storage policy migrations are triggered by ozone admin storagepolicies satisfyStoragePolicy. Modifying the storage policy of a bucket/prefix/key does not directly trigger storage policy migrations. -- Support manually setting storage policies for buckets/prefixes/keys and implementing asynchronous migrations. -- Responsible for managing storage policy-related migration work. -- Responsible for responding to user requests and checking whether the specified resources satisfy the storage policy. +Display Datanode usage information by storage type: -## Permissions Management +- `ozone admin datanode usageinfo` -- Changing the storage policy of buckets/prefixes/keys requires administrator permissions. +Display cluster-wide space usage information by storage type: -## Storage Policy Supported Replicas Types +- `ozone admin storagepolicy usageinfo` -- Support replica type storage policies. -- Support EC type storage policies. - - EC types cannot support some storage policies that mix different storage media. +Check whether a key's storage policy is satisfied: -## Storage Policy Supported Bucket Types +- `ozone admin storagepolicy check` -- Support all types of buckets. +## Container Balancer -## Compatible HDFS Commands +- Datanodes report the status information of all containers to SCM through ContainerReplicaProto, including the + storage type of the container replica itself and the storage type of the volume where it resides. +- Container Balancer supports balancing based on the space usage of each storage type. Each storage type is calculated + independently. Container replicas are migrated to volumes of the matching storage type based on the container + replica's storage type. +- Container Balancer does not perform fallback during migration. It will not migrate a container replica to a volume + of a different storage type. -HDFS storage policy-related commands are not supported. +## ReplicationManager -Do not support specified storage policy when creating a file using HDFS Filesystem interface, the Hadoop FileSystem (org.apache.hadoop.fs.FileSystem) does not support passing the storage policy by Create method. +- Datanodes report the status information of all containers to SCM through ContainerReplicaProto, including the + storage type of the container replica itself and the storage type of the volume where it resides. +- Datanodes report their storage space information to SCM through StorageReportProto, including the space usage and + storage type of all volumes (the default storage type for a Datanode volume is DISK). +- ReplicationManager introduces two new container health states: MIS_STORAGE_TYPE_WITH_VOLUME and + MIS_STORAGE_TYPE_WITH_CONTAINER. + - MIS_STORAGE_TYPE_WITH_VOLUME: Container replica whose storage type mismatches its volume's storage type. + - ReplicationManager will select a matching volume and migrate the container replica for this health state. + - MIS_STORAGE_TYPE_WITH_CONTAINER: Container replica whose storage type mismatches the container's storage tier. + - ReplicationManager will send a command to update the storage type of the container replica on the Datanode. + - ReplicationManager will first handle MIS_STORAGE_TYPE_WITH_CONTAINER and then handle + MIS_STORAGE_TYPE_WITH_VOLUME. +- ReplicationManager uses the SCM container's storage tier as the source of truth to check whether the container + replicas and the volumes where replicas reside match the storage tier. + - If the SCM container's storage tier is null, no storage tier checks will be performed. Containers created before + the storage policy feature was introduced will have a null storage tier. + - To manage existing containers' replicas (e.g., to prevent these container replicas from occupying SSD type + volume space), the container storage tier needs to be set. Once set, ReplicationManager will manage the + container replicas based on the storage tier. Otherwise, container replicas may reside on volumes of any + storage type. +- When ReplicationManager replicates, migrates, or reconstructs container replicas (e.g., for UNDER_REPLICATED or + MIS_REPLICATED containers), it selects matching Datanodes based on the storage type corresponding to the SCM + container's storage tier. +- When ReplicationManager replicates, migrates, or reconstructs container replicas, if it cannot find a matching + Datanode, it will try to use the fallback storage types defined by the SCM container's storage tier. If the storage + tier has no defined fallback storage types, or no Datanodes matching the fallback storage types can be found, the + operation fails. -- Ozone needs to record the access frequency of keys/buckets. +## DiskBalancerService -## Recon Side Feature +- DiskBalancerService calculates space usage and migrates containers between volumes of the same storage type within a + Datanode. -Support displaying relevant storage tier information in Recon. Need to design more detailed. // TODO +# Configurations -# Architecture Design +New configuration keys introduced by the storage policy feature: -This section aims to describe the architecture and design concepts of the system without delving into specific implementation details. Its purpose is to validate the feasibility and rationality of the technical route, as well as to assess the workload and technical dependencies of related tasks. +- `ozone.default.storagepolicy` + Default value: WARM. The default storage policy for writing keys when neither the key nor the bucket has a storage + policy set. -## Writing Key Process +- `ozone.default.storageTier` + Default value: DISK. The default storage tier. -Writing data to the specified tier: +- `ozone.s3.default.storagepolicy` + Default value: STANDARD_IA. The default S3 storage class when StorageClass is not specified in the S3 request and + the bucket does not have a storage policy set. STANDARD_IA maps to Ozone storage policy WARM. -- When a user creates a key, they set the storage policy and fallbackStrategy parameters and send the request to the OM: - - CreateKey - - CreateFile - - CreateMultipartKey -- Community