This topic describes how to run the cp command to copy an Object Storage Service (OSS) object to other directories within the same bucket or to other buckets that are located in the same region as the source object. The cp command can be used only to copy objects but not parts that are generated by multipart copy tasks in progress.

Note Sample command lines in this topic are based on the 64-bit Linux system. For other systems, replace ./ossutil64 in the commands with the corresponding binary name. For more information, see ossutil.

Command syntax

./ossutil64 cp cloud_url cloud_url
[-r, --recursive]
[-f --force]
[-u --update]
[--disable-ignore-error]
[--only-current-dir]
[--bigfile-threshold <value>]
[--part-size <value>]
[--checkpoint-dir <value>]
[--encoding-type <value>]
[--include <value>]
[--exclude <value>]
[--meta <value>]
[--acl <value>]
[--disable-crc64]
[--payer <value>]
[-j, --job <value>]
[--parallel <value>]
[--version-id <value>]

The following table describes the options that you can configure to run the cp command.

Option Description
cloud_url Specifies the paths of the source object and the destination object. Format: oss://bucketname/objectname. For example, to copy the source object named srcobject.jpg from the destination bucket named examplebucket to the destination object named destobject.jpg in the same bucket, specify the path of the source object as oss://examplebucket/srcobject.jpg and the path of the destination object as oss://examplebucket/destobject.jpg.
-r, --recursive Specifies recursive operations. If you specify this option in a command, ossutil performs operations on all objects in a bucket that meet the specified condition. If you do not specify this option in a command, ossutil performs operations only on the specified object.
-f --force Specifies the operation to forcibly perform without prompts for confirmation.
-u, --update Specifies that ossutil copies the source object only when the object does not exist in the destination bucket or when the last modified time of the source object is later than that of the destination object.
--disable-ignore-error Specifies that errors are not ignored during batch operations.
--only-current-dir Specifies that only objects in the current directory are copied. Subdirectories in the current directory and objects in these subdirectories are not copied.
-bigfile-threshold Specifies the size of objects to be copied by using resumable upload. Objects that are larger than the value of this option are copied by using resumable upload. Unit: bytes.

Default value: 104857600 (100 MB).

Valid values: 0 to 9223372036854775807.

--part-size Specifies the part size. Unit: bytes. By default, ossutil determines the part size based on the object size.

Valid values: 1 to 9223372036854775807.

--checkpoint-dir Specifies the directory in which the log information of resumable upload tasks is stored. If a resumable upload task fails, ossutil automatically creates the .ossutil_checkpoint directory and records the checkpoint information in the directory. After the resumable upload task is completed, ossutil deletes this directory. Therefore, if you specify a checkpoint directory, make sure that the directory can be deleted.
--encoding-type Specifies the method used to encode the names of objects. Valid values: url. If you do not specify this option, the names of objects are not encoded.
--include Specifies that the command applies to all objects that meet the specified conditions.
--exclude Specifies that the command applies to all objects that do not meet the specified conditions.
--meta Specifies that the metadata of an object is in the header:value#header:value format. Example: Cache-Control:no-cache#Content-Encoding:gzip. For more information about object metadata, see set-meta.
--acl Specifies the access control list (ACL) of an object. Default value: private. Valid values:
  • default: The ACL of the objects is the same as the ACL of the bucket in which the objects are stored.
  • private: Only the bucket owner can perform read and write operations on the objects in the bucket. Other users cannot access the objects in the bucket.
  • public-read: Only the bucket owner can perform write operations on the objects in the bucket. Other users, including anonymous users, can perform only read operations on the objects in the bucket. This may result in unrestricted access to the data in your bucket and unexpected high fees. If a user uploads prohibited data or information, your legitimate interests and rights may be infringed. Therefore, we recommend that you do not set the object ACL to this value except in special cases.
  • public-read-write: All users, including anonymous users, can perform read and write operations on the objects in the bucket. This may result in unrestricted access to the data in your bucket and unexpected high fees. Exercise caution when you set this option to public-read-write.
--disable-crc64 Specifies that CRC-64 is disabled. By default, ossutil enables CRC-64 during data transmission.
--payer Specifies the payer of the traffic and request fees incurred when the command is run. If you want the requester who accesses the resources in the specified path to pay for the traffic and request fees, set this option to requester.
-j, --jobs Specifies the number of concurrent tasks performed across multiple objects. Valid values: 1 to 10000. Default value: 3.
--parallel Specifies the number of concurrent tasks run to upload a single object. Valid values: 1 to 10000. By default, if you do not specify this option, ossutil sets the value of this option based on the operation type and the object size.
--version-id Specifies the version ID of the object that you want to copy. You can use this option for objects only in a bucket that has versioning enabled.

