Concurrent PBT for join plan cache

XMLWordPrintableJSON

    • Type: Improvement
    • Resolution: Unresolved
    • Priority: Major - P3
    • None
    • Affects Version/s: None
    • Component/s: None
    • Query Optimization
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      Implement a PBT that has a smaller number of threads (e.g. 2-3) where 1 thread is doing random DDLs and other threads are running join queries. The property we should assert on is that for run of a given query that return results, they all return the same results, and for queries where a DDL affects its plan (e.g. dropped index that was in-use), that they do not crash.

      We should have a static dataset (not generated by PBT), and it should be appropriately sized so that we have both fast and some slow running join queries.

      We should run this PBT on a sanitizer build to detect and race conditions.

            Assignee:
            Ben Shteinfeld
            Reporter:
            Naafiyan Ahmed
            Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

              Created:
              Updated: