postgresql/src/include/replication/walreceiver.h

214 lines
7.1 KiB
C
Raw Normal View History

/*-------------------------------------------------------------------------
*
* walreceiver.h
* Exports from replication/walreceiverfuncs.c.
*
* Portions Copyright (c) 2010-2016, PostgreSQL Global Development Group
*
2010-09-20 22:08:53 +02:00
* src/include/replication/walreceiver.h
*
*-------------------------------------------------------------------------
*/
#ifndef _WALRECEIVER_H
#define _WALRECEIVER_H
#include "access/xlog.h"
#include "access/xlogdefs.h"
#include "fmgr.h"
Allow a streaming replication standby to follow a timeline switch. Before this patch, streaming replication would refuse to start replicating if the timeline in the primary doesn't exactly match the standby. The situation where it doesn't match is when you have a master, and two standbys, and you promote one of the standbys to become new master. Promoting bumps up the timeline ID, and after that bump, the other standby would refuse to continue. There's significantly more timeline related logic in streaming replication now. First of all, when a standby connects to primary, it will ask the primary for any timeline history files that are missing from the standby. The missing files are sent using a new replication command TIMELINE_HISTORY, and stored in standby's pg_xlog directory. Using the timeline history files, the standby can follow the latest timeline present in the primary (recovery_target_timeline='latest'), just as it can follow new timelines appearing in an archive directory. START_REPLICATION now takes a TIMELINE parameter, to specify exactly which timeline to stream WAL from. This allows the standby to request the primary to send over WAL that precedes the promotion. The replication protocol is changed slightly (in a backwards-compatible way although there's little hope of streaming replication working across major versions anyway), to allow replication to stop when the end of timeline reached, putting the walsender back into accepting a replication command. Many thanks to Amit Kapila for testing and reviewing various versions of this patch.
2012-12-13 18:00:00 +01:00
#include "storage/latch.h"
#include "storage/spin.h"
#include "pgtime.h"
/* user-settable parameters */
2011-04-10 17:42:00 +02:00
extern int wal_receiver_status_interval;
extern int wal_receiver_timeout;
extern bool hot_standby_feedback;
/*
* MAXCONNINFO: maximum size of a connection string.
*
* XXX: Should this move to pg_config_manual.h?
*/
#define MAXCONNINFO 1024
/* Can we allow the standby to accept replication connection from another standby? */
#define AllowCascadeReplication() (EnableHotStandby && max_wal_senders > 0)
/*
* Values for WalRcv->walRcvState.
*/
typedef enum
{
2010-02-26 03:01:40 +01:00
WALRCV_STOPPED, /* stopped and mustn't start up again */
WALRCV_STARTING, /* launched, but the process hasn't
* initialized yet */
Allow a streaming replication standby to follow a timeline switch. Before this patch, streaming replication would refuse to start replicating if the timeline in the primary doesn't exactly match the standby. The situation where it doesn't match is when you have a master, and two standbys, and you promote one of the standbys to become new master. Promoting bumps up the timeline ID, and after that bump, the other standby would refuse to continue. There's significantly more timeline related logic in streaming replication now. First of all, when a standby connects to primary, it will ask the primary for any timeline history files that are missing from the standby. The missing files are sent using a new replication command TIMELINE_HISTORY, and stored in standby's pg_xlog directory. Using the timeline history files, the standby can follow the latest timeline present in the primary (recovery_target_timeline='latest'), just as it can follow new timelines appearing in an archive directory. START_REPLICATION now takes a TIMELINE parameter, to specify exactly which timeline to stream WAL from. This allows the standby to request the primary to send over WAL that precedes the promotion. The replication protocol is changed slightly (in a backwards-compatible way although there's little hope of streaming replication working across major versions anyway), to allow replication to stop when the end of timeline reached, putting the walsender back into accepting a replication command. Many thanks to Amit Kapila for testing and reviewing various versions of this patch.
2012-12-13 18:00:00 +01:00
WALRCV_STREAMING, /* walreceiver is streaming */
WALRCV_WAITING, /* stopped streaming, waiting for orders */
WALRCV_RESTARTING, /* asked to restart streaming */
2010-02-26 03:01:40 +01:00
WALRCV_STOPPING /* requested to stop, but still running */
} WalRcvState;
/* Shared memory area for management of walreceiver process */
typedef struct
{
/*
* PID of currently active walreceiver process, its current state and
2010-07-06 21:19:02 +02:00
* start time (actually, the time at which it was requested to be
* started).