The -j, --jobs and --parallel options in the preceding command syntax can be used to adjust performance if the default number of concurrent tasks does not meet your performance requirements. By default, ossutil calculates the number of concurrent operations based on the object size. When you copy multiple large objects, the actual number of concurrent tasks is calculated by multiplying the number of concurrent jobs by the number of concurrent operations.

  • We recommend that you adjust the number of concurrent tasks to a value smaller than 100 if the resources such as network bandwidth, memory, and CPU of your ECS instance or server are limited. If resources such as the network bandwidth, memory, and CPU are available, you can increase the number of concurrent operations.
  • If the number of concurrent tasks is excessively large, the copy performance of ossutil may decrease, or an end-of-file (EOF) error may occur because thread resources are switched and threads compete for resources. To resolve this issue, you must adjust the values of the -j, --jobs and --parallel options based on the actual available resources. To perform stress testing, specify a small value for the two options before you incrementally increase them to the optimal values.

Sample environment

In this topic, objects are copied between different directories or buckets in Linux. You can modify the parameter settings in the examples based on your operating system and environment. Sample environment:

  • Operating system: Linux
  • Source bucket: examplebucket1
  • Directory 1 in the source bucket: srcfolder1
  • Directory 2 in the source bucket: srcfolder2
  • Destination bucket: examplebucket2
  • Directory in the destination bucket: desfolder

Usage notes

If the source object is a directory whose path is not ended with a forward slash (/), all subdirectories and objects in the directory that include the directory name in their names are copied to the destination bucket. If the source object is a directory whose path is ended with a forward slash (/), all subdirectories and objects in the directory that do not include the directory name in their names are copied to the destination bucket.

For example, the following objects are stored in the srcfolder1 directory within the source bucket named examplebucket1:

srcfolder1/exampleobject1.txt
srcfolder1/exampleobject2.png
srcfolder1/dir1/
srcfolder1/dir1/exampleobject3.jpg
srcfolder1/dir2/
srcfolder1/dir2/exampleobject4.jpg
  • Copy objects from a path that is not ended with a forward slash (/)
    ./ossutil64 cp oss://examplebucket1/srcfolder1  oss://examplebucket2 -r

    The following objects are copied from the source bucket to the destination bucket named examplebucket2:

    srcfolder1/exampleobject1.txt
    srcfolder1/exampleobject2.png
    srcfolder1/dir1/
    srcfolder1/dir1/exampleobject3.jpg
    srcfolder1/dir2/
    srcfolder1/dir2/exampleobject4.jpg
  • Copy objects from a path that is ended with a forward slash (/)
    ./ossutil64 cp oss://examplebucket1/srcfolder1/  oss://examplebucket2 -r

    The following objects are copied from the source bucket to the destination bucket named examplebucket2:

    exampleobject1.txt
    exampleobject2.png
    dir1/
    dir1/exampleobject3.jpg
    dir2/
    dir2/exampleobject4.jpg

Examples for simple copy

  • Copy a single object
    ./ossutil64 cp oss://examplebucket1/examplefile.txt oss://examplebucket1/srcfolder2/                                 
  • Copy incremental objects
    If you specify the --update option when you copy multiple objects, ossutil copies the objects only when the destination objects do not exist, or the last modified time of the source objects is later than that of the destination objects. You can run the following command to copy incremental objects:
    ./ossutil64 cp oss://examplebucket1/srcfolder1/ oss://examplebucket2/path2/ -r --update
    This option can be used to copy objects that fail to be copied or to skip copied objects when you copy incremental objects.
  • Rename objects
    ./ossutil64 cp oss://examplebucket1/examplefile.txt oss://examplebucket1/example.txt                        

    When you run the cp command to rename an object, the original object still exists. You can delete the original object after you rename the object.

  • Copy only objects in the current directory and ignore subdirectories
    ./ossutil64 cp oss://examplebucket1/srcfolder1/ oss://examplebucket1/srcfolder2/ --only-current-dir -r
  • Modify the metadata of an object
    When you copy an object, you can use the --meta option to modify the object metadata in the header:value#header:value... format.
    ./ossutil64 cp oss://examplebucket1/examplefile.txt oss://examplebucket1/ --meta=Cache-Control:no-cache
  • Copy an object from a bucket that has pay-by-requester enabled to a bucket that has pay-by-requester disabled
    ./ossutil64 cp oss://examplebucket1/examplefile.txt oss://examplebucket2/desfolder/  --payer=requester

Copy multiple objects that meet specified conditions

To copy objects that match specified conditions, you can use the --include and --exclude parameters.
  • Copy all objects that are not in the JPG format
    ./ossutil64 cp oss://examplebucket1/srcfolder1/ oss://examplebucket2/desfolder/ --exclude "*.jpg" -r
  • Copy all objects that contain abc in their names and are not in the JPG or TXT format
    ./ossutil64 cp oss://examplebucket1/srcfolder1/ oss://examplebucket2/desfolder/ --include "*abc*" --exclude "*.jpg" --exclude "*.txt" -r

Modify the storage class of an object

When you overwrite an object, you can use the --meta option to convert the storage class of the object. OSS supports the following storage classes:

  • Standard
  • Infrequent Access (IA)
  • Archive

For more information about storage classes, see Overview.

