/*----------------
* BitmapShouldInitializeSharedState
*
* The first process to come here and see the state to the BM_INITIAL
* will become the leader for the parallel bitmap scan and will be
* responsible for populating the TIDBitmap. The other processes will
* be blocked by the condition variable until the leader wakes them up.
* ---------------
*/
static bool
BitmapShouldInitializeSharedState(ParallelBitmapHeapState *pstate)
{
SharedBitmapState state;
while (1)
{
SpinLockAcquire(&pstate->mutex);
state = pstate->state;
if (pstate->state == BM_INITIAL)
pstate->state = BM_INPROGRESS;
SpinLockRelease(&pstate->mutex);
/* Exit if bitmap is done, or if we're the leader. */
if (state != BM_INPROGRESS)
break;
/* Wait for the leader to wake us up. */
ConditionVariableSleep(&pstate->cv, WAIT_EVENT_PARALLEL_BITMAP_SCAN);
}
ConditionVariableCancelSleep();
return (state == BM_INITIAL);
}
/*
* _bt_parallel_seize() -- Begin the process of advancing the scan to a new
* page. Other scans must wait until we call bt_parallel_release() or
* bt_parallel_done().
*
* The return value is true if we successfully seized the scan and false
* if we did not. The latter case occurs if no pages remain for the current
* set of scankeys.
*
* If the return value is true, *pageno returns the next or current page
* of the scan (depending on the scan direction). An invalid block number
* means the scan hasn't yet started, and P_NONE means we've reached the end.
* The first time a participating process reaches the last page, it will return
* true and set *pageno to P_NONE; after that, further attempts to seize the
* scan will return false.
*
* Callers should ignore the value of pageno if the return value is false.
*/
bool
_bt_parallel_seize(IndexScanDesc scan, BlockNumber *pageno)
{
BTScanOpaque so = (BTScanOpaque) scan->opaque;
BTPS_State pageStatus;
bool exit_loop = false;
bool status = true;
ParallelIndexScanDesc parallel_scan = scan->parallel_scan;
BTParallelScanDesc btscan;
*pageno = P_NONE;
btscan = (BTParallelScanDesc) OffsetToPointer((void *) parallel_scan,
parallel_scan->ps_offset);
while (1)
{
SpinLockAcquire(&btscan->btps_mutex);
pageStatus = btscan->btps_pageStatus;
if (so->arrayKeyCount < btscan->btps_arrayKeyCount)
{
/* Parallel scan has already advanced to a new set of scankeys. */
status = false;
}
else if (pageStatus == BTPARALLEL_DONE)
{
/*
* We're done with this set of scankeys. This may be the end, or
* there could be more sets to try.
*/
status = false;
}
else if (pageStatus != BTPARALLEL_ADVANCING)
{
/*
* We have successfully seized control of the scan for the purpose
* of advancing it to a new page!
*/
btscan->btps_pageStatus = BTPARALLEL_ADVANCING;
*pageno = btscan->btps_scanPage;
exit_loop = true;
}
SpinLockRelease(&btscan->btps_mutex);
if (exit_loop || !status)
break;
ConditionVariableSleep(&btscan->btps_cv, WAIT_EVENT_BTREE_PAGE);
}
ConditionVariableCancelSleep();
return status;
}
/*
* Prepare to wait on a given condition variable.
*
* This can optionally be called before entering a test/sleep loop.
* Doing so is more efficient if we'll need to sleep at least once.
* However, if the first test of the exit condition is likely to succeed,
* it's more efficient to omit the ConditionVariablePrepareToSleep call.
* See comments in ConditionVariableSleep for more detail.
*
* Caution: "before entering the loop" means you *must* test the exit
* condition between calling ConditionVariablePrepareToSleep and calling
* ConditionVariableSleep. If that is inconvenient, omit calling
* ConditionVariablePrepareToSleep.
*/
void
ConditionVariablePrepareToSleep(ConditionVariable *cv)
{
int pgprocno = MyProc->pgprocno;
/*
* If first time through in this process, create a WaitEventSet, which
* we'll reuse for all condition variable sleeps.
