* Portions Copyright (c) 2010-2010, PostgreSQL Global Development Group
*
* IDENTIFICATION
- * $PostgreSQL: pgsql/src/backend/replication/walsender.c,v 1.28 2010/07/06 19:18:57 momjian Exp $
+ * $PostgreSQL: pgsql/src/backend/replication/walsender.c,v 1.32 2010/09/15 06:51:19 heikki Exp $
*
*-------------------------------------------------------------------------
*/
#include "postgres.h"
+#include <signal.h>
#include <unistd.h>
#include "access/xlog_internal.h"
#include "storage/fd.h"
#include "storage/ipc.h"
#include "storage/pmsignal.h"
+#include "storage/proc.h"
#include "tcop/tcopprot.h"
#include "utils/guc.h"
#include "utils/memutils.h"
/* My slot in the shared memory array */
static WalSnd *MyWalSnd = NULL;
+/* Array of WalSndWaiter in shared memory */
+static WalSndWaiter *WalSndWaiters;
+
/* Global state */
bool am_walsender = false; /* Am I a walsender process ? */
/* User-settable parameters for walsender */
int max_wal_senders = 0; /* the maximum number of concurrent walsenders */
int WalSndDelay = 200; /* max sleep time between some actions */
+int replication_timeout = 0; /* maximum time to wait for the Ack from the standby */
-#define NAPTIME_PER_CYCLE 100000L /* max sleep time between cycles
- * (100ms) */
+/*
+ * Buffer for WAL sending
+ *
+ * WalSndOutBuffer is a work area in which the output message is constructed.
+ * It's used in just so we can avoid re-palloc'ing the buffer on each cycle.
+ * It must be of size 6 + sizeof(WalDataMessageHeader) + MAX_SEND_SIZE.
+ */
+static char *WalSndOutBuffer;
+static int WalSndOutHead; /* head of pending output */
+static int WalSndOutTail; /* tail of pending output */
/*
* These variables are used similarly to openLogFile/Id/Seg/Off,
*/
static XLogRecPtr sentPtr = {0, 0};
+/*
+ * How far have we completed replication already? This is also
+ * advertised in MyWalSnd->ackdPtr. This is not used in asynchronous
+ * replication case.
+ */
+static XLogRecPtr ackdPtr = {0, 0};
+
+/* Replication mode requested by connected standby */
+static int rplMode = REPLICATION_MODE_ASYNC;
+
/* Flags set by signal handlers for later service in main loop */
static volatile sig_atomic_t got_SIGHUP = false;
static volatile sig_atomic_t shutdown_requested = false;
static volatile sig_atomic_t ready_to_stop = false;
+/* Flag set by signal handler of backends for replication */
+static volatile sig_atomic_t replication_done = false;
+
/* Signal handlers */
static void WalSndSigHupHandler(SIGNAL_ARGS);
static void WalSndShutdownHandler(SIGNAL_ARGS);
static void WalSndQuickDieHandler(SIGNAL_ARGS);
+static void WalSndXLogSendHandler(SIGNAL_ARGS);
+static void WalSndLastCycleHandler(SIGNAL_ARGS);
/* Prototypes for private functions */
static int WalSndLoop(void);
static void WalSndHandshake(void);
static void WalSndKill(int code, Datum arg);
static void XLogRead(char *buf, XLogRecPtr recptr, Size nbytes);
-static bool XLogSend(char *msgbuf, bool *caughtup);
-static void CheckClosedConnection(void);
+static bool XLogSend(bool *caughtup, bool *pending);
+static void ProcessStreamMsgs(StringInfo inMsg);
+
+static void RegisterWalSndWaiter(BackendId backendId, XLogRecPtr record,
+ Latch *latch);
+static void WakeupWalSndWaiters(XLogRecPtr record);
+static XLogRecPtr GetOldestAckdPtr(void);
/* Main entry point for walsender process */
/* Handle handshake messages before streaming */
WalSndHandshake();
+ /* Initialize shared memory status */
+ {
+ /* use volatile pointer to prevent code rearrangement */
+ volatile WalSnd *walsnd = MyWalSnd;
