lightdb enterprise postgres wal详解之整体体系及pg_waldump、pg_switch_wal的实现

　　要理解WAL，首先要知道WAL文件、WAL块以及WAL记录的物理布局。

https://www.pgcon.org/2012/schedule/attachments/258_212_Internals%20Of%20PostgreSQL%20Wal.pdf

　　启动时redo重放流程：

[zjh@hs-10-20-30-193 ~]$ pstack 166980
#0  0x00007f0682909dc0 in __lseek_nocancel () from /lib64/libpthread.so.0
#1  0x00000000007b02c2 in _mdnblocks (reln=0x1e92f30, forknum=MAIN_FORKNUM, seg=0x1e623c8) at md.c:1326
#2  mdnblocks (reln=0x1e92f30, forknum=MAIN_FORKNUM) at md.c:801
#3  0x000000000052c9e5 in XLogReadBufferExtended (rnode=..., forknum=MAIN_FORKNUM, blkno=88, mode=mode@entry=RBM_NORMAL) at xlogutils.c:465
#4  0x000000000052cc8e in XLogReadBufferForRedoExtended (record=record@entry=0x1e73ff8, block_id=<optimized out>, mode=mode@entry=RBM_NORMAL, get_cleanup_lock=get_cleanup_lock@entry=false, buf=buf@entry=0x7ffd4f0e1988) at xlogutils.c:396
#5  0x000000000052ce20 in XLogReadBufferForRedo (record=record@entry=0x1e73ff8, block_id=<optimized out>, buf=buf@entry=0x7ffd4f0e1988) at xlogutils.c:298
#6  0x00000000004c6b6c in heap_xlog_update (record=0x1e73ff8, hot_update=hot_update@entry=true) at heapam.c:8569
#7  0x00000000004cd1d2 in heap_redo (record=0x1e73ff8) at heapam.c:9008
#8  0x0000000000520c77 in StartupXLOG () at xlog.c:7825
#9  0x000000000073b05d in StartupProcessMain () at startup.c:223
#10 0x0000000000534455 in AuxiliaryProcessMain (argc=argc@entry=2, argv=argv@entry=0x7ffd4f0e7840) at bootstrap.c:497
#11 0x0000000000737b70 in StartChildProcess (type=StartupProcess) at postmaster.c:6093
#12 0x000000000073a458 in PostmasterMain (argc=argc@entry=3, argv=argv@entry=0x1e2dd20) at postmaster.c:1560
#13 0x000000000048d4cd in main (argc=3, argv=0x1e2dd20) at main.c:232
===========================

#0  0x00007f0680b3bb23 in __select_nocancel () from /lib64/libc.so.6
#1  0x0000000000914f7a in pg_usleep (microsec=microsec@entry=10) at pgsleep.c:56
#2  0x000000000078312c in MarkBufferDirty (buffer=569) at bufmgr.c:1661
#3  0x00000000004cdd60 in heap_xlog_clean (record=0x1e73ff8) at heapam.c:7921
#4  heap2_redo (record=0x1e73ff8) at heapam.c:9032
#5  0x0000000000520c77 in StartupXLOG () at xlog.c:7825
#6  0x000000000073b05d in StartupProcessMain () at startup.c:223
#7  0x0000000000534455 in AuxiliaryProcessMain (argc=argc@entry=2, argv=argv@entry=0x7ffd4f0e7840) at bootstrap.c:497
#8  0x0000000000737b70 in StartChildProcess (type=StartupProcess) at postmaster.c:6093
#9  0x000000000073a458 in PostmasterMain (argc=argc@entry=3, argv=argv@entry=0x1e2dd20) at postmaster.c:1560
#10 0x000000000048d4cd in main (argc=3, argv=0x1e2dd20) at main.c:232

　　除了源码，了解WAL最好的方式是通过lt_waldump入手：

[lightdb@lightdb1 bin]$ ./lt_waldump --help
lt_waldump decodes and displays LightDB write-ahead logs for debugging.

Usage:
  lt_waldump [OPTION]... [STARTSEG [ENDSEG]]

Options:
  -b, --bkp-details      output detailed information about backup blocks
  -e, --end=RECPTR       stop reading at WAL location RECPTR
  -f, --follow           keep retrying after reaching end of WAL
  -n, --limit=N          number of records to display
  -p, --path=PATH        directory in which to find log segment files or a
                         directory with a ./pg_wal that contains such files
                         (default: current directory, ./pg_wal, $PGDATA/pg_wal)
  -q, --quiet            do not print any output, except for errors
  -r, --rmgr=RMGR        only show records generated by resource manager RMGR;
                         use --rmgr=list to list valid resource manager names
  -s, --start=RECPTR     start reading at WAL location RECPTR
  -t, --timeline=TLI     timeline from which to read log records
                         (default: 1 or the value used in STARTSEG)
  -V, --version          output version information, then exit
  -x, --xid=XID          only show records with transaction ID XID
  -z, --stats[=record]   show statistics instead of records
                         (optionally, show per-record statistics)
  -?, --help             show this help, then exit

Report bugs to <https://github.com/hslightdb>.
LightDB home page: <https://www.hs.net/lightdb>
[lightdb@lightdb1 bin]$ ./lt_waldump --rmgr=list
XLOG
Transaction
Storage
CLOG
Database
Tablespace
MultiXact
RelMap
Standby
Heap2
Heap
Btree
Hash
Gin
Gist
Sequence
SPGist
BRIN
CommitTs
ReplicationOrigin
Generic
LogicalMessage

　　对应的资源管理器定义结构如下：

/*
 * List of resource manager entries.  Note that order of entries defines the
 * numerical values of each rmgr's ID, which is stored in WAL records.  New
 * entries should be added at the end, to avoid changing IDs of existing
 * entries.
 *
 * Changes to this list possibly need an XLOG_PAGE_MAGIC bump.
 */

/* symbol name, textual name, redo, desc, identify, startup, cleanup */
PG_RMGR(RM_XLOG_ID, "XLOG", xlog_redo, xlog_desc, xlog_identify, NULL, NULL, NULL)
PG_RMGR(RM_XACT_ID, "Transaction", xact_redo, xact_desc, xact_identify, NULL, NULL, NULL)
PG_RMGR(RM_SMGR_ID, "Storage", smgr_redo, smgr_desc, smgr_identify, NULL, NULL, NULL)
PG_RMGR(RM_CLOG_ID, "CLOG", clog_redo, clog_desc, clog_identify, NULL, NULL, NULL)
PG_RMGR(RM_DBASE_ID, "Database", dbase_redo, dbase_desc, dbase_identify, NULL, NULL, NULL)
PG_RMGR(RM_TBLSPC_ID, "Tablespace", tblspc_redo, tblspc_desc, tblspc_identify, NULL, NULL, NULL)
PG_RMGR(RM_MULTIXACT_ID, "MultiXact", multixact_redo, multixact_desc, multixact_identify, NULL, NULL, NULL)
PG_RMGR(RM_RELMAP_ID, "RelMap", relmap_redo, relmap_desc, relmap_identify, NULL, NULL, NULL)
PG_RMGR(RM_STANDBY_ID, "Standby", standby_redo, standby_desc, standby_identify, NULL, NULL, NULL)
PG_RMGR(RM_HEAP2_ID, "Heap2", heap2_redo, heap2_desc, heap2_identify, NULL, NULL, heap_mask)
PG_RMGR(RM_HEAP_ID, "Heap", heap_redo, heap_desc, heap_identify, NULL, NULL, heap_mask)
PG_RMGR(RM_BTREE_ID, "Btree", btree_redo, btree_desc, btree_identify, btree_xlog_startup, btree_xlog_cleanup, btree_mask)
PG_RMGR(RM_HASH_ID, "Hash", hash_redo, hash_desc, hash_identify, NULL, NULL, hash_mask)
PG_RMGR(RM_GIN_ID, "Gin", gin_redo, gin_desc, gin_identify, gin_xlog_startup, gin_xlog_cleanup, gin_mask)
PG_RMGR(RM_GIST_ID, "Gist", gist_redo, gist_desc, gist_identify, gist_xlog_startup, gist_xlog_cleanup, gist_mask)
PG_RMGR(RM_SEQ_ID, "Sequence", seq_redo, seq_desc, seq_identify, NULL, NULL, seq_mask)
PG_RMGR(RM_SPGIST_ID, "SPGist", spg_redo, spg_desc, spg_identify, spg_xlog_startup, spg_xlog_cleanup, spg_mask)
PG_RMGR(RM_BRIN_ID, "BRIN", brin_redo, brin_desc, brin_identify, NULL, NULL, brin_mask)
PG_RMGR(RM_COMMIT_TS_ID, "CommitTs", commit_ts_redo, commit_ts_desc, commit_ts_identify, NULL, NULL, NULL)
PG_RMGR(RM_REPLORIGIN_ID, "ReplicationOrigin", replorigin_redo, replorigin_desc, replorigin_identify, NULL, NULL, NULL)
PG_RMGR(RM_GENERIC_ID, "Generic", generic_redo, generic_desc, generic_identify, NULL, NULL, generic_mask)
PG_RMGR(RM_LOGICALMSG_ID, "LogicalMessage", logicalmsg_redo, logicalmsg_desc, logicalmsg_identify, NULL, NULL, NULL)