Ozone: - - CreateStreamKey - - CreateStreamFile - - CreateMultipartStreamKey -- OM adds the storage policy information to the allocateBlock request sent to the SCM. The actual storage strategy should be calculated based on the storage strategy requirements above. -- SCM, upon receiving the request, selects the Pipeline based on the storage policy and allocates a Container. - - The storage type of each Container replicas should be added to for the Datanodes in the Pipeline, just like the replicaIndex for the EC key. - - If the expected storage tier Pipeline cannot be found, attempt to find the fallback tier defined by the storage policy if the fallbackStrategy is allowed. If it cannot be found, the creation process fails. - - If the SCM needs to create a new Container for a specific storage tier, SCM needs to record the Container’s storage tier for the Container. So the ContainerData should add a storage tier field. -- After receiving the Pipeline with the specified storage type information, the client writes the Chunk replica with the storage type information. -- Datanode writes data to the specified Container replica or creates a Container replica in the specified volume based on the storage type . -- OMKeyInfo needs to persist in the storage policy field, if the storage policy does not be specified, the storage policy should be null. +- `hdds.datanode.default.storagetype` + Default value: DISK. The default storage type used when creating a container replica on a Datanode without a + specified storage type. This typically applies to write requests from older clients that do not support the storage + policy feature. -## DN Container Replica Creation +- `ozone.scm.container.allow.null.storage.tier` + Default value: false. Whether to allow selecting containers with a null storage tier during container allocation. + This can be used during the upgrade period to allow new keys to be written to existing containers that do not yet Review Comment: Could you please clarify this configuration. I don't understand why we need it during the upgrade period. ########## hadoop-hdds/docs/content/design/storage-policy.md: ########## @@ -19,379 +20,588 @@ status: draft # Terminology -## Terminology +## Definitions - Storage Policy: Defines where key data replicas should be stored in specific storage tiers. -- Storage Type: The types of disks/Container replicas in a Datanode, storage type could include RAM_DISK, SSD, HDD, ARCHIVE, etc. -- Storage Tier: A set of Container replicas in a cluster that satisfy the storage policy. -- Volume: In this document, unless otherwise specified, a volume refers to the volume of a Datanode.. -- prefix: The prefix in this article, unless otherwise specified, refers to the prefix of the storage policy type, not the ACL prefix. The prefix of the storage policy type is used to configure the prefix of the storage policy for the specified prefix. +- Storage Type: The type of each Datanode volume or container replica. Each Datanode volume can be configured with a + storage type, including SSD, DISK, and ARCHIVE. +- Storage Tier: A specific storage tier is composed of all replicas of a container based on their storage type. For + example, a 3-replica SSD tier consists of 3 replicas of SSD type. +- Volume: In this document, unless otherwise specified, a volume refers to the volume of a Datanode. +- Key: In this document, a key refers to an object in Ozone, including entries in both the KeyTable and FileTable. ## Storage Policy vs Storage Type vs Storage Tier  -The relation of Storage Policy, Storage Type and Storage Tier +The relationship between Storage Policy, Storage Type, and Storage Tier: -- The storage policy is the property of key/bucket/ prefix (Managed by OM); -- The storage tier is the property of Pipeline and Container (Managed by SCM); -- The storage type is the property of volume and Container replicas (Managed by DN); -- Only the storage policy can be modified by the user directly via ozone command; +- The storage policy is the property of key/bucket (managed by OM). +- The storage tier is the property of Pipeline and Container (managed by SCM). +- The storage type is the property of volume and container replica (managed by DN). +- Only the storage policy can be modified by the user directly via the ozone command. Example: -For a keyA, its storage policy is Hot, its Container 1 tier is SSD tier, and Container 1 has three replicas, all of which are of the SSD storage type. +For a keyA, its storage policy is Hot, Its Container tier is SSD tier, the Container has three replicas, all of which +are of the SSD storage type. # User Scenarios -- User A needs a bucket that supports high-performance IO, so create a bucket with the storage policy set to Hot. Data written by User A to bucket will automatically be distributed across the SSD disks in the cluster. -- User B needs higher IO performance for the directory/prefix /project/metadata, so set the storage policy for the prefix /project/metadata to Hot. Subsequently, data written to /project/metadata will be automatically distributed across the SSD disks in the cluster. -- User C has already written key1 to the cluster and requires better IO performance. The storage policy for key1 can be set to Hot, and then a migration can be triggered to move key1 to the SSD disks. -- Use D use command `aws s3 cp myfile.txt s3://my-bucket/myfile.txt --storage-class XXX` upload a file the Ozone SSD tier +- User A needs a bucket that supports high-performance IO, so they create a bucket with the storage policy set to Hot. + Data written by User A to the bucket will automatically be distributed across SSD disks in the cluster. +- User B needs higher IO performance for a specific key. They write a key with the storage policy set to Hot. The + key's data will be distributed across SSD disks in the cluster. +- User C uses the command `aws s3 cp myfile.txt s3://my-bucket/myfile.txt --storage-class STANDARD` to upload a file + to the Ozone SSD tier. The key's data will be distributed across SSD disks in the cluster. + +# Goals + +- Storage Policy: Introduce storage policy and related concepts. Define multiple storage policies and support S3 + storage class. +- Storage Policy Writing: Allow writing keys/files to specified storage tiers based on storage policy. Support S3, + API, and shell command interfaces. +- Storage Policy Update: Enable setting and unsetting storage policies for buckets, and setting storage tiers for + containers. +- Storage Policy Display: Support displaying the storage policy attribute of buckets and keys. Support displaying the + storage tier of SCM containers and pipelines. Support displaying Datanode storage type usage information. Support + checking whether the key storage policy is satisfied. +- Container Balancer: Support migrating container replicas between Datanodes to volumes of the matching storage type. + For example, SSD type container replicas will be migrated to SSD type volumes, and will not be migrated to DISK + type volumes. +- ReplicationManager: Support managing the storage type of container replicas to ensure that container replicas on + Datanodes reside on the correct volumes. Ensure that the storage types of container replicas forming a storage + tier are correct. For example, a 3-replica SSD storage tier container in SCM should consist of 3 SSD type container + replicas, and each container replica should reside on an SSD type volume. +- DiskBalancerService: Support migrating container replicas within a Datanode to volumes of the matching storage type. + For example, SSD type container replicas will be migrated to SSD type volumes, and will not be migrated to DISK + type volumes. + +# Design + +## Supported Storage Policies + +- Supported storage policies: Hot / Warm / Cold +- Supported storage tiers: SSD / DISK / ARCHIVE / EMPTY +- Supported storage types: SSD / DISK / ARCHIVE +- Supported bucket layouts: FILE_SYSTEM_OPTIMIZED, OBJECT_STORE, LEGACY +- S3 storage classes: STANDARD / STANDARD_IA / GLACIER + +### Storage Policy Map to Storage Tier -# Current Status +| Storage Policy | Storage Tier for Write | Fallback Tier for Write | +|----------------|------------------------|-------------------------| +| Hot | SSD | DISK | +| Warm | DISK | EMPTY | +| Cold | ARCHIVE | EMPTY | -- Ozone currently has some support for tiered storage such as storage type, and some parts of this article may already be implemented. -- Currently, in Ozone, when a key is created, the key's Block can appear on any volume of a Datanode. When a key is created, SCM first needs to allocate a Block for the key through Pipelines. The Client then writes the Block to the corresponding Datanode based on the Pipeline information. In this process, the smallest element managed by the SCM Pipeline is the Datanode, and when the Datanode creates a Container, the Container may appear on any volume with enough remaining space. Under the current architecture, Ozone does not support writing data to specific disks +- Storage Tier for Write: The primary storage tier where data is written when a storage policy is specified. +- Fallback Tier for Write: If the specified storage policy cannot be satisfied with the primary storage tier, SCM + will attempt to use this fallback tier to meet the policy requirements. EMPTY means no fallback is available. -# Goal Requirements Specification +### Storage Tier Map to Storage Type -### **Support for Storage Policy Writing and Management** +| Tier | Storage Type of Pipeline | One Replica Container Storage Type | Three Replica Container Storage Type | EC Container Replicas Storage Type | +|---------|--------------------------|-------------------------------------|--------------------------------------|-------------------------------------| +| SSD | SSD | SSD | 3 SSD | n SSD | +| DISK | DISK | DISK | 3 DISK | n DISK | +| ARCHIVE | ARCHIVE | ARCHIVE | 3 ARCHIVE | n ARCHIVE | +| EMPTY | - | - | - | - | -- **Writing keys**: Allow keys to be written to specified storage tiers based on storage policies. -- **Policy Management**: Enable setting, unsetting, and inheriting storage policies for keys, prefixes, and buckets. Inherit policies based on the longest matching prefix or bucket if no specific policy is set. +### Fallback Storage Type for Container Replica Replication/Migration -### **Support for Data Migration Across Different Storage Policies** +| Container Replica Storage Type | Fallback Storage Types (ordered) [1] | +|--------------------------------|--------------------------------------| +| SSD | DISK, ARCHIVE | +| DISK | ARCHIVE | +| ARCHIVE | none | -- **Data Migration**: Support data migration across different storage policies via manual triggers, ensuring data is moved to the appropriate storage tiers. +- Fallback Storage Type: During the container replica replication or migration process, if SCM cannot find a suitable + volume type that matches the original container replica's storage type, it will attempt to use the fallback storage + types in order. -### **Adaptation of AWS S3 StorageClass** +[1] A container replica does not know the storage policy of the key or the storage tier of the SCM container it belongs +to. The container replica only knows its own expected storage type, which is why the column name is "Fallback Storage +Types" rather than "Fallback Storage Tier". -- **S3 StorageClass Mapping**: Map AWS S3 storage classes to Ozone storage policies, supporting related API operations (PutObject, CopyObject, Multipart Upload, GetObject, HeadObject, ListObjects). +### AWS S3 StorageClass -### **Management and Monitoring Tools** +| AWS S3 StorageClass | Ozone Storage Policy | +|---------------------|----------------------| +| STANDARD [1] | Hot | +| STANDARD_IA | Warm | +| GLACIER | Cold | +| DEEP_ARCHIVE | Warm | -- **Storage Policy Commands**: Provide tools to view storage policies of containers, datanode usage, and pipeline information. -- **Metrics and Monitoring**: Enable visibility into storage policy compliance, container storage types, and space information across different storage policies. +> AWS StorageClass Valid Values: STANDARD | REDUCED_REDUNDANCY | STANDARD_IA | ONEZONE_IA | INTELLIGENT_TIERING | +> GLACIER | DEEP_ARCHIVE | OUTPOSTS | GLACIER_IR | SNOW | EXPRESS_ONEZONE +> According to AWS S3 documentation, STANDARD is the highest performance S3 StorageClass, but its name is STANDARD, +> which is not straightforward to map to the Ozone SSD tier. -### **Future Enhancements** +[1] The field names here reuse the AWS S3 field names, but the actual semantics differ from AWS S3. For example, in +Ozone, STANDARD represents the Hot storage policy, while in AWS S3, STANDARD has different semantics. -- **Intelligent Storage Policies**: Plan to support automatic data migration based on access frequency, similar to S3 Intelligent-Tiering. -- **Bucket StorageClass Lifecycle Rules: Support setting storage policies Lifecycle Rules at the bucket level.