+
+ SpinLockAcquire(&walsnd->mutex);
+ walsnd->sentPtr = sentPtr;
+ SpinLockRelease(&walsnd->mutex);
+ }
+
/* Main loop of walsender */
return WalSndLoop();
}
/* Handle the very limited subset of commands expected in this phase */
switch (firstchar)
{
+ char rplModeStr[6];
+
case 'Q': /* Query message */
{
const char *query_string;
ReadyForQuery(DestRemote);
/* ReadyForQuery did pq_flush for us */
}
- else if (sscanf(query_string, "START_REPLICATION %X/%X",
- &recptr.xlogid, &recptr.xrecoff) == 2)
+ else if (sscanf(query_string, "START_REPLICATION %X/%X MODE %5s",
+ &recptr.xlogid, &recptr.xrecoff, rplModeStr) == 3)
{
StringInfoData buf;
(errcode(ERRCODE_CANNOT_CONNECT_NOW),
errmsg("standby connections not allowed because wal_level=minimal")));
- /* Send a CopyOutResponse message, and start streaming */
- pq_beginmessage(&buf, 'H');
- pq_sendbyte(&buf, 0);
- pq_sendint(&buf, 0, 2);
+ /* Verify that the specified replication mode is valid */
+ {
+ const struct config_enum_entry *entry;
+
+ for (entry = replication_mode_options; entry && entry->name; entry++)
+ {
+ if (strcmp(rplModeStr, entry->name) == 0)
+ {
+ rplMode = entry->val;
+ break;
+ }
+ }
+ if (entry == NULL || entry->name == NULL)
+ ereport(FATAL,
+ (errcode(ERRCODE_PROTOCOL_VIOLATION),
+ errmsg("invalid replication mode: %s", rplModeStr)));
+ }
+ MyWalSnd->rplMode = rplMode;
+
+ /* Send a CopyXLogResponse message, and start streaming */
+ pq_beginmessage(&buf, 'W');
pq_endmessage(&buf);
pq_flush();
/*
- * Initialize position to the received one, then the
+ * Initialize positions to the received one, then the
* xlog records begin to be shipped from that position
*/
- sentPtr = recptr;
+ sentPtr = ackdPtr = recptr;
/* break out of the loop */
replication_started = true;
}
/*
- * Check if the remote end has closed the connection.
+ * Process messages received from the standby.
+ *
+ * ereports on error.
*/
static void
-CheckClosedConnection(void)
+ProcessStreamMsgs(StringInfo inMsg)
{
- unsigned char firstchar;
- int r;
+ bool acked = false;
- r = pq_getbyte_if_available(&firstchar);
- if (r < 0)
- {
- /* unexpected error or EOF */
- ereport(COMMERROR,
- (errcode(ERRCODE_PROTOCOL_VIOLATION),
- errmsg("unexpected EOF on standby connection")));
- proc_exit(0);
- }
- if (r == 0)
+ /* Loop to process successive complete messages available */
+ for (;;)
{
- /* no data available without blocking */
- return;
- }
+ unsigned char firstchar;
+ int r;
+
+ r = pq_getbyte_if_available(&firstchar);
+ if (r < 0)
+ {
+ /* unexpected error or EOF */
+ ereport(COMMERROR,
+ (errcode(ERRCODE_PROTOCOL_VIOLATION),
+ errmsg("unexpected EOF on standby connection")));
+ proc_exit(0);
+ }
+ if (r == 0)
+ {
+ /* no data available without blocking */
+ break;
+ }
+
+ /* Handle the very limited subset of commands expected in this phase */
+ switch (firstchar)
+ {
+ case 'd': /* CopyData message */
+ {
+ unsigned char rpltype;
+
+ /*
+ * Read the message contents. This is expected to be done without
+ * blocking because we've been able to get message type code.
+ */
+ if (pq_getmessage(inMsg, 0))
+ proc_exit(0); /* suitable message already logged */
+
+ /* Read the replication message type from CopyData message */
+ rpltype = pq_getmsgbyte(inMsg);
+ switch (rpltype)
+ {
+ case 'l':
+ {
+ WalAckMessageData *msgdata;
+
+ msgdata = (WalAckMessageData *) pq_getmsgbytes(inMsg, sizeof(WalAckMessageData));
+
+ /*
+ * Update local status.
+ *
+ * The ackd ptr received from standby should not
+ * go backwards.