　　对应的，每种access method通常都有对应的xlog（它主要对应block data/main data的物理组织结构）结构，如下：

　　具体到heapam中，新增、修改、删除、multi insert等等都是不同的结构定义，具体可见heapam_xlog.h。

　　先看pg_waldump的解析日志：

insert一条带主键的记录
rmgr: XLOG len (rec/tot): 49/ 8169, tx: 0, lsn: 0/4CBB9D78, prev 0/4CBB9D40, desc: FPI_FOR_HINT , blkref #0: rel 1663/13580/2619 blk 16 FPW
rmgr: Heap2 len (rec/tot): 56/ 56, tx: 0, lsn: 0/4CBBBD80, prev 0/4CBB9D78, desc: CLEAN remxid 538, blkref #0: rel 1663/13580/2619 blk 16
rmgr: Standby len (rec/tot): 50/ 50, tx: 0, lsn: 0/4CBBBDB8, prev 0/4CBBBD80, desc: RUNNING_XACTS nextXid 592 latestCompletedXid 591 oldestRunningXid 592
rmgr: Heap len (rec/tot): 54/ 3510, tx: 592, lsn: 0/4CBBBDF0, prev 0/4CBBBDB8, desc: INSERT off 78 flags 0x00, blkref #0: rel 1663/13580/32899 blk 5405 FPW
rmgr: Btree len (rec/tot): 53/ 5641, tx: 592, lsn: 0/4CBBCBC0, prev 0/4CBBBDF0, desc: INSERT_LEAF off 8, blkref #0【本wal记录内块号】: rel 1663【表空间】/13580【数据库】/41089【对象】 blk 848【文件内块号】 FPW
rmgr: Standby len (rec/tot): 54/ 54, tx: 0, lsn: 0/4CBBE1E8, prev 0/4CBBCBC0, desc: RUNNING_XACTS nextXid 593 latestCompletedXid 591 oldestRunningXid 592; 1 xacts: 592
pg_waldump: fatal: error in WAL record at 0/4CBBE1E8: invalid record length at 0/4CBBE220: wanted 24, got 0

rec<tot基本上是因为发生了FPW，如果开启了FPW，则压缩是非常有价值的。
rmgr: Standby len (rec/tot): 54/ 54, tx: 0, lsn: 0/4CBBE220, prev 0/4CBBE1E8, desc: RUNNING_XACTS nextXid 593 latestCompletedXid 591 oldestRunningXid 592; 1 xacts: 592
rmgr: XLOG len (rec/tot): 114/ 114, tx: 0, lsn: 0/4CBBE258, prev 0/4CBBE220, desc: CHECKPOINT_ONLINE redo 0/4CBBE220; tli 1; prev tli 1; fpw true; xid 0:593; oid 57492; multi 1; offset 0; oldest xid 478 in DB 1; oldest multi 1 in DB 1; oldest/newest commit timestamp xid: 0/0; oldest running xid 592; online
rmgr: Standby len (rec/tot): 54/ 54, tx: 0, lsn: 0/4CBBE2D0, prev 0/4CBBE258, desc: RUNNING_XACTS nextXid 593 latestCompletedXid 591 oldestRunningXid 592; 1 xacts: 592
rmgr: Transaction len (rec/tot): 34/ 34, tx: 592, lsn: 0/4CBBE308, prev 0/4CBBE2D0, desc: COMMIT 2021-07-17 15:44:51.835849 CST
rmgr: Standby len (rec/tot): 50/ 50, tx: 0, lsn: 0/4CBBE330, prev 0/4CBBE308, desc: RUNNING_XACTS nextXid 593 latestCompletedXid 592 oldestRunningXid 593

blkref 说明：分配给一个wal记录的缓冲默认情况下为5个块，由下列常量定义。

/*
 * The minimum size of the WAL construction working area. If you need to
 * register more than XLR_NORMAL_MAX_BLOCK_ID block references or have more
 * than XLR_NORMAL_RDATAS data chunks in a single WAL record, you must call
 * XLogEnsureRecordSpace() first to allocate more working memory.
 */
#define XLR_NORMAL_MAX_BLOCK_ID        4
#define XLR_NORMAL_RDATAS            20

　　正常都是从block id=0开始，如下：

　　所以可以看到除非text/bytea等大字段类型，一般情况下blkref都为0。

=====

　　wal记录可能包含两部分，FPI和非FPI，通过DecodedBkpBlock.has_img区分某个wal record的各个块是FPI还是非FPI。

typedef struct
{
    /* Is this block ref in use? */
    bool        in_use;

    /* Identify the block this refers to */
    RelFileNode rnode;
    ForkNumber    forknum;
    BlockNumber blkno;

    /* copy of the fork_flags field from the XLogRecordBlockHeader */
    uint8        flags;

    /* Information on full-page image, if any */
    bool        has_image;        /* has image, even for consistency checking */
    bool        apply_image;    /* has image that should be restored */
    char       *bkp_image;
    uint16        hole_offset;
    uint16        hole_length;
    uint16        bimg_len;
    uint8        bimg_info;

    /* Buffer holding the rmgr-specific data associated with this block */
    bool        has_data;
    char       *data;
    uint16        data_len;
    uint16        data_bufsz;
} DecodedBkpBlock;

struct XLogReaderState
{
    /*
     * Operational callbacks
     */
    XLogReaderRoutine routine;

    /* ----------------------------------------
     * Public parameters
     * ----------------------------------------
     */

    /*
     * System identifier of the xlog files we're about to read.  Set to zero
     * (the default value) if unknown or unimportant.
     */
    uint64        system_identifier;

    /*
     * Opaque data for callbacks to use.  Not used by XLogReader.
     */
    void       *private_data;

    /*
     * Start and end point of last record read.  EndRecPtr is also used as the
     * position to read next.  Calling XLogBeginRead() sets EndRecPtr to the
     * starting position and ReadRecPtr to invalid.
     */
    XLogRecPtr    ReadRecPtr;        /* start of last record read */
    XLogRecPtr    EndRecPtr;        /* end+1 of last record read */


    /* ----------------------------------------
     * Decoded representation of current record
     *
     * Use XLogRecGet* functions to investigate the record; these fields
     * should not be accessed directly.
     * ----------------------------------------
     */
    XLogRecord *decoded_record; /* currently decoded record */

    char       *main_data;        /* record's main data portion */
    uint32        main_data_len;    /* main data portion's length */
    uint32        main_data_bufsz;    /* allocated size of the buffer */

    RepOriginId record_origin;

    /* information about blocks referenced by the record. */
    DecodedBkpBlock blocks[XLR_MAX_BLOCK_ID + 1];

    int            max_block_id;    /* highest block_id in use (-1 if none) */

    /* ----------------------------------------
     * private/internal state
     * ----------------------------------------
     */

    /*
     * Buffer for currently read page (XLOG_BLCKSZ bytes, valid up to at least
     * readLen bytes)
     */
    char       *readBuf;
    uint32        readLen;

    /* last read XLOG position for data currently in readBuf */
    WALSegmentContext segcxt;
    WALOpenSegment seg;
    uint32        segoff;

    /*
     * beginning of prior page read, and its TLI.  Doesn't necessarily
     * correspond to what's in readBuf; used for timeline sanity checks.
     */
    XLogRecPtr    latestPagePtr;
    TimeLineID    latestPageTLI;

    /* beginning of the WAL record being read. */
    XLogRecPtr    currRecPtr;
    /* timeline to read it from, 0 if a lookup is required */
    TimeLineID    currTLI;

    /*
     * Safe point to read to in currTLI if current TLI is historical
     * (tliSwitchPoint) or InvalidXLogRecPtr if on current timeline.
     *
     * Actually set to the start of the segment containing the timeline switch
     * that ends currTLI's validity, not the LSN of the switch its self, since
     * we can't assume the old segment will be present.
     */
    XLogRecPtr    currTLIValidUntil;

    /*
     * If currTLI is not the most recent known timeline, the next timeline to
     * read from when currTLIValidUntil is reached.
     */
    TimeLineID    nextTLI;

    /*
     * Buffer for current ReadRecord result (expandable), used when a record
     * crosses a page boundary.
     */
    char       *readRecordBuf;
    uint32        readRecordBufSize;

    /* Buffer to hold error message */
    char       *errormsg_buf;
};

wal记录固定长度头部的定义

/*
 * The overall layout of an XLOG record is:
 *        Fixed-size header (XLogRecord struct)
 *        XLogRecordBlockHeader struct
 *        XLogRecordBlockHeader struct
 *        ...
 *        XLogRecordDataHeader[Short|Long] struct
 *        block data
 *        block data
 *        ...
 *        main data
 *
 * There can be zero or more XLogRecordBlockHeaders, and 0 or more bytes of
 * rmgr-specific data not associated with a block.  XLogRecord structs
 * always start on MAXALIGN boundaries in the WAL files, but the rest of
 * the fields are not aligned.
 *
 * The XLogRecordBlockHeader, XLogRecordDataHeaderShort and
 * XLogRecordDataHeaderLong structs all begin with a single 'id' byte. It's
 * used to distinguish between block references, and the main data structs.
 */
typedef struct XLogRecord
{
    uint32        xl_tot_len;        /* total len of entire record */
    TransactionId xl_xid;        /* xact id */
    XLogRecPtr    xl_prev;        /* ptr to previous record in log */
    uint8        xl_info;        /* flag bits, see below */   分为前4字节（rmgr保存各个rmgr特有的一些标记信息，如一致性检查，典型的例子是xlog_identify(uint8 info)）和后4字节
    RmgrId        xl_rmid;        /* resource manager for this record */
    /* 2 bytes of padding here, initialize to zero */
    pg_crc32c    xl_crc;            /* CRC for this record */

    /* XLogRecordBlockHeaders and XLogRecordDataHeader follow, no padding */

} XLogRecord;

理解XLOG的这个布局是非常重要的，因为waldump解析出来的每条记录和语句不是一一对应的，而是和修改的物理对象一一对应（否则做不到物理apply）。

waldump读的逻辑基本时还是依赖xlogreader.c完成，waldump只是做了具体的实现，和sql执行过程一样，所有的状态由XLogReaderState整个周期内维护。

/*
 * pg_switch_wal: switch to next xlog file
 *
 * Permission checking for this function is managed through the normal
 * GRANT system.
 */
Datum
pg_switch_wal(PG_FUNCTION_ARGS)
{
    XLogRecPtr    switchpoint;

    if (RecoveryInProgress())
        ereport(ERROR,
                (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
                 errmsg("recovery is in progress"),
                 errhint("WAL control functions cannot be executed during recovery.")));

    switchpoint = RequestXLogSwitch(false);  // 具体的实现

    /*
     * As a convenience, return the WAL location of the switch record
     */
    PG_RETURN_LSN(switchpoint);
}

注：pg_switch_wal是一个内置函数。在pg_proc.dat中注册（genbki.pl据此生成postgres.bki，gen_fmgrtab.pl生成fmgroids.h, fmgrprotos.h, and fmgrtab.c）。

{ oid => '2848', descr => 'switch to new wal file',
  proname => 'pg_switch_wal', provolatile => 'v', prorettype => 'pg_lsn',
  proargtypes => '', prosrc => 'pg_switch_wal' },

select * from pg_proc where proname='pg_switch_wal';
Name           |Value              |
---------------+-------------------+
oid            |2848               |
proname        |pg_switch_wal      |
pronamespace   |11                 |
proowner       |10                 |
prolang        |12                 |
procost        |1.0                |
prorows        |0.0                |
provariadic    |0                  |
prosupport     |-                  |
prokind        |f                  |
prosecdef      |false              |
proleakproof   |false              |
proisstrict    |true               |
proretset      |false              |
provolatile    |v                  |
proparallel    |s                  |
pronargs       |0                  |
pronargdefaults|0                  |
prorettype     |3220               |
proargtypes    |{}                 |
proallargtypes |NULL               |
proargmodes    |NULL               |
proargnames    |NULL               |
proargdefaults |                   |
protrftypes    |NULL               |
prosrc         |pg_switch_wal      |
probin         |                   |
proconfig      |NULL               |
proacl         |{lightdb=X/lightdb}|

下面来分析wal文件的生成与切换。

pg_switch_wal moves to the next write-ahead log file, allowing the current file to be archived (assuming you are using continuous archiving). The return value is the ending write-ahead log location + 1 within the just-completed write-ahead log file. If there has been no write-ahead log activity since the last write-ahead log switch, pg_switch_wal does nothing and returns the start location of the write-ahead log file currently in use.

postgres=# select pg_switch_wal();
pg_switch_wal
---------------
0/4CBBE468
(1 row)

[zjh@hs-10-20-30-193 data]$ ll pg_wal/
total 2097192
-rw------- 1 zjh zjh 1073741824 Jul 17 16:00 000000010000000000000001
-rw------- 1 zjh zjh 1073741824 Jul 17 16:01 000000010000000000000002
drwx------ 2 zjh zjh 6 Jun 14 19:05 archive_status
-rw-rw-r-- 1 zjh zjh 37691 Jul 15 16:06 gmon.out

postgres=# select pg_switch_wal();
pg_switch_wal
---------------
0/80000078
(1 row)

