#include "postgres.h"
#include "access/xlog.h"
#include "libpq/pqsignal.h"
#include "miscadmin.h"
#include "pgstat.h"
#include "postmaster/auxprocess.h"
#include "postmaster/bgwriter.h"
#include "postmaster/interrupt.h"
#include "storage/aio_subsys.h"
#include "storage/buf_internals.h"
#include "storage/bufmgr.h"
#include "storage/condition_variable.h"
#include "storage/fd.h"
#include "storage/lwlock.h"
#include "storage/proc.h"
#include "storage/procsignal.h"
#include "storage/smgr.h"
#include "storage/standby.h"
#include "utils/memutils.h"
#include "utils/resowner.h"
#include "utils/timestamp.h"

Include dependency graph for bgwriter.c:

Macros
#define	HIBERNATE_FACTOR 50

#define	LOG_SNAPSHOT_INTERVAL_MS 15000

Functions
void	BackgroundWriterMain (const void *startup_data, size_t startup_data_len)

Variables
int	BgWriterDelay = 200

static TimestampTz	last_snapshot_ts

static XLogRecPtr	last_snapshot_lsn = InvalidXLogRecPtr

Macro Definition Documentation

◆ HIBERNATE_FACTOR

#define HIBERNATE_FACTOR 50

Definition at line 64 of file bgwriter.c.

◆ LOG_SNAPSHOT_INTERVAL_MS

#define LOG_SNAPSHOT_INTERVAL_MS 15000

Definition at line 70 of file bgwriter.c.

Function Documentation

◆ BackgroundWriterMain()

void BackgroundWriterMain	(	const void *	startup_data,
		size_t	startup_data_len
	)

Definition at line 88 of file bgwriter.c.

{
    sigjmp_buf  local_sigjmp_buf;
    MemoryContext bgwriter_context;
    bool        prev_hibernate;
    WritebackContext wb_context;
 
    Assert(startup_data_len == 0);
 
    AuxiliaryProcessMainCommon();
 
    /*
     * Properly accept or ignore signals that might be sent to us.
     */
    pqsignal(SIGHUP, SignalHandlerForConfigReload);
    pqsignal(SIGINT, SIG_IGN);
    pqsignal(SIGTERM, SignalHandlerForShutdownRequest);
    /* SIGQUIT handler was already set up by InitPostmasterChild */
    pqsignal(SIGALRM, SIG_IGN);
    pqsignal(SIGPIPE, SIG_IGN);
    pqsignal(SIGUSR1, procsignal_sigusr1_handler);
    pqsignal(SIGUSR2, SIG_IGN);
 
    /*
     * Reset some signals that are accepted by postmaster but not here
     */
    pqsignal(SIGCHLD, SIG_DFL);
 
    /*
     * We just started, assume there has been either a shutdown or
     * end-of-recovery snapshot.
     */
    last_snapshot_ts = GetCurrentTimestamp();
 
    /*
     * Create a memory context that we will do all our work in.  We do this so
     * that we can reset the context during error recovery and thereby avoid
     * possible memory leaks.  Formerly this code just ran in
     * TopMemoryContext, but resetting that would be a really bad idea.
     */
    bgwriter_context = AllocSetContextCreate(TopMemoryContext,
                                             "Background Writer",
                                             ALLOCSET_DEFAULT_SIZES);
    MemoryContextSwitchTo(bgwriter_context);
 
    WritebackContextInit(&wb_context, &bgwriter_flush_after);
 
    /*
     * If an exception is encountered, processing resumes here.
     *
     * You might wonder why this isn't coded as an infinite loop around a
     * PG_TRY construct.  The reason is that this is the bottom of the
     * exception stack, and so with PG_TRY there would be no exception handler
     * in force at all during the CATCH part.  By leaving the outermost setjmp
     * always active, we have at least some chance of recovering from an error
     * during error recovery.  (If we get into an infinite loop thereby, it
     * will soon be stopped by overflow of elog.c's internal state stack.)
     *
     * Note that we use sigsetjmp(..., 1), so that the prevailing signal mask
     * (to wit, BlockSig) will be restored when longjmp'ing to here.  Thus,
     * signals other than SIGQUIT will be blocked until we complete error
     * recovery.  It might seem that this policy makes the HOLD_INTERRUPTS()
     * call redundant, but it is not since InterruptPending might be set
     * already.
     */
    if (sigsetjmp(local_sigjmp_buf, 1) != 0)
    {
        /* Since not using PG_TRY, must reset error stack by hand */
        error_context_stack = NULL;
 
        /* Prevent interrupts while cleaning up */
        HOLD_INTERRUPTS();
 
        /* Report the error to the server log */
        EmitErrorReport();
 
        /*
         * These operations are really just a minimal subset of
         * AbortTransaction().  We don't have very many resources to worry
         * about in bgwriter, but we do have LWLocks, buffers, and temp files.
         */
        LWLockReleaseAll();
        ConditionVariableCancelSleep();
        pgaio_error_cleanup();
        UnlockBuffers();
        ReleaseAuxProcessResources(false);
        AtEOXact_Buffers(false);
        AtEOXact_SMgr();
        AtEOXact_Files(false);
        AtEOXact_HashTables(false);
 
        /*
         * Now return to normal top-level context and clear ErrorContext for
         * next time.
         */
        MemoryContextSwitchTo(bgwriter_context);
        FlushErrorState();
 
        /* Flush any leaked data in the top-level context */
        MemoryContextReset(bgwriter_context);
 
        /* re-initialize to avoid repeated errors causing problems */
        WritebackContextInit(&wb_context, &bgwriter_flush_after);
 