+ */
+ if (XLByteLE(ackdPtr, msgdata->ackEnd))
+ ackdPtr = msgdata->ackEnd;
+ else
+ ereport(FATAL,
+ (errmsg("replication completion location went back from "
+ "%X/%X to %X/%X",
+ ackdPtr.xlogid, ackdPtr.xrecoff,
+ msgdata->ackEnd.xlogid, msgdata->ackEnd.xrecoff)));
+
+ acked = true; /* also need to update shared position */
+ break;
+ }
+ default:
+ ereport(FATAL,
+ (errcode(ERRCODE_PROTOCOL_VIOLATION),
+ errmsg("invalid replication message type %d",
+ rpltype)));
+ }
+ break;
+ }
- /* Handle the very limited subset of commands expected in this phase */
- switch (firstchar)
- {
/*
* 'X' means that the standby is closing down the socket.
*/
- case 'X':
- proc_exit(0);
+ case 'X':
+ proc_exit(0);
- default:
- ereport(FATAL,
- (errcode(ERRCODE_PROTOCOL_VIOLATION),
- errmsg("invalid standby closing message type %d",
- firstchar)));
+ default:
+ ereport(FATAL,
+ (errcode(ERRCODE_PROTOCOL_VIOLATION),
+ errmsg("invalid standby message type %d",
+ firstchar)));
+ }
}
+
+ if (acked)
+ {
+ /* use volatile pointer to prevent code rearrangement */
+ volatile WalSnd *walsnd = MyWalSnd;
+
+ SpinLockAcquire(&walsnd->mutex);
+ walsnd->ackdPtr = ackdPtr;
+ SpinLockRelease(&walsnd->mutex);
+ }
+
+ /* Wake up the backends that this walsender had been blocking */
+ WakeupWalSndWaiters(ackdPtr);
}
/* Main loop of walsender process */
static int
WalSndLoop(void)
{
- char *output_message;
+ StringInfoData input_message;
bool caughtup = false;
+ bool pending = false;
+
+ initStringInfo(&input_message);
/*
* Allocate buffer that will be used for each output message. We do this
* just once to reduce palloc overhead. The buffer must be made large
* enough for maximum-sized messages.
*/
- output_message = palloc(1 + sizeof(WalDataMessageHeader) + MAX_SEND_SIZE);
+ WalSndOutBuffer = palloc(6 + sizeof(WalDataMessageHeader) + MAX_SEND_SIZE);
+ WalSndOutHead = WalSndOutTail = 0;
/* Loop forever, unless we get an error */
for (;;)
{
- long remain; /* remaining time (us) */
-
/*
* Emergency bailout if postmaster has died. This is to avoid the
* necessity for manual cleanup of all postmaster children.
*/
if (ready_to_stop)
{
- if (!XLogSend(output_message, &caughtup))
+ if (!XLogSend(&caughtup, &pending))
break;
- if (caughtup)
+ if (caughtup && !pending)
shutdown_requested = true;
}
}
/*
- * If we had sent all accumulated WAL in last round, nap for the
- * configured time before retrying.
- *
- * On some platforms, signals won't interrupt the sleep. To ensure we
- * respond reasonably promptly when someone signals us, break down the
- * sleep into NAPTIME_PER_CYCLE increments, and check for interrupts
- * after each nap.
+ * If we had sent all accumulated WAL in last round or could not
+ * flush pending WAL in output buffer because the socket was not
+ * writable, nap for the configured time before retrying.
*/
- if (caughtup)
+ if (caughtup || pending)
{
- remain = WalSndDelay * 1000L;
- while (remain > 0)
+ /*
+ * Even if we wrote all the WAL that was available when we started
+ * sending, more might have arrived while we were sending this
+ * batch. We had the latch set while sending, so we have not
+ * received any signals from that time. Let's arm the latch
+ * again, and after that check that we're still up-to-date.
+ */
+ ResetLatch(&MyWalSnd->latch);
+
+ if (!XLogSend(&caughtup, &pending))
+ break;
+ if ((caughtup || pending) && !got_SIGHUP && !ready_to_stop &&
+ !shutdown_requested)
{
- /* Check for interrupts */
- if (got_SIGHUP || shutdown_requested || ready_to_stop)
- break;
+ bool check_timeout;
+ long sleeptime;
+ int res;
+
+ /*
+ * XXX: We don't really need the periodic wakeups anymore,
+ * WaitLatchOrSocket should reliably wake up as soon as
+ * something interesting happens.
