[DRILL-4181] Improve memory limit assignment for Sort operators

• Type: Improvement
• Status: In Progress
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 1.4.0
• Assignee: Unassigned
• Reporter: Aman Sinha
• Created: 09/Dec/15 18:56
• Updated: 26/Feb/19 04:10

Description

Currently, the max_query_memory_per_node is divided among all the Sort operators in the query plan, regardless of whether the Sort operators are executing at the same time. Since Sort is a blocking operator, the number of concurrent Sorts running is limited. Thus, the Sort on both sides of a MergeJoin could be concurrently executing but a Sort that occurs above the MergeJoin cannot start until the MergeJoin produces a row which is only when both the child Sorts have produced a row.

Due to the conservative estimate, we have seen queries such as TPC-H Q8 with forced MergeJoin run out-of-memory even with high max_query_memory_per_node. This query plan has 15 Sort operators and with max_width = 23, max_query_memory_per_node = 20GB, each Sort gets only 63MB (20GB/(23*15)). We should improve the algorithm for computing the memory limit.

Issue Links

• Is contained by: DRILL-5716 Queue-based memory assignment for buffering operators