[zjh@hs-10-20-30-193 data]$ ll pg_wal/
total 3145768
-rw------- 1 zjh zjh 1073741824 Jul 17 16:00 000000010000000000000001
-rw------- 1 zjh zjh 1073741824 Jul 17 16:01 000000010000000000000002
-rw------- 1 zjh zjh 1073741824 Jul 17 16:01 000000010000000000000003
drwx------ 2 zjh zjh 6 Jun 14 19:05 archive_status
-rw-rw-r-- 1 zjh zjh 37691 Jul 15 16:06 gmon.out

postgres=# select pg_switch_wal();
pg_switch_wal
---------------
0/C0000078
(1 row)

postgres=# select pg_switch_wal();
pg_switch_wal
---------------
1/0 ？？这个值什么时候会出现？
(1 row)

[zjh@hs-10-20-30-193 data]$ ll pg_wal/
total 4194344
-rw------- 1 zjh zjh 1073741824 Jul 17 16:00 000000010000000000000001
-rw------- 1 zjh zjh 1073741824 Jul 17 16:01 000000010000000000000002
-rw------- 1 zjh zjh 1073741824 Jul 17 16:01 000000010000000000000003
-rw------- 1 zjh zjh 1073741824 Jul 17 16:01 000000010000000100000000
drwx------ 2 zjh zjh 6 Jun 14 19:05 archive_status
-rw-rw-r-- 1 zjh zjh 37691 Jul 15 16:06 gmon.out

postgres=# select pg_switch_wal();
pg_switch_wal
---------------
1/78 ？？这个值什么时候会出现？
(1 row)

postgres=# select pg_switch_wal();
pg_switch_wal
---------------
1/40000000
(1 row)

postgres=# select pg_current_wal_lsn();
pg_current_wal_lsn
--------------------
1/40000148
(1 row)

postgres=# insert into t_sample select* from t_sample limit 1000000;

rmgr: XLOG len (rec/tot): 114/ 114, tx: 0, lsn: 1/40000098, prev 1/40000060, desc: CHECKPOINT_ONLINE redo 1/40000060; tli 1; prev tli 1; fpw true; xid 0:593; oid 57492; m
ulti 1; offset 0; oldest xid 478 in DB 1; oldest multi 1 in DB 1; oldest/newest commit timestamp xid: 0/0; oldest running xid 593; online
rmgr: Standby len (rec/tot): 50/ 50, tx: 0, lsn: 1/40000110, prev 1/40000098, desc: RUNNING_XACTS nextXid 593 latestCompletedXid 592 oldestRunningXid 593
=== wal_level >= WAL_LEVEL_REPLICA时，会写Standby。
=== 下一行开始不停的生成
rmgr: Heap len (rec/tot): 54/ 3554, tx: 593, lsn: 1/40000148, prev 1/40000110, desc: INSERT off 79 flags 0x00, blkref #0: rel 1663/13580/32899 blk 5405 FPW # full page write
rmgr: Btree len (rec/tot): 53/ 5661, tx: 593, lsn: 1/40000F30, prev 1/40000148, desc: INSERT_LEAF off 9, blkref #0: rel 1663/13580/41089 blk 848 FPW
rmgr: Heap len (rec/tot): 64/ 64, tx: 593, lsn: 1/40002568, prev 1/40000F30, desc: INSERT off 80 flags 0x00, blkref #0: rel 1663/13580/32899 blk 5405
rmgr: Btree len (rec/tot): 64/ 64, tx: 593, lsn: 1/400025A8, prev 1/40002568, desc: INSERT_LEAF off 10, blkref #0: rel 1663/13580/41089 blk 848
rmgr: Heap len (rec/tot): 64/ 64, tx: 593, lsn: 1/400025E8, prev 1/400025A8, desc: INSERT off 81 flags 0x00, blkref #0: rel 1663/13580/32899 blk 5405
rmgr: Btree len (rec/tot): 64/ 64, tx: 593, lsn: 1/40002628, prev 1/400025E8, desc: INSERT_LEAF off 11, blkref #0: rel 1663/13580/41089 blk 848

正常情况下len都比较小，即使插入8k长度记录也一样（会被TOAST+压缩，所以很小），要模拟大记录，要设置列存储方式为plain。此时就可以看到大的rec记录。

rmgr: Heap        len (rec/tot):   8159/  8159, tx:    2215040, lsn: E/A0128F10, prev E/A0126F18, desc: INSERT+INIT off 1 flags 0x00, blkref #0: rel 1663/15154/434247 blk 28
rmgr: Heap        len (rec/tot):   8159/  8159, tx:    2215040, lsn: E/A012AF08, prev E/A0128F10, desc: INSERT+INIT off 1 flags 0x00, blkref #0: rel 1663/15154/434247 blk 29
rmgr: Heap        len (rec/tot):   8159/  8159, tx:    2215040, lsn: E/A012CF00, prev E/A012AF08, desc: INSERT+INIT off 1 flags 0x00, blkref #0: rel 1663/15154/434247 blk 30
rmgr: Heap        len (rec/tot):   8159/  8159, tx:    2215040, lsn: E/A012EEF8, prev E/A012CF00, desc: INSERT+INIT off 1 flags 0x00, blkref #0: rel 1663/15154/434247 blk 31
rmgr: Transaction len (rec/tot):     34/    34, tx:    2215040, lsn: E/A0130EF0, prev E/A012EEF8, desc: COMMIT 2021-11-07 19:14:27.878160 CST

waldump也能够直接打印统计信息，如下：

lightdb$ lt_waldump --stats -p $LTDATA/lt_wal -s 0/3A69C530 -e 0/3BE87658
Type              N      (%)   Record size      (%)   FPI size      (%)  
----              -      ---   -----------      ---   --------      ---  
XLOG           1721 (  1,03)         84329 (  0,77)   13916104 (100,00)       
Transaction   27235 ( 16,32)        926070 (  8,46)          0 (  0,00)         
Storage           1 (  0,00)            42 (  0,00)          0 (  0,00)             
CLOG              1 (  0,00)            30 (  0,00)          0 (  0,00)             
Standby           4 (  0,00)           240 (  0,00)          0 (  0,00)            
Heap2         27522 ( 16,49)       1726352 ( 15,76)          0 (  0,00)        
Heap         109691 ( 65,71)       8169121 ( 74,59)          0 (  0,00)        
Btree           756 (  0,45)         45380 (  0,41)          0 (  0,00)          
           --------               --------            --------                
Total        166931               10951564 [44,04%]   13916104 [55,96%]

timeline https://postgreshelp.com/postgresql-timelines/切换，跟oracle的resetlogs一样，在不完全恢复后增加1

=== rmgr，所有有物理记录的都是资源，定义在rmgrlist.h接口中。src/include/access/rmgr.h，src/bin/pg_waldump/rmgrdesc.h
typedef struct RmgrDescData
{
const char *rm_name;
void (*rm_desc) (StringInfo buf, XLogReaderState *record);
const char *(*rm_identify) (uint8 info);
} RmgrDescData;

#define PG_RMGR(symname,name,redo,desc,identify,startup,cleanup,mask) \
{ name, desc, identify},

const RmgrDescData RmgrDescTable[RM_MAX_ID + 1] = {
#include "access/rmgrlist.h" # 居然还有这种用法？？？？预处理器果然NB
};

src/bin/pg_waldump/pg_waldump.c是pg_waldump入口

desc定义在src/backend/access/rmgrdesc/xxxdesc.c中，每个资源管理器一个。

postgres=# select pg_current_wal_lsn();
pg_current_wal_lsn
--------------------
1/47BB62D0 -- 7BB62D0（10进制等于‭129721040‬，约123M）
(1 row)

postgres=# insert into t_sample select* from t_sample limit 1000000;
INSERT 0 1000000
postgres=# select pg_current_wal_lsn();
pg_current_wal_lsn
--------------------
1/521B5A00 -- ‭A5FF730‬（10进制等于‭174061360‬‬，约168M）
(1 row)

