Line data Source code
1 : /*-------------------------------------------------------------------------
2 : *
3 : * pg_backup_directory.c
4 : *
5 : * A directory format dump is a directory, which contains a "toc.dat" file
6 : * for the TOC, and a separate