*/
if (cv_wait_event_set == NULL)
{
WaitEventSet *new_event_set;
new_event_set = CreateWaitEventSet(TopMemoryContext, 2);
AddWaitEventToSet(new_event_set, WL_LATCH_SET, PGINVALID_SOCKET,
MyLatch, NULL);
AddWaitEventToSet(new_event_set, WL_EXIT_ON_PM_DEATH, PGINVALID_SOCKET,
NULL, NULL);
/* Don't set cv_wait_event_set until we have a correct WES. */
cv_wait_event_set = new_event_set;
}
/*
* If some other sleep is already prepared, cancel it; this is necessary
* because we have just one static variable tracking the prepared sleep,
* and also only one cvWaitLink in our PGPROC. It's okay to do this
* because whenever control does return to the other test-and-sleep loop,
* its ConditionVariableSleep call will just re-establish that sleep as
* the prepared one.
*/
if (cv_sleep_target != NULL)
ConditionVariableCancelSleep();
/* Record the condition variable on which we will sleep. */
cv_sleep_target = cv;
/*
* Reset my latch before adding myself to the queue, to ensure that we
* don't miss a wakeup that occurs immediately.
*/
ResetLatch(MyLatch);
/* Add myself to the wait queue. */
SpinLockAcquire(&cv->mutex);
proclist_push_tail(&cv->wakeup, pgprocno, cvWaitLink);
SpinLockRelease(&cv->mutex);
}
/*
* Main entry point for checkpointer process
*
* This is invoked from AuxiliaryProcessMain, which has already created the
* basic execution environment, but not enabled signals yet.
*/
void
CheckpointerMain(void)
{
sigjmp_buf local_sigjmp_buf;
MemoryContext checkpointer_context;
CheckpointerShmem->checkpointer_pid = MyProcPid;
/*
* Properly accept or ignore signals the postmaster might send us
*
* Note: we deliberately ignore SIGTERM, because during a standard Unix
* system shutdown cycle, init will SIGTERM all processes at once. We
* want to wait for the backends to exit, whereupon the postmaster will
* tell us it's okay to shut down (via SIGUSR2).
*/
pqsignal(SIGHUP, ChkptSigHupHandler); /* set flag to read config
* file */
pqsignal(SIGINT, ReqCheckpointHandler); /* request checkpoint */
pqsignal(SIGTERM, SIG_IGN); /* ignore SIGTERM */
pqsignal(SIGQUIT, chkpt_quickdie); /* hard crash time */
pqsignal(SIGALRM, SIG_IGN);
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, chkpt_sigusr1_handler);
pqsignal(SIGUSR2, ReqShutdownHandler); /* request shutdown */
/*
* Reset some signals that are accepted by postmaster but not here
*/
pqsignal(SIGCHLD, SIG_DFL);
pqsignal(SIGTTIN, SIG_DFL);
pqsignal(SIGTTOU, SIG_DFL);
pqsignal(SIGCONT, SIG_DFL);
pqsignal(SIGWINCH, SIG_DFL);
/* We allow SIGQUIT (quickdie) at all times */
sigdelset(&BlockSig, SIGQUIT);
/*
* Initialize so that first time-driven event happens at the correct time.
*/
last_checkpoint_time = last_xlog_switch_time = (pg_time_t) time(NULL);
/*
* Create a resource owner to keep track of our resources (currently only
* buffer pins).
*/
CurrentResourceOwner = ResourceOwnerCreate(NULL, "Checkpointer");
/*
* 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.
*/
checkpointer_context = AllocSetContextCreate(TopMemoryContext,
"Checkpointer",
ALLOCSET_DEFAULT_SIZES);
MemoryContextSwitchTo(checkpointer_context);
/*
* If an exception is encountered, processing resumes here.
*
* See notes in postgres.c about the design of this coding.
*/
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 checkpointer, but we do have LWLocks, buffers, and temp
* files.
*/
LWLockReleaseAll();
ConditionVariableCancelSleep();
pgstat_report_wait_end();
AbortBufferIO();
UnlockBuffers();
/* buffer pins are released here: */
ResourceOwnerRelease(CurrentResourceOwner,
RESOURCE_RELEASE_BEFORE_LOCKS,
false, true);
/* we needn't bother with the other ResourceOwnerRelease phases */
AtEOXact_Buffers(false);
AtEOXact_SMgr();
AtEOXact_Files();
AtEOXact_HashTables(false);
/* Warn any waiting backends that the checkpoint failed. */
if (ckpt_active)
{
SpinLockAcquire(&CheckpointerShmem->ckpt_lck);
CheckpointerShmem->ckpt_failed++;
CheckpointerShmem->ckpt_done = CheckpointerShmem->ckpt_started;
SpinLockRelease(&CheckpointerShmem->ckpt_lck);
ckpt_active = false;
}
/*
* Now return to normal top-level context and clear ErrorContext for
* next time.
