database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-11-10 17:42:29 +03:00
Code Activity

During WAL recovery, when reading a page that we intend to overwrite completely

Browse Source 
from the WAL data, don't bother to physically read it; just have bufmgr.c
return a zeroed-out buffer instead.  This speeds recovery significantly,
and also avoids unnecessary failures when a page-to-be-overwritten has corrupt
page headers on disk.  This replaces a former kluge that accomplished the
latter by pretending zero_damaged_pages was always ON during WAL recovery;
which was OK when the kluge was put in, but is unsafe when restoring a WAL
log that was written with full_page_writes off.

Heikki Linnakangas
This commit is contained in:
Tom Lane
2007-05-02 23:18:03 +00:00
parent 8ec943856a
commit 8c3cc86e7b
3 changed files with 49 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

48
src/backend/storage/buffer/bufmgr.c
View File

@@ -8,7 +8,7 @@
*
*
* IDENTIFICATION
* $PostgreSQL: pgsql/src/backend/storage/buffer/bufmgr.c,v 1.216 2007/03/30 18:34:55 mha Exp $
* $PostgreSQL: pgsql/src/backend/storage/buffer/bufmgr.c,v 1.217 2007/05/02 23:18:03 tgl Exp $
*
*-------------------------------------------------------------------------
*/
@@ -17,6 +17,12 @@
* and pin it so that no one can destroy it while this process
* is using it.
*
* ReadOrZeroBuffer() -- like ReadBuffer, but if the page is not already in
* cache we don't read it, but just return a zeroed-out buffer. Useful
* when the caller intends to fill the page from scratch, since this
* saves I/O and avoids unnecessary failure if the page-on-disk has
* corrupt page headers.
*
* ReleaseBuffer() -- unpin a buffer
*
* MarkBufferDirty() -- mark a pinned buffer's contents as "dirty".
@@ -87,6 +93,8 @@ static volatile BufferDesc *PinCountWaitBuf = NULL;
extern PgStat_MsgBgWriter BgWriterStats;
static Buffer ReadBuffer_common(Relation reln, BlockNumber blockNum,
bool zeroPage);
static bool PinBuffer(volatile BufferDesc *buf);
static void PinBuffer_Locked(volatile BufferDesc *buf);
static void UnpinBuffer(volatile BufferDesc *buf,
@@ -120,6 +128,27 @@ static void AtProcExit_Buffers(int code, Datum arg);
*/
Buffer
ReadBuffer(Relation reln, BlockNumber blockNum)
{
return ReadBuffer_common(reln, blockNum, false);
}
/*
* ReadOrZeroBuffer -- like ReadBuffer, but if the page isn't in buffer
* cache already, it's filled with zeros instead of reading it from
* disk. The caller is expected to overwrite the whole buffer,
* so that the current page contents are not interesting.
*/
Buffer
ReadOrZeroBuffer(Relation reln, BlockNumber blockNum)
{
return ReadBuffer_common(reln, blockNum, true);
}
/*
* ReadBuffer_common -- common logic for ReadBuffer and ReadOrZeroBuffer
*/
static Buffer
ReadBuffer_common(Relation reln, BlockNumber blockNum, bool zeroPage)
{
volatile BufferDesc *bufHdr;
Block bufBlock;
@@ -253,17 +282,18 @@ ReadBuffer(Relation reln, BlockNumber blockNum)
}
else
{
smgrread(reln->rd_smgr, blockNum, (char *) bufBlock);
/*
* Read in the page, unless the caller intends to overwrite it
* and just wants us to allocate a buffer.
*/
if (zeroPage)
MemSet((char *) bufBlock, 0, BLCKSZ);
else
smgrread(reln->rd_smgr, blockNum, (char *) bufBlock);
/* check for garbage data */
if (!PageHeaderIsValid((PageHeader) bufBlock))
{
/*
* During WAL recovery, the first access to any data page should
* overwrite the whole page from the WAL; so a clobbered page
* header is not reason to fail. Hence, when InRecovery we may
* always act as though zero_damaged_pages is ON.
*/
if (zero_damaged_pages || InRecovery)
if (zero_damaged_pages)
{
ereport(WARNING,
(errcode(ERRCODE_DATA_CORRUPTED),