Make WAL segment size configurable at initdb time.

For performance reasons a larger segment size than the default 16MB
can be useful. A larger segment size has two main benefits: Firstly,
in setups using archiving, it makes it easier to write scripts that
can keep up with higher amounts of WAL, secondly, the WAL has to be
written and synced to disk less frequently.

But at the same time large segment size are disadvantageous for
smaller databases. So far the segment size had to be configured at
compile time, often making it unrealistic to choose one fitting to a
particularly load. Therefore change it to a initdb time setting.

This includes a breaking changes to the xlogreader.h API, which now
requires the current segment size to be configured.  For that and
similar reasons a number of binaries had to be taught how to recognize
the current segment size.

Author: Beena Emerson, editorialized by Andres Freund
Reviewed-By: Andres Freund, David Steele, Kuntal Ghosh, Michael
    Paquier, Peter Eisentraut, Robert Hass, Tushar Ahuja
Discussion: https://postgr.es/m/CAOG9ApEAcQ--1ieKbhFzXSQPw_YLmepaa4hNdnY5+ZULpt81Mw@mail.gmail.com

1 files changed, 151 insertions, 104 deletions

diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 96ebf32a58a..051347163b5 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -110,6 +110,8 @@ int			wal_retrieve_retry_interval = 5000;
 bool		XLOG_DEBUG = false;
 #endif
 
+int			wal_segment_size = DEFAULT_XLOG_SEG_SIZE;
+
 /*
  * Number of WAL insertion locks to use. A higher value allows more insertions
  * to happen concurrently, but adds some CPU overhead to flushing the WAL,
@@ -731,14 +733,16 @@ static ControlFileData *ControlFile = NULL;
 	(((recptr) / XLOG_BLCKSZ) % (XLogCtl->XLogCacheBlck + 1))
 
 /*
- * These are the number of bytes in a WAL page and segment usable for WAL data.
+ * These are the number of bytes in a WAL page usable for WAL data.
  */
 #define UsableBytesInPage (XLOG_BLCKSZ - SizeOfXLogShortPHD)
-#define UsableBytesInSegment ((XLOG_SEG_SIZE / XLOG_BLCKSZ) * UsableBytesInPage - (SizeOfXLogLongPHD - SizeOfXLogShortPHD))
 
 /* Convert min_wal_size_mb and max wal_size_mb to equivalent segment count */
-#define ConvertToXSegs(x)	\
-	(x / (XLOG_SEG_SIZE / (1024 * 1024)))
+#define ConvertToXSegs(x, segsize)	\
+	(x / ((segsize) / (1024 * 1024)))
+
+/* The number of bytes in a WAL segment usable for WAL data. */
+static int	UsableBytesInSegment;
 
 /*
  * Private, possibly out-of-date copy of shared LogwrtResult.
@@ -1137,7 +1141,9 @@ XLogInsertRecord(XLogRecData *rdata,
 			EndPos = StartPos + SizeOfXLogRecord;
 			if (StartPos / XLOG_BLCKSZ != EndPos / XLOG_BLCKSZ)
 			{
-				if (EndPos % XLOG_SEG_SIZE == EndPos % XLOG_BLCKSZ)
+				uint64		offset = XLogSegmentOffset(EndPos, wal_segment_size);
+
+				if (offset == EndPos % XLOG_BLCKSZ)
 					EndPos += SizeOfXLogLongPHD;
 				else
 					EndPos += SizeOfXLogShortPHD;
@@ -1170,7 +1176,7 @@ XLogInsertRecord(XLogRecData *rdata,
 			appendBinaryStringInfo(&recordBuf, rdata->data, rdata->len);
 
 		if (!debug_reader)
-			debug_reader = XLogReaderAllocate(NULL, NULL);
+			debug_reader = XLogReaderAllocate(wal_segment_size, NULL, NULL);
 
 		if (!debug_reader)
 		{
@@ -1296,7 +1302,7 @@ ReserveXLogSwitch(XLogRecPtr *StartPos, XLogRecPtr *EndPos, XLogRecPtr *PrevPtr)
 	startbytepos = Insert->CurrBytePos;
 
 	ptr = XLogBytePosToEndRecPtr(startbytepos);
-	if (ptr % XLOG_SEG_SIZE == 0)
+	if (XLogSegmentOffset(ptr, wal_segment_size) == 0)
 	{
 		SpinLockRelease(&Insert->insertpos_lck);
 		*EndPos = *StartPos = ptr;
@@ -1309,8 +1315,8 @@ ReserveXLogSwitch(XLogRecPtr *StartPos, XLogRecPtr *EndPos, XLogRecPtr *PrevPtr)
 	*StartPos = XLogBytePosToRecPtr(startbytepos);
 	*EndPos = XLogBytePosToEndRecPtr(endbytepos);
 
-	segleft = XLOG_SEG_SIZE - ((*EndPos) % XLOG_SEG_SIZE);
-	if (segleft != XLOG_SEG_SIZE)
+	segleft = wal_segment_size - XLogSegmentOffset(*EndPos, wal_segment_size);
+	if (segleft != wal_segment_size)
 	{
 		/* consume the rest of the segment */
 		*EndPos += segleft;
@@ -1323,7 +1329,7 @@ ReserveXLogSwitch(XLogRecPtr *StartPos, XLogRecPtr *EndPos, XLogRecPtr *PrevPtr)
 
 	*PrevPtr = XLogBytePosToRecPtr(prevbytepos);
 
