Add writeback to pg_stat_io

28e626bde00 added the concept of IOOps but neglected to include writeback
operations. ac8d53dae5 added time spent doing these I/O operations. Without
counting writeback, checkpointer write time in the log often differed
substantially from that in pg_stat_io. To fix this, add IOOp IOOP_WRITEBACK
and track writeback in pg_stat_io.

Bumps catversion.

Author: Melanie Plageman <melanieplageman@gmail.com>
Reviewed-by: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
Reported-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/20230419172326.dhgyo4wrrhulovt6%40awork3.anarazel.de

4 files changed, 40 insertions, 13 deletions

diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 48aacf66ee..d0c932ad0e 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -1131,6 +1131,8 @@ SELECT
        b.read_time,
        b.writes,
        b.write_time,
+       b.writebacks,
+       b.writeback_time,
        b.extends,
        b.extend_time,
        b.op_bytes,
diff --git a/src/backend/storage/buffer/bufmgr.c b/src/backend/storage/buffer/bufmgr.c
index 31f63c180b..0bbf09564a 100644
--- a/src/backend/storage/buffer/bufmgr.c
+++ b/src/backend/storage/buffer/bufmgr.c
@@ -1685,7 +1685,7 @@ again:
 		FlushBuffer(buf_hdr, NULL, IOOBJECT_RELATION, io_context);
 		LWLockRelease(content_lock);
 
-		ScheduleBufferTagForWriteback(&BackendWritebackContext,
+		ScheduleBufferTagForWriteback(&BackendWritebackContext, io_context,
 									  &buf_hdr->tag);
 	}
 
@@ -2725,8 +2725,11 @@ BufferSync(int flags)
 		CheckpointWriteDelay(flags, (double) num_processed / num_to_scan);
 	}
 
-	/* issue all pending flushes */
-	IssuePendingWritebacks(&wb_context);
+	/*
+	 * Issue all pending flushes. Only checkpointer calls BufferSync(), so
+	 * IOContext will always be IOCONTEXT_NORMAL.
+	 */
+	IssuePendingWritebacks(&wb_context, IOCONTEXT_NORMAL);
 
 	pfree(per_ts_stat);
 	per_ts_stat = NULL;
@@ -3110,7 +3113,11 @@ SyncOneBuffer(int buf_id, bool skip_recently_used, WritebackContext *wb_context)
 
 	UnpinBuffer(bufHdr);
 
-	ScheduleBufferTagForWriteback(wb_context, &tag);
+	/*
+	 * SyncOneBuffer() is only called by checkpointer and bgwriter, so
+	 * IOContext will always be IOCONTEXT_NORMAL.
+	 */
+	ScheduleBufferTagForWriteback(wb_context, IOCONTEXT_NORMAL, &tag);
 
 	return result | BUF_WRITTEN;
 }
@@ -5445,7 +5452,8 @@ WritebackContextInit(WritebackContext *context, int *max_pending)
  * Add buffer to list of pending writeback requests.
  */
 void
-ScheduleBufferTagForWriteback(WritebackContext *wb_context, BufferTag *tag)
+ScheduleBufferTagForWriteback(WritebackContext *wb_context, IOContext io_context,
+							  BufferTag *tag)
 {
 	PendingWriteback *pending;
 
@@ -5471,7 +5479,7 @@ ScheduleBufferTagForWriteback(WritebackContext *wb_context, BufferTag *tag)
 	 * is now disabled.
 	 */
 	if (wb_context->nr_pending >= *wb_context->max_pending)
-		IssuePendingWritebacks(wb_context);
+		IssuePendingWritebacks(wb_context, io_context);
 }
 
