Copy large files in S3 using the AWS S3 Multipart API.
$ npm install aws-s3-object-multipart-copyconst { S3 } = require('aws-sdk')
const copy = require('aws-s3-object-multipart-copy')
const s3 = new S3()
const source = 's3://source-bucket/path'
const destination = 's3://destination-bucket/path'
// async
copy(source, destination, { s3 })
.then(() => { console.log('done') })
// async with emitter
copy(source, destination, { s3 })
.on('progress', console.log)
.then(() => { console.log('done') })const { Session, Source } = require('aws-s3-object-multipart-copy')
const { SingleBar } = require('cli-progress')
const { S3 } = require('aws-sdk')
const s3 = new S3()
const progress = new SingleBar()
const session = new Session({ s3 })
const source = Source.from(session, 's3://bucket/path')
progress.start(100, 0
session.add(source, 's3://destination/path')
session.run().on('progress', (e) => {
progress.update(e.value.upload.progress)
})Copy source into destination where source or destination can be
a URI or an instance of Source and Destination respectively. opts
can be:
{
s3: null, // an instance of `AWS.S3` <required>
retries: 4, // number of max retries for failed uploads [optional]
partSize: 5 * 1024 * 1024, // the default part size for all uploads in this session [optional]
concurrency: os.cpus().length * os.cpus().length // the upload concurrency [optional]
}Computes the part size for a given contentLength. This is useful if
you want to compute the partSize ahead of time. This module will
compute the correct partSize for very large files if this number is
too small.
const s3 = new S3()
const session = new Session({ s3 }
const source = Source.from(session, 's3://bucket/path')
await source.ready()
const partSize = computePartSize(source.contentLength)The Session class is a container for a multipart copy request.
Create a new Session instance where opts can be:
{
s3: null, // an instance of `AWS.S3` <required>
retries: 4, // number of max retries for failed uploads [optional]
partSize: 5 * 1024 * 1024, // the default part size for all uploads in this session [optional]
concurrency: os.cpus().length * os.cpus().length // the upload concurrency [optional]
}Add a source and destination pair to the session queue.
session.add('s3://source/path', 's3://destination/path')Run the session.
Aborts the running session blocking until the lock is released.
if (oops) {
// blocks until all requests have been aborted
await session.abort()
}then() implementation to proxy to current active session promise.
await sessioncatch() implementation to proxy to current active session promise.
session.catch((err) => {
// handle error
})Emitted when a part is uploaded. The object emitted is the same from the
'progress' event in the Batch
module. The value of event.value is a Part instance containing
information about the upload (ETag, part number) and a pointer to the
MultipartUpload instance at event.value.upload which contains
information like how many parts have been uploaded, the progress as a
percentage, and how many parts are pending.
session.on('progress', (event) => {
console.log(event.value.upload.id, event.value.upload.progress)
})Emitted when an error occurs during the life time of a running session.
session.run().on('error', (err) => {
// handle err
})Emitted when the session has finished running successfully.
session.run().on('end', () => {
// session run finished successfully
})The Source class is a container for a source object in a bucket.
Create a new Source instance where session is an instance of
Session and uriOrOpts can be a S3 URI (s3://bucket/...) or an
object specifying a bucket and key ({bucket: 'bucket', key: 'path/to/file'}).
const source = Source.from(session, 's3://bucket/path')
// or
const source = Source.from(session, { bucket: 'bucket', key: 'path/to/file'})The source's key path in the S3 bucket.
The source's bucket in S3.
The size in bytes of the source in S3.
Wait for the source to be ready (loaded metadata).
await source.ready()The Destination class is a container for a destination object in a bucket.
Create a new Destination instance where session is an instance of
Session and uriOrOpts can be a S3 URI (s3://bucket/...) or an
object specifying a bucket and key ({bucket: 'bucket', key: 'path/to/file'}).
const destination = Destination.from(session, 's3://bucket/path')
// or
const destination = Destination.from(session, { bucket: 'bucket', key: 'path/to/file'})The destination's key path in the S3 bucket.
The destination's bucket in S3.
Wait for the destination to be ready (loaded metadata).
await destination.ready()The MultipartUpload class is a container for a multipart upload request
tracking uploaded parts, progress, etc.
Create a new MultipartUpload instance from a session where source and
destination are Source and Destination instances respectively.
const upload = new MultipartUpload(session, source, destination)The identifier generated for this upload by AWS (UploadID).
The destination key of the upload.
The Part instances for each uploaded part in this multipart upload.
Each Part contains the ETag for the upload request.
The total number of parts to upload.
The destination bucket of the upload.
The total number of pending parts to upload.
The progress as a percentage of the multipart upload.
TODO (I am sorry)
MIT