diff --git a/doc/src/sgml/protocol.sgml b/doc/src/sgml/protocol.sgml index 73f26b432da0943ecbc9477fce139f9f832acca3..0775b7a8f4eecf3fa28123a1683e6920fa82c360 100644 --- a/doc/src/sgml/protocol.sgml +++ b/doc/src/sgml/protocol.sgml @@ -1460,7 +1460,7 @@ The commands accepted in walsender mode are: </varlistentry> <varlistentry> - <term>BASE_BACKUP [<literal>LABEL</literal> <replaceable>'label'</replaceable>] [<literal>PROGRESS</literal>] [<literal>FAST</literal>]</term> + <term>BASE_BACKUP [<literal>LABEL</literal> <replaceable>'label'</replaceable>] [<literal>PROGRESS</literal>] [<literal>FAST</literal>] [<literal>WAL</literal>]</term> <listitem> <para> Instructs the server to start streaming a base backup. @@ -1505,6 +1505,18 @@ The commands accepted in walsender mode are: </para> </listitem> </varlistentry> + + <varlistentry> + <term><literal>WAL</literal></term> + <listitem> + <para> + Include the necessary WAL segments in the backup. This will include + all the files between start and stop backup in the + <filename>pg_xlog</filename> directory of the base directory tar + file. + </para> + </listitem> + </varlistentry> </variablelist> </para> <para> @@ -1561,7 +1573,10 @@ The commands accepted in walsender mode are: </listitem> <listitem> <para> - <filename>pg_xlog</> (including subdirectories) + <filename>pg_xlog</>, including subdirectories. If the backup is run + with wal files included, a synthesized version of pg_xlog will be + included, but it will only contain the files necessary for the + backup to work, not the rest of the contents. </para> </listitem> </itemizedlist> diff --git a/doc/src/sgml/ref/pg_basebackup.sgml b/doc/src/sgml/ref/pg_basebackup.sgml index 321c8cade1cccad5d4e10b1f2e863e53a1bb213a..f4f78fbbfc5e85777f566735e0145a6823b3f3e4 100644 --- a/doc/src/sgml/ref/pg_basebackup.sgml +++ b/doc/src/sgml/ref/pg_basebackup.sgml @@ -144,6 +144,31 @@ PostgreSQL documentation </listitem> </varlistentry> + <varlistentry> + <term><option>-x</option></term> + <term><option>--xlog</option></term> + <listitem> + <para> + Includes the required transaction log files (WAL files) in the + backup. This will include all transaction logs generated during + the backup. If this option is specified, it is possible to start + a postmaster directly in the extracted directory without the need + to consult the log archive, thus making this a completely standalone + backup. + </para> + <note> + <para> + The transaction log files are collected at the end of the backup. + Therefore, it is necessary for the + <xref linkend="guc-wal-keep-segments"> parameter to be set high + enough that the log is not removed before the end of the backup. + If the log has been rotated when it's time to transfer it, the + backup will fail and be unusable. + </para> + </note> + </listitem> + </varlistentry> + <varlistentry> <term><option>-Z <replaceable class="parameter">level</replaceable></option></term> <term><option>--compress=<replaceable class="parameter">level</replaceable></option></term> @@ -164,7 +189,7 @@ PostgreSQL documentation <variablelist> <varlistentry> <term><option>-c <replaceable class="parameter">fast|spread</replaceable></option></term> - <term><option>--checkpoint <replaceable class="parameter">fast|spread</replaceable></option></term> + <term><option>--checkpoint=<replaceable class="parameter">fast|spread</replaceable></option></term> <listitem> <para> Sets checkpoint mode to fast or spread (default). @@ -191,7 +216,10 @@ PostgreSQL documentation Enables progress reporting. Turning this on will deliver an approximate progress report during the backup. Since the database may change during the backup, this is only an approximation and may not end at exactly - <literal>100%</literal>. + <literal>100%</literal>. In particular, when WAL log is included in the + backup, the total amount of data cannot be estimated in advance, and + in this case the progress report will only count towards the total + amount of data without WAL. </para> <para> When this is enabled, the backup will