Bulk index source reads data for all indices

[danielmitterdorfer]

Consider the following (incomplete) track snippet:

{
  "indices": [
    {
      "name": "elasticlogs_generated",
      "types": [
        {
          "name": "type",
          "mapping": "mappings.json",
          "documents": "generated_elasticlogs_1M.txt.bz2",
          "document-count": 1000000,
        }
      ]
    },
    {
      "name": "elasticlogs_live",
      "types": [
        {
          "name": "type",
          "mapping": "mappings.json",
          "documents": "live_elasticlogs_1M.txt.bz2",
          "document-count": 1000000,
        }
      ]
    }
  ],
  "operations": [
    {
      "name": "index-generated",
      "operation-type": "index",
      "index": "elasticlogs_generated",
      "bulk-size": 1000
    }
  ]
}

• Expectation: index-generated should index 1 million documents
• Actual result: it indexes 2 million documents

Analysis: The parameter source gets data about all indices but does not filter against the provided index, hence it reads both indices.

Note: This is not a problem for our official tracks as they don't use Rally in such a way.