Use a distinct carver request_id

[theopolis]

This is half of the solution for #6727. We do not want to use the last-most-recent distributed request ID for all carve requests. Carve requests can be initiated multiple ways, through a scheduled query, distributed query, shell query, etc.

This also improve the UX of the carves table. It will report the single row requested when carve = 1 is used.

A follow up change should explore storing metadata with queries such that the virtual table can retrieve and use this data.

[zwass]

I think this is working as expected:

{
  "host": "zachs-macbook-pro.local",
  "rows": [
    {
      "carve": "1",
      "carve_guid": "f072809b-48ea-401b-8f5a-c27584b6e5ee",
      "host_hostname": "zachs-macbook-pro.local",
      "path": "/etc/hosts",
      "request_id": "73c58a9f-a168-450f-8f57-a5c31529a248",
      "sha256": "",
      "size": "-1",
      "status": "SCHEDULED",
      "time": "1613943074"
    }
  ]
}