*/
2010-02-26 03:01:40 +01:00
pid_t pid;
WalRcvState walRcvState;
2010-02-26 03:01:40 +01:00
pg_time_t startTime;
/*
* receiveStart and receiveStartTLI indicate the first byte position and
* timeline that will be received. When startup process starts the
* walreceiver, it sets these to the point where it wants the streaming to
* begin.
*/
XLogRecPtr receiveStart;
Allow a streaming replication standby to follow a timeline switch. Before this patch, streaming replication would refuse to start replicating if the timeline in the primary doesn't exactly match the standby. The situation where it doesn't match is when you have a master, and two standbys, and you promote one of the standbys to become new master. Promoting bumps up the timeline ID, and after that bump, the other standby would refuse to continue. There's significantly more timeline related logic in streaming replication now. First of all, when a standby connects to primary, it will ask the primary for any timeline history files that are missing from the standby. The missing files are sent using a new replication command TIMELINE_HISTORY, and stored in standby's pg_xlog directory. Using the timeline history files, the standby can follow the latest timeline present in the primary (recovery_target_timeline='latest'), just as it can follow new timelines appearing in an archive directory. START_REPLICATION now takes a TIMELINE parameter, to specify exactly which timeline to stream WAL from. This allows the standby to request the primary to send over WAL that precedes the promotion. The replication protocol is changed slightly (in a backwards-compatible way although there's little hope of streaming replication working across major versions anyway), to allow replication to stop when the end of timeline reached, putting the walsender back into accepting a replication command. Many thanks to Amit Kapila for testing and reviewing various versions of this patch.
2012-12-13 18:00:00 +01:00
TimeLineID receiveStartTLI;
/*
* receivedUpto-1 is the last byte position that has already been
* received, and receivedTLI is the timeline it came from. At the first
Allow a streaming replication standby to follow a timeline switch. Before this patch, streaming replication would refuse to start replicating if the timeline in the primary doesn't exactly match the standby. The situation where it doesn't match is when you have a master, and two standbys, and you promote one of the standbys to become new master. Promoting bumps up the timeline ID, and after that bump, the other standby would refuse to continue. There's significantly more timeline related logic in streaming replication now. First of all, when a standby connects to primary, it will ask the primary for any timeline history files that are missing from the standby. The missing files are sent using a new replication command TIMELINE_HISTORY, and stored in standby's pg_xlog directory. Using the timeline history files, the standby can follow the latest timeline present in the primary (recovery_target_timeline='latest'), just as it can follow new timelines appearing in an archive directory. START_REPLICATION now takes a TIMELINE parameter, to specify exactly which timeline to stream WAL from. This allows the standby to request the primary to send over WAL that precedes the promotion. The replication protocol is changed slightly (in a backwards-compatible way although there's little hope of streaming replication working across major versions anyway), to allow replication to stop when the end of timeline reached, putting the walsender back into accepting a replication command. Many thanks to Amit Kapila for testing and reviewing various versions of this patch.
2012-12-13 18:00:00 +01:00
* startup of walreceiver, these are set to receiveStart and
* receiveStartTLI. After that, walreceiver updates these whenever it
* flushes the received WAL to disk.
*/
XLogRecPtr receivedUpto;
Allow a streaming replication standby to follow a timeline switch. Before this patch, streaming replication would refuse to start replicating if the timeline in the primary doesn't exactly match the standby. The situation where it doesn't match is when you have a master, and two standbys, and you promote one of the standbys to become new master. Promoting bumps up the timeline ID, and after that bump, the other standby would refuse to continue. There's significantly more timeline related logic in streaming replication now. First of all, when a standby connects to primary, it will ask the primary for any timeline history files that are missing from the standby. The missing files are sent using a new replication command TIMELINE_HISTORY, and stored in standby's pg_xlog directory. Using the timeline history files, the standby can follow the latest timeline present in the primary (recovery_target_timeline='latest'), just as it can follow new timelines appearing in an archive directory. START_REPLICATION now takes a TIMELINE parameter, to specify exactly which timeline to stream WAL from. This allows the standby to request the primary to send over WAL that precedes the promotion. The replication protocol is changed slightly (in a backwards-compatible way although there's little hope of streaming replication working across major versions anyway), to allow replication to stop when the end of timeline reached, putting the walsender back into accepting a replication command. Many thanks to Amit Kapila for testing and reviewing various versions of this patch.
