Multipart Upload APIs (SDK for Node.js)
To upload a large file, multipart upload is recommended. Multipart upload is applicable to many scenarios. Below are some examples.
- Files to be uploaded are larger than 100 MB.
- The network connection to the OBS server breaks often.
- Sizes of files to be uploaded are uncertain.
Multipart upload consists of three phases:
- Initiating a Multipart Upload (SDK for Node.js)
- Uploading a Part (SDK for Node.js)
- Assembling Parts (SDK for Node.js) or Aborting a Multipart Upload (SDK for Node.js)
// Import the OBS library. // Use npm to install the client. const ObsClient = require('esdk-obs-nodejs'); // Use the source code to install the client. // var ObsClient = require('./lib/obs'); const fs = require('fs'); // Create an ObsClient instance. var obsClient = new ObsClient({ // Obtain an AK/SK pair using environment variables or import an AK/SK pair in other ways. Using hard coding may result in leakage. // Obtain an AK/SK pair on the management console. For details, see https://support.huaweicloud.com/intl/en-us/usermanual-ca/ca_01_0003.html. access_key_id: process.env.ACCESS_KEY_ID, secret_access_key: process.env.SECRET_ACCESS_KEY, // (Optional) If you use a temporary AK/SK pair and a security token to access OBS, you are advised not to use hard coding, which may result in information leakage. You can obtain an AK/SK pair using environment variables or import an AK/SK pair in other ways. // security_token: process.env.SECURITY_TOKEN, // Enter the endpoint corresponding to the region where the bucket is located. CN-Hong Kong is used here as an example. Replace it with the one currently in use. server: "https://obs.ap-southeast-1.myhuaweicloud.com" }); // Specify the bucket name. const Bucket = 'bucketname' // Specify an object (example/objectname in this example). const Key = 'objectname' // Specify the part size. const DEFAULT_PART_SIZE = 9 * 1024 * 1024; // Set the number of concurrent parts to 20. const CONCURRENCY = 20 // Prepare multipart upload parameters. const preparePartParams = (Bucket, Key, UploadId) => (sampleFile, partSize = DEFAULT_PART_SIZE) => { try { const fileSize = fs.lstatSync(sampleFile).size; const partCount = fileSize % partSize === 0 ? Math.floor(fileSize / partSize) : Math.floor(fileSize / partSize) + 1; const uploadPartParams = []; // Specify the concurrent upload. for (let i = 0; i < partCount; i++) { // Start position of parts in the file let Offset = i * partSize; // Part size let currPartSize = (i + 1 === partCount) ? fileSize - Offset : partSize; // Part number let PartNumber = i + 1; uploadPartParams.push({ Bucket, Key, PartNumber, UploadId, Offset, SourceFile: sampleFile, PartSize: currPartSize, }); }; return ({ uploadPartParams, fileSize }); } catch (error) { console.log(error) }; }; /** * uploadSuccessSize: Size of the parts that have been uploaded * uploadSuccessCount: Number of the parts that have been uploaded * concurrency: Current concurrency */ let uploadSuccessSize = 0; let uploadSuccessCount = 0; let concurrency = 0 const parts = []; // Upload parts. const uploadPart = (uploadPartParam, otherUploadPartInfo) => { const partCount = otherUploadPartInfo.partCount; const fileSize = otherUploadPartInfo.fileSize; concurrency++; return obsClient .uploadPart(uploadPartParam) .then(result => { const { PartNumber, PartSize } = uploadPartParam; if (result.CommonMsg.Status < 300) { uploadSuccessCount++; uploadSuccessSize += PartSize; // Print the upload progress. console.log(`the current concurrent count is ${concurrency} | uploaded segment: ${uploadSuccessCount}/${partCount}. the progress is ${((uploadSuccessSize / fileSize) * 100).toFixed(2)}% | the partNumber ${PartNumber} upload successed.`); parts.push({ PartNumber, ETag: result.InterfaceResult.ETag }); } else { console.log(result.CommonMsg.Code, parts); }; concurrency--; }).catch(function (err) { console.log(err); throw err; }) }; // Multipart upload const uploadFile = (sourceFile) => { // Initiate the multipart upload task. obsClient.initiateMultipartUpload({ Bucket, Key }).then(res => { const Status = res.CommonMsg.Status; const UploadId = res.InterfaceResult.UploadId; if (typeof Status === 'number' && Status > 300) { console.log(`initiateMultipartUpload failed! Status:${Status}`); return; }; const partParams = preparePartParams(Bucket, Key, UploadId)(sourceFile); const uploadPartParams = partParams.uploadPartParams; const fileSize = partParams.fileSize; const partCount = uploadPartParams.length; const otherUploadPartInfo = { fileSize, partCount }; // Call the parallel upload function. parallelFunc(uploadPartParams, (param) => uploadPart(param, otherUploadPartInfo), CONCURRENCY) .then(() => { // Assemble parts. obsClient.completeMultipartUpload({ Bucket, Key, UploadId, Parts: parts.sort((a, b) => a.PartNumber - b.PartNumber) }, (err, result) => { if (err) { console.log('Error-->' + err); } else { console.log('Status-->' + result.CommonMsg.Status); }; }); }); }).catch(function (err) { console.log(err) }); }; /** * Implement the parallel function execution. * @param {Array} params Parameter array of the callback function * @param {Promise} promiseFn Callback function * @param {number} limit Number of parallel parts */ const parallelFunc = (params, promiseFn, limit) => { return new Promise((resolve) => { let concurrency = 0; let finished = 0; const count = params.length; const run = (param) => { concurrency++; promiseFn(param) .then(() => { concurrency--; drainQueue(); finished++; if (finished === count) { resolve(); }; }); }; const drainQueue = () => { while (params.length > 0 && concurrency < limit) { var param = params.shift(); run(param); }; }; drainQueue(); }); }; uploadFile('localfile');
When uploading a large file in parts, you need to use the Offset and PartSize parameters to specify the start and end positions of each part in the file.
If the concurrency value is too great, timeout may occur due to network instability. In such case, you need to reduce that value.
Feedback
Was this page helpful?
Provide feedbackThank you very much for your feedback. We will continue working to improve the documentation.See the reply and handling status in My Cloud VOC.
For any further questions, feel free to contact us through the chatbot.
Chatbot