注：LSN和十进制LSN的对应关系如下：

[zjh@hs-10-20-30-193 data]$ ll pg_wal/ --- 但是WAL文件多生成了1个
total 5242920
-rw------- 1 zjh zjh 1073741824 Jul 17 16:23 000000010000000100000001
-rw------- 1 zjh zjh 1073741824 Jul 17 16:00 000000010000000100000002
-rw------- 1 zjh zjh 1073741824 Jul 17 16:01 000000010000000100000003
-rw------- 1 zjh zjh 1073741824 Jul 17 16:01 000000010000000200000000
-rw------- 1 zjh zjh 1073741824 Jul 17 16:02 000000010000000200000001
drwx------ 2 zjh zjh 6 Jun 14 19:05 archive_status
-rw-rw-r-- 1 zjh zjh 37691 Jul 15 16:06 gmon.out

postgres=# select pg_current_wal_lsn();
pg_current_wal_lsn
--------------------
1/521B5A00
(1 row)

postgres=# select pg_walfile_name('1/521B5A70'); -- 这就说明虽然执行了switch file，但是WAL还是往000000010000000100000001在写？？这不就有问题么？
pg_walfile_name
--------------------------
000000010000000100000001
(1 row)

postgres=# select pg_walfile_name('1/521B5A00');
pg_walfile_name
--------------------------
000000010000000100000001
(1 row)

postgres=# select pg_switch_wal();
pg_switch_wal
---------------
1/521B5B38
(1 row)

postgres=# select pg_switch_wal();
pg_switch_wal
---------------
1/80000078
(1 row)

postgres=# select pg_walfile_name('1/80000078');
pg_walfile_name
--------------------------
000000010000000100000002
(1 row)
## 即使再切换，也没有生成新的WAL文件，是因为没有开启归档，同时已经满足了min_wal_size的要求
max_wal_size = 64GB
min_wal_size = 5GB

============现在开始分析wal文件的规则=================

postgres=# select pg_current_wal_flush_lsn(),pg_current_wal_insert_lsn(),pg_current_wal_lsn(),pg_walfile_name_offset(pg_current_wal_lsn());
pg_current_wal_flush_lsn | pg_current_wal_insert_lsn | pg_current_wal_lsn | pg_walfile
_name_offset
--------------------------+---------------------------+--------------------+-----------------
-------------------
E/601CDF60 | E/601CDF60 | E/601CDF60 | (000000030000000
E00000003,1892192)

在pg的wal中，有个三个位置，write位置、insert位置以及flush位置。三者的定义和差异为：

insert位置是指从逻辑角度而言，已经被占用的位置，也就是已经分配给wal writer或进程
write位置是已经调用了fwrite，但是还没有调用fflush，但是服务器外部可能可见
flush是指已经调用了fflush，只有fflush之后的wal，才是真正完成的。

↓flush的位置 ↓write的位置 ↓insert的位置

00000000000000001000000000000000000100000000000000000000000100000000000000000000000000000000000000000000000000000000000000

从切换可知，wal文件名一定是递增的。

[lightdb@iZvv70fyhxdyvjgzmr7ikpZ 13.3-21.2]$ cd data/defaultCluster/pg_wal/
[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ ll
total 2097180
-rw------- 1 lightdb lightdb        42 Nov  4 23:05 00000002.history
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:26 000000030000000C00000007
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:35 000000030000000D00000000
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:25 000000030000000D00000001
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:26 000000030000000D00000002
-rw------- 1 lightdb lightdb        85 Nov  4 23:17 00000003.history
drwx------ 2 lightdb lightdb      4096 Nov  7 11:26 archive_status
[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ ls -lt
total 2097180
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:35 000000030000000D00000000
drwx------ 2 lightdb lightdb      4096 Nov  7 11:26 archive_status
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:26 000000030000000C00000007
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:26 000000030000000D00000002
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:25 000000030000000D00000001
-rw------- 1 lightdb lightdb        85 Nov  4 23:17 00000003.history
-rw------- 1 lightdb lightdb        42 Nov  4 23:05 00000002.history
[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ ls -lt
total 2097180
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:35 000000030000000D00000000
drwx------ 2 lightdb lightdb      4096 Nov  7 11:26 archive_status
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:26 000000030000000C00000007
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:26 000000030000000D00000002
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:25 000000030000000D00000001
-rw------- 1 lightdb lightdb        85 Nov  4 23:17 00000003.history
-rw------- 1 lightdb lightdb        42 Nov  4 23:05 00000002.history
[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ ls -lt
total 2097180
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:36 000000030000000D00000001
drwx------ 2 lightdb lightdb      4096 Nov  7 11:36 archive_status
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:36 000000030000000D00000000
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:26 000000030000000C00000007
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:26 000000030000000D00000002
-rw------- 1 lightdb lightdb        85 Nov  4 23:17 00000003.history
-rw------- 1 lightdb lightdb        42 Nov  4 23:05 00000002.history
[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ ls -lt
total 2097180
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:36 000000030000000D00000001
drwx------ 2 lightdb lightdb      4096 Nov  7 11:36 archive_status
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:36 000000030000000D00000000
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:26 000000030000000C00000007
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:26 000000030000000D00000002
-rw------- 1 lightdb lightdb        85 Nov  4 23:17 00000003.history
-rw------- 1 lightdb lightdb        42 Nov  4 23:05 00000002.history
[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ ls -lt
total 2097180
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:36 000000030000000D00000001
drwx------ 2 lightdb lightdb      4096 Nov  7 11:36 archive_status
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:36 000000030000000D00000000
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:26 000000030000000C00000007
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:26 000000030000000D00000002
-rw------- 1 lightdb lightdb        85 Nov  4 23:17 00000003.history
-rw------- 1 lightdb lightdb        42 Nov  4 23:05 00000002.history
[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ ls -lt
total 2097180
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:37 000000030000000D00000002
drwx------ 2 lightdb lightdb      4096 Nov  7 11:37 archive_status
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:37 000000030000000D00000001
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:36 000000030000000D00000000
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:26 000000030000000C00000007
-rw------- 1 lightdb lightdb        85 Nov  4 23:17 00000003.history
-rw------- 1 lightdb lightdb        42 Nov  4 23:05 00000002.history
[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ ls -lt
total 2097180
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:37 000000030000000D00000003
drwx------ 2 lightdb lightdb      4096 Nov  7 11:37 archive_status
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:37 000000030000000D00000006
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:37 000000030000000D00000005
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:36 000000030000000D00000004
-rw------- 1 lightdb lightdb        85 Nov  4 23:17 00000003.history
-rw------- 1 lightdb lightdb        42 Nov  4 23:05 00000002.history
[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ ls -lt
total 2097180
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:38 000000030000000D00000004
drwx------ 2 lightdb lightdb      4096 Nov  7 11:38 archive_status
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:38 000000030000000D00000003
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:37 000000030000000D00000006
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:37 000000030000000D00000005
-rw------- 1 lightdb lightdb        85 Nov  4 23:17 00000003.history
-rw------- 1 lightdb lightdb        42 Nov  4 23:05 00000002.history
[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ ls -lt
total 2097180
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:38 000000030000000D00000005
drwx------ 2 lightdb lightdb      4096 Nov  7 11:38 archive_status
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:38 000000030000000D00000007
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:38 000000030000000E00000000
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:37 000000030000000D00000006
-rw------- 1 lightdb lightdb        85 Nov  4 23:17 00000003.history
-rw------- 1 lightdb lightdb        42 Nov  4 23:05 00000002.history
[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ ls -lt
total 2097180
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:39 000000030000000D00000006
drwx------ 2 lightdb lightdb      4096 Nov  7 11:39 archive_status
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:39 000000030000000D00000005
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:38 000000030000000D00000007
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:38 000000030000000E00000000
-rw------- 1 lightdb lightdb        85 Nov  4 23:17 00000003.history
-rw------- 1 lightdb lightdb        42 Nov  4 23:05 00000002.history
[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ ls -lt
total 2097180
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:40 000000030000000D00000007
drwx------ 2 lightdb lightdb      4096 Nov  7 11:40 archive_status
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:40 000000030000000E00000002
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:39 000000030000000E00000001
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:38 000000030000000E00000000
-rw------- 1 lightdb lightdb        85 Nov  4 23:17 00000003.history
-rw------- 1 lightdb lightdb        42 Nov  4 23:05 00000002.history
[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ ls -lt
total 2097180
-rw------- 1 lightdb lightdb 536870912 Nov  7 12:30 000000030000000D00000007
drwx------ 2 lightdb lightdb      4096 Nov  7 11:40 archive_status
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:40 000000030000000E00000002
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:39 000000030000000E00000001
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:38 000000030000000E00000000
-rw------- 1 lightdb lightdb        85 Nov  4 23:17 00000003.history
-rw------- 1 lightdb lightdb        42 Nov  4 23:05 00000002.history
[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ ls -lt
total 2097180
-rw------- 1 lightdb lightdb 536870912 Nov  7 12:30 000000030000000E00000000
drwx------ 2 lightdb lightdb      4096 Nov  7 12:30 archive_status
-rw------- 1 lightdb lightdb 536870912 Nov  7 12:30 000000030000000D00000007
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:40 000000030000000E00000002
-rw------- 1 lightdb lightdb 536870912 Nov  7 11:39 000000030000000E00000001
-rw------- 1 lightdb lightdb        85 Nov  4 23:17 00000003.history
-rw------- 1 lightdb lightdb        42 Nov  4 23:05 00000002.history