*/
MemoryContextSwitchTo(checkpointer_context);
FlushErrorState();
/* Flush any leaked data in the top-level context */
MemoryContextResetAndDeleteChildren(checkpointer_context);
/* 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);
/*
* Close all open files after any error. This is helpful on Windows,
* where holding deleted files open causes various strange errors.
* It's not clear we need it elsewhere, but shouldn't hurt.
*/
smgrcloseall();
}
/* We can now handle ereport(ERROR) */
PG_exception_stack = &local_sigjmp_buf;
/*
* Unblock signals (they were blocked when the postmaster forked us)
*/
PG_SETMASK(&UnBlockSig);
/*
* Ensure all shared memory values are set correctly for the config. Doing
* this here ensures no race conditions from other concurrent updaters.
*/
UpdateSharedMemoryConfig();
/*
* Advertise our latch that backends can use to wake us up while we're
* sleeping.
*/
ProcGlobal->checkpointerLatch = &MyProc->procLatch;
/*
* Loop forever
*/
for (;;)
{
bool do_checkpoint = false;
int flags = 0;
pg_time_t now;
int elapsed_secs;
int cur_timeout;
int rc;
/* Clear any already-pending wakeups */
ResetLatch(MyLatch);
/*
* Process any requests or signals received recently.
*/
AbsorbFsyncRequests();
if (got_SIGHUP)
{
got_SIGHUP = false;
ProcessConfigFile(PGC_SIGHUP);
/*
* Checkpointer is the last process to shut down, so we ask it to
* hold the keys for a range of other tasks required most of which
* have nothing to do with checkpointing at all.
*
* For various reasons, some config values can change dynamically
* so the primary copy of them is held in shared memory to make
* sure all backends see the same value. We make Checkpointer
* responsible for updating the shared memory copy if the
* parameter setting changes because of SIGHUP.
*/
UpdateSharedMemoryConfig();
}
if (checkpoint_requested)
{
checkpoint_requested = false;
do_checkpoint = true;
BgWriterStats.m_requested_checkpoints++;
}
if (shutdown_requested)
{
/*
* From here on, elog(ERROR) should end with exit(1), not send
* control back to the sigsetjmp block above
*/
ExitOnAnyError = true;
/* Close down the database */
ShutdownXLOG(0, 0);
/* Normal exit from the checkpointer is here */
proc_exit(0); /* done */
}
/*
* Force a checkpoint if too much time has elapsed since the last one.
* Note that we count a timed checkpoint in stats only when this
* occurs without an external request, but we set the CAUSE_TIME flag
* bit even if there is also an external request.
*/
now = (pg_time_t) time(NULL);
elapsed_secs = now - last_checkpoint_time;
if (elapsed_secs >= CheckPointTimeout)
{
if (!do_checkpoint)
BgWriterStats.m_timed_checkpoints++;
do_checkpoint = true;
flags |= CHECKPOINT_CAUSE_TIME;
}
/*
* Do a checkpoint if requested.
*/
if (do_checkpoint)
{
bool ckpt_performed = false;
bool do_restartpoint;
/*
* Check if we should perform a checkpoint or a restartpoint. As a
* side-effect, RecoveryInProgress() initializes TimeLineID if
* it's not set yet.
*/
do_restartpoint = RecoveryInProgress();
/*
* Atomically fetch the request flags to figure out what kind of a
* checkpoint we should perform, and increase the started-counter
* to acknowledge that we've started a new checkpoint.
*/
SpinLockAcquire(&CheckpointerShmem->ckpt_lck);
flags |= CheckpointerShmem->ckpt_flags;
CheckpointerShmem->ckpt_flags = 0;
CheckpointerShmem->ckpt_started++;
SpinLockRelease(&CheckpointerShmem->ckpt_lck);
/*
* The end-of-recovery checkpoint is a real checkpoint that's
* performed while we're still in recovery.
