Re: POC: postgres_fdw insert batching

Поиск
Список
Период
Сортировка
От Zhihong Yu
Тема Re: POC: postgres_fdw insert batching
Дата
Msg-id CALNJ-vSt7Fn=oj4NxAyCDQQF-1yxmP=LYyQwciApg4L+QdYJQw@mail.gmail.com
обсуждение исходный текст
Ответ на Re: POC: postgres_fdw insert batching  (Tomas Vondra <tomas.vondra@enterprisedb.com>)
Ответы Re: POC: postgres_fdw insert batching  (Tomas Vondra <tomas.vondra@enterprisedb.com>)
Список pgsql-hackers
Дерево обсуждения
Hi,
The assignment to resultRelInfo is done when junk_filter_needed is true:

        if (junk_filter_needed)
        {
            resultRelInfo = mtstate->resultRelInfo;

Should the code for determining batch size access mtstate->resultRelInfo directly ?

diff --git a/src/backend/executor/nodeModifyTable.c b/src/backend/executor/nodeModifyTable.c
index 9c36860704..a6a814454d 100644
--- a/src/backend/executor/nodeModifyTable.c
+++ b/src/backend/executor/nodeModifyTable.c
@@ -2798,17 +2798,17 @@ ExecInitModifyTable(ModifyTable *node, EState *estate, int eflags)
      * size (a FDW may support batching, but it may be disabled for the
      * server/table).
      */
-    if (!resultRelInfo->ri_usesFdwDirectModify &&
+    if (!mtstate->resultRelInfo->ri_usesFdwDirectModify &&
         operation == CMD_INSERT &&
-        resultRelInfo->ri_FdwRoutine != NULL &&
-        resultRelInfo->ri_FdwRoutine->GetForeignModifyBatchSize &&
-        resultRelInfo->ri_FdwRoutine->ExecForeignBatchInsert)
-        resultRelInfo->ri_BatchSize =
-            resultRelInfo->ri_FdwRoutine->GetForeignModifyBatchSize(resultRelInfo);
+        mtstate->resultRelInfo->ri_FdwRoutine != NULL &&
+        mtstate->resultRelInfo->ri_FdwRoutine->GetForeignModifyBatchSize &&
+        mtstate->resultRelInfo->ri_FdwRoutine->ExecForeignBatchInsert)
+        mtstate->resultRelInfo->ri_BatchSize =
+            mtstate->resultRelInfo->ri_FdwRoutine->GetForeignModifyBatchSize(mtstate->resultRelInfo);
     else
-        resultRelInfo->ri_BatchSize = 1;
+        mtstate->resultRelInfo->ri_BatchSize = 1;

-    Assert(resultRelInfo->ri_BatchSize >= 1);
+    Assert(mtstate->resultRelInfo->ri_BatchSize >= 1);

     /*
      * Lastly, if this is not the primary (canSetTag) ModifyTable node, add it

Cheers

On Wed, Jan 20, 2021 at 3:52 PM Tomas Vondra <tomas.vondra@enterprisedb.com> wrote:
Hmm, seems that florican doesn't like this :-(

https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=florican&dt=2021-01-20%2023%3A08%3A15

It's a i386 machine running FreeBSD, so not sure what exactly it's picky
about. But when I tried running this under valgrind, I get some strange
failures in the new chunk in ExecInitModifyTable:

   /*
    * Determine if the FDW supports batch insert and determine the batch
    * size (a FDW may support batching, but it may be disabled for the
    * server/table).
    */
   if (!resultRelInfo->ri_usesFdwDirectModify &&
       operation == CMD_INSERT &&
       resultRelInfo->ri_FdwRoutine != NULL &&
       resultRelInfo->ri_FdwRoutine->GetForeignModifyBatchSize &&
       resultRelInfo->ri_FdwRoutine->ExecForeignBatchInsert)
       resultRelInfo->ri_BatchSize =

resultRelInfo->ri_FdwRoutine->GetForeignModifyBatchSize(resultRelInfo);
   else
       resultRelInfo->ri_BatchSize = 1;

   Assert(resultRelInfo->ri_BatchSize >= 1);

It seems as if the resultRelInfo is not initialized, or something like
that. I wouldn't be surprised if the 32-bit machine was pickier and
failing because of that.

A sample of the valgrind log is attached. It's pretty much just
repetitions of these three reports.

regards

--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

В списке pgsql-hackers по дате отправления:

Предыдущее
От: David Rowley
Дата:
Сообщение: Heap's backwards scan scans the incorrect pages with heap_setscanlimits()
Следующее
От: Tomas Vondra
Дата:
Сообщение: Re: POC: postgres_fdw insert batching