[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ lt_waldump -s E/61045FB0  000000030000000E00000003
rmgr: Transaction len (rec/tot):     34/    34, tx:    2075705, lsn: E/61052460, prev E/610523C0, desc: COMMIT 2021-11-07 14:42:26.209643 CST
rmgr: Heap        len (rec/tot):    155/   155, tx:    2075706, lsn: E/61052488, prev E/61052460, desc: HOT_UPDATE off 18 xmax 2075706 flags 0x60 ; new off 19 xmax 0, blkref #0: rel 1663/20678/16575 blk 13
rmgr: Transaction len (rec/tot):     34/    34, tx:    2075706, lsn: E/61052528, prev E/61052488, desc: COMMIT 2021-11-07 14:42:26.210615 CST
rmgr: Heap        len (rec/tot):     78/    78, tx:    2075707, lsn: E/61052550, prev E/61052528, desc: HOT_UPDATE off 24 xmax 2075707 flags 0x60 ; new off 25 xmax 0, blkref #0: rel 1663/15154/17153 blk 19
rmgr: Heap        len (rec/tot):    155/   155, tx:    2075707, lsn: E/610525A0, prev E/61052550, desc: HOT_UPDATE off 25 xmax 2075707 flags 0x60 ; new off 26 xmax 0, blkref #0: rel 1663/15154/17153 blk 19
rmgr: Transaction len (rec/tot):     34/    34, tx:    2075707, lsn: E/61052640, prev E/610525A0, desc: COMMIT 2021-11-07 14:42:26.211391 CST
rmgr: XLOG        len (rec/tot):     24/    24, tx:          0, lsn: E/61052668, prev E/61052640, desc: SWITCH --执行了切换命令

## 不指定-s也可以用来查看某个wal文件的起始lsn、事务号，反过来可以看到上个文件的最后lsn和事务号
[lightdb@iZvv70fyhxdyvjgzmr7ikpZ pg_wal]$ lt_waldump -n 5 000000030000000E00000004
rmgr: Heap        len (rec/tot):    111/   111, tx:    2075708, lsn: E/80000028【第一条记录刚好从XLogLongPageHeaderData之后开始】, prev E/61052668, desc: INSERT off 75 flags 0x00, blkref #0: rel 1663/17542/17577 blk 1200
rmgr: Heap        len (rec/tot):    163/   163, tx:    2075710, lsn: E/80000098, prev E/80000028, desc: HOT_UPDATE off 19 xmax 2075710 flags 0x60 ; new off 20 xmax 0, blkref #0: rel 1663/20678/16575 blk 13
rmgr: Heap        len (rec/tot):    163/   163, tx:    2075711, lsn: E/80000140, prev E/80000098, desc: HOT_UPDATE off 32 xmax 2075711 flags 0x60 ; new off 33 xmax 0, blkref #0: rel 1663/17542/16575 blk 21
rmgr: Transaction len (rec/tot):     34/    34, tx:    2075710, lsn: E/800001E8, prev E/80000140, desc: COMMIT 2021-11-07 14:42:35.151189 CST
rmgr: Heap        len (rec/tot):    163/   163, tx:    2075709, lsn: E/80000210, prev E/800001E8, desc: HOT_UPDATE off 50 xmax 2075709 flags 0x60 ; new off 51 xmax 0, blkref #0: rel 1663/17614/16575 blk 19

postgres=# SELECT * FROM pg_walfile_name_offset('E/80000028');
file_name | file_offset
--------------------------+-------------
000000030000000E00000004 | 40
(1 row)

　　如果开启了归档的话，文件在满足min_wal_size后被删除之前一定会被归档。

　　回到切换的源码实现RequestXLogSwitch。

XLogRecPtr
RequestXLogSwitch(bool mark_unimportant)
{
    XLogRecPtr    RecPtr;

    /* XLOG SWITCH has no data */
    XLogBeginInsert();

    if (mark_unimportant)
        XLogSetRecordFlags(XLOG_MARK_UNIMPORTANT);
    RecPtr = XLogInsert(RM_XLOG_ID, XLOG_SWITCH);

    return RecPtr;
}

XLogRecPtr
XLogInsert(RmgrId rmid, uint8 info)
{
    XLogRecPtr    EndPos;

    /* XLogBeginInsert() must have been called. */
    if (!begininsert_called)
        elog(ERROR, "XLogBeginInsert was not called");

    /*
     * The caller can set rmgr bits, XLR_SPECIAL_REL_UPDATE and
     * XLR_CHECK_CONSISTENCY; the rest are reserved for use by me.
     */
    if ((info & ~(XLR_RMGR_INFO_MASK |
                  XLR_SPECIAL_REL_UPDATE |
                  XLR_CHECK_CONSISTENCY)) != 0)
        elog(PANIC, "invalid xlog info mask %02X", info);

    TRACE_POSTGRESQL_WAL_INSERT(rmid, info);

    /*
     * In bootstrap mode, we don't actually log anything but XLOG resources;
     * return a phony record pointer.
     */
    if (IsBootstrapProcessingMode() && rmid != RM_XLOG_ID)
    {
        XLogResetInsertion();
        EndPos = SizeOfXLogLongPHD; /* start of 1st chkpt record */
        return EndPos;
    }

    do
    {
        XLogRecPtr    RedoRecPtr;
        bool        doPageWrites;
        XLogRecPtr    fpw_lsn;
        XLogRecData *rdt;
        int            num_fpi = 0;

        /*
         * Get values needed to decide whether to do full-page writes. Since
         * we don't yet have an insertion lock, these could change under us,
         * but XLogInsertRecord will recheck them once it has a lock.
         */
        GetFullPageWriteInfo(&RedoRecPtr, &doPageWrites);

        rdt = XLogRecordAssemble(rmid, info, RedoRecPtr, doPageWrites,
                                 &fpw_lsn, &num_fpi);