-	Assert((*EndPos) % XLOG_SEG_SIZE == 0);
+	Assert(XLogSegmentOffset(*EndPos, wal_segment_size) == 0);
 	Assert(XLogRecPtrToBytePos(*EndPos) == endbytepos);
 	Assert(XLogRecPtrToBytePos(*StartPos) == startbytepos);
 	Assert(XLogRecPtrToBytePos(*PrevPtr) == prevbytepos);
@@ -1501,7 +1507,7 @@ CopyXLogRecordToWAL(int write_len, bool isLogSwitch, XLogRecData *rdata,
 			pagehdr->xlp_info |= XLP_FIRST_IS_CONTRECORD;
 
 			/* skip over the page header */
-			if (CurrPos % XLogSegSize == 0)
+			if (XLogSegmentOffset(CurrPos, wal_segment_size) == 0)
 			{
 				CurrPos += SizeOfXLogLongPHD;
 				currpos += SizeOfXLogLongPHD;
@@ -1532,16 +1538,16 @@ CopyXLogRecordToWAL(int write_len, bool isLogSwitch, XLogRecData *rdata,
 	 * allocated and zeroed in the WAL buffers so that when the caller (or
 	 * someone else) does XLogWrite(), it can really write out all the zeros.
 	 */
-	if (isLogSwitch && CurrPos % XLOG_SEG_SIZE != 0)
+	if (isLogSwitch && XLogSegmentOffset(CurrPos, wal_segment_size) != 0)
 	{
 		/* An xlog-switch record doesn't contain any data besides the header */
 		Assert(write_len == SizeOfXLogRecord);
 
 		/*
 		 * We do this one page at a time, to make sure we don't deadlock
-		 * against ourselves if wal_buffers < XLOG_SEG_SIZE.
+		 * against ourselves if wal_buffers < wal_segment_size.
 		 */
-		Assert(EndPos % XLogSegSize == 0);
+		Assert(XLogSegmentOffset(EndPos, wal_segment_size) == 0);
 
 		/* Use up all the remaining space on the first page */
 		CurrPos += freespace;
@@ -1866,10 +1872,10 @@ GetXLogBuffer(XLogRecPtr ptr)
 		 * the page header.
 		 */
 		if (ptr % XLOG_BLCKSZ == SizeOfXLogShortPHD &&
-			ptr % XLOG_SEG_SIZE > XLOG_BLCKSZ)
+			XLogSegmentOffset(ptr, wal_segment_size) > XLOG_BLCKSZ)
 			initializedUpto = ptr - SizeOfXLogShortPHD;
 		else if (ptr % XLOG_BLCKSZ == SizeOfXLogLongPHD &&
-				 ptr % XLOG_SEG_SIZE < XLOG_BLCKSZ)
+				 XLogSegmentOffset(ptr, wal_segment_size) < XLOG_BLCKSZ)
 			initializedUpto = ptr - SizeOfXLogLongPHD;
 		else
 			initializedUpto = ptr;
@@ -1939,7 +1945,7 @@ XLogBytePosToRecPtr(uint64 bytepos)
 		seg_offset += fullpages * XLOG_BLCKSZ + bytesleft + SizeOfXLogShortPHD;
 	}
 
-	XLogSegNoOffsetToRecPtr(fullsegs, seg_offset, result);
+	XLogSegNoOffsetToRecPtr(fullsegs, seg_offset, result, wal_segment_size);
 
 	return result;
 }
@@ -1985,7 +1991,7 @@ XLogBytePosToEndRecPtr(uint64 bytepos)
 			seg_offset += fullpages * XLOG_BLCKSZ + bytesleft + SizeOfXLogShortPHD;
 	}
 
-	XLogSegNoOffsetToRecPtr(fullsegs, seg_offset, result);
+	XLogSegNoOffsetToRecPtr(fullsegs, seg_offset, result, wal_segment_size);
 
 	return result;
 }
@@ -2001,9 +2007,9 @@ XLogRecPtrToBytePos(XLogRecPtr ptr)
 	uint32		offset;
 	uint64		result;
 
-	XLByteToSeg(ptr, fullsegs);
+	XLByteToSeg(ptr, fullsegs, wal_segment_size);
 
-	fullpages = (ptr % XLOG_SEG_SIZE) / XLOG_BLCKSZ;
+	fullpages = (XLogSegmentOffset(ptr, wal_segment_size)) / XLOG_BLCKSZ;
 	offset = ptr % XLOG_BLCKSZ;
 
 	if (fullpages == 0)
@@ -2168,12 +2174,12 @@ AdvanceXLInsertBuffer(XLogRecPtr upto, bool opportunistic)
 		/*
 		 * If first page of an XLOG segment file, make it a long header.
 		 */
-		if ((NewPage->xlp_pageaddr % XLogSegSize) == 0)
+		if ((XLogSegmentOffset(NewPage->xlp_pageaddr, wal_segment_size)) == 0)
 		{
 			XLogLongPageHeader NewLongPage = (XLogLongPageHeader) NewPage;
 
 			NewLongPage->xlp_sysid = ControlFile->system_identifier;
-			NewLongPage->xlp_seg_size = XLogSegSize;
+			NewLongPage->xlp_seg_size = wal_segment_size;
 			NewLongPage->xlp_xlog_blcksz = XLOG_BLCKSZ;
 			NewPage->xlp_info |= XLP_LONG_HEADER;
 		}
@@ -2220,7 +2226,8 @@ CalculateCheckpointSegments(void)
 	 *	  number of segments consumed between checkpoints.
 	 *-------
 	 */
-	target = (double) ConvertToXSegs(max_wal_size_mb) / (2.0 + CheckPointCompletionTarget);
+	target = (double) ConvertToXSegs(max_wal_size_mb, wal_segment_size) /
+		(2.0 + CheckPointCompletionTarget);
 