*/
if (flags & CHECKPOINT_END_OF_RECOVERY)
do_restartpoint = false;
/*
* We will warn if (a) too soon since last checkpoint (whatever
* caused it) and (b) somebody set the CHECKPOINT_CAUSE_XLOG flag
* since the last checkpoint start. Note in particular that this
* implementation will not generate warnings caused by
* CheckPointTimeout < CheckPointWarning.
*/
if (!do_restartpoint &&
(flags & CHECKPOINT_CAUSE_XLOG) &&
elapsed_secs < CheckPointWarning)
ereport(LOG,
(errmsg_plural("checkpoints are occurring too frequently (%d second apart)",
"checkpoints are occurring too frequently (%d seconds apart)",
elapsed_secs,
elapsed_secs),
errhint("Consider increasing the configuration parameter \"max_wal_size\".")));
/*
* Initialize checkpointer-private variables used during
* checkpoint.
*/
ckpt_active = true;
if (do_restartpoint)
ckpt_start_recptr = GetXLogReplayRecPtr(NULL);
else
ckpt_start_recptr = GetInsertRecPtr();
ckpt_start_time = now;
ckpt_cached_elapsed = 0;
/*
* Do the checkpoint.
*/
if (!do_restartpoint)
{
CreateCheckPoint(flags);
ckpt_performed = true;
}
else
ckpt_performed = CreateRestartPoint(flags);
/*
* After any checkpoint, close all smgr files. This is so we
* won't hang onto smgr references to deleted files indefinitely.
*/
smgrcloseall();
/*
* Indicate checkpoint completion to any waiting backends.
*/
SpinLockAcquire(&CheckpointerShmem->ckpt_lck);
CheckpointerShmem->ckpt_done = CheckpointerShmem->ckpt_started;
SpinLockRelease(&CheckpointerShmem->ckpt_lck);
if (ckpt_performed)
{
/*
* Note we record the checkpoint start time not end time as
* last_checkpoint_time. This is so that time-driven
* checkpoints happen at a predictable spacing.
*/
last_checkpoint_time = now;
}
else
{
/*
* We were not able to perform the restartpoint (checkpoints
* throw an ERROR in case of error). Most likely because we
* have not received any new checkpoint WAL records since the
* last restartpoint. Try again in 15 s.
*/
last_checkpoint_time = now - CheckPointTimeout + 15;
}
ckpt_active = false;
}
/* Check for archive_timeout and switch xlog files if necessary. */
CheckArchiveTimeout();
/*
* Send off activity statistics to the stats collector. (The reason
* why we re-use bgwriter-related code for this is that the bgwriter
* and checkpointer used to be just one process. It's probably not
* worth the trouble to split the stats support into two independent
* stats message types.)
*/
pgstat_send_bgwriter();
/*
* Sleep until we are signaled or it's time for another checkpoint or
* xlog file switch.
*/
now = (pg_time_t) time(NULL);
elapsed_secs = now - last_checkpoint_time;
if (elapsed_secs >= CheckPointTimeout)
continue; /* no sleep for us ... */
cur_timeout = CheckPointTimeout - elapsed_secs;
if (XLogArchiveTimeout > 0 && !RecoveryInProgress())
{
elapsed_secs = now - last_xlog_switch_time;
if (elapsed_secs >= XLogArchiveTimeout)
continue; /* no sleep for us ... */
cur_timeout = Min(cur_timeout, XLogArchiveTimeout - elapsed_secs);
}
rc = WaitLatch(MyLatch,
WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
cur_timeout * 1000L /* convert to ms */,
WAIT_EVENT_CHECKPOINTER_MAIN);
/*
* Emergency bailout if postmaster has died. This is to avoid the
* necessity for manual cleanup of all postmaster children.
*/
if (rc & WL_POSTMASTER_DEATH)
exit(1);
}
}
/*
* Find a previously created slot and mark it as used by this backend.
*/
void
ReplicationSlotAcquire(const char *name, bool nowait)
{
ReplicationSlot *slot;
int active_pid;
int i;
retry:
Assert(MyReplicationSlot == NULL);
/*
* Search for the named slot and mark it active if we find it. If the
* slot is already active, we exit the loop with active_pid set to the PID
* of the backend that owns it.
*/
active_pid = 0;
slot = NULL;
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (i = 0; i < max_replication_slots; i++)
{
ReplicationSlot *s = &ReplicationSlotCtl->replication_slots[i];
if (s->in_use && strcmp(name, NameStr(s->data.name)) == 0)
{
/*
* This is the slot we want. We don't know yet if it's active, so
* get ready to sleep on it in case it is. (We may end up not
* sleeping, but we don't want to do this while holding the
* spinlock.)
*/
ConditionVariablePrepareToSleep(&s->active_cv);
SpinLockAcquire(&s->mutex);
active_pid = s->active_pid;
if (active_pid == 0)
active_pid = s->active_pid = MyProcPid;
SpinLockRelease(&s->mutex);
slot = s;
break;
}
}
LWLockRelease(ReplicationSlotControlLock);
/* If we did not find the slot, error out. */
if (slot == NULL)
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
errmsg("replication slot \"%s\" does not exist", name)));
/*
* If we found the slot but it's already active in another backend, we
* either error out or retry after a short wait, as caller specified.
*/
if (active_pid != MyProcPid)
{
if (nowait)
ereport(ERROR,
(errcode(ERRCODE_OBJECT_IN_USE),
errmsg("replication slot \"%s\" is active for PID %d",
name, active_pid)));
/* Wait here until we get signaled, and then restart */
ConditionVariableSleep(&slot->active_cv,
WAIT_EVENT_REPLICATION_SLOT_DROP);
ConditionVariableCancelSleep();
goto retry;
}
else
ConditionVariableCancelSleep(); /* no sleep needed after all */
/* Let everybody know we've modified this slot */
ConditionVariableBroadcast(&slot->active_cv);
/* We made this slot active, so it's ours now. */
MyReplicationSlot = slot;
}
/*
* Arrive at this barrier, wait for all other attached participants to arrive
* too and then return. Increments the current phase. The caller must be
* attached.
*
* While waiting, pg_stat_activity shows a wait_event_class and wait_event
* controlled by the wait_event_info passed in, which should be a value from
* one of the WaitEventXXX enums defined in pgstat.h.
*
* Return true in one arbitrarily chosen participant. Return false in all
* others. The return code can be used to elect one participant to execute a
* phase of work that must be done serially while other participants wait.
*/
bool
BarrierArriveAndWait(Barrier *barrier, uint32 wait_event_info)
{
bool release = false;
bool elected;
int start_phase;
int next_phase;
SpinLockAcquire(&barrier->mutex);
start_phase = barrier->phase;
next_phase = start_phase + 1;
++barrier->arrived;
if (barrier->arrived == barrier->participants)
{
release = true;
barrier->arrived = 0;
barrier->phase = next_phase;
barrier->elected = next_phase;
}
SpinLockRelease(&barrier->mutex);
/*
* If we were the last expected participant to arrive, we can release our
* peers and return true to indicate that this backend has been elected to
* perform any serial work.
*/
if (release)
{
ConditionVariableBroadcast(&barrier->condition_variable);
return true;
}
/*
* Otherwise we have to wait for the last participant to arrive and
* advance the phase.
*/
elected = false;
ConditionVariablePrepareToSleep(&barrier->condition_variable);
for (;;)
{
/*
* We know that phase must either be start_phase, indicating that we
* need to keep waiting, or next_phase, indicating that the last
* participant that we were waiting for has either arrived or detached
* so that the next phase has begun. The phase cannot advance any
* further than that without this backend's participation, because
* this backend is attached.
*/
SpinLockAcquire(&barrier->mutex);
Assert(barrier->phase == start_phase || barrier->phase == next_phase);
release = barrier->phase == next_phase;
if (release && barrier->elected != next_phase)
{
/*
* Usually the backend that arrives last and releases the other
* backends is elected to return true (see above), so that it can
* begin processing serial work while it has a CPU timeslice.
* However, if the barrier advanced because someone detached, then
* one of the backends that is awoken will need to be elected.
*/
barrier->elected = barrier->phase;
elected = true;
}
SpinLockRelease(&barrier->mutex);
if (release)
break;
ConditionVariableSleep(&barrier->condition_variable, wait_event_info);
}
ConditionVariableCancelSleep();
return elected;
}
/*
* Main entry point for bgwriter process
*
* This is invoked from AuxiliaryProcessMain, which has already created the
* basic execution environment, but not enabled signals yet.
*/
void
BackgroundWriterMain(void)
{
sigjmp_buf local_sigjmp_buf;
MemoryContext bgwriter_context;
bool prev_hibernate;
WritebackContext wb_context;
/*
* Properly accept or ignore signals the postmaster might send us.
*
* bgwriter doesn't participate in ProcSignal signalling, but a SIGUSR1
* handler is still needed for latch wakeups.
*/
pqsignal(SIGHUP, BgSigHupHandler); /* set flag to read config file */
pqsignal(SIGINT, SIG_IGN);
pqsignal(SIGTERM, ReqShutdownHandler); /* shutdown */
pqsignal(SIGQUIT, bg_quickdie); /* hard crash time */
pqsignal(SIGALRM, SIG_IGN);
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, bgwriter_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN);
/*
* Reset some signals that are accepted by postmaster but not here
*/
pqsignal(SIGCHLD, SIG_DFL);
pqsignal(SIGTTIN, SIG_DFL);
pqsignal(SIGTTOU, SIG_DFL);
pqsignal(SIGCONT, SIG_DFL);
pqsignal(SIGWINCH, SIG_DFL);
/* We allow SIGQUIT (quickdie) at all times */
sigdelset(&BlockSig, SIGQUIT);
/*
* Create a resource owner to keep track of our resources (currently only
* buffer pins).
*/
CurrentResourceOwner = ResourceOwnerCreate(NULL, "Background Writer");
/*
* 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.
*
* See notes in postgres.c about the design of this coding.
*/
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();
AbortBufferIO();
UnlockBuffers();
/* buffer pins are released here: */
ResourceOwnerRelease(CurrentResourceOwner,
RESOURCE_RELEASE_BEFORE_LOCKS,
false, true);
/* we needn't bother with the other ResourceOwnerRelease phases */
AtEOXact_Buffers(false);
AtEOXact_SMgr();
AtEOXact_Files();
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 */
MemoryContextResetAndDeleteChildren(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);
/*
* Close all open files after any error. This is helpful on Windows,
* where holding deleted files open causes various strange errors.
* It's not clear we need it elsewhere, but shouldn't hurt.
*/
smgrcloseall();
/* 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)
*/
PG_SETMASK(&UnBlockSig);
/*
* Reset hibernation state after any error.
*/
prev_hibernate = false;
/*
* Loop forever
*/
for (;;)
{
bool can_hibernate;
int rc;
/* Clear any already-pending wakeups */
ResetLatch(MyLatch);
if (got_SIGHUP)
{
got_SIGHUP = false;
ProcessConfigFile(PGC_SIGHUP);
}
if (shutdown_requested)
{
/*
* From here on, elog(ERROR) should end with exit(1), not send
* control back to the sigsetjmp block above
*/
ExitOnAnyError = true;
/* Normal exit from the bgwriter is here */
proc_exit(0); /* done */
}
/*
* Do one cycle of dirty-buffer writing.
*/
can_hibernate = BgBufferSync(&wb_context);
/*
* Send off activity statistics to the stats collector
*/
pgstat_send_bgwriter();
if (FirstCallSinceLastCheckpoint())
{
/*
* After any checkpoint, close all smgr files. This is so we
* won't hang onto smgr references to deleted files indefinitely.
*/
smgrcloseall();
}
/*
* 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_snap_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_POSTMASTER_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(MyProc->pgprocno);
/* Sleep ... */
rc = WaitLatch(MyLatch,
WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
BgWriterDelay * HIBERNATE_FACTOR,
WAIT_EVENT_BGWRITER_HIBERNATE);
/* Reset the notification request in case we timed out */
StrategyNotifyBgWriter(-1);
}
/*
* Emergency bailout if postmaster has died. This is to avoid the
* necessity for manual cleanup of all postmaster children.
*/
if (rc & WL_POSTMASTER_DEATH)
exit(1);
prev_hibernate = can_hibernate;
}
}
/*
* Main entry point for walwriter process
*
* This is invoked from AuxiliaryProcessMain, which has already created the
* basic execution environment, but not enabled signals yet.
*/
void
WalWriterMain(void)
{
sigjmp_buf local_sigjmp_buf;
MemoryContext walwriter_context;
int left_till_hibernate;
bool hibernating;
/*
* Properly accept or ignore signals the postmaster might send us
*
* We have no particular use for SIGINT at the moment, but seems
* reasonable to treat like SIGTERM.