  • Convert the storage class of the specified object to Archive
    ./ossutil64 cp oss://examplebucket1/srcfolder1/examplefile.txt oss://examplebucket1/srcfolder1/examplefile.txt --meta X-oss-Storage-Class:Archive
  • Convert the storage class of all objects in the specified directory to Standard
    ./ossutil64 cp oss://examplebucket1/srcfolder1/ oss://examplebucket1/srcfolder1/ --meta X-oss-Storage-Class:Standard -r
    Notice
    • The storage class of an Archive or Cold Archive object cannot be converted to other storage classes by running the cp command. You must first run the restore (restore objects) command to restore the object, and then run the cp command to convert the storage class of the object.
    • When you run the cp command to overwrite an object, fees may be charged. If an object of the IA, Archive, or Cold Archive storage class is stored for less than the minimum storage duration and is overwritten, you are charged for the minimum storage duration, which includes the remaining duration. For more information, see Storage fees.

Modify the tags of an object

You can run the following command to which the --tagging option is added to modify the tags of an object. Separate the tags with ampersands (&). You can run the following command to modify the tags of an object:
./ossutil64 cp oss://examplebucket1/examplefile.txt oss://examplebucket1/ --tagging "abc=1&bcd=2&……"

For more information about object tags, see object-tagging (add, modify, query, and delete object tags).

Copy and encrypt an object

You can specify the server-side encryption method when you copy an object and store the encrypted object in a bucket. For more information about server-side encryption, see Server-side encryption.
  • Copy an object and set the server-side encryption method to AES-256
    ./ossutil64 cp oss://examplebucket1/examplefile.txt oss://examplebucket1/srcfolder2/ --meta=x-oss-server-side-encryption:AES256
  • Copy an object and set the server-side encryption method to KMS
    ./ossutil64 cp oss://examplebucket1/examplefile.txt oss://examplebucket2/desfolder/ --meta=x-oss-server-side-encryption:KMS
    Notice When you use Key Management Service (KMS) to encrypt an object, OSS creates a customer master key (CMK) in the KMS console for the object. You are charged a small amount of fees when the KMS API operation is called. For more information, see Billing.
  • Copy an object, set the encryption method of the object to SSE-KMS, and specify a CMK ID
    ./ossutil64 cp oss://examplebucket1/examplefile.txt oss://examplebucket2/desfolder/ --meta=x-oss-server-side-encryption:KMS#x-oss-server-side-encryption-key-id:7bd6e2fe-cd0e-483e-acb0-f4b9e1******

Recover objects in a versioned bucket

After you enable versioning for a bucket, objects that are overwritten or deleted in the bucket are saved as previous versions. You can run the following cp command to which the --version-id option is added to overwrite a previous version of an object as the latest version. Example:
./ossutil64 cp oss://examplebucket1/examplefile.txt oss://examplebucket2/ --version-id  CAEQARiBgID8rumR2hYiIGUyOTAyZGY2MzU5MjQ5ZjlhYzQzZjNlYTAyZDE3MDRk
Notice
  • To use the --version-id option, you must run the ls --all-versions command to query version IDs of the object.
  • The --version-id option can be used for objects only in versioned buckets. For more information about how to enable versioning for a bucket, see bucket-versioning.

Common options

If you use ossutil to switch to a bucket that is located in a different region, you can use the -e option to specify the endpoint of the region in which the specified bucket resides. If you use ossutil to switch to a bucket that belongs to another Alibaba Cloud account, you can use the -i option to specify the AccessKey ID of the specified account, and use the -k option to specify the AccessKey secret of the specified account.

For example, you can run the following command to copy the srcobject.png object in the root directory of the source bucket named examplebucket to the destination bucket named destbucket. The source bucket is located in the China (Shanghai) region and is owned by another Alibaba Cloud account.
./ossutil64 cp oss://examplebucket/srcobject.png  oss://destbucket  -e oss-cn-shanghai.aliyuncs.com -i LTAI4Fw2NbDUCV8zYUzA****  -k 67DLVBkH7EamOjy2W5RVAHUY9H****

For more information about other common options that you can use for the cp command, see Common options.