 #define ST_SORT sort_pending_writebacks
@@ -5489,8 +5497,9 @@ ScheduleBufferTagForWriteback(WritebackContext *wb_context, BufferTag *tag)
  * error out - it's just a hint.
  */
 void
-IssuePendingWritebacks(WritebackContext *wb_context)
+IssuePendingWritebacks(WritebackContext *wb_context, IOContext io_context)
 {
+	instr_time	io_start;
 	int			i;
 
 	if (wb_context->nr_pending == 0)
@@ -5503,6 +5512,8 @@ IssuePendingWritebacks(WritebackContext *wb_context)
 	sort_pending_writebacks(wb_context->pending_writebacks,
 							wb_context->nr_pending);
 
+	io_start = pgstat_prepare_io_time();
+
 	/*
 	 * Coalesce neighbouring writes, but nothing else. For that we iterate
 	 * through the, now sorted, array of pending flushes, and look forward to
@@ -5556,6 +5567,13 @@ IssuePendingWritebacks(WritebackContext *wb_context)
 		smgrwriteback(reln, BufTagGetForkNum(&tag), tag.blockNum, nblocks);
 	}
 
+	/*
+	 * Assume that writeback requests are only issued for buffers containing
+	 * blocks of permanent relations.
+	 */
+	pgstat_count_io_op_time(IOOBJECT_RELATION, io_context,
+							IOOP_WRITEBACK, io_start, wb_context->nr_pending);
+
 	wb_context->nr_pending = 0;
 }
 
diff --git a/src/backend/utils/activity/pgstat_io.c b/src/backend/utils/activity/pgstat_io.c
index 25735eb6c0..eb7d35d422 100644
--- a/src/backend/utils/activity/pgstat_io.c
+++ b/src/backend/utils/activity/pgstat_io.c
@@ -425,6 +425,14 @@ pgstat_tracks_io_op(BackendType bktype, IOObject io_object,
 		return false;
 
 	/*
+	 * Temporary tables are not logged and thus do not require fsync'ing.
+	 * Writeback is not requested for temporary tables.
+	 */
+	if (io_object == IOOBJECT_TEMP_RELATION &&
+		(io_op == IOOP_FSYNC || io_op == IOOP_WRITEBACK))
+		return false;
+
+	/*
 	 * Some IOOps are not valid in certain IOContexts and some IOOps are only
 	 * valid in certain contexts.
 	 */
@@ -448,12 +456,6 @@ pgstat_tracks_io_op(BackendType bktype, IOObject io_object,
 	if (strategy_io_context && io_op == IOOP_FSYNC)
 		return false;
 
-	/*
-	 * Temporary tables are not logged and thus do not require fsync'ing.
-	 */
-	if (io_context == IOCONTEXT_NORMAL &&
-		io_object == IOOBJECT_TEMP_RELATION && io_op == IOOP_FSYNC)
-		return false;
 
 	return true;
 }
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 0d57b6a7c0..70da0a2de1 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1268,6 +1268,8 @@ typedef enum io_stat_col
 	IO_COL_READ_TIME,
 	IO_COL_WRITES,
 	IO_COL_WRITE_TIME,
+	IO_COL_WRITEBACKS,
+	IO_COL_WRITEBACK_TIME,
 	IO_COL_EXTENDS,
 	IO_COL_EXTEND_TIME,
 	IO_COL_CONVERSION,
@@ -1303,6 +1305,8 @@ pgstat_get_io_op_index(IOOp io_op)
 			return IO_COL_REUSES;
 		case IOOP_WRITE:
 			return IO_COL_WRITES;
+		case IOOP_WRITEBACK:
+			return IO_COL_WRITEBACKS;
 	}
 
 	elog(ERROR, "unrecognized IOOp value: %d", io_op);
@@ -1322,6 +1326,7 @@ pgstat_get_io_time_index(IOOp io_op)
 	{
 		case IOOP_READ:
 		case IOOP_WRITE:
+		case IOOP_WRITEBACK:
 		case IOOP_EXTEND:
 		case IOOP_FSYNC:
 			return pgstat_get_io_op_index(io_op) + 1;