2012-12-13 18:00:00 +01:00
TimeLineID receivedTLI;
/*
* latestChunkStart is the starting byte position of the current "batch"
* of received WAL. It's actually the same as the previous value of
* receivedUpto before the last flush to disk. Startup process can use
* this to detect whether it's keeping up or not.
*/
XLogRecPtr latestChunkStart;
/*
* Time of send and receive of any message received.
*/
TimestampTz lastMsgSendTime;
TimestampTz lastMsgReceiptTime;
/*
* Latest reported end of WAL on the sender
*/
XLogRecPtr latestWalEnd;
TimestampTz latestWalEndTime;
/*
* connection string; initially set to connect to the primary, and later
* clobbered to hide security-sensitive fields.
*/
char conninfo[MAXCONNINFO];
/*
* replication slot name; is also used for walreceiver to connect with the
* primary
*/
char slotname[NAMEDATALEN];
2010-02-26 03:01:40 +01:00
slock_t mutex; /* locks shared variables shown above */
Allow a streaming replication standby to follow a timeline switch. Before this patch, streaming replication would refuse to start replicating if the timeline in the primary doesn't exactly match the standby. The situation where it doesn't match is when you have a master, and two standbys, and you promote one of the standbys to become new master. Promoting bumps up the timeline ID, and after that bump, the other standby would refuse to continue. There's significantly more timeline related logic in streaming replication now. First of all, when a standby connects to primary, it will ask the primary for any timeline history files that are missing from the standby. The missing files are sent using a new replication command TIMELINE_HISTORY, and stored in standby's pg_xlog directory. Using the timeline history files, the standby can follow the latest timeline present in the primary (recovery_target_timeline='latest'), just as it can follow new timelines appearing in an archive directory. START_REPLICATION now takes a TIMELINE parameter, to specify exactly which timeline to stream WAL from. This allows the standby to request the primary to send over WAL that precedes the promotion. The replication protocol is changed slightly (in a backwards-compatible way although there's little hope of streaming replication working across major versions anyway), to allow replication to stop when the end of timeline reached, putting the walsender back into accepting a replication command. Many thanks to Amit Kapila for testing and reviewing various versions of this patch.
2012-12-13 18:00:00 +01:00
/*
* force walreceiver reply? This doesn't need to be locked; memory
* barriers for ordering are sufficient.
*/
bool force_reply;
/* set true once conninfo is ready to display (obfuscated pwds etc) */
bool ready_to_display;
Allow a streaming replication standby to follow a timeline switch. Before this patch, streaming replication would refuse to start replicating if the timeline in the primary doesn't exactly match the standby. The situation where it doesn't match is when you have a master, and two standbys, and you promote one of the standbys to become new master. Promoting bumps up the timeline ID, and after that bump, the other standby would refuse to continue. There's significantly more timeline related logic in streaming replication now. First of all, when a standby connects to primary, it will ask the primary for any timeline history files that are missing from the standby. The missing files are sent using a new replication command TIMELINE_HISTORY, and stored in standby's pg_xlog directory. Using the timeline history files, the standby can follow the latest timeline present in the primary (recovery_target_timeline='latest'), just as it can follow new timelines appearing in an archive directory. START_REPLICATION now takes a TIMELINE parameter, to specify exactly which timeline to stream WAL from. This allows the standby to request the primary to send over WAL that precedes the promotion. The replication protocol is changed slightly (in a backwards-compatible way although there's little hope of streaming replication working across major versions anyway), to allow replication to stop when the end of timeline reached, putting the walsender back into accepting a replication command. Many thanks to Amit Kapila for testing and reviewing various versions of this patch.
2012-12-13 18:00:00 +01:00
/*
* Latch used by startup process to wake up walreceiver after telling it
* where to start streaming (after setting receiveStart and
* receiveStartTLI), and also to tell it to send apply feedback to the
* primary whenever specially marked commit records are applied.
* This is normally mapped to procLatch when walreceiver is running.