 	/* round down */
 	CheckPointSegments = (int) target;
@@ -2260,8 +2267,10 @@ XLOGfileslop(XLogRecPtr PriorRedoPtr)
 	 * correspond to. Always recycle enough segments to meet the minimum, and
 	 * remove enough segments to stay below the maximum.
 	 */
-	minSegNo = PriorRedoPtr / XLOG_SEG_SIZE + ConvertToXSegs(min_wal_size_mb) - 1;
-	maxSegNo = PriorRedoPtr / XLOG_SEG_SIZE + ConvertToXSegs(max_wal_size_mb) - 1;
+	minSegNo = PriorRedoPtr / wal_segment_size +
+		ConvertToXSegs(min_wal_size_mb, wal_segment_size) - 1;
+	maxSegNo = PriorRedoPtr / wal_segment_size +
+		ConvertToXSegs(max_wal_size_mb, wal_segment_size) - 1;
 
 	/*
 	 * Between those limits, recycle enough segments to get us through to the
@@ -2290,7 +2299,8 @@ XLOGfileslop(XLogRecPtr PriorRedoPtr)
 	/* add 10% for good measure. */
 	distance *= 1.10;
 
-	recycleSegNo = (XLogSegNo) ceil(((double) PriorRedoPtr + distance) / XLOG_SEG_SIZE);
+	recycleSegNo = (XLogSegNo) ceil(((double) PriorRedoPtr + distance) /
+									wal_segment_size);
 
 	if (recycleSegNo < minSegNo)
 		recycleSegNo = minSegNo;
@@ -2314,7 +2324,7 @@ XLogCheckpointNeeded(XLogSegNo new_segno)
 {
 	XLogSegNo	old_segno;
 
-	XLByteToSeg(RedoRecPtr, old_segno);
+	XLByteToSeg(RedoRecPtr, old_segno, wal_segment_size);
 
 	if (new_segno >= old_segno + (uint64) (CheckPointSegments - 1))
 		return true;
@@ -2392,7 +2402,8 @@ XLogWrite(XLogwrtRqst WriteRqst, bool flexible)
 		LogwrtResult.Write = EndPtr;
 		ispartialpage = WriteRqst.Write < LogwrtResult.Write;
 
-		if (!XLByteInPrevSeg(LogwrtResult.Write, openLogSegNo))
+		if (!XLByteInPrevSeg(LogwrtResult.Write, openLogSegNo,
+							 wal_segment_size))
 		{
 			/*
 			 * Switch to new logfile segment.  We cannot have any pending
@@ -2401,7 +2412,8 @@ XLogWrite(XLogwrtRqst WriteRqst, bool flexible)
 			Assert(npages == 0);
 			if (openLogFile >= 0)
 				XLogFileClose();
-			XLByteToPrevSeg(LogwrtResult.Write, openLogSegNo);
+			XLByteToPrevSeg(LogwrtResult.Write, openLogSegNo,
+							wal_segment_size);
 
 			/* create/use new log file */
 			use_existent = true;
@@ -2412,7 +2424,8 @@ XLogWrite(XLogwrtRqst WriteRqst, bool flexible)
 		/* Make sure we have the current logfile open */
 		if (openLogFile < 0)
 		{
-			XLByteToPrevSeg(LogwrtResult.Write, openLogSegNo);
+			XLByteToPrevSeg(LogwrtResult.Write, openLogSegNo,
+							wal_segment_size);
 			openLogFile = XLogFileOpen(openLogSegNo);
 			openLogOff = 0;
 		}
@@ -2422,7 +2435,8 @@ XLogWrite(XLogwrtRqst WriteRqst, bool flexible)
 		{
 			/* first of group */
 			startidx = curridx;
-			startoffset = (LogwrtResult.Write - XLOG_BLCKSZ) % XLogSegSize;
+			startoffset = XLogSegmentOffset(LogwrtResult.Write - XLOG_BLCKSZ,
+											wal_segment_size);
 		}
 		npages++;
 
@@ -2435,7 +2449,7 @@ XLogWrite(XLogwrtRqst WriteRqst, bool flexible)
 		last_iteration = WriteRqst.Write <= LogwrtResult.Write;
 
 		finishing_seg = !ispartialpage &&
-			(startoffset + npages * XLOG_BLCKSZ) >= XLogSegSize;
+			(startoffset + npages * XLOG_BLCKSZ) >= wal_segment_size;
 
 		if (last_iteration ||
 			curridx == XLogCtl->XLogCacheBlck ||
@@ -2562,11 +2576,13 @@ XLogWrite(XLogwrtRqst WriteRqst, bool flexible)
 			sync_method != SYNC_METHOD_OPEN_DSYNC)
 		{
 			if (openLogFile >= 0 &&
-				!XLByteInPrevSeg(LogwrtResult.Write, openLogSegNo))
+				!XLByteInPrevSeg(LogwrtResult.Write, openLogSegNo,
+								 wal_segment_size))
 				XLogFileClose();
 			if (openLogFile < 0)
 			{
-				XLByteToPrevSeg(LogwrtResult.Write, openLogSegNo);
+				XLByteToPrevSeg(LogwrtResult.Write, openLogSegNo,
+								wal_segment_size);
 				openLogFile = XLogFileOpen(openLogSegNo);
 				openLogOff = 0;
 			}
@@ -2982,7 +2998,8 @@ XLogBackgroundFlush(void)
 	{
 		if (openLogFile >= 0)
 		{
-			if (!XLByteInPrevSeg(LogwrtResult.Write, openLogSegNo))
+			if (!XLByteInPrevSeg(LogwrtResult.Write, openLogSegNo,
+								 wal_segment_size))
 			{
 				XLogFileClose();
 			}
@@ -3161,7 +3178,7 @@ XLogFileInit(XLogSegNo logsegno, bool *use_existent, bool use_lock)
 	int			fd;
 	int			nbytes;
 