+ */
- /* Sleep and check that the connection is still alive */
- pg_usleep(remain > NAPTIME_PER_CYCLE ? NAPTIME_PER_CYCLE : remain);
- CheckClosedConnection();
+ /*
+ * Check for replication timeout if it's enabled and we need
+ * to wait until the socket has become writable to flush
+ * pending WAL in output buffer or until the Ack message
+ * from the standby has become available.
+ */
+ if (replication_timeout > 0 &&
+ (pending ||
+ (rplMode != REPLICATION_MODE_ASYNC &&
+ XLByteLT(ackdPtr, sentPtr))))
+ {
+ sleeptime = replication_timeout;
+ check_timeout = true;
+ }
+ else
+ {
+ sleeptime = WalSndDelay;
+ check_timeout = false;
+ }
+
+ /* Sleep */
+ res = WaitLatchOrSocket(&MyWalSnd->latch, MyProcPort->sock,
+ true, (WalSndOutTail > 0),
+ sleeptime * 1000L);
- remain -= NAPTIME_PER_CYCLE;
+ if (res == 0 && check_timeout)
+ {
+ /*
+ * Since typically expiration of replication timeout means
+ * communication problem, we don't send the error message
+ * to the standby.
+ */
+ ereport(COMMERROR,
+ (errmsg("terminating walsender process due to replication timeout")));
+ break;
+ }
}
- }
- /* Attempt to send the log once every loop */
- if (!XLogSend(output_message, &caughtup))
- break;
+ /* Process messages received from the standby */
+ ProcessStreamMsgs(&input_message);
+ }
+ else
+ {
+ /* Attempt to send the log once every loop */
+ if (!XLogSend(&caughtup, &pending))
+ break;
+ }
}
/*
}
else
{
- /* found */
- MyWalSnd = (WalSnd *) walsnd;
+ /*
+ * Found a free slot. Reserve it for us.
+ */
walsnd->pid = MyProcPid;
- MemSet(&MyWalSnd->sentPtr, 0, sizeof(XLogRecPtr));
+ MemSet(&walsnd->sentPtr, 0, sizeof(XLogRecPtr));
+ MemSet(&walsnd->ackdPtr, 0, sizeof(XLogRecPtr));
SpinLockRelease(&walsnd->mutex);
+ /* don't need the lock anymore */
+ OwnLatch((Latch *) &walsnd->latch);
+ MyWalSnd = (WalSnd *) walsnd;
+
break;
}
}
{
Assert(MyWalSnd != NULL);
+ /* Wake up the backends that this walsender had been blocking */
+ MyWalSnd->rplMode = REPLICATION_MODE_ASYNC;
+ WakeupWalSndWaiters(GetOldestAckdPtr());
+
/*
* Mark WalSnd struct no longer in use. Assume that no lock is required
* for this.
*/
MyWalSnd->pid = 0;
+ DisownLatch(&MyWalSnd->latch);
/* WalSnd struct isn't mine anymore */
MyWalSnd = NULL;
* Read up to MAX_SEND_SIZE bytes of WAL that's been flushed to disk,
* but not yet sent to the client, and send it.
*
- * msgbuf is a work area in which the output message is constructed. It's
- * passed in just so we can avoid re-palloc'ing the buffer on each cycle.
- * It must be of size 1 + sizeof(WalDataMessageHeader) + MAX_SEND_SIZE.
- *
* If there is no unsent WAL remaining, *caughtup is set to true, otherwise
* *caughtup is set to false.
*
+ * If there is pending WAL in output buffer, *pending is set to true,
+ * otherwise *pending is set to false.
+ *
* Returns true if OK, false if trouble.
*/
static bool
-XLogSend(char *msgbuf, bool *caughtup)
+XLogSend(bool *caughtup, bool *pending)
{
XLogRecPtr SendRqstPtr;
XLogRecPtr startptr;
- XLogRecPtr endptr;
+ static XLogRecPtr endptr;
Size nbytes;
+ uint32 n32;
+ int res;
WalDataMessageHeader msghdr;
+ /* Attempt to flush pending WAL in output buffer */
+ if (*pending)
+ {
+ if (WalSndOutHead != WalSndOutTail)
+ {
+ res = pq_putbytes_if_writable(WalSndOutBuffer + WalSndOutHead,
+ WalSndOutTail - WalSndOutHead);
+ if (res == EOF)
+ return false;
+ WalSndOutHead += res;
+ if (WalSndOutHead != WalSndOutTail)
+ return true;
+ }
+
+ res = pq_flush_if_writable();
+ if (res == EOF)
+ return false;
+ if (res == 0)
+ return true;
+
+ goto updt;
+ }
+
/*
* Attempt to send all data that's already been written out and fsync'd to
* disk. We cannot go further than what's been written out given the
/*
* OK to read and send the slice.
