Bug #45840 read_buffer_size allocated for each partition when

 "insert into.. select * from"

When inserting into a partitioned table using 'insert into
<target> select * from <src>', read_buffer_size bytes of memory
are allocated for each partition in the target table.

This resulted in large memory consumption when the number of
partitions are high.

This patch introduces a new method which tries to estimate the
buffer size required for each partition and limits the maximum
buffer size used to maximum of 10 * read_buffer_size, 
11 * read_buffer_size in case of monotonic partition functions.

sql/ha_partition.cc:
  Introduced a method ha_partition::estimate_read_buffer_size
  to estimate buffer size required for each partition. 
  Method ha_partition::start_part_bulk_insert updated
  to update the read_buffer_size before calling bulk upload
  in storage engines.
  Added thd in ha_partition::start_part_bulk_insert method signature.
sql/ha_partition.h:
  Introduced a method ha_partition::estimate_read_buffer_size.
  Added thd in ha_partition::start_part_bulk_insert method signature.

1 files changed, 2 insertions, 1 deletions

diff --git a/sql/ha_partition.h b/sql/ha_partition.h
index f47dfe8f621..1c863d6c294 100644
--- a/sql/ha_partition.h
+++ b/sql/ha_partition.h
@@ -367,7 +367,8 @@ public:
   virtual int end_bulk_insert();
 private:
   ha_rows guess_bulk_insert_rows();
-  void start_part_bulk_insert(uint part_id);
+  void start_part_bulk_insert(THD *thd, uint part_id);
+  long estimate_read_buffer_size(long original_size);
 public:
 
   virtual bool is_fatal_error(int error, uint flags)