start by enumerating the size of diff --git a/src/backend/replication/basebackup.c b/src/backend/replication/basebackup.c index 251ed8e083b21704c9fc7822d2d6b14c142277f8..d0248f6dccc5a93cd22c9096327d65a7041357a1 100644 --- a/src/backend/replication/basebackup.c +++ b/src/backend/replication/basebackup.c @@ -37,6 +37,7 @@ typedef struct const char *label; bool progress; bool fastcheckpoint; + bool includewal; } basebackup_options; @@ -46,11 +47,17 @@ static void _tarWriteHeader(char *filename, char *linktarget, struct stat * statbuf); static void send_int8_string(StringInfoData *buf, int64 intval); static void SendBackupHeader(List *tablespaces); -static void SendBackupDirectory(char *location, char *spcoid); static void base_backup_cleanup(int code, Datum arg); static void perform_base_backup(basebackup_options *opt, DIR *tblspcdir); static void parse_basebackup_options(List *options, basebackup_options *opt); +/* + * Size of each block sent into the tar stream for larger files. + * + * XLogSegSize *MUST* be evenly dividable by this + */ +#define TAR_SEND_SIZE 32768 + typedef struct { char *oid; @@ -78,7 +85,10 @@ base_backup_cleanup(int code, Datum arg) static void perform_base_backup(basebackup_options *opt, DIR *tblspcdir) { - do_pg_start_backup(opt->label, opt->fastcheckpoint); + XLogRecPtr startptr; + XLogRecPtr endptr; + + startptr = do_pg_start_backup(opt->label, opt->fastcheckpoint); PG_ENSURE_ERROR_CLEANUP(base_backup_cleanup, (Datum) 0); { @@ -87,12 +97,6 @@ perform_base_backup(basebackup_options *opt, DIR *tblspcdir) struct dirent *de; tablespaceinfo *ti; - - /* Add a node for the base directory */ - ti = palloc0(sizeof(tablespaceinfo)); - ti->size = opt->progress ? sendDir(".", 1, true) : -1; - tablespaces = lappend(tablespaces, ti); - /* Collect information about all tablespaces */ while ((de = ReadDir(tblspcdir, "pg_tblspc")) != NULL) { @@ -120,6 +124,10 @@ perform_base_backup(basebackup_options *opt, DIR *tblspcdir) tablespaces = lappend(tablespaces, ti); } + /* Add a node for the base directory at the end */ + ti = palloc0(sizeof(tablespaceinfo)); + ti->size = opt->progress ? sendDir(".", 1, true) : -1; + tablespaces = lappend(tablespaces, ti); /* Send tablespace header */ SendBackupHeader(tablespaces); @@ -128,13 +136,102 @@ perform_base_backup(basebackup_options *opt, DIR *tblspcdir) foreach(lc, tablespaces) { tablespaceinfo *ti = (tablespaceinfo *) lfirst(lc); + StringInfoData buf; + + /* Send CopyOutResponse message */ + pq_beginmessage(&buf, 'H'); + pq_sendbyte(&buf, 0); /* overall format */ + pq_sendint(&buf, 0, 2); /* natts */ + pq_endmessage(&buf); + + sendDir(ti->path == NULL ? "." : ti->path, + ti->path == NULL ? 1 : strlen(ti->path), + false); - SendBackupDirectory(ti->path, ti->oid); + /* + * If we're including WAL, and this is the main data directory we + * don't terminate the tar stream here. Instead, we will append + * the xlog files below and terminate it then. This is safe since + * the main data directory is always sent *last*. + */ + if (opt->includewal && ti->path == NULL) + { + Assert(lnext(lc) == NULL); + } + else + pq_putemptymessage('c'); /* CopyDone */ } } PG_END_ENSURE_ERROR_CLEANUP(base_backup_cleanup, (Datum) 0); - do_pg_stop_backup(); + endptr = do_pg_stop_backup(); + + if (opt->includewal) + { + /* + * We've left the last tar file "open", so we can now append the + * required WAL files to it. + */ + uint32 logid, + logseg; + uint32 endlogid, + endlogseg; + struct stat statbuf; + + MemSet(&statbuf, 0, sizeof(statbuf)); + statbuf.st_mode = S_IRUSR | S_IWUSR; +#ifndef WIN32 + statbuf.st_uid = geteuid(); + statbuf.st_gid = getegid(); +#endif + statbuf.st_size = XLogSegSize; + statbuf.st_mtime = time(NULL); + + XLByteToSeg(startptr, logid, logseg); + XLByteToPrevSeg(endptr, endlogid, endlogseg); + + while (true) + { + /* Send another xlog segment */ + char fn[MAXPGPATH]; + int i; + + XLogFilePath(fn, ThisTimeLineID, logid, logseg); + _tarWriteHeader(fn, NULL, &statbuf); + + /* Send the actual WAL file contents, block-by-block */ + for (i = 0; i < XLogSegSize / TAR_SEND_SIZE; i++) + { + char buf[TAR_SEND_SIZE]; + XLogRecPtr ptr; + + ptr.xlogid = logid; + ptr.xrecoff = logseg * XLogSegSize + TAR_SEND_SIZE * i; + + XLogRead(buf, ptr, TAR_SEND_SIZE); + if (pq_putmessage('d', buf, TAR_SEND_SIZE)) + ereport(ERROR, + (errmsg("base backup could not send data, aborting backup"))); + } + + /* + * Files are always fixed size, and always end on a 512 byte + * boundary, so padding is never necessary. + */ + + + /* Advance to the next WAL file */ + NextLogSeg(logid, logseg); + + /* Have we reached our stop position yet? */ + if (logid > endlogid || + (logid == endlogid && logseg > endlogseg)) + break; + } + + /* Send CopyDone message for the last tar file */ + pq_putemptymessage('c'); + } } /* @@ -147,6 +244,7 @@ parse_basebackup_options(List *options, basebackup_options *opt) bool o_label = false; bool o_progress = false; bool o_fast = false; + bool o_wal = false; MemSet(opt, 0, sizeof(*opt)); foreach(lopt, options) @@ -180,6 +278,15 @@ parse_basebackup_options(List *options, basebackup_options *opt) opt->fastcheckpoint = true; o_fast = true; } + else if (strcmp(defel->defname, "wal") == 0) + { + if (o_wal) + ereport(ERROR, + (errcode(ERRCODE_SYNTAX_ERROR), + errmsg("duplicate option \"%s\"", defel->defname))); + opt->includewal = true; + o_wal = true; + } else elog(ERROR, "option \"%s\" not recognized", defel->defname); @@ -316,26 +423,6 @@ SendBackupHeader(List *tablespaces) pq_puttextmessage('C', "SELECT"); } -static void -SendBackupDirectory(char *location, char *spcoid) -{ - StringInfoData buf; - - /* Send CopyOutResponse message */ - pq_beginmessage(&buf, 'H'); - pq_sendbyte(&buf, 0); /* overall format */ - pq_sendint(&buf, 0, 2); /* natts */ - pq_endmessage(&buf); - - /* tar up the data directory if NULL, otherwise the tablespace */ - sendDir(location == NULL ? "." : location, - location == NULL ? 1 : strlen(location), - false); - - /* Send CopyDone message */ - pq_putemptymessage('c'); -} - static int64 sendDir(char *path, int basepathlen, bool sizeonly) @@ -506,7 +593,7 @@ static void sendFile(char *filename, int basepathlen, struct stat * statbuf) { FILE *fp; - char buf[32768]; + char buf[TAR_SEND_SIZE]; size_t cnt; pgoff_t len = 0; size_t pad; diff --git a/src/backend/replication/repl_gram.y b/src/backend/replication/repl_gram.y index 879a0bd7db0d9248027cd5869ccc34567185301f..e1a4a51a0e9696c43d19a699e9a68ae3afcc856b 100644 --- a/src/backend/replication/repl_gram.y +++ b/src/backend/replication/repl_gram.y @@ -71,6 +71,7 @@ Node *replication_parse_result; %token K_LABEL %token K_PROGRESS %token K_FAST +%token K_WAL %token K_START_REPLICATION %type <node> command @@ -106,7 +107,7 @@ identify_system: ; /* - * BASE_BACKUP [LABEL <label>] [PROGRESS] [FAST] + * BASE_BACKUP [LABEL '<label>'] [PROGRESS] [FAST] [WAL] */ base_backup: K_BASE_BACKUP base_backup_opt_list @@ -136,7 +137,12 @@ base_backup_opt: $$ = makeDefElem("fast", (Node *)makeInteger(TRUE)); } - + | K_WAL + { + $$ = makeDefElem("wal", + (Node *)makeInteger(TRUE)); + } + ; /* * START_REPLICATION %X/%X diff --git a/src/backend/replication/repl_scanner.l b/src/backend/replication/repl_scanner.l index e6dfb041b6ad5445fc048dc8a84d265a7f5865f3..87e77d975a904842c1dcbf97cd50531825ab8df6 100644 --- a/src/backend/replication/repl_scanner.l +++ b/src/backend/replication/repl_scanner.l @@ -61,6 +61,7 @@ FAST { return K_FAST; } IDENTIFY_SYSTEM { return K_IDENTIFY_SYSTEM; } LABEL { return K_LABEL; } PROGRESS { return K_PROGRESS; } +WAL { return K_WAL; } START_REPLICATION { return K_START_REPLICATION; } "," { return ','; } ";" { return ';'; } diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c index 44efa9fc25e8d37a94e2d7e72264c0794a96a211..f70458e01a257d53d6daf7307f630fa9a830f22b 100644 --- a/src/backend/replication/walsender.c +++ b/src/backend/replication/walsender.c @@ -105,7 +105,6 @@ static int WalSndLoop(void); static void InitWalSnd(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 void IdentifySystem(void); @@ -649,8 +648,13 @@ WalSndKill(int code, Datum arg) * * XXX probably this should be improved to suck data directly from the * WAL buffers when possible. + * + * Will open, and keep open, one WAL segment stored in the global file + * descriptor sendFile. This means if XLogRead is used once, there will + * always be one descriptor left open until the process ends, but never + * more than one. */ -static void +void XLogRead(char *buf, XLogRecPtr recptr, Size nbytes) { XLogRecPtr startRecPtr = recptr; diff --git a/src/bin/pg_basebackup/pg_basebackup.c b/src/bin/pg_basebackup/pg_basebackup.c index 536303461a68c83aa53cc88a73e69f3e950a8174..ef2718a435ed74cf0f420bce3c0e41921150b0db 100644 --- a/src/bin/pg_basebackup/pg_basebackup.c +++ b/src/bin/pg_basebackup/pg_basebackup.c @@ -33,6 +33,7 @@ char *label = "pg_basebackup base backup"; bool showprogress = false; int verbose = 0; int compresslevel = 0; +bool includewal = false; bool fastcheckpoint = false; char *dbhost = NULL; char *dbuser = NULL; @@ -124,6 +125,7 @@ usage(void) printf(_("\nOptions controlling the output:\n")); printf(_(" -D, --pgdata=directory receive base backup into directory\n")); printf(_(" -F, --format=p|t output format (plain, tar)\n")); + printf(_(" -x, --xlog include required WAL files in backup\n")); printf(_(" -Z, --compress=0-9 compress tar output\n")); printf(_("\nGeneral options:\n")); printf(_(" -c, --checkpoint=fast|spread\n" @@ -200,16 +202,20 @@ verify_dir_is_empty_or_create(char *dirname) static void progress_report(int tablespacenum, char *fn) { + int percent = (int) ((totaldone / 1024) * 100 / totalsize); + if (percent > 100) + percent = 100; + if (verbose) fprintf(stderr, INT64_FORMAT "/" INT64_FORMAT " kB (%i%%) %i/%i tablespaces (%-30s)\r", totaldone / 1024, totalsize, - (int) ((totaldone / 1024) * 100 / totalsize), + percent, tablespacenum, tablespacecount, fn); else fprintf(stderr, INT64_FORMAT "/" INT64_FORMAT " kB (%i%%) %i/%i tablespaces\r", totaldone / 1024, totalsize, - (int) ((totaldone / 1024) * 100 / totalsize), + percent, tablespacenum, tablespacecount); } @@ -746,9 +752,10 @@ BaseBackup() conn = GetConnection(); PQescapeStringConn(conn, escaped_label, label, sizeof(escaped_label), &i); - snprintf(current_path, sizeof(current_path), "BASE_BACKUP LABEL '%s' %s %s", + snprintf(current_path, sizeof(current_path), "BASE_BACKUP LABEL '%s' %s %s %s", escaped_label, showprogress ? "PROGRESS" : "", + includewal ? "WAL" : "", fastcheckpoint ? "FAST" : ""); if (PQsendQuery(conn, current_path) == 0) @@ -789,7 +796,7 @@ BaseBackup() * first once since it can be relocated, and it will be checked before * we do anything anyway. */ - if (format == 'p' && i > 0) + if (format == 'p' && !PQgetisnull(res, i, 1)) verify_dir_is_empty_or_create(PQgetvalue(res, i, 1)); } @@ -848,6 +855,7 @@ main(int argc, char **argv) {"pgdata", required_argument, NULL, 'D'}, {"format", required_argument, NULL, 'F'}, {"checkpoint", required_argument, NULL, 'c'}, + {"xlog", no_argument, NULL, 'x'}, {"compress", required_argument, NULL, 'Z'}, {"label", required_argument, NULL, 'l'}, {"host", required_argument, NULL, 'h'}, @@ -881,7 +889,7 @@ main(int argc, char **argv) } } - while ((c = getopt_long(argc, argv, "D:F:l:Z:c:h:p:U:wWvP", + while ((c = getopt_long(argc, argv, "D:F:l:Z:c:h:p:U:xwWvP", long_options, &option_index)) != -1) { switch (c) @@ -901,6 +909,9 @@ main(int argc, char **argv) exit(1); } break; + case 'x': + includewal = true; + break; case 'l': label = xstrdup(optarg); break; diff --git a/src/include/replication/walsender.h b/src/include/replication/walsender.h index bd9e19320fb75f3c9d5b374a9cbcf7cfb641743e..9a196ab1c8bb69d4a7013e416d7a23792a7bb954 100644 --- a/src/include/replication/walsender.h +++ b/src/include/replication/walsender.h @@ -67,6 +67,7 @@ extern Size WalSndShmemSize(void); extern void WalSndShmemInit(void); extern void WalSndWakeup(void); extern void WalSndSetState(WalSndState state); +extern void XLogRead(char *buf, XLogRecPtr recptr, Size nbytes); extern Datum pg_stat_get_wal_senders(PG_FUNCTION_ARGS);