*/
- msgbuf[0] = 'w';
+ WalSndOutBuffer[0] = 'd';
+ WalSndOutBuffer[5] = 'w';
+ WalSndOutHead = 0;
+ WalSndOutTail = 6 + sizeof(WalDataMessageHeader) + nbytes;
+
+ n32 = htonl((uint32) WalSndOutTail - 1);
+ memcpy(WalSndOutBuffer + 1, &n32, 4);
/*
* Read the log directly into the output buffer to avoid extra memcpy
* calls.
*/
- XLogRead(msgbuf + 1 + sizeof(WalDataMessageHeader), startptr, nbytes);
+ XLogRead(WalSndOutBuffer + 6 + sizeof(WalDataMessageHeader), startptr, nbytes);
/*
* We fill the message header last so that the send timestamp is taken as
msghdr.walEnd = SendRqstPtr;
msghdr.sendTime = GetCurrentTimestamp();
- memcpy(msgbuf + 1, &msghdr, sizeof(WalDataMessageHeader));
+ memcpy(WalSndOutBuffer + 6, &msghdr, sizeof(WalDataMessageHeader));
- pq_putmessage('d', msgbuf, 1 + sizeof(WalDataMessageHeader) + nbytes);
+ res = pq_putbytes_if_writable(WalSndOutBuffer, WalSndOutTail);
+ if (res == EOF)
+ return false;
+
+ WalSndOutHead = res;
+ if (WalSndOutHead != WalSndOutTail)
+ {
+ *caughtup = false;
+ *pending = true;
+ return true;
+ }
/* Flush pending output to the client */
- if (pq_flush())
+ res = pq_flush_if_writable();
+ if (res == EOF)
return false;
+ if (res == 0)
+ {
+ *caughtup = false;
+ *pending = true;
+ return true;
+ }
+
+updt:
+ WalSndOutHead = WalSndOutTail = 0;
+ *pending = false;
sentPtr = endptr;
WalSndSigHupHandler(SIGNAL_ARGS)
{
got_SIGHUP = true;
+ if (MyWalSnd)
+ SetLatch(&MyWalSnd->latch);
}
/* SIGTERM: set flag to shut down */
WalSndShutdownHandler(SIGNAL_ARGS)
{
shutdown_requested = true;
+ if (MyWalSnd)
+ SetLatch(&MyWalSnd->latch);
}
/*
exit(2);
}
+/* SIGUSR1: set flag to send WAL records */
+static void
+WalSndXLogSendHandler(SIGNAL_ARGS)
+{
+ latch_sigusr1_handler();
+}
+
/* SIGUSR2: set flag to do a last cycle and shut down afterwards */
static void
WalSndLastCycleHandler(SIGNAL_ARGS)
{
ready_to_stop = true;
+ if (MyWalSnd)
+ SetLatch(&MyWalSnd->latch);
}
/* Set up signal handlers */
pqsignal(SIGQUIT, WalSndQuickDieHandler); /* hard crash time */
pqsignal(SIGALRM, SIG_IGN);
pqsignal(SIGPIPE, SIG_IGN);
- pqsignal(SIGUSR1, SIG_IGN); /* not used */
+ pqsignal(SIGUSR1, WalSndXLogSendHandler); /* request WAL sending */
pqsignal(SIGUSR2, WalSndLastCycleHandler); /* request a last cycle and
* shutdown */
size = offsetof(WalSndCtlData, walsnds);
size = add_size(size, mul_size(max_wal_senders, sizeof(WalSnd)));
+ /*
+ * If replication is enabled, we have a data structure called
+ * WalSndWaiters, created in shared memory.