** -- **Recon Support**: Enhance Recon to display relevant storage tier information. +## Component Changes -# Detailed Requirements Specification +### Datanode Container Replica -## Storage Policy and Storage Types +A storage type field is added to container replicas on Datanodes, which is persisted in the container's metadata YAML +file. -### Supported Storage Types +### Bucket, Key -- Specify the Storage Type for each volume through configuration. If no Storage Type is specified, the default value will be DISK. -- Support Storage Type:SSD / DISK / ARCHIVE / RAM_DISK +A storage policy attribute is added to buckets and keys on OM. -### Supported Storage Policies +### SCM Container, Pipeline -Support storage policy: Hot , Warm, Cold +A storage tier attribute is added to containers and pipelines on SCM. A pipeline can support multiple storage tiers. +For example, if all Datanodes in a pipeline have both SSD and DISK type volumes, the pipeline's supported storage tier +attributes will include both SSD and DISK. -### Storage Policies Map To Storage Tiers +## Datanode Volume Storage Type -| Storage Policy | Storage Tier for Write | Fallback Tier for Write | -| --- | --- | --- | -| Hot | SSD | DISK | -| Warm | DISK | none | -| Cold | ARCHIVE | none | -- **Storage Tier For Write**: The priority storage tier where data is written when storage policy is specified. -- **Fallback Tier for Write**: If the specified storage policy cannot be satisfied with the priority storage tier, the SCM will attempt to use this fallback tier to meet the policy requirements. +- Referencing HDFS [1], use configuration to define the storage type of each volume. If no storage type information is + configured for a volume, the storage type will be DISK. + - For example: + ```xml + <property> + <name>hdds.datanode.dir</name> + <value>[SSD]/mnt/disk/0/ozone,[DISK]/mnt/disk/1/ozone</value> + </property> + ``` + Volume /mnt/disk/0/ozone will be SSD storage type, and Volume /mnt/disk/1/ozone will be DISK storage type. -### Storage Tier Map To Storage Type +[1] Refer to https://hadoop.apache.org/docs/stable/hadoop-project-dist/hadoop-hdfs/ArchivalStorage.html section +Configuration. -| Tier | StorageType of Pipeline | One Replication -Container Replicas Storage Type | Three replication -Container Replicas Storage Type | EC -Container Replicas Storage Type | -| --- | --- | --- | --- | --- | -| SSD | SSD | SSD | 3 SSD | n SSD | -| DISK | DISK | DISK | 3 DISK | n DISK | -| ARCHIVE | ARCHIVE | ARCHIVE | 3 ARCHIVE | n ARCHIVE | +## Writing Keys with Storage Policy -### Fallback Storage Type For Container replicas Replication/Migration +### Ozone Filesystem -| Container Replicas Type | Container Replicas Fallback Storage Type [1] | -| --- | --- | -| SSD | DISK | -| DISK | none | -| ARCHIVE | none | -- Container Replicas Fallback Storage Type: During the Container replicas replication or migration process, if the SCM cannot find a suitable volume type that matches the original Container replica's storage type, it will attempt to use this fallback storage tier. +- Support specifying a storage policy when writing a key. + - If a storage policy is specified when writing a key, the key storage policy is the specified storage policy. + - If a storage policy is not specified when writing a key, the storage policy of the key will depend on the bucket. + If the bucket does not have a storage policy set, the default storage policy is determined by the configuration + `ozone.default.storagepolicy`, the default value is WARM. -[1] For a Container replicas, it will not know the Storage Policy of the Container’s key or the tier of the SCM Container located, the Container replicas just know its own expected storage type, So column name is “Fallback Storage Type” +- Fallback strategy: + - When writing a key, if data cannot be written to the specified storage tier, it can be written to the fallback + storage tier, provided the fallback storage tier is not EMPTY and fallback is allowed. + - Fallback control: Fallback can be controlled at the bucket granularity. If allowed, the key will attempt to write + to the fallback storage tier. The default is to allow fallback. -## Support for Ozone Storage Policy Writing and Management +### S3 -### Support storage policy writing +- If StorageClass is specified in the request, the storage policy is derived from the StorageClass in the request. +- If StorageClass is not specified in the request, the storage policy of the bucket is used. +- If the storage policy of the bucket is null, the storage policy is determined by the configuration + `ozone.s3.default.storagepolicy`, the default value is STANDARD_IA, which maps to Ozone storage policy WARM. -- Support specifying a storage policy when writing a key. - - If a storage policy is specified when writing a key, the key storage policy is the specified storage policy. - - If no storage policy is specified, the default behavior refers to the "Inheritance of storage policy" section. - - If a key neither inheriting any storage policy nor specified a storage policy when writing a key, then the key storage policy will be default storage policy (can refers to the "default storage policy" section) - - If the priority storage policy is not satisfied, support writing to