*/
pqsignal(SIGHUP, WalSigHupHandler); /* set flag to read config file */
pqsignal(SIGINT, WalShutdownHandler); /* request shutdown */
pqsignal(SIGTERM, WalShutdownHandler); /* request shutdown */
pqsignal(SIGQUIT, wal_quickdie); /* hard crash time */
pqsignal(SIGALRM, SIG_IGN);
pqsignal(SIGPIPE, SIG_IGN);
pqsignal(SIGUSR1, walwriter_sigusr1_handler);
pqsignal(SIGUSR2, SIG_IGN); /* not used */
/*
* Reset some signals that are accepted by postmaster but not here
*/
pqsignal(SIGCHLD, SIG_DFL);
pqsignal(SIGTTIN, SIG_DFL);
pqsignal(SIGTTOU, SIG_DFL);
pqsignal(SIGCONT, SIG_DFL);
pqsignal(SIGWINCH, SIG_DFL);
/* We allow SIGQUIT (quickdie) at all times */
sigdelset(&BlockSig, SIGQUIT);
/*
* Create a resource owner to keep track of our resources (not clear that
* we need this, but may as well have one).
*/
CurrentResourceOwner = ResourceOwnerCreate(NULL, "Wal Writer");
/*
* 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.
*/
walwriter_context = AllocSetContextCreate(TopMemoryContext,
"Wal Writer",
ALLOCSET_DEFAULT_SIZES);
MemoryContextSwitchTo(walwriter_context);
/*
* If an exception is encountered, processing resumes here.
*
* This code is heavily based on bgwriter.c, q.v.
*/
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 walwriter, but we do have LWLocks, and perhaps buffers?
*/
LWLockReleaseAll();
ConditionVariableCancelSleep();
pgstat_report_wait_end();
AbortBufferIO();
UnlockBuffers();
/* buffer pins are released here: */
ResourceOwnerRelease(CurrentResourceOwner,
RESOURCE_RELEASE_BEFORE_LOCKS,
false, true);
/* we needn't bother with the other ResourceOwnerRelease phases */
AtEOXact_Buffers(false);
AtEOXact_SMgr();
AtEOXact_Files();
AtEOXact_HashTables(false);
/*
* Now return to normal top-level context and clear ErrorContext for
* next time.
*/
MemoryContextSwitchTo(walwriter_context);
FlushErrorState();
/* Flush any leaked data in the top-level context */
MemoryContextResetAndDeleteChildren(walwriter_context);
/* 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);
/*
* Close all open files after any error. This is helpful on Windows,
* where holding deleted files open causes various strange errors.
* It's not clear we need it elsewhere, but shouldn't hurt.
*/
smgrcloseall();
}
/* We can now handle ereport(ERROR) */
PG_exception_stack = &local_sigjmp_buf;
/*
* Unblock signals (they were blocked when the postmaster forked us)
*/
PG_SETMASK(&UnBlockSig);
/*
* Reset hibernation state after any error.
*/
left_till_hibernate = LOOPS_UNTIL_HIBERNATE;
hibernating = false;
SetWalWriterSleeping(false);
/*
* Advertise our latch that backends can use to wake us up while we're
* sleeping.
*/
ProcGlobal->walwriterLatch = &MyProc->procLatch;
/*
* Loop forever
*/
for (;;)
{
long cur_timeout;
int rc;
/*
* Advertise whether we might hibernate in this cycle. We do this
* before resetting the latch to ensure that any async commits will
* see the flag set if they might possibly need to wake us up, and
* that we won't miss any signal they send us. (If we discover work
* to do in the last cycle before we would hibernate, the global flag
* will be set unnecessarily, but little harm is done.) But avoid
* touching the global flag if it doesn't need to change.
*/
if (hibernating != (left_till_hibernate <= 1))
{
hibernating = (left_till_hibernate <= 1);
SetWalWriterSleeping(hibernating);
}
/* Clear any already-pending wakeups */
ResetLatch(MyLatch);
/*
* Process any requests or signals received recently.
*/
if (got_SIGHUP)
{
got_SIGHUP = false;
ProcessConfigFile(PGC_SIGHUP);
}
if (shutdown_requested)
{
/* Normal exit from the walwriter is here */
proc_exit(0); /* done */
}
/*
* Do what we're here for; then, if XLogBackgroundFlush() found useful
* work to do, reset hibernation counter.