+ */
+ if (max_wal_senders > 0)
+ size = add_size(size, mul_size(MaxBackends, sizeof(WalSndWaiter)));
+
return size;
}
{
bool found;
int i;
+ Size size = add_size(offsetof(WalSndCtlData, walsnds),
+ mul_size(max_wal_senders, sizeof(WalSnd)));
WalSndCtl = (WalSndCtlData *)
- ShmemInitStruct("Wal Sender Ctl", WalSndShmemSize(), &found);
+ ShmemInitStruct("Wal Sender Ctl", size, &found);
if (!found)
{
/* First time through, so initialize */
- MemSet(WalSndCtl, 0, WalSndShmemSize());
+ MemSet(WalSndCtl, 0, size);
for (i = 0; i < max_wal_senders; i++)
{
WalSnd *walsnd = &WalSndCtl->walsnds[i];
SpinLockInit(&walsnd->mutex);
+ InitSharedLatch(&walsnd->latch);
+ }
+ }
+
+ /* Create or attach to the WalSndWaiters array too, if needed */
+ if (max_wal_senders > 0)
+ {
+ WalSndWaiters = (WalSndWaiter *)
+ ShmemInitStruct("WalSndWaiters",
+ mul_size(MaxBackends, sizeof(WalSndWaiter)),
+ &found);
+ WalSndCtl->maxWaiters = MaxBackends;
+ }
+}
+
+/* Wake up all walsenders */
+void
+WalSndWakeup(void)
+{
+ int i;
+
+ for (i = 0; i < max_wal_senders; i++)
+ SetLatch(&WalSndCtl->walsnds[i].latch);
+}
+
+/*
+ * Ensure that replication has been completed up to the given position.
+ */
+void
+WaitXLogSend(XLogRecPtr record)
+{
+ int i;
+ bool mustwait = false;
+
+ Assert(max_wal_senders > 0);
+
+ for (i = 0; i < max_wal_senders; i++)
+ {
+ /* use volatile pointer to prevent code rearrangement */
+ volatile WalSnd *walsnd = &WalSndCtl->walsnds[i];
+ XLogRecPtr recptr;
+
+ /* Don't need to wait for asynchronous walsender */
+ if (walsnd->pid == 0 ||
+ walsnd->rplMode <= REPLICATION_MODE_ASYNC)
+ continue;
+
+ SpinLockAcquire(&walsnd->mutex);
+ recptr = walsnd->ackdPtr;
+ SpinLockRelease(&walsnd->mutex);
+
+ if (recptr.xlogid == 0 && recptr.xrecoff == 0)
+ continue;
+
+ /* Quick exit if already known replicated */
+ if (XLByteLE(record, recptr))
+ return;
+
+ mustwait = true;
+ }
+
+ /*
+ * Don't need to wait for replication if there is no synchronous
+ * standby
+ */
+ if (!mustwait)
+ return;
+
+ /*
+ * Register myself into the wait list and sleep until replication
+ * has been completed up to the given position and the walsender
+ * signals me.
+ *
+ * If replication has been completed up to the latest position
+ * before the registration, walsender might be unable to send the
+ * signal immediately. We must wake up the walsender after the
+ * registration.
+ */
+ ResetLatch(&MyProc->latch);
+ RegisterWalSndWaiter(MyBackendId, record, &MyProc->latch);
+ WalSndWakeup();
+
+ for (;;)
+ {
+ WaitLatch(&MyProc->latch, 1000000L);
+
+ /* If done already, we finish waiting */
+ if (replication_done)
+ {
+ replication_done = false;
+ return;
}
}
}
/*
- * This isn't currently used for anything. Monitoring tools might be
- * interested in the future, and we'll need something like this in the
- * future for synchronous replication.
+ * Register the given backend into the wait list.
*/
-#ifdef NOT_USED
+static void
+RegisterWalSndWaiter(BackendId backendId, XLogRecPtr record, Latch *latch)
+{
+ /* use volatile pointer to prevent code rearrangement */
+ volatile WalSndCtlData *walsndctl = WalSndCtl;
+ int i;
+ int count = 0;
+
+ LWLockAcquire(WalSndWaiterLock, LW_EXCLUSIVE);
+
+ /* Out of slots. This should not happen. */
+ if (walsndctl->numWaiters + 1 > walsndctl->maxWaiters)
+ elog(PANIC, "out of replication waiters slots");
+
+ /*
+ * The given position is expected to be relatively new in the
+ * wait list. Since the entries in the list are sorted in an
+ * increasing order of XLogRecPtr, we can shorten the time it
+ * takes to find an insert slot by scanning the list backwards.
