a lambda function to automatically take ownership of any objects written into an s3 bucket. Inspired by https://gist.github.com/joech4n/953c1cd6a36698c5d120

lake-s3-object-take-ownership.js

console.log('Loading event');
var aws = require('aws-sdk');
var s3 = new aws.S3({apiVersion: '2006-03-01'});
var sqs = new aws.SQS({apiVersion: '2012-11-05'});

exports.handler = function(event, context, callback) {
  s3.listBuckets(function(err,data) {
    if (err) {
      console.log('ERROR: Problem getting list of buckets. This should have something to do with incorrect IAM permissions for the lambda.');
      errorMessage = 'ERROR: Error from S3: '+err;
      console.log(errorMessage);
      context.done('error', errorMessage);
    } else {
      var canonicalAccountId = data.Owner.ID;

      console.log('Received event:');
      console.log(JSON.stringify(event, null, '  '));

      // TODO: This is only a hack. The delay needs to be removed after Talend jobs fix the ACL issue
      console.log('Delaying for 180 seconds...');
      setTimeout(function() {
        // This converts the SNS message into the same format as from S3 directly
        var s3Event = event.Records[0].Sns.Message;
        
        // Get the data from the event
        var bucket = s3Event.Records[0].s3.bucket.name;
        var key = s3Event.Records[0].s3.object.key;

        var getOwnerParams = {
          Bucket: bucket, /* required */
          Key: decodeURIComponent(key), /* required */
        };

        // Gets the owner of the S3 Object
        s3.getObjectAcl(getOwnerParams, function(err, data) {
          if (err) {
            console.log('ERROR: Problem getting the ACL. This should have something to do with the object not applying the bucket-owner-full-control ACL.');
            errorMessage = 'ERROR: Error from S3: '+err;
            console.log(errorMessage);
            context.done('error', errorMessage);
          } else {
            // If the owner is not the correct, then overwrite itself, else do nothing
            var objectOwnerId = data.Owner.ID;
            if (objectOwnerId != canonicalAccountId) {
              console.log("INFO: Object owner not correct. overwriting...");
              overwriteS3Object();
            } else {
              console.log("INFO: Object owner correct. Pushing event into BucketMapper SQS...");

              var queueUrl = process.env.BUCKETMAPPER_QUEUE_URL;
              var bucketMapperQueueParams = {
                QueueUrl: queueUrl,
                MessageBody: JSON.stringify(s3Event, null, '  ')
              };
              console.log('Writing to bucketmapper queue for bucket:' + bucket + ', key: ' + key);
              sqs.sendMessage(bucketMapperQueueParams, function (err, data) {
                if (err) {
                  var errorMessage = 'ERROR: Error creating SQS message: for bucket:' + bucket + ', key: ' + key + '. ' + err;
                  console.log(errorMessage);
                  callback(errorMessage);
                } else {
                  var successMessage = 'SUCCESS: Created SQS message for bucket:' + bucket + ', key: ' + key;
                  console.log(successMessage);
                  callback(null, successMessage);
                }
              });
            }
          }
        });

        function overwriteS3Object() {
          // Configure params for CopyObject: http://docs.aws.amazon.com/AWSJavaScriptSDK/latest/AWS/S3.html#copyObject-property
          var copyParams = {
            Bucket: bucket,                 /* required */
            CopySource: bucket + '/' + key, /* required */
            Key: decodeURIComponent(key),   /* required; URL-decoded because event URL-encodes the key.
                                           For some reason, CopySource does not require this */
            StorageClass: 'STANDARD',       /* Must include this to be able to overwrite a file */
            ServerSideEncryption: 'AES256' /* server side encryption */
          };

          s3.copyObject(copyParams,
            function(err,data) {
              if (err) {
                console.log('ERROR: Problem overwriting ' + bucket + '/' +key + '. Make sure they exist and your bucket is in the same region as this function.');
                errorMessage = 'ERROR: Error from S3: '+err;
                console.log(errorMessage);
                context.done('error', errorMessage);
              }
              else {
                console.log('SUCCESS: Overwritten ' + bucket + '/' + key);
                context.done(null,'');
              }
            }
          );
        }
      }, 180000);
    }
  });
};

Hi, could you point me in the direction why the timeout hack is introduced in this script?

https://gist.github.com/williamtsoi1/104531c65852a852399a3dc1096a2dcc#file-lake-s3-object-take-ownership-js-L19

+1