the fallback tier if the fallbackStrategy is “allow” +### Supported APIs -### Support fallback strategy configuration +#### Ozone Filesystem -- fallbackStrategy - - Allow (default): In this case, the behavior is similar to HDFS, with automatic fallback, and it does not trigger errors or additional alerts; - - Prohibit: Prohibit fallback; if a tier that satisfies the storage policy cannot be found, the write operation fails directly. +- createKey +- createStreamKey +- createFile +- createStreamFile +- initiateMultipartUpload -### Inheritance of storage policies +#### S3 Request -- If no storage policy is specified (undefined storage policy) when writing a key, the key's storage policy inherits the longest matching prefix. If there is no matching prefix, it inherits the storage policy of the bucket. If the bucket has no effective storage policy [1], the key's storage policy will be the default storage policy . -- If a key is created with an effective storage policy, the storage policy of the key will not change with the storage policy changing of the bucket or prefix. +Refer to +[Using Amazon S3 storage classes - Amazon Simple Storage Service](https://docs.aws.amazon.com/AmazonS3/latest/userguide/storage-class-intro.html) -[1] Effective storage policy means a non-empty storage policy. +- PutObject: + - Support specifying the StorageClass parameter in the PutObject request to determine the storage policy for the + object. +- CopyObject: + - Support specifying a new storage policy (StorageClass) in the CopyObject request and applying the new storage + policy when copying the object from the source location to the target location. If no new storage policy is + specified, inherit the storage policy of the source object. +- Multipart Upload: + - Support specifying the StorageClass parameter in the CreateMultipartUpload request and following the StorageClass + parameter of CreateMultipartUpload in UploadPart. +- GetObject: + - Return the current storage policy of the object in the GetObject response. +- HeadObject: + - Return the metadata of the object, including its storage policy (StorageClass). +- ListObjects: + - Include the storage policy (StorageClass) information of each object in the ListObjects response. -### undefined storage policy +### Write Key Process -- If the user does not specify any storage policy when creating a key, the user's storage policy is undefined storage policy. -- Even if the user's key inherits the storage policy of the bucket/prefix, the user's storage policy is still undefined storage policy. -- Undefined storage policy does not mean the key no storage policy. if the key inherits a storage policy, the key actual storage policy is the inherited storage policy. -- The undefined storage policy will change as the changing of the prefix/bucket storage policy, including when the key is renamed to a prefix with a different storage policy. +#### Normal Write (No Fallback) -### default storage policy + -- If a key neither inheriting any storage policy nor specified a storage policy when writing a key, then the key storage policy will be default storage policy -- The default storage policy is the storage policy for existing keys before the storage policy feature is launched. That is, all keys have at least the default storage policy, even if the key was created before the storage policy feature was launched. -- If the user has not configured a default storage policy, the default storage policy should be Warm. -- The default storage policy can be configured. +- Client sends CreateKey request to OM with storage policy Hot. +- OM requests SCM to allocate a container with storage policy Hot. +- SCM allocates a container and pipeline with storage tier SSD according to storage policy Hot. +- Client determines the storage type of the chunk to be SSD according to the storage tier SSD. +- DN creates the container or writes the chunk on the specified storage medium according to the storage type. -## Storage policy management +#### Fallback Write -### key storage policy + -- Support setting and unsetting the storage policy for keys. - - After unsetting the storage policy, the actual storage policy of the key refers to the "Inheritance of storage policy" section, or is the default storage policy if key do not inherit any storage policy. -- Support displaying the storage policy in the key list/info results(Include whether the storage policy is the default storage policy). +- Client sends CreateKey request to OM with storage policy Hot. +- OM requests SCM to allocate a container with storage policy Hot. +- SCM attempts to allocate a container and pipeline with storage tier SSD according to storage policy Hot, but finds + that there are no available containers or pipelines with SSD tier. SCM then selects the fallback tier DISK. +- Client determines the storage type of the chunk to be DISK according to the fallback storage tier DISK. +- DN creates the container or writes the chunk on the specified storage medium according to the storage type. -### Bucket storage policy management +### Datanode Container Replica Creation -- Support setting and unsetting storage policies for buckets. - - After unsetting the storage policy, the storage policy is the default storage policy. -- Support specifying a storage policy when creating a bucket. - - If no storage policy is specified, the storage policy is the default storage policy. -- Support displaying the storage policy in the bucket list/info results (Include whether the storage policy is the default storage policy). +- When the Ozone client writes data to a container replica on a Datanode, the storage type is included in the request. +- When the Datanode receives a write request for a container replica and the container does not exist, the Datanode + creates the container by selecting an appropriate volume (VolumeChoosingPolicy) based on the storage type in the + request. +- If the storage type is not specified when creating a container replica on the Datanode, the container replica will be + created on a volume of the storage type specified