-	XLogFilePath(path, ThisTimeLineID, logsegno);
+	XLogFilePath(path, ThisTimeLineID, logsegno, wal_segment_size);
 
 	/*
 	 * Try to use existent file (checkpoint maker may have created it already)
@@ -3215,7 +3232,7 @@ XLogFileInit(XLogSegNo logsegno, bool *use_existent, bool use_lock)
 	 */
 	zbuffer = (char *) MAXALIGN(zbuffer_raw);
 	memset(zbuffer, 0, XLOG_BLCKSZ);
-	for (nbytes = 0; nbytes < XLogSegSize; nbytes += XLOG_BLCKSZ)
+	for (nbytes = 0; nbytes < wal_segment_size; nbytes += XLOG_BLCKSZ)
 	{
 		errno = 0;
 		pgstat_report_wait_start(WAIT_EVENT_WAL_INIT_WRITE);
@@ -3332,7 +3349,7 @@ XLogFileCopy(XLogSegNo destsegno, TimeLineID srcTLI, XLogSegNo srcsegno,
 	/*
 	 * Open the source file
 	 */
-	XLogFilePath(path, srcTLI, srcsegno);
+	XLogFilePath(path, srcTLI, srcsegno, wal_segment_size);
 	srcfd = OpenTransientFile(path, O_RDONLY | PG_BINARY, 0);
 	if (srcfd < 0)
 		ereport(ERROR,
@@ -3357,7 +3374,7 @@ XLogFileCopy(XLogSegNo destsegno, TimeLineID srcTLI, XLogSegNo srcsegno,
 	/*
 	 * Do the data copying.
 	 */
-	for (nbytes = 0; nbytes < XLogSegSize; nbytes += sizeof(buffer))
+	for (nbytes = 0; nbytes < wal_segment_size; nbytes += sizeof(buffer))
 	{
 		int			nread;
 
@@ -3467,7 +3484,7 @@ InstallXLogFileSegment(XLogSegNo *segno, char *tmppath,
 	char		path[MAXPGPATH];
 	struct stat stat_buf;
 
-	XLogFilePath(path, ThisTimeLineID, *segno);
+	XLogFilePath(path, ThisTimeLineID, *segno, wal_segment_size);
 
 	/*
 	 * We want to be sure that only one process does this at a time.
@@ -3493,7 +3510,7 @@ InstallXLogFileSegment(XLogSegNo *segno, char *tmppath,
 				return false;
 			}
 			(*segno)++;
-			XLogFilePath(path, ThisTimeLineID, *segno);
+			XLogFilePath(path, ThisTimeLineID, *segno, wal_segment_size);
 		}
 	}
 
@@ -3524,7 +3541,7 @@ XLogFileOpen(XLogSegNo segno)
 	char		path[MAXPGPATH];
 	int			fd;
 
-	XLogFilePath(path, ThisTimeLineID, segno);
+	XLogFilePath(path, ThisTimeLineID, segno, wal_segment_size);
 
 	fd = BasicOpenFile(path, O_RDWR | PG_BINARY | get_sync_bit(sync_method),
 					   S_IRUSR | S_IWUSR);
@@ -3551,7 +3568,7 @@ XLogFileRead(XLogSegNo segno, int emode, TimeLineID tli,
 	char		path[MAXPGPATH];
 	int			fd;
 
-	XLogFileName(xlogfname, tli, segno);
+	XLogFileName(xlogfname, tli, segno, wal_segment_size);
 
 	switch (source)
 	{
@@ -3563,7 +3580,7 @@ XLogFileRead(XLogSegNo segno, int emode, TimeLineID tli,
 
 			restoredFromArchive = RestoreArchivedFile(path, xlogfname,
 													  "RECOVERYXLOG",
-													  XLogSegSize,
+													  wal_segment_size,
 													  InRedo);
 			if (!restoredFromArchive)
 				return -1;
@@ -3571,7 +3588,7 @@ XLogFileRead(XLogSegNo segno, int emode, TimeLineID tli,
 
 		case XLOG_FROM_PG_WAL:
 		case XLOG_FROM_STREAM:
-			XLogFilePath(path, tli, segno);
+			XLogFilePath(path, tli, segno, wal_segment_size);
 			restoredFromArchive = false;
 			break;
 
@@ -3690,7 +3707,7 @@ XLogFileReadAnyTLI(XLogSegNo segno, int emode, int source)
 	}
 
 	/* Couldn't find it.  For simplicity, complain about front timeline */
-	XLogFilePath(path, recoveryTargetTLI, segno);
+	XLogFilePath(path, recoveryTargetTLI, segno, wal_segment_size);
 	errno = ENOENT;
 	ereport(emode,
 			(errcode_for_file_access(),
@@ -3741,9 +3758,11 @@ PreallocXlogFiles(XLogRecPtr endptr)
 	XLogSegNo	_logSegNo;
 	int			lf;
 	bool		use_existent;
+	uint64		offset;
 
-	XLByteToPrevSeg(endptr, _logSegNo);
-	if ((endptr - 1) % XLogSegSize >= (uint32) (0.75 * XLogSegSize))
+	XLByteToPrevSeg(endptr, _logSegNo, wal_segment_size);
+	offset = XLogSegmentOffset(endptr - 1, wal_segment_size);
+	if (offset >= (uint32) (0.75 * wal_segment_size))
 	{
 		_logSegNo++;
 		use_existent = true;
@@ -3774,7 +3793,7 @@ CheckXLogRemoved(XLogSegNo segno, TimeLineID tli)
 	{
 		char		filename[MAXFNAMELEN];
 
