Re-indexing with ingestSegment firehose

Currently, we have hourly data stored in a cube, and want to periodically roll up data older than 2 weeks old into daily granularity. I am posting the below task. The task runs, and returns a success code, but when I query the data, it still has hourly granularity, and has not been rolled up.
The logs suggest that the datasource is being read from and that the new segments are being persisted, but the data does not reflect the changes

{

“type”: “index”,

“spec”: {

“dataSchema”: {

“metricsSpec”: [

{

“fieldName”: “impressions”,

“type”: “longSum”,

“name”: “impressions”

},

{

“fieldName”: “mouseovers”,

“type”: “longSum”,

“name”: “mouseovers”

},

{

“fieldName”: “clicks”,

“type”: “longSum”,

“name”: “clicks”

}

],

“granularitySpec”: {

“queryGranularity”: “DAY”,

“segmentGranularity”: “DAY”,

“type”: “uniform”,

“intervals”: [

“2017-01-03/2017-01-04”

]

},

“parser”: {

“parseSpec”: {

“dimensionsSpec”: {

“dimensionExclusions”: ,

“spatialDimensions”:

},

“columns”: [

“dim1”,

“dim2”,

“dim3”,

“impressions”,

“mouseovers”,

“clicks”

],

“format”: “json”

},

“type”: “string”

},

“dataSource”: “datasource”

},

“tuningConfig”: {

“rowFlushBoundary”: 500000,

“buildV9Directly”: “false”,

“targetPartitionSize”: 5000000,

“indexSpec”: {

“bitmap”: {

“type”: “concise”

}

},

“numShards”: -1,

“type”: “index”

},

“ioConfig”: {

“firehose”: {

“interval”: “2017-01-03/2017-01-04”,

“type”: “ingestSegment”,

“dataSource”: “datasource”

},

“type”: “index”

}

}

}