by the configuration `hdds.datanode.default.storagetype` (default + is DISK). This typically only occurs when a client from an older version that does not support the storage policy + feature writes data to the Datanode. +- The Datanode does not perform fallback when creating container replicas. If a volume of the storage type specified in + the client's write request cannot be found, an exception will be thrown. Since SCM selects appropriate Datanodes to + create pipelines and allocate containers based on Datanode reports, this situation should not occur during normal + operation. -### Prefix management +### SCM Pipeline Creation and Selection -- Support creating, deleting, setting, getting, and listing prefixes. -- The display of the prefix storage policy should display whether the storage policy is the default storage policy -- A prefix can only have one type of policy. -- Prefixes do not support unsetting storage policies; deleting a prefix is equivalent to unsetting the storage policy. +- Datanodes report their storage space information to SCM through StorageReportProto, including the space usage and + storage type of all volumes (the default storage type for a Datanode volume is DISK). +- SCM calculates the supported volume types and capacities for each Datanode based on StorageReportProto. +- When creating a container, SCM selects appropriate Datanodes to create a pipeline based on the storage tier + corresponding to the storage policy in the request. + - For example, the Hot storage policy corresponds to the SSD storage tier. For a 3-replica SSD storage tier, the + pipeline will consist of 3 Datanodes that have SSD type volumes. + - For the Hot storage policy with EC 6+3 replication, the pipeline will consist of 9 Datanodes that have SSD + type volumes. +- If SCM cannot find enough suitable Datanodes to create a pipeline, it will try to create a pipeline using the + fallback tier of the storage policy. If the fallback tier is EMPTY, or the fallback tier also cannot find enough + suitable Datanodes, the creation fails. +- SCM does not mix storage types within a storage tier. An SSD storage tier pipeline will only consist of Datanodes + with SSD type volumes, and will not include Datanodes with only DISK type volumes. +- The storage tier is an attribute of the pipeline, indicating the storage tiers supported by that pipeline. A pipeline + can support multiple storage tiers. For example, if Datanodes all have both SSD and DISK type volumes, the pipeline + created on these Datanodes will have supported storage tier attributes of both SSD and DISK. +- When allocating a container through an existing pipeline, SCM filters matching pipelines based on their supported + storage tiers. -### Support for persistent storage of storage policy changes: -- Use ozone admin storagepolicies satisfyStoragePolicy to trigger the migration of corresponding changes, and mark the corresponding storage policy changes as completed. +- BackgroundPipelineCreator will attempt to create pipelines for all storage tiers. If the Datanodes in the cluster + can support a certain storage tier, BackgroundPipelineCreator will automatically create pipelines for that tier. +- The pipeline count limit is calculated independently for each storage tier. -### FSO type buckets: +## Storage Policy Update -- Use prefixes to implement directory-level storage policy management, not directly support setting storage policies for directories. -- Do not support setting storage policies for directory-type keys. +Setting the bucket storage policy is supported. Existing buckets can be updated with a new storage policy. -## Adaptation of AWS S3 +- `ozone sh bucket update --storagepolicy <storagePolicyStr>` + Allowed values: HOT, WARM, COLD, or null to unset. -### Adaptation of AWS S3 StorageClass +Setting the SCM container storage tier attribute is supported. This is mainly used to update existing containers in the +cluster so that they have the specified storage tier attribute. -Not all the StorageClass will be support by the Ozone +- `ozone admin container setstoragetier --storage-tier=<storageTierStr>` + Allowed values: SSD, DISK, ARCHIVE, or null to unset. -A possible solution +## Storage Policy Display -| AWS S3 StorageClass | Ozone StoragePolicy | -| --- | --- | -| STANDARD | Hot | -| STANDARD_IA | Warm | -| GLACIER | COLD | +Display the bucket storage policy attribute: -> According to AWS S3 documentation, STANDARD is the highest performance S3 StorageClass, but its name is STANDARD, which is not easy to convert it to OZONE SSD +- `ozone sh bucket info` -> AWS StorageClass Valid Values: STANDARD | REDUCED_REDUNDANCY | STANDARD_IA | ONEZONE_IA | INTELLIGENT_TIERING | GLACIER | DEEP_ARCHIVE | OUTPOSTS | GLACIER_IR | SNOW | EXPRESS_ONEZONE +Display the key storage policy attribute: -### Adaptation of AWS S3 Related API +- `ozone sh key info` -refer to -[Using Amazon S3 storage classes - Amazon Simple Storage Service](https://docs.aws.amazon.com/AmazonS3/latest/userguide/storage-class-intro.html) +Display the container storage tier, the storage type of its replicas, and the storage type of the volumes where the +replicas reside: -- PutObject: - - Support specifying the StorageClass parameter in the PutObject request to determine the storage policy for the object. -- CopyObject: - - Support specifying a new storage policy (StorageClass) in the CopyObject request and applying the new storage policy when migrating the object from the source location to the target location. If no new storage policy is specified, inherit the storage policy of the source object. -- Multipart Upload - - Support specifying the StorageClass parameter in the CreateMultipartUpload request and following the StorageClass parameter of CreateMultipartUpload in UploadPart. -- GetObject operation: - - Return the current storage policy of the object in the GetObject response. -- HeadObject: - - Support the HeadObject request to return the metadata of the object, including its storage policy (StorageClass). -- ListObjects: - - Include the storage policy (StorageClass) information of each object in the ListObjects response. -- Bucket StorageClass Lifecycle Rules: - - Support setting storage policies Lifecycle Rules at the bucket level and automatically managing the storage policy conversion of objects through policies and lifecycle rules. For example, automatically transfer objects from SSD to HDD or from HDD to NVMe SSD based on the object's age or access frequency. -- ~~RestoreObject operation: [Not Supported]~~ +- `ozone admin container info` -## Support for Storage Policy Management Commands/Metrics +List containers filtered by storage tier: -Lists commands that need to be added/adapted to storage policies, but may not be all commands +- `ozone admin container list --storage-tier=<storageTierStr>` -- Storage policy management - - `ozone admin storagepolicies list`, lists all supported policies in the current cluster. - - `ozone admin storagepolicies get ${key}`, get the key storage policy (include the inherited storage policy) - - `ozone admin storagepolicies check ${key}`, check whether the key storage policy is satisfied - - `ozone admin storagepolicies satisfyStoragePolicy -bucket / -prefix / -key`, triggers migrations to satisfy the corresponding data's storage policy. - - `ozone admin storagepolicies satisfyStoragePolicy status -bucket / -prefix / -key`, checks the migration status of the storage policy. - - `ozone admin storagepolicies checkSatisfyStoragePolicy -bucket / -prefix / -key`, checks whether the specified resource's storage policy is satisfied. - - Containers may be migrated to a fallback tier. At this time, the storage policy of the corresponding resource is not satisfied. -- `ozone admin container info`, supports displaying the storage policy of Containers. -- `ozone admin datanode usageinfo`, supports displaying space information according to different storage policies. -- `ozone admin pipeline list`, supports displaying the storage tier supported by Pipelines. -- `ozone admin pipeline create`, supports creating Pipelines that support specified storage tier. -- Datanode volume related Metrics support displaying the storage policies of volumes. -- SCM supports displaying the data space information of different storage policies in the cluster. +Display the storage tiers supported by pipelines associated with Datanodes: -## Storage Policy Satisfier Service +- `ozone admin datanode list` -- Support data migration across different storage policies. - - Storage policy migrations are triggered by ozone admin storagepolicies satisfyStoragePolicy. Modifying the storage policy of a bucket/prefix/key does not directly trigger storage policy migrations. -- Support manually setting storage policies for buckets/prefixes/keys and implementing asynchronous migrations. -- Responsible for managing storage policy-related migration work. -- Responsible for responding to user requests and checking whether the specified resources satisfy the storage policy. +Display Datanode usage information by storage type: -## Permissions Management +- `ozone admin datanode usageinfo` -- Changing the storage policy of buckets/prefixes/keys requires administrator permissions. +Display cluster-wide space usage information by storage type: -## Storage Policy Supported Replicas Types +- `ozone admin storagepolicy usageinfo` -- Support replica type storage policies. -- Support EC type storage policies. - - EC types cannot support some storage policies that mix different storage media. +Check whether a key's storage policy is satisfied: -## Storage Policy Supported Bucket Types +- `ozone admin storagepolicy check` -- Support all types of buckets. +## Container Balancer -## Compatible HDFS Commands +- Datanodes report the status information of all containers to SCM through ContainerReplicaProto, including the + storage type of the container replica itself and the storage type of the volume where it resides. +- Container Balancer supports balancing based on the space usage of each storage type. Each storage type is calculated + independently. Container replicas are migrated to volumes of the matching storage type based on the container + replica's storage type. +- Container Balancer does not perform fallback during migration. It will not migrate a container replica to a volume + of a different storage type. -HDFS storage policy-related commands are not supported. +## ReplicationManager -Do not support specified storage policy when creating a file using HDFS Filesystem interface, the Hadoop FileSystem (org.apache.hadoop.fs.FileSystem) does not support passing the storage policy by Create method. +- Datanodes report the status information of all containers to SCM through ContainerReplicaProto, including the + storage type of the container replica itself and the storage type of the volume where it resides. +- Datanodes report their storage space information to SCM through StorageReportProto, including the space usage and + storage type of all volumes (the default storage type for a Datanode volume is DISK). +- ReplicationManager introduces two new container health states: MIS_STORAGE_TYPE_WITH_VOLUME and + MIS_STORAGE_TYPE_WITH_CONTAINER. + - MIS_STORAGE_TYPE_WITH_VOLUME: Container replica whose storage type mismatches its volume's storage type. + - ReplicationManager will select a matching volume and migrate the container replica for this health state. + - MIS_STORAGE_TYPE_WITH_CONTAINER: Container replica whose storage type mismatches the container's storage tier. + - ReplicationManager will send a command to update the storage type of the container replica on the Datanode. + - ReplicationManager will first handle MIS_STORAGE_TYPE_WITH_CONTAINER and then handle + MIS_STORAGE_TYPE_WITH_VOLUME. +- ReplicationManager uses the SCM container's storage tier as the source of truth to check whether the container + replicas and the volumes where replicas reside match the storage tier. + - If the SCM container's storage tier is null, no storage tier checks will be performed. Containers created before + the storage policy feature was introduced will have a