        /* Now we can allow interrupts again */
        RESUME_INTERRUPTS();
 
        /*
         * Sleep at least 1 second after any error.  A write error is likely
         * to be repeated, and we don't want to be filling the error logs as
         * fast as we can.
         */
        pg_usleep(1000000L);
 
        /* Report wait end here, when there is no further possibility of wait */
        pgstat_report_wait_end();
    }
 
    /* We can now handle ereport(ERROR) */
    PG_exception_stack = &local_sigjmp_buf;
 
    /*
     * Unblock signals (they were blocked when the postmaster forked us)
     */
    sigprocmask(SIG_SETMASK, &UnBlockSig, NULL);
 
    /*
     * Reset hibernation state after any error.
     */
    prev_hibernate = false;
 
    /*
     * Loop forever
     */
    for (;;)
    {
        bool        can_hibernate;
        int         rc;
 
        /* Clear any already-pending wakeups */
        ResetLatch(MyLatch);
 
        ProcessMainLoopInterrupts();
 
        /*
         * Do one cycle of dirty-buffer writing.
         */
        can_hibernate = BgBufferSync(&wb_context);
 
        /* Report pending statistics to the cumulative stats system */
        pgstat_report_bgwriter();
        pgstat_report_wal(true);
 
        if (FirstCallSinceLastCheckpoint())
        {
            /*
             * After any checkpoint, free all smgr objects.  Otherwise we
             * would never do so for dropped relations, as the bgwriter does
             * not process shared invalidation messages or call
             * AtEOXact_SMgr().
             */
            smgrdestroyall();
        }
 
        /*
         * Log a new xl_running_xacts every now and then so replication can
         * get into a consistent state faster (think of suboverflowed
         * snapshots) and clean up resources (locks, KnownXids*) more
         * frequently. The costs of this are relatively low, so doing it 4
         * times (LOG_SNAPSHOT_INTERVAL_MS) a minute seems fine.
         *
         * We assume the interval for writing xl_running_xacts is
         * significantly bigger than BgWriterDelay, so we don't complicate the
         * overall timeout handling but just assume we're going to get called
         * often enough even if hibernation mode is active. It's not that
         * important that LOG_SNAPSHOT_INTERVAL_MS is met strictly. To make
         * sure we're not waking the disk up unnecessarily on an idle system
         * we check whether there has been any WAL inserted since the last
         * time we've logged a running xacts.
         *
         * We do this logging in the bgwriter as it is the only process that
         * is run regularly and returns to its mainloop all the time. E.g.
         * Checkpointer, when active, is barely ever in its mainloop and thus
         * makes it hard to log regularly.
         */
        if (XLogStandbyInfoActive() && !RecoveryInProgress())
        {
            TimestampTz timeout = 0;
            TimestampTz now = GetCurrentTimestamp();
 
            timeout = TimestampTzPlusMilliseconds(last_snapshot_ts,
                                                  LOG_SNAPSHOT_INTERVAL_MS);
 
            /*
             * Only log if enough time has passed and interesting records have
             * been inserted since the last snapshot.  Have to compare with <=
             * instead of < because GetLastImportantRecPtr() points at the
             * start of a record, whereas last_snapshot_lsn points just past
             * the end of the record.
             */
            if (now >= timeout &&
                last_snapshot_lsn <= GetLastImportantRecPtr())
            {
                last_snapshot_lsn = LogStandbySnapshot();
                last_snapshot_ts = now;
            }
        }
 
        /*
         * Sleep until we are signaled or BgWriterDelay has elapsed.
         *
         * Note: the feedback control loop in BgBufferSync() expects that we
         * will call it every BgWriterDelay msec.  While it's not critical for
         * correctness that that be exact, the feedback loop might misbehave
         * if we stray too far from that.  Hence, avoid loading this process
         * down with latch events that are likely to happen frequently during
         * normal operation.
         */
        rc = WaitLatch(MyLatch,
                       WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
                       BgWriterDelay /* ms */ , WAIT_EVENT_BGWRITER_MAIN);
 
        /*
         * If no latch event and BgBufferSync says nothing's happening, extend
         * the sleep in "hibernation" mode, where we sleep for much longer
         * than bgwriter_delay says.  Fewer wakeups save electricity.  When a
         * backend starts using buffers again, it will wake us up by setting
         * our latch.  Because the extra sleep will persist only as long as no
         * buffer allocations happen, this should not distort the behavior of
         * BgBufferSync's control loop too badly; essentially, it will think
         * that the system-wide idle interval didn't exist.
         *
         * There is a race condition here, in that a backend might allocate a
         * buffer between the time BgBufferSync saw the alloc count as zero
         * and the time we call StrategyNotifyBgWriter.  While it's not
         * critical that we not hibernate anyway, we try to reduce the odds of
         * that by only hibernating when BgBufferSync says nothing's happening
         * for two consecutive cycles.  Also, we mitigate any possible
         * consequences of a missed wakeup by not hibernating forever.
         */
        if (rc == WL_TIMEOUT && can_hibernate && prev_hibernate)
        {
            /* Ask for notification at next buffer allocation */
            StrategyNotifyBgWriter(MyProcNumber);
            /* Sleep ... */
            (void) WaitLatch(MyLatch,
                             WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
                             BgWriterDelay * HIBERNATE_FACTOR,
                             WAIT_EVENT_BGWRITER_HIBERNATE);
            /* Reset the notification request in case we timed out */
            StrategyNotifyBgWriter(-1);
        }
 
        prev_hibernate = can_hibernate;
    }
}