        EndPos = XLogInsertRecord(rdt, fpw_lsn, curinsert_flags, num_fpi);    ## 统一实现写wal记录和切换的函数
    } while (EndPos == InvalidXLogRecPtr);

    XLogResetInsertion();

    return EndPos;
}

　　checkpoint_timeout=1800(默认300）checkpoint完全应该自动，根据恢复时间要求来自动调整，跟oracle一样（因为checkpoint以来写了多少WAL，启动恢复相应数量的WAL需要多久也知道，所以就能算出来）。超过max_wal_size之后也会执行强制checkpoint（因为wal被重用）

　　pg_start_backup和pg_basebackup,CREATE DATABASE / DROP DATABASE执行的时候，停止pg的时候也会执行checkpoint。
　　事务隔离性使得写wal可以随便并发。

　　wal_log_hints会导致checkpoint后，页如果发生了即使不重要的数据改动,也遵循全页写机制。如果wal_checksum启用了，这个参数会自动强制启用。
　　full_page_writes为了避免部分写（8k被内核拆分为4k，被磁盘拆分为512字节） https://www.2ndquadrant.com/en/blog/on-the-impact-of-full-page-writes/，所以只要恢复一部分数据块即可，还不用先读取data file。这样的话checkpoint的频率以及shared buffer的大小（增量检查点很重要）就非常重要了。
　　checksum在每次写到OS以及读的时候计算，不是每次修改都计算，所以那负载就好多了（offload cpu checksum），所以OLTP开启吧https://postgreshelp.com/postgresql-checksum/，DSS和批处理不要开启。
　　pg_rewind（PostgreSQL 9.5官方包含，用于主从故障切换），一般推荐使用repmgr rejoin机制。

WAL的写入

　　一直以为wal日志记录的写入是由wal writer负责的，debug下来发现主要是backend进程写的。如下：

2021-07-17 22:39:08.196670T psql zjh@postgres localhost(58634) client backend INSERT 00000 [2021-07-17 22:39:01 CST] 598 [59686] LOG:  00000: INSERT @ 1/C18B8330:  - Heap/INSERT: off 57 flags 0x00
2021-07-17 22:39:08.196670T psql zjh@postgres localhost(58634) client backend INSERT 00000 [2021-07-17 22:39:01 CST] 598 [59686] LOCATION:  XLogInsertRecord, xlog.c:1245
2021-07-17 22:39:08.196670T psql zjh@postgres localhost(58634) client backend INSERT 00000 [2021-07-17 22:39:01 CST] 598 [59686] STATEMENT:  insert into t_sample select* from t_sample limit 100000;

2021-07-17 22:39:08.196692T psql zjh@postgres localhost(58634) client backend INSERT 00000 [2021-07-17 22:39:01 CST] 598 [59686] LOG:  00000: INSERT @ 1/C18B8370:  - Btree/INSERT_LEAF: off 259
2021-07-17 22:39:08.196692T psql zjh@postgres localhost(58634) client backend INSERT 00000 [2021-07-17 22:39:01 CST] 598 [59686] LOCATION:  XLogInsertRecord, xlog.c:1245
2021-07-17 22:39:08.196692T psql zjh@postgres localhost(58634) client backend INSERT 00000 [2021-07-17 22:39:01 CST] 598 [59686] STATEMENT:  insert into t_sample select* from t_sample limit 100000;

2021-07-17 23:59:15.533598T psql zjh@postgres localhost(48414) client backend INSERT 00000 [2021-07-17 23:59:11 CST] 599 [131103] DEBUG:  00000: CommitTransaction(1) name: unnamed; blockState
: STARTED; state: INPROGRESS, xid/subid/cid: 599/1/0 (used)
2021-07-17 23:59:15.533598T psql zjh@postgres localhost(48414) client backend INSERT 00000 [2021-07-17 23:59:11 CST] 599 [131103] LOCATION:  ShowTransactionStateRec, xact.c:5351
2021-07-17 23:59:15.533598T psql zjh@postgres localhost(48414) client backend INSERT 00000 [2021-07-17 23:59:11 CST] 599 [131103] STATEMENT:  insert into t_sample select* from t_sample limit 
10;
2021-07-17 23:59:15.533642T psql zjh@postgres localhost(48414) client backend INSERT 00000 [2021-07-17 23:59:11 CST] 599 [131103] LOG:  00000: INSERT @ 1/C18BA970:  - Transaction/COMMIT: 2021
-07-17 23:59:15.533614+08
2021-07-17 23:59:15.533642T psql zjh@postgres localhost(48414) client backend INSERT 00000 [2021-07-17 23:59:11 CST] 599 [131103] LOCATION:  XLogInsertRecord, xlog.c:1245
2021-07-17 23:59:15.533642T psql zjh@postgres localhost(48414) client backend INSERT 00000 [2021-07-17 23:59:11 CST] 599 [131103] STATEMENT:  insert into t_sample select* from t_sample limit 
10;
2021-07-17 23:59:15.533658T psql zjh@postgres localhost(48414) client backend INSERT 00000 [2021-07-17 23:59:11 CST] 599 [131103] LOG:  00000: xlog flush request 1/C18BA970; write 1/C18B8538;
 flush 1/C18B8538
2021-07-17 23:59:15.533658T psql zjh@postgres localhost(48414) client backend INSERT 00000 [2021-07-17 23:59:11 CST] 599 [131103] LOCATION:  XLogFlush, xlog.c:2873
2021-07-17 23:59:15.533658T psql zjh@postgres localhost(48414) client backend INSERT 00000 [2021-07-17 23:59:11 CST] 599 [131103] STATEMENT:  insert into t_sample select* from t_sample limit 
10;

        /*
         * If this is the single and first tuple on page, we can reinit the
         * page instead of restoring the whole thing.  Set flag, and hide
         * buffer references from XLogInsert.
         */
        if (ItemPointerGetOffsetNumber(&(heaptup->t_self)) == FirstOffsetNumber &&
            PageGetMaxOffsetNumber(page) == FirstOffsetNumber)
        {
            info |= XLOG_HEAP_INIT_PAGE;
            bufflags |= REGBUF_WILL_INIT;
        }

        xlrec.offnum = ItemPointerGetOffsetNumber(&heaptup->t_self);
        xlrec.flags = 0;
        if (all_visible_cleared)
            xlrec.flags |= XLH_INSERT_ALL_VISIBLE_CLEARED;
        if (options & HEAP_INSERT_SPECULATIVE)
            xlrec.flags |= XLH_INSERT_IS_SPECULATIVE;
        Assert(ItemPointerGetBlockNumber(&heaptup->t_self) == BufferGetBlockNumber(buffer));

        /*
         * For logical decoding, we need the tuple even if we're doing a full
         * page write, so make sure it's included even if we take a full-page
         * image. (XXX We could alternatively store a pointer into the FPW).
         */
        if (RelationIsLogicallyLogged(relation) &&
            !(options & HEAP_INSERT_NO_LOGICAL))
        {
            xlrec.flags |= XLH_INSERT_CONTAINS_NEW_TUPLE;
            bufflags |= REGBUF_KEEP_DATA;
        }

        XLogBeginInsert();
        XLogRegisterData((char *) &xlrec, SizeOfHeapInsert);

        xlhdr.t_infomask2 = heaptup->t_data->t_infomask2;
        xlhdr.t_infomask = heaptup->t_data->t_infomask;
        xlhdr.t_hoff = heaptup->t_data->t_hoff;