-		XLogFileName(filename, tli, segno);
+		XLogFileName(filename, tli, segno, wal_segment_size);
 		ereport(ERROR,
 				(errcode_for_file_access(),
 				 errmsg("requested WAL segment %s has already been removed",
@@ -3811,7 +3830,7 @@ UpdateLastRemovedPtr(char *filename)
 	uint32		tli;
 	XLogSegNo	segno;
 
-	XLogFromFileName(filename, &tli, &segno);
+	XLogFromFileName(filename, &tli, &segno, wal_segment_size);
 
 	SpinLockAcquire(&XLogCtl->info_lck);
 	if (segno > XLogCtl->lastRemovedSegNo)
@@ -3845,7 +3864,7 @@ RemoveOldXlogFiles(XLogSegNo segno, XLogRecPtr PriorRedoPtr, XLogRecPtr endptr)
 	 * doesn't matter, we ignore that in the comparison. (During recovery,
 	 * ThisTimeLineID isn't set, so we can't use that.)
 	 */
-	XLogFileName(lastoff, 0, segno);
+	XLogFileName(lastoff, 0, segno, wal_segment_size);
 
 	elog(DEBUG2, "attempting to remove WAL segments older than log file %s",
 		 lastoff);
@@ -3906,7 +3925,7 @@ RemoveNonParentXlogFiles(XLogRecPtr switchpoint, TimeLineID newTLI)
 	char		switchseg[MAXFNAMELEN];
 	XLogSegNo	endLogSegNo;
 
-	XLByteToPrevSeg(switchpoint, endLogSegNo);
+	XLByteToPrevSeg(switchpoint, endLogSegNo, wal_segment_size);
 
 	xldir = AllocateDir(XLOGDIR);
 	if (xldir == NULL)
@@ -3918,7 +3937,7 @@ RemoveNonParentXlogFiles(XLogRecPtr switchpoint, TimeLineID newTLI)
 	/*
 	 * Construct a filename of the last segment to be kept.
 	 */
-	XLogFileName(switchseg, newTLI, endLogSegNo);
+	XLogFileName(switchseg, newTLI, endLogSegNo, wal_segment_size);
 
 	elog(DEBUG2, "attempting to remove WAL segments newer than log file %s",
 		 switchseg);
@@ -3974,7 +3993,7 @@ RemoveXlogFile(const char *segname, XLogRecPtr PriorRedoPtr, XLogRecPtr endptr)
 	/*
 	 * Initialize info about where to try to recycle to.
 	 */
-	XLByteToSeg(endptr, endlogSegNo);
+	XLByteToSeg(endptr, endlogSegNo, wal_segment_size);
 	if (PriorRedoPtr == InvalidXLogRecPtr)
 		recycleSegNo = endlogSegNo + 10;
 	else
@@ -4192,9 +4211,11 @@ ReadRecord(XLogReaderState *xlogreader, XLogRecPtr RecPtr, int emode,
 			XLogSegNo	segno;
 			int32		offset;
 
-			XLByteToSeg(xlogreader->latestPagePtr, segno);
-			offset = xlogreader->latestPagePtr % XLogSegSize;
-			XLogFileName(fname, xlogreader->readPageTLI, segno);
+			XLByteToSeg(xlogreader->latestPagePtr, segno, wal_segment_size);
+			offset = XLogSegmentOffset(xlogreader->latestPagePtr,
+									   wal_segment_size);
+			XLogFileName(fname, xlogreader->readPageTLI, segno,
+						 wal_segment_size);
 			ereport(emode_for_corrupt_record(emode,
 											 RecPtr ? RecPtr : EndRecPtr),
 					(errmsg("unexpected timeline ID %u in log segment %s, offset %u",
@@ -4399,7 +4420,7 @@ WriteControlFile(void)
 	ControlFile->blcksz = BLCKSZ;
 	ControlFile->relseg_size = RELSEG_SIZE;
 	ControlFile->xlog_blcksz = XLOG_BLCKSZ;
-	ControlFile->xlog_seg_size = XLOG_SEG_SIZE;
+	ControlFile->xlog_seg_size = wal_segment_size;
 
 	ControlFile->nameDataLen = NAMEDATALEN;
 	ControlFile->indexMaxKeys = INDEX_MAX_KEYS;
@@ -4467,6 +4488,7 @@ ReadControlFile(void)
 {
 	pg_crc32c	crc;
 	int			fd;
+	static char wal_segsz_str[20];
 
 	/*
 	 * Read data...
@@ -4569,13 +4591,6 @@ ReadControlFile(void)
 						   " but the server was compiled with XLOG_BLCKSZ %d.",
 						   ControlFile->xlog_blcksz, XLOG_BLCKSZ),
 				 errhint("It looks like you need to recompile or initdb.")));
-	if (ControlFile->xlog_seg_size != XLOG_SEG_SIZE)
-		ereport(FATAL,
-				(errmsg("database files are incompatible with server"),
-				 errdetail("The database cluster was initialized with XLOG_SEG_SIZE %d,"
-						   " but the server was compiled with XLOG_SEG_SIZE %d.",
-						   ControlFile->xlog_seg_size, XLOG_SEG_SIZE),
-				 errhint("It looks like you need to recompile or initdb.")));
 	if (ControlFile->nameDataLen != NAMEDATALEN)
 		ereport(FATAL,
 				(errmsg("database files are incompatible with server"),
@@ -4637,6 +4652,32 @@ ReadControlFile(void)
 				 errhint("It looks like you need to recompile or initdb.")));
 #endif
 
