From 3bbc4f09c0b8249ce76ea19b3a422e1f765978a1 Mon Sep 17 00:00:00 2001 From: Bharath Rupireddy Date: Thu, 17 Mar 2022 12:53:00 +0000 Subject: [PATCH v44] Add checkpoint stats of snapshot and mapping files of pg_logical dir At times, there can be many snapshot and mapping files under pg_logical dir that the checkpoint might have to delete/fsync based on the cutoff LSN which can increase the checkpoint time. Add stats related to these files to better understand the delays or time spent by the checkpointer processing them. Add these new log message only when necessary i.e. at least one snapshot or mapping files is processed during the checkpoint. --- src/backend/access/heap/rewriteheap.c | 6 +++ src/backend/access/transam/xlog.c | 42 +++++++++++++++++++++ src/backend/replication/logical/snapbuild.c | 4 ++ src/include/access/xlog.h | 13 +++++++ 4 files changed, 65 insertions(+) diff --git a/src/backend/access/heap/rewriteheap.c b/src/backend/access/heap/rewriteheap.c index 2a53826736..59cc3fa567 100644 --- a/src/backend/access/heap/rewriteheap.c +++ b/src/backend/access/heap/rewriteheap.c @@ -1210,6 +1210,8 @@ CheckPointLogicalRewriteHeap(void) if (cutoff != InvalidXLogRecPtr && redo < cutoff) cutoff = redo; + CheckpointStats.repl_map_cutoff_lsn = cutoff; + mappings_dir = AllocateDir("pg_logical/mappings"); while ((mapping_de = ReadDir(mappings_dir, "pg_logical/mappings")) != NULL) { @@ -1247,6 +1249,8 @@ CheckPointLogicalRewriteHeap(void) ereport(ERROR, (errcode_for_file_access(), errmsg("could not remove file \"%s\": %m", path))); + + CheckpointStats.repl_map_files_rmvd_cnt++; } else { @@ -1280,6 +1284,8 @@ CheckPointLogicalRewriteHeap(void) (errcode_for_file_access(), errmsg("could not close file \"%s\": %m", path))); } + + CheckpointStats.repl_map_files_syncd_cnt++; } FreeDir(mappings_dir); diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c index f436471b27..1dd31d4cc5 100644 --- a/src/backend/access/transam/xlog.c +++ b/src/backend/access/transam/xlog.c @@ -6115,6 +6115,41 @@ LogCheckpointEnd(bool restartpoint) CheckpointStats.ckpt_sync_rels; average_msecs = (long) ((average_sync_time + 999) / 1000); + /* + * Let's emit snapshot and mapping files stats only when necessary, that to + * as separate log messages, to not clutter the "checkpoint/restartpoint + * complete" message. + */ + if (CheckpointStats.repl_snap_files_rmvd_cnt > 0) + { + long t_msecs; + + t_msecs = TimestampDifferenceMilliseconds(CheckpointStats.repl_snap_start_t, + CheckpointStats.repl_snap_end_t); + + ereport(LOG, + (errmsg("logical decoding snapshot file(s) removed=" UINT64_FORMAT ", time=%ld.%03d s, cutoff LSN=%X/%X", + CheckpointStats.repl_snap_files_rmvd_cnt, + t_msecs / 1000, (int) (t_msecs % 1000), + LSN_FORMAT_ARGS(CheckpointStats.repl_snap_cutoff_lsn)))); + } + + if (CheckpointStats.repl_map_files_rmvd_cnt > 0 || + CheckpointStats.repl_map_files_syncd_cnt > 0) + { + long t_msecs; + + t_msecs = TimestampDifferenceMilliseconds(CheckpointStats.repl_map_start_t, + CheckpointStats.repl_map_end_t); + + ereport(LOG, + (errmsg("logical decoding rewrite mapping file(s) removed=" UINT64_FORMAT ", synced=" UINT64_FORMAT ", time=%ld.%03d s, cutoff LSN=%X/%X", + CheckpointStats.repl_map_files_rmvd_cnt, + CheckpointStats.repl_map_files_syncd_cnt, + t_msecs / 1000, (int) (t_msecs % 1000), + LSN_FORMAT_ARGS(CheckpointStats.repl_map_cutoff_lsn)))); + } + if (restartpoint) ereport(LOG, (errmsg("restartpoint complete: wrote %d buffers (%.1f%%); " @@ -6811,8 +6846,15 @@ CheckPointGuts(XLogRecPtr checkPointRedo, int flags) { CheckPointRelationMap(); CheckPointReplicationSlots(); + + CheckpointStats.repl_snap_start_t = GetCurrentTimestamp(); CheckPointSnapBuild(); + CheckpointStats.repl_snap_end_t = GetCurrentTimestamp(); + + CheckpointStats.repl_map_start_t = GetCurrentTimestamp(); CheckPointLogicalRewriteHeap(); + CheckpointStats.repl_map_end_t = GetCurrentTimestamp(); + CheckPointReplicationOrigin(); /* Write out all dirty data in SLRUs and the main buffer pool */ diff --git a/src/backend/replication/logical/snapbuild.c b/src/backend/replication/logical/snapbuild.c index 83fca8a77d..e652acfbe6 100644 --- a/src/backend/replication/logical/snapbuild.c +++ b/src/backend/replication/logical/snapbuild.c @@ -1941,6 +1941,8 @@ CheckPointSnapBuild(void) if (redo < cutoff) cutoff = redo; + CheckpointStats.repl_snap_cutoff_lsn = cutoff; + snap_dir = AllocateDir("pg_logical/snapshots"); while ((snap_de = ReadDir(snap_dir, "pg_logical/snapshots")) != NULL) { @@ -1997,6 +1999,8 @@ CheckPointSnapBuild(void) path))); continue; } + + CheckpointStats.repl_snap_files_rmvd_cnt++; } } FreeDir(snap_dir); diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h index 09f6464331..32f8116388 100644 --- a/src/include/access/xlog.h +++ b/src/include/access/xlog.h @@ -172,6 +172,19 @@ typedef struct CheckpointStatsData * times, which is not necessarily the * same as the total elapsed time for the * entire sync phase. */ + + /* Statistics of snapshot files under "pg_logical/snapshots" */ + uint64 repl_snap_files_rmvd_cnt; + XLogRecPtr repl_snap_cutoff_lsn; + TimestampTz repl_snap_start_t; + TimestampTz repl_snap_end_t; + + /* Statistics of map files under "pg_logical/mappings" */ + uint64 repl_map_files_syncd_cnt; + uint64 repl_map_files_rmvd_cnt; + XLogRecPtr repl_map_cutoff_lsn; + TimestampTz repl_map_start_t; + TimestampTz repl_map_end_t; } CheckpointStatsData; extern CheckpointStatsData CheckpointStats; -- 2.25.1