        /*
         * note we mark xlhdr as belonging to buffer; if XLogInsert decides to
         * write the whole page to the xlog, we don't need to store
         * xl_heap_header in the xlog.
         */
        XLogRegisterBuffer(0, buffer, REGBUF_STANDARD | bufflags);
        XLogRegisterBufData(0, (char *) &xlhdr, SizeOfHeapHeader);
        /* PG73FORMAT: write bitmap [+ padding] [+ oid] + data */
        XLogRegisterBufData(0,
                            (char *) heaptup->t_data + SizeofHeapTupleHeader,
                            heaptup->t_len - SizeofHeapTupleHeader);

        /* filtering by origin on a row level is much more efficient */
        XLogSetRecordFlags(XLOG_INCLUDE_ORIGIN);

        recptr = XLogInsert(RM_HEAP_ID, info);

        PageSetLSN(page, recptr);

　　所有写入逻辑通过xloginsert.c中的XLogInsert函数调用XLogInsertRecord实现，这样的话wal offset的位置如何分配就非常重要了（毕竟会有连接死掉，超过max_wal_size时还有事务未提交等情况）。

　　至于wal writer。异步模式下，wal writer负责定时（wal_writer_delay控制）负责调用XLogBackgroundFlush(void)

2021-07-17 22:39:29.119288T  @  walwriter  00000 [2021-07-15 09:26:52 CST] 0 [133565] LOG:  00000: xlog bg flush request write 1/C18B8450; flush: 1/C18B8450, current is write 1/C18B8418; flush 1/C18B8418
2021-07-17 22:39:29.119288T  @  walwriter  00000 [2021-07-15 09:26:52 CST] 0 [133565] LOCATION:  XLogBackgroundFlush, xlog.c:3125
2021-07-17 22:39:29.119559T  @  walwriter  00000 [2021-07-15 09:26:52 CST] 0 [133565] DEBUG:  00000: initialized 92 pages, up to 1/C1CB8000
2021-07-17 22:39:29.119559T  @  walwriter  00000 [2021-07-15 09:26:52 CST] 0 [133565] LOCATION:  AdvanceXLInsertBuffer, xlog.c:2283

AdvanceXLInsertBuffer
- 　　GetXLogBuffer
  - 　　　　CopyXLogRecordToWAL(3 usages)
    - 　　　　　　XLogInsertRecord
      - 　　　　　　　　XLogInsert
XLogBackgroundFlush
- 　　WalSndWaitForWal
  - 　　logical_read_xlog_page
- 　　WalWriterMain
  - 　　　　AuxiliaryProcessMain
    - 　　　　　　main
    - 　　　　　　StartChildProcess

WAL解析

　　要正确解析wal，需要使用一个大小为最大理论上支持的并发数的wal entry数组做解析缓存。维护一个全局当前已经解析的lsn个apply的lsn。除了get外，解析和应用也可以完全并行。

XLOG/WAL的主要状态

/*
 * Total shared-memory state for XLOG.
 */
typedef struct XLogCtlData
{
    XLogCtlInsert Insert;

    /* Protected by info_lck: */
    XLogwrtRqst LogwrtRqst;
    XLogRecPtr    RedoRecPtr;        /* a recent copy of Insert->RedoRecPtr */
    FullTransactionId ckptFullXid;    /* nextFullXid of latest checkpoint */
    XLogRecPtr    asyncXactLSN;    /* LSN of newest async commit/abort */
    XLogRecPtr    replicationSlotMinLSN;    /* oldest LSN needed by any slot */

    XLogSegNo    lastRemovedSegNo;    /* latest removed/recycled XLOG segment */

    /* Fake LSN counter, for unlogged relations. Protected by ulsn_lck. */
    XLogRecPtr    unloggedLSN;
    slock_t        ulsn_lck;

    /* Time and LSN of last xlog segment switch. Protected by WALWriteLock. */
    pg_time_t    lastSegSwitchTime;
    XLogRecPtr    lastSegSwitchLSN;

    /*
     * Protected by info_lck and WALWriteLock (you must hold either lock to
     * read it, but both to update)
     */
    XLogwrtResult LogwrtResult;

    /*
     * Latest initialized page in the cache (last byte position + 1).
     *
     * To change the identity of a buffer (and InitializedUpTo), you need to
     * hold WALBufMappingLock.  To change the identity of a buffer that's
     * still dirty, the old page needs to be written out first, and for that
     * you need WALWriteLock, and you need to ensure that there are no
     * in-progress insertions to the page by calling
     * WaitXLogInsertionsToFinish().
     */
    XLogRecPtr    InitializedUpTo;

    /*
     * These values do not change after startup, although the pointed-to pages
     * and xlblocks values certainly do.  xlblocks values are protected by
     * WALBufMappingLock.
     */
    char       *pages;            /* buffers for unwritten XLOG pages */
    XLogRecPtr *xlblocks;        /* 1st byte ptr-s + XLOG_BLCKSZ */
    int            XLogCacheBlck;    /* highest allocated xlog buffer index */

    /*
     * Shared copy of ThisTimeLineID. Does not change after end-of-recovery.
     * If we created a new timeline when the system was started up,
     * PrevTimeLineID is the old timeline's ID that we forked off from.
     * Otherwise it's equal to ThisTimeLineID.
     */
    TimeLineID    ThisTimeLineID;
    TimeLineID    PrevTimeLineID;

    /*
     * SharedRecoveryState indicates if we're still in crash or archive
     * recovery.  Protected by info_lck.
     */
    RecoveryState SharedRecoveryState;

    /*
     * SharedHotStandbyActive indicates if we allow hot standby queries to be
     * run.  Protected by info_lck.
     */
    bool        SharedHotStandbyActive;

    /*
     * SharedPromoteIsTriggered indicates if a standby promotion has been
     * triggered.  Protected by info_lck.
     */
    bool        SharedPromoteIsTriggered;

    /*
     * WalWriterSleeping indicates whether the WAL writer is currently in
     * low-power mode (and hence should be nudged if an async commit occurs).
     * Protected by info_lck.
     */
    bool        WalWriterSleeping;

    /*
     * recoveryWakeupLatch is used to wake up the startup process to continue
     * WAL replay, if it is waiting for WAL to arrive or failover trigger file
     * to appear.
     */
    Latch        recoveryWakeupLatch;

    /*
     * During recovery, we keep a copy of the latest checkpoint record here.
     * lastCheckPointRecPtr points to start of checkpoint record and
     * lastCheckPointEndPtr points to end+1 of checkpoint record.  Used by the
     * checkpointer when it wants to create a restartpoint.
     *
     * Protected by info_lck.
     */
    XLogRecPtr    lastCheckPointRecPtr;
    XLogRecPtr    lastCheckPointEndPtr;
    CheckPoint    lastCheckPoint;

    /*
     * lastReplayedEndRecPtr points to end+1 of the last record successfully
     * replayed. When we're currently replaying a record, ie. in a redo
     * function, replayEndRecPtr points to the end+1 of the record being
     * replayed, otherwise it's equal to lastReplayedEndRecPtr.
     */
    XLogRecPtr    lastReplayedEndRecPtr;
    TimeLineID    lastReplayedTLI;
    XLogRecPtr    replayEndRecPtr;
    TimeLineID    replayEndTLI;
    /* timestamp of last COMMIT/ABORT record replayed (or being replayed) */
    TimestampTz recoveryLastXTime;

    /*
     * timestamp of when we started replaying the current chunk of WAL data,
     * only relevant for replication or archive recovery
     */
    TimestampTz currentChunkStartTime;
    /* Are we requested to pause recovery? */
    bool        recoveryPause;

    /*
     * lastFpwDisableRecPtr points to the start of the last replayed
     * XLOG_FPW_CHANGE record that instructs full_page_writes is disabled.
     */
    XLogRecPtr    lastFpwDisableRecPtr;

    slock_t        info_lck;        /* locks shared variables shown above */
} XLogCtlData;