+	wal_segment_size = ControlFile->xlog_seg_size;
+
+	if (!IsValidWalSegSize(wal_segment_size))
+		ereport(ERROR, (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+						errmsg("WAL segment size must be a power of two between 1MB and 1GB, but the control file specifies %d bytes",
+							   wal_segment_size)));
+
+	snprintf(wal_segsz_str, sizeof(wal_segsz_str), "%d", wal_segment_size);
+	SetConfigOption("wal_segment_size", wal_segsz_str, PGC_INTERNAL,
+					PGC_S_OVERRIDE);
+
+	/* check and update variables dependent on wal_segment_size */
+	if (ConvertToXSegs(min_wal_size_mb, wal_segment_size) < 2)
+		ereport(ERROR, (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+						errmsg("\"min_wal_size\" must be at least twice \"wal_segment_size\".")));
+
+	if (ConvertToXSegs(max_wal_size_mb, wal_segment_size) < 2)
+		ereport(ERROR, (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+						errmsg("\"max_wal_size\" must be at least twice \"wal_segment_size\".")));
+
+	UsableBytesInSegment =
+		(wal_segment_size / XLOG_BLCKSZ * UsableBytesInPage) -
+		(SizeOfXLogLongPHD - SizeOfXLogShortPHD);
+
+	CalculateCheckpointSegments();
+
 	/* Make the initdb settings visible as GUC variables, too */
 	SetConfigOption("data_checksums", DataChecksumsEnabled() ? "yes" : "no",
 					PGC_INTERNAL, PGC_S_OVERRIDE);
@@ -4757,8 +4798,8 @@ XLOGChooseNumBuffers(void)
 	int			xbuffers;
 
 	xbuffers = NBuffers / 32;
-	if (xbuffers > XLOG_SEG_SIZE / XLOG_BLCKSZ)
-		xbuffers = XLOG_SEG_SIZE / XLOG_BLCKSZ;
+	if (xbuffers > (wal_segment_size / XLOG_BLCKSZ))
+		xbuffers = (wal_segment_size / XLOG_BLCKSZ);
 	if (xbuffers < 8)
 		xbuffers = 8;
 	return xbuffers;
@@ -5034,7 +5075,7 @@ BootStrapXLOG(void)
 	 * segment with logid=0 logseg=1. The very first WAL segment, 0/0, is not
 	 * used, so that we can use 0/0 to mean "before any valid WAL segment".
 	 */
-	checkPoint.redo = XLogSegSize + SizeOfXLogLongPHD;
+	checkPoint.redo = wal_segment_size + SizeOfXLogLongPHD;
 	checkPoint.ThisTimeLineID = ThisTimeLineID;
 	checkPoint.PrevTimeLineID = ThisTimeLineID;
 	checkPoint.fullPageWrites = fullPageWrites;
@@ -5065,10 +5106,10 @@ BootStrapXLOG(void)
 	page->xlp_magic = XLOG_PAGE_MAGIC;
 	page->xlp_info = XLP_LONG_HEADER;
 	page->xlp_tli = ThisTimeLineID;
-	page->xlp_pageaddr = XLogSegSize;
+	page->xlp_pageaddr = wal_segment_size;
 	longpage = (XLogLongPageHeader) page;
 	longpage->xlp_sysid = sysidentifier;
-	longpage->xlp_seg_size = XLogSegSize;
+	longpage->xlp_seg_size = wal_segment_size;
 	longpage->xlp_xlog_blcksz = XLOG_BLCKSZ;
 
 	/* Insert the initial checkpoint record */
@@ -5550,8 +5591,8 @@ exitArchiveRecovery(TimeLineID endTLI, XLogRecPtr endOfLog)
 	 * they are the same, but if the switch happens exactly at a segment
 	 * boundary, startLogSegNo will be endLogSegNo + 1.
 	 */
-	XLByteToPrevSeg(endOfLog, endLogSegNo);
-	XLByteToSeg(endOfLog, startLogSegNo);
+	XLByteToPrevSeg(endOfLog, endLogSegNo, wal_segment_size);
+	XLByteToSeg(endOfLog, startLogSegNo, wal_segment_size);
 
 	/*
 	 * Initialize the starting WAL segment for the new timeline. If the switch
@@ -5569,7 +5610,7 @@ exitArchiveRecovery(TimeLineID endTLI, XLogRecPtr endOfLog)
 		 * avoid emplacing a bogus file.
 		 */
 		XLogFileCopy(endLogSegNo, endTLI, endLogSegNo,
-					 endOfLog % XLOG_SEG_SIZE);
+					 XLogSegmentOffset(endOfLog, wal_segment_size));
 	}
 	else
 	{
@@ -5593,7 +5634,7 @@ exitArchiveRecovery(TimeLineID endTLI, XLogRecPtr endOfLog)
 	 * Let's just make real sure there are not .ready or .done flags posted
 	 * for the new segment.
 	 */
-	XLogFileName(xlogfname, ThisTimeLineID, startLogSegNo);
+	XLogFileName(xlogfname, ThisTimeLineID, startLogSegNo, wal_segment_size);
 	XLogArchiveCleanup(xlogfname);
 
 	/*
@@ -6390,7 +6431,7 @@ StartupXLOG(void)
 
 	/* Set up XLOG reader facility */
 	MemSet(&private, 0, sizeof(XLogPageReadPrivate));
-	xlogreader = XLogReaderAllocate(&XLogPageRead, &private);
+	xlogreader = XLogReaderAllocate(wal_segment_size, &XLogPageRead, &private);
 	if (!xlogreader)
 		ereport(ERROR,
 				(errcode(ERRCODE_OUT_OF_MEMORY),
@@ -7523,7 +7564,7 @@ StartupXLOG(void)
 		XLogRecPtr	pageBeginPtr;
 
 		pageBeginPtr = EndOfLog - (EndOfLog % XLOG_BLCKSZ);
-		Assert(readOff == pageBeginPtr % XLogSegSize);
+		Assert(readOff == XLogSegmentOffset(pageBeginPtr, wal_segment_size));
 
 		firstIdx = XLogRecPtrToBufIdx(EndOfLog);
 