Allow a streaming replication standby to follow a timeline switch. Before this patch, streaming replication would refuse to start replicating if the timeline in the primary doesn't exactly match the standby. The situation where it doesn't match is when you have a master, and two standbys, and you promote one of the standbys to become new master. Promoting bumps up the timeline ID, and after that bump, the other standby would refuse to continue. There's significantly more timeline related logic in streaming replication now. First of all, when a standby connects to primary, it will ask the primary for any timeline history files that are missing from the standby. The missing files are sent using a new replication command TIMELINE_HISTORY, and stored in standby's pg_xlog directory. Using the timeline history files, the standby can follow the latest timeline present in the primary (recovery_target_timeline='latest'), just as it can follow new timelines appearing in an archive directory. START_REPLICATION now takes a TIMELINE parameter, to specify exactly which timeline to stream WAL from. This allows the standby to request the primary to send over WAL that precedes the promotion. The replication protocol is changed slightly (in a backwards-compatible way although there's little hope of streaming replication working across major versions anyway), to allow replication to stop when the end of timeline reached, putting the walsender back into accepting a replication command. Many thanks to Amit Kapila for testing and reviewing various versions of this patch.
2012-12-13 18:00:00 +01:00
*/
Latch *latch;
} WalRcvData;
extern WalRcvData *WalRcv;
struct WalReceiverConn;
typedef struct WalReceiverConn WalReceiverConn;
Allow a streaming replication standby to follow a timeline switch. Before this patch, streaming replication would refuse to start replicating if the timeline in the primary doesn't exactly match the standby. The situation where it doesn't match is when you have a master, and two standbys, and you promote one of the standbys to become new master. Promoting bumps up the timeline ID, and after that bump, the other standby would refuse to continue. There's significantly more timeline related logic in streaming replication now. First of all, when a standby connects to primary, it will ask the primary for any timeline history files that are missing from the standby. The missing files are sent using a new replication command TIMELINE_HISTORY, and stored in standby's pg_xlog directory. Using the timeline history files, the standby can follow the latest timeline present in the primary (recovery_target_timeline='latest'), just as it can follow new timelines appearing in an archive directory. START_REPLICATION now takes a TIMELINE parameter, to specify exactly which timeline to stream WAL from. This allows the standby to request the primary to send over WAL that precedes the promotion. The replication protocol is changed slightly (in a backwards-compatible way although there's little hope of streaming replication working across major versions anyway), to allow replication to stop when the end of timeline reached, putting the walsender back into accepting a replication command. Many thanks to Amit Kapila for testing and reviewing various versions of this patch.
2012-12-13 18:00:00 +01:00
/* libpqwalreceiver hooks */
typedef WalReceiverConn *(*walrcv_connect_fn) (const char *conninfo, bool logical,
const char *appname);
typedef char *(*walrcv_get_conninfo_fn) (WalReceiverConn *conn);
typedef char *(*walrcv_identify_system_fn) (WalReceiverConn *conn,
TimeLineID *primary_tli);
typedef void (*walrcv_readtimelinehistoryfile_fn) (WalReceiverConn *conn,
TimeLineID tli,
char **filename,
char **content, int *size);
typedef bool (*walrcv_startstreaming_fn) (WalReceiverConn *conn,
TimeLineID tli,
XLogRecPtr startpoint,
const char *slotname);
typedef void (*walrcv_endstreaming_fn) (WalReceiverConn *conn,
TimeLineID *next_tli);
typedef int (*walrcv_receive_fn) (WalReceiverConn *conn, char **buffer,
pgsocket *wait_fd);
typedef void (*walrcv_send_fn) (WalReceiverConn *conn, const char *buffer,
int nbytes);
typedef void (*walrcv_disconnect_fn) (WalReceiverConn *conn);
typedef struct WalReceiverFunctionsType
{
walrcv_connect_fn connect;
walrcv_get_conninfo_fn get_conninfo;
walrcv_identify_system_fn identify_system;
walrcv_readtimelinehistoryfile_fn readtimelinehistoryfile;
walrcv_startstreaming_fn startstreaming;
walrcv_endstreaming_fn endstreaming;
walrcv_receive_fn receive;
walrcv_send_fn send;
walrcv_disconnect_fn disconnect;
} WalReceiverFunctionsType;
extern PGDLLIMPORT WalReceiverFunctionsType *WalReceiverFunctions;
#define walrcv_connect(conninfo, logical, appname) \
WalReceiverFunctions->connect(conninfo, logical, appname)