null storage tier. + - To manage existing containers' replicas (e.g., to prevent these container replicas from occupying SSD type + volume space), the container storage tier needs to be set. Once set, ReplicationManager will manage the + container replicas based on the storage tier. Otherwise, container replicas may reside on volumes of any + storage type. +- When ReplicationManager replicates, migrates, or reconstructs container replicas (e.g., for UNDER_REPLICATED or + MIS_REPLICATED containers), it selects matching Datanodes based on the storage type corresponding to the SCM + container's storage tier. +- When ReplicationManager replicates, migrates, or reconstructs container replicas, if it cannot find a matching + Datanode, it will try to use the fallback storage types defined by the SCM container's storage tier. If the storage + tier has no defined fallback storage types, or no Datanodes matching the fallback storage types can be found, the + operation fails. -- Ozone needs to record the access frequency of keys/buckets. +## DiskBalancerService -## Recon Side Feature +- DiskBalancerService calculates space usage and migrates containers between volumes of the same storage type within a + Datanode. -Support displaying relevant storage tier information in Recon. Need to design more detailed. // TODO +# Configurations -# Architecture Design +New configuration keys introduced by the storage policy feature: -This section aims to describe the architecture and design concepts of the system without delving into specific implementation details. Its purpose is to validate the feasibility and rationality of the technical route, as well as to assess the workload and technical dependencies of related tasks. +- `ozone.default.storagepolicy` + Default value: WARM. The default storage policy for writing keys when neither the key nor the bucket has a storage + policy set. -## Writing Key Process +- `ozone.default.storageTier` + Default value: DISK. The default storage tier. -Writing data to the specified tier: +- `ozone.s3.default.storagepolicy` + Default value: STANDARD_IA. The default S3 storage class when StorageClass is not specified in the S3 request and + the bucket does not have a storage policy set. STANDARD_IA maps to Ozone storage policy WARM. -- When a user creates a key, they set the storage policy and fallbackStrategy parameters and send the request to the OM: - - CreateKey - - CreateFile - - CreateMultipartKey -- Community Ozone: - - CreateStreamKey - - CreateStreamFile - - CreateMultipartStreamKey -- OM adds the storage policy information to the allocateBlock request sent to the SCM. The actual storage strategy should be calculated based on the storage strategy requirements above. -- SCM, upon receiving the request, selects the Pipeline based on the storage policy and allocates a Container. - - The storage type of each Container replicas should be added to for the Datanodes in the Pipeline, just like the replicaIndex for the EC key. - - If the expected storage tier Pipeline cannot be found, attempt to find the fallback tier defined by the storage policy if the fallbackStrategy is allowed. If it cannot be found, the creation process fails. - - If the SCM needs to create a new Container for a specific storage tier, SCM needs to record the Container’s storage tier for the Container. So the ContainerData should add a storage tier field. -- After receiving the Pipeline with the specified storage type information, the client writes the Chunk replica with the storage type information. -- Datanode writes data to the specified Container replica or creates a Container replica in the specified volume based on the storage type . -- OMKeyInfo needs to persist in the storage policy field, if the storage policy does not be specified, the storage policy should be null. +- `hdds.datanode.default.storagetype` + Default value: DISK. The default storage type used when creating a container replica on a Datanode without a + specified storage type. This typically applies to write requests from older clients that do not support the storage + policy feature. -## DN Container Replica Creation +- `ozone.scm.container.allow.null.storage.tier` + Default value: false. Whether to allow selecting containers with a null storage tier during container allocation. + This can be used during the upgrade period to allow new keys to be written to existing containers that do not yet + have a storage tier set. + +For a pure SSD cluster (where all Datanode volumes are configured with SSD storage type), the following configurations +should be updated accordingly: + +- `ozone.default.storagepolicy` should be set to HOT (Hot maps to SSD storage tier). +- `ozone.default.storageTier` should be set to SSD. +- `ozone.s3.default.storagepolicy` should be set to STANDARD (STANDARD maps to Hot storage policy). +- `hdds.datanode.default.storagetype` should be set to SSD. + +Otherwise, when writing a key without specifying a storage policy, the system will default to the WARM storage policy, +which corresponds to the DISK storage tier. Since WARM does not define a fallback tier (its fallback is EMPTY), SCM +cannot fall back to another tier, and the write will fail because no DISK type volumes are available in the cluster. + +Note that this situation only occurs when the volumes are explicitly configured as SSD in the `hdds.datanode.dir` +configuration (e.g., `[SSD]/mnt/disk/0/ozone`). Even if the underlying physical disks are SSDs, as long as the volume +configuration does not specify a storage type, the Datanode will treat them as DISK type volumes by default, and no +configuration changes are needed. + +# Protobuf Changes + +Major protobuf changes: + +hdds.proto + +```protobuf +enum StorageType { + DISK_TYPE = 1; + SSD_TYPE = 2; + ARCHIVE_TYPE = 3; + RAM_DISK_TYPE = 4; + PROVIDED_TYPE = 5; +} + +enum StorageTierProto { + UNKNOWN_TIER = 0; Review Comment: Why do enums start at 0 by default? This differs from our current Protobuf implementation in Ozone. Is this a new standard? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