*/
if (XLogBackgroundFlush())
left_till_hibernate = LOOPS_UNTIL_HIBERNATE;
else if (left_till_hibernate > 0)
left_till_hibernate--;
/*
* Sleep until we are signaled or WalWriterDelay has elapsed. If we
* haven't done anything useful for quite some time, lengthen the
* sleep time so as to reduce the server's idle power consumption.
*/
if (left_till_hibernate > 0)
cur_timeout = WalWriterDelay; /* in ms */
else
cur_timeout = WalWriterDelay * HIBERNATE_FACTOR;
rc = WaitLatch(MyLatch,
WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
cur_timeout,
WAIT_EVENT_WAL_WRITER_MAIN);
/*
* Emergency bailout if postmaster has died. This is to avoid the
* necessity for manual cleanup of all postmaster children.
*/
if (rc & WL_POSTMASTER_DEATH)
exit(1);
}
}
/*
* Wake up all processes sleeping on the given CV.
*
* This guarantees to wake all processes that were sleeping on the CV
* at time of call, but processes that add themselves to the list mid-call
* will typically not get awakened.
*/
void
ConditionVariableBroadcast(ConditionVariable *cv)
{
int pgprocno = MyProc->pgprocno;
PGPROC *proc = NULL;
bool have_sentinel = false;
/*
* In some use-cases, it is common for awakened processes to immediately
* re-queue themselves. If we just naively try to reduce the wakeup list
* to empty, we'll get into a potentially-indefinite loop against such a
* process. The semantics we really want are just to be sure that we have
* wakened all processes that were in the list at entry. We can use our
* own cvWaitLink as a sentinel to detect when we've finished.
*
* A seeming flaw in this approach is that someone else might signal the
* CV and in doing so remove our sentinel entry. But that's fine: since
* CV waiters are always added and removed in order, that must mean that
* every previous waiter has been wakened, so we're done. We'll get an
* extra "set" on our latch from the someone else's signal, which is
* slightly inefficient but harmless.
*
* We can't insert our cvWaitLink as a sentinel if it's already in use in
* some other proclist. While that's not expected to be true for typical
* uses of this function, we can deal with it by simply canceling any
* prepared CV sleep. The next call to ConditionVariableSleep will take
* care of re-establishing the lost state.
*/
if (cv_sleep_target != NULL)
ConditionVariableCancelSleep();
/*
* Inspect the state of the queue. If it's empty, we have nothing to do.
* If there's exactly one entry, we need only remove and signal that
* entry. Otherwise, remove the first entry and insert our sentinel.
*/
SpinLockAcquire(&cv->mutex);
/* While we're here, let's assert we're not in the list. */
Assert(!proclist_contains(&cv->wakeup, pgprocno, cvWaitLink));
if (!proclist_is_empty(&cv->wakeup))
{
proc = proclist_pop_head_node(&cv->wakeup, cvWaitLink);
if (!proclist_is_empty(&cv->wakeup))
{
proclist_push_tail(&cv->wakeup, pgprocno, cvWaitLink);
have_sentinel = true;
}
}
SpinLockRelease(&cv->mutex);
/* Awaken first waiter, if there was one. */
if (proc != NULL)
SetLatch(&proc->procLatch);
while (have_sentinel)
{
/*
* Each time through the loop, remove the first wakeup list entry, and
* signal it unless it's our sentinel. Repeat as long as the sentinel
* remains in the list.
*
* Notice that if someone else removes our sentinel, we will waken one
* additional process before exiting. That's intentional, because if
* someone else signals the CV, they may be intending to waken some
* third process that added itself to the list after we added the
* sentinel. Better to give a spurious wakeup (which should be
* harmless beyond wasting some cycles) than to lose a wakeup.
*/
proc = NULL;
SpinLockAcquire(&cv->mutex);
if (!proclist_is_empty(&cv->wakeup))
proc = proclist_pop_head_node(&cv->wakeup, cvWaitLink);
have_sentinel = proclist_contains(&cv->wakeup, pgprocno, cvWaitLink);
SpinLockRelease(&cv->mutex);
if (proc != NULL && proc != MyProc)
SetLatch(&proc->procLatch);
}
}