#define walrcv_get_conninfo(conn) \
WalReceiverFunctions->get_conninfo(conn)
#define walrcv_identify_system(conn, primary_tli) \
WalReceiverFunctions->identify_system(conn, primary_tli)
#define walrcv_readtimelinehistoryfile(conn, tli, filename, content, size) \
WalReceiverFunctions->readtimelinehistoryfile(conn, tli, filename, content, size)
#define walrcv_startstreaming(conn, tli, startpoint, slotname) \
WalReceiverFunctions->startstreaming(conn, tli, startpoint, slotname)
#define walrcv_endstreaming(conn, next_tli) \
WalReceiverFunctions->endstreaming(conn, next_tli)
#define walrcv_receive(conn, buffer, wait_fd) \
WalReceiverFunctions->receive(conn, buffer, wait_fd)
#define walrcv_send(conn, buffer, nbytes) \
WalReceiverFunctions->send(conn, buffer, nbytes)
#define walrcv_disconnect(conn) \
WalReceiverFunctions->disconnect(conn)
/* prototypes for functions in walreceiver.c */
extern void WalReceiverMain(void) pg_attribute_noreturn();
extern Datum pg_stat_get_wal_receiver(PG_FUNCTION_ARGS);
/* prototypes for functions in walreceiverfuncs.c */
extern Size WalRcvShmemSize(void);
extern void WalRcvShmemInit(void);
extern void ShutdownWalRcv(void);
Allow a streaming replication standby to follow a timeline switch. Before this patch, streaming replication would refuse to start replicating if the timeline in the primary doesn't exactly match the standby. The situation where it doesn't match is when you have a master, and two standbys, and you promote one of the standbys to become new master. Promoting bumps up the timeline ID, and after that bump, the other standby would refuse to continue. There's significantly more timeline related logic in streaming replication now. First of all, when a standby connects to primary, it will ask the primary for any timeline history files that are missing from the standby. The missing files are sent using a new replication command TIMELINE_HISTORY, and stored in standby's pg_xlog directory. Using the timeline history files, the standby can follow the latest timeline present in the primary (recovery_target_timeline='latest'), just as it can follow new timelines appearing in an archive directory. START_REPLICATION now takes a TIMELINE parameter, to specify exactly which timeline to stream WAL from. This allows the standby to request the primary to send over WAL that precedes the promotion. The replication protocol is changed slightly (in a backwards-compatible way although there's little hope of streaming replication working across major versions anyway), to allow replication to stop when the end of timeline reached, putting the walsender back into accepting a replication command. Many thanks to Amit Kapila for testing and reviewing various versions of this patch.
2012-12-13 18:00:00 +01:00
extern bool WalRcvStreaming(void);
extern bool WalRcvRunning(void);
extern void RequestXLogStreaming(TimeLineID tli, XLogRecPtr recptr,
const char *conninfo, const char *slotname);
Allow a streaming replication standby to follow a timeline switch. Before this patch, streaming replication would refuse to start replicating if the timeline in the primary doesn't exactly match the standby. The situation where it doesn't match is when you have a master, and two standbys, and you promote one of the standbys to become new master. Promoting bumps up the timeline ID, and after that bump, the other standby would refuse to continue. There's significantly more timeline related logic in streaming replication now. First of all, when a standby connects to primary, it will ask the primary for any timeline history files that are missing from the standby. The missing files are sent using a new replication command TIMELINE_HISTORY, and stored in standby's pg_xlog directory. Using the timeline history files, the standby can follow the latest timeline present in the primary (recovery_target_timeline='latest'), just as it can follow new timelines appearing in an archive directory. START_REPLICATION now takes a TIMELINE parameter, to specify exactly which timeline to stream WAL from. This allows the standby to request the primary to send over WAL that precedes the promotion. The replication protocol is changed slightly (in a backwards-compatible way although there's little hope of streaming replication working across major versions anyway), to allow replication to stop when the end of timeline reached, putting the walsender back into accepting a replication command. Many thanks to Amit Kapila for testing and reviewing various versions of this patch.
2012-12-13 18:00:00 +01:00
extern XLogRecPtr GetWalRcvWriteRecPtr(XLogRecPtr *latestChunkStart, TimeLineID *receiveTLI);
extern int GetReplicationApplyDelay(void);
extern int GetReplicationTransferLatency(void);
extern void WalRcvForceReply(void);
#endif /* _WALRECEIVER_H */