@@ -7672,13 +7713,14 @@ StartupXLOG(void)
 		 * restored from the archive to begin with, it's expected to have a
 		 * .done file).
 		 */
-		if (EndOfLog % XLOG_SEG_SIZE != 0 && XLogArchivingActive())
+		if (XLogSegmentOffset(EndOfLog, wal_segment_size) != 0 &&
+			XLogArchivingActive())
 		{
 			char		origfname[MAXFNAMELEN];
 			XLogSegNo	endLogSegNo;
 
-			XLByteToPrevSeg(EndOfLog, endLogSegNo);
-			XLogFileName(origfname, EndOfLogTLI, endLogSegNo);
+			XLByteToPrevSeg(EndOfLog, endLogSegNo, wal_segment_size);
+			XLogFileName(origfname, EndOfLogTLI, endLogSegNo, wal_segment_size);
 
 			if (!XLogArchiveIsReadyOrDone(origfname))
 			{
@@ -7686,7 +7728,7 @@ StartupXLOG(void)
 				char		partialfname[MAXFNAMELEN];
 				char		partialpath[MAXPGPATH];
 
-				XLogFilePath(origpath, EndOfLogTLI, endLogSegNo);
+				XLogFilePath(origpath, EndOfLogTLI, endLogSegNo, wal_segment_size);
 				snprintf(partialfname, MAXFNAMELEN, "%s.partial", origfname);
 				snprintf(partialpath, MAXPGPATH, "%s.partial", origpath);
 
@@ -8192,6 +8234,9 @@ InitXLOGAccess(void)
 	ThisTimeLineID = XLogCtl->ThisTimeLineID;
 	Assert(ThisTimeLineID != 0 || IsBootstrapProcessingMode());
 
+	/* set wal_segment_size */
+	wal_segment_size = ControlFile->xlog_seg_size;
+
 	/* Use GetRedoRecPtr to copy the RedoRecPtr safely */
 	(void) GetRedoRecPtr();
 	/* Also update our copy of doPageWrites. */
@@ -8522,7 +8567,7 @@ UpdateCheckPointDistanceEstimate(uint64 nbytes)
 	 * more.
 	 *
 	 * When checkpoints are triggered by max_wal_size, this should converge to
-	 * CheckpointSegments * XLOG_SEG_SIZE,
+	 * CheckpointSegments * wal_segment_size,
 	 *
 	 * Note: This doesn't pay any attention to what caused the checkpoint.
 	 * Checkpoints triggered manually with CHECKPOINT command, or by e.g.
@@ -8721,7 +8766,7 @@ CreateCheckPoint(int flags)
 	freespace = INSERT_FREESPACE(curInsert);
 	if (freespace == 0)
 	{
-		if (curInsert % XLogSegSize == 0)
+		if (XLogSegmentOffset(curInsert, wal_segment_size) == 0)
 			curInsert += SizeOfXLogLongPHD;
 		else
 			curInsert += SizeOfXLogShortPHD;
@@ -8955,7 +9000,7 @@ CreateCheckPoint(int flags)
 		/* Update the average distance between checkpoints. */
 		UpdateCheckPointDistanceEstimate(RedoRecPtr - PriorRedoPtr);
 
-		XLByteToSeg(PriorRedoPtr, _logSegNo);
+		XLByteToSeg(PriorRedoPtr, _logSegNo, wal_segment_size);
 		KeepLogSeg(recptr, &_logSegNo);
 		_logSegNo--;
 		RemoveOldXlogFiles(_logSegNo, PriorRedoPtr, recptr);
@@ -9283,7 +9328,7 @@ CreateRestartPoint(int flags)
 		/* Update the average distance between checkpoints/restartpoints. */
 		UpdateCheckPointDistanceEstimate(RedoRecPtr - PriorRedoPtr);
 
-		XLByteToSeg(PriorRedoPtr, _logSegNo);
+		XLByteToSeg(PriorRedoPtr, _logSegNo, wal_segment_size);
 
 		/*
 		 * Get the current end of xlog replayed or received, whichever is
@@ -9378,7 +9423,7 @@ KeepLogSeg(XLogRecPtr recptr, XLogSegNo *logSegNo)
 	XLogSegNo	segno;
 	XLogRecPtr	keep;
 
-	XLByteToSeg(recptr, segno);
+	XLByteToSeg(recptr, segno, wal_segment_size);
 	keep = XLogGetReplicationSlotMinimumLSN();
 
 	/* compute limit for wal_keep_segments first */
@@ -9396,7 +9441,7 @@ KeepLogSeg(XLogRecPtr recptr, XLogSegNo *logSegNo)
 	{
 		XLogSegNo	slotSegNo;
 
-		XLByteToSeg(keep, slotSegNo);
+		XLByteToSeg(keep, slotSegNo, wal_segment_size);
 
 		if (slotSegNo <= 0)
 			segno = 1;
@@ -10179,7 +10224,7 @@ XLogFileNameP(TimeLineID tli, XLogSegNo segno)
 {
 	char	   *result = palloc(MAXFNAMELEN);
 
-	XLogFileName(result, tli, segno);
+	XLogFileName(result, tli, segno, wal_segment_size);
 	return result;
 }
 
@@ -10433,8 +10478,8 @@ do_pg_start_backup(const char *backupidstr, bool fast, TimeLineID *starttli_p,
 			WALInsertLockRelease();
 		} while (!gotUniqueStartpoint);
 
