Details

    • Type: Sub-task Sub-task
    • Status: Closed
    • Priority: Major Major
    • Resolution: Won't Fix
    • Affects Version/s: None
    • Fix Version/s: None
    • Component/s: None
    • Labels:
      None

      Description

      TestTableIndex and TestTableMapReduce rely on MakeMultiRegionTable, which in turn calls addContent in HBaseTestCase. This method produces 17k rows in the table, and something like 10 regions. That's a lot more than we actually need to prove the functionality of TTI and TTMR.

      Can we reduce the number of rows that we use by shortening the run of addContent? Is there any risk in changing addContent globally to produce less data?

        Activity

        Hide
        stack added a comment -

        When MultiRegionTable was refactored, added 27 regions rather than 10 but less data overall (One run of addContent). Do you still want less regions and data added?

        Show
        stack added a comment - When MultiRegionTable was refactored, added 27 regions rather than 10 but less data overall (One run of addContent). Do you still want less regions and data added?
        Hide
        Bryan Duxbury added a comment -

        So how long do these tests run now?

        Show
        Bryan Duxbury added a comment - So how long do these tests run now?
        Hide
        stack added a comment -

        TTMR takes 56s and TTI 54s on my laptop.

        Show
        stack added a comment - TTMR takes 56s and TTI 54s on my laptop.
        Hide
        Bryan Duxbury added a comment -

        That's a notable improvement. We can live with that for the time being.

        To be clear, my previous suggestion was not to use addContent at all. Instead, use the TestRegionRebalancing approach, and just forge N regions with hardly any rows in them.

        Show
        Bryan Duxbury added a comment - That's a notable improvement. We can live with that for the time being. To be clear, my previous suggestion was not to use addContent at all. Instead, use the TestRegionRebalancing approach, and just forge N regions with hardly any rows in them.
        Hide
        stack added a comment -

        Will I close the issue B?

        Show
        stack added a comment - Will I close the issue B?

          People

          • Assignee:
            Unassigned
            Reporter:
            Bryan Duxbury
          • Votes:
            0 Vote for this issue
            Watchers:
            0 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved:

              Development