SQL Return only duplicate rows

First, identify the duplicates. Second, join back to extract these rows.

A non-aggregated (or non-window/ranking) self join forms a partial cross join and gives the square of duplicates for any set of keys. Including non-duplicates too. 1 x 1 = 1 after all.

SELECT
    t2.*
FROM
    (
    SELECT 
       StateId, OrderId, OrderTime, PermitId
    FROM
       myTable
    GROUP BY
       StateId, OrderId, OrderTime, PermitId
    HAVING
       COUNT(*) >= 2
    ) T1
    JOIN
    mytable T2 ON T1.StateId = T2.StateId AND T1.OrderId = T2.OrderId AND
                   T1.OrderTime = T2.OrderTime AND T1.PermitId = T2.PermitId