-		XLByteToSeg(startpoint, _logSegNo);
-		XLogFileName(xlogfilename, starttli, _logSegNo);
+		XLByteToSeg(startpoint, _logSegNo, wal_segment_size);
+		XLogFileName(xlogfilename, starttli, _logSegNo, wal_segment_size);
 
 		/*
 		 * Construct tablespace_map file
@@ -10985,8 +11030,8 @@ do_pg_stop_backup(char *labelfile, bool waitforarchive, TimeLineID *stoptli_p)
 		 */
 		RequestXLogSwitch(false);
 
-		XLByteToPrevSeg(stoppoint, _logSegNo);
-		XLogFileName(stopxlogfilename, stoptli, _logSegNo);
+		XLByteToPrevSeg(stoppoint, _logSegNo, wal_segment_size);
+		XLogFileName(stopxlogfilename, stoptli, _logSegNo, wal_segment_size);
 
 		/* Use the log timezone here, not the session timezone */
 		stamp_time = (pg_time_t) time(NULL);
@@ -10997,9 +11042,9 @@ do_pg_stop_backup(char *labelfile, bool waitforarchive, TimeLineID *stoptli_p)
 		/*
 		 * Write the backup history file
 		 */
-		XLByteToSeg(startpoint, _logSegNo);
+		XLByteToSeg(startpoint, _logSegNo, wal_segment_size);
 		BackupHistoryFilePath(histfilepath, stoptli, _logSegNo,
-							  (uint32) (startpoint % XLogSegSize));
+							  startpoint, wal_segment_size);
 		fp = AllocateFile(histfilepath, "w");
 		if (!fp)
 			ereport(ERROR,
@@ -11053,12 +11098,12 @@ do_pg_stop_backup(char *labelfile, bool waitforarchive, TimeLineID *stoptli_p)
 		((!backup_started_in_recovery && XLogArchivingActive()) ||
 		 (backup_started_in_recovery && XLogArchivingAlways())))
 	{
-		XLByteToPrevSeg(stoppoint, _logSegNo);
-		XLogFileName(lastxlogfilename, stoptli, _logSegNo);
+		XLByteToPrevSeg(stoppoint, _logSegNo, wal_segment_size);
+		XLogFileName(lastxlogfilename, stoptli, _logSegNo, wal_segment_size);
 
-		XLByteToSeg(startpoint, _logSegNo);
+		XLByteToSeg(startpoint, _logSegNo, wal_segment_size);
 		BackupHistoryFileName(histfilename, stoptli, _logSegNo,
-							  (uint32) (startpoint % XLogSegSize));
+							  startpoint, wal_segment_size);
 
 		seconds_before_warning = 60;
 		waits = 0;
@@ -11501,14 +11546,15 @@ XLogPageRead(XLogReaderState *xlogreader, XLogRecPtr targetPagePtr, int reqLen,
 	uint32		targetPageOff;
 	XLogSegNo	targetSegNo PG_USED_FOR_ASSERTS_ONLY;
 
-	XLByteToSeg(targetPagePtr, targetSegNo);
-	targetPageOff = targetPagePtr % XLogSegSize;
+	XLByteToSeg(targetPagePtr, targetSegNo, wal_segment_size);
+	targetPageOff = XLogSegmentOffset(targetPagePtr, wal_segment_size);
 
 	/*
 	 * See if we need to switch to a new segment because the requested record
 	 * is not in the currently open one.
 	 */
-	if (readFile >= 0 && !XLByteInSeg(targetPagePtr, readSegNo))
+	if (readFile >= 0 &&
+		!XLByteInSeg(targetPagePtr, readSegNo, wal_segment_size))
 	{
 		/*
 		 * Request a restartpoint if we've replayed too much xlog since the
@@ -11529,7 +11575,7 @@ XLogPageRead(XLogReaderState *xlogreader, XLogRecPtr targetPagePtr, int reqLen,
 		readSource = 0;
 	}
 
-	XLByteToSeg(targetPagePtr, readSegNo);
+	XLByteToSeg(targetPagePtr, readSegNo, wal_segment_size);
 
 retry:
 	/* See if we need to retrieve more data */
@@ -11569,7 +11615,8 @@ retry:
 		if (((targetPagePtr) / XLOG_BLCKSZ) != (receivedUpto / XLOG_BLCKSZ))
 			readLen = XLOG_BLCKSZ;
 		else
-			readLen = receivedUpto % XLogSegSize - targetPageOff;
+			readLen = XLogSegmentOffset(receivedUpto, wal_segment_size) -
+				targetPageOff;
 	}
 	else
 		readLen = XLOG_BLCKSZ;
@@ -11580,7 +11627,7 @@ retry:
 	{
 		char		fname[MAXFNAMELEN];
 
-		XLogFileName(fname, curFileTLI, readSegNo);
+		XLogFileName(fname, curFileTLI, readSegNo, wal_segment_size);
 		ereport(emode_for_corrupt_record(emode, targetPagePtr + reqLen),
 				(errcode_for_file_access(),
 				 errmsg("could not seek in log segment %s to offset %u: %m",
@@ -11594,7 +11641,7 @@ retry:
 		char		fname[MAXFNAMELEN];
 
 		pgstat_report_wait_end();
-		XLogFileName(fname, curFileTLI, readSegNo);
+		XLogFileName(fname, curFileTLI, readSegNo, wal_segment_size);
 		ereport(emode_for_corrupt_record(emode, targetPagePtr + reqLen),
 				(errcode_for_file_access(),
 				 errmsg("could not read from log segment %s, offset %u: %m",