+ */
+ for (i = walsndctl->numWaiters; i > 0; i--)
+ {
+ if (XLByteLE(WalSndWaiters[i - 1].record, record))
+ break;
+ count++;
+ }
+
+ /* Shuffle the list if needed */
+ if (count > 0)
+ memmove(&WalSndWaiters[i + 1], &WalSndWaiters[i],
+ count * sizeof(WalSndWaiter));
+
+ WalSndWaiters[i].backendId = backendId;
+ WalSndWaiters[i].record = record;
+ WalSndWaiters[i].latch = latch;
+ walsndctl->numWaiters++;
+
+ LWLockRelease(WalSndWaiterLock);
+}
+
+/*
+ * Wake up the backends waiting until replication has been completed
+ * up to the position older than or equal to the given one.
+ *
+ * Wake up all waiters if InvalidXLogRecPtr is given.
+ */
+static void
+WakeupWalSndWaiters(XLogRecPtr record)
+{
+ /* use volatile pointer to prevent code rearrangement */
+ volatile WalSndCtlData *walsndctl = WalSndCtl;
+ int i;
+ int count = 0;
+ bool all_wakeup = (record.xlogid == 0 && record.xrecoff == 0);
+
+ LWLockAcquire(WalSndWaiterLock, LW_EXCLUSIVE);
+
+ for (i = 0; i < walsndctl->numWaiters; i++)
+ {
+ /* use volatile pointer to prevent code rearrangement */
+ volatile WalSndWaiter *waiter = &WalSndWaiters[i];
+
+ if (all_wakeup || XLByteLE(waiter->record, record))
+ {
+ SetProcLatch(waiter->latch, PROCSIG_REPLICATION_INTERRUPT,
+ waiter->backendId);
+ count++;
+ }
+ else
+ {
+ /*
+ * If the backend waiting for the Ack position newer than
+ * the given one is found, we don't need to search the wait
+ * list any more. This is because the waiters in the list
+ * are guaranteed to be sorted in an increasing order of
+ * XLogRecPtr.
+ */
+ break;
+ }
+ }
+
+ /* If there are still some waiters, left-justify them in the list */
+ walsndctl->numWaiters -= count;
+ if (walsndctl->numWaiters > 0 && count > 0)
+ memmove(&WalSndWaiters[0], &WalSndWaiters[i],
+ walsndctl->numWaiters * sizeof(WalSndWaiter));
+
+ LWLockRelease(WalSndWaiterLock);
+}
+
/*
- * Returns the oldest Send position among walsenders. Or InvalidXLogRecPtr
- * if none.
+ * Returns the oldest Ack position in synchronous walsenders. Or
+ * InvalidXLogRecPtr if none.
*/
-XLogRecPtr
-GetOldestWALSendPointer(void)
+static XLogRecPtr
+GetOldestAckdPtr(void)
{
XLogRecPtr oldest = {0, 0};
- int i;
- bool found = false;
+ int i;
+ bool found = false;
for (i = 0; i < max_wal_senders; i++)
{
/* use volatile pointer to prevent code rearrangement */
volatile WalSnd *walsnd = &WalSndCtl->walsnds[i];
- XLogRecPtr recptr;
+ XLogRecPtr recptr;
- if (walsnd->pid == 0)
+ /*
+ * Ignore the Ack position that asynchronous walsender has
+ * since it has never received any Ack.
+ */
+ if (walsnd->pid == 0 ||
+ walsnd->rplMode <= REPLICATION_MODE_ASYNC)
continue;
SpinLockAcquire(&walsnd->mutex);
- recptr = walsnd->sentPtr;
+ recptr = walsnd->ackdPtr;
SpinLockRelease(&walsnd->mutex);
+ /*
+ * Ignore the Ack position that the walsender which has not
+ * received any Ack yet has.
+ */
if (recptr.xlogid == 0 && recptr.xrecoff == 0)
continue;
return oldest;
}
-#endif
+/*
+ * This is called when PROCSIG_REPLICATION_INTERRUPT is received.
+ */
+void
+HandleReplicationInterrupt(void)
+{
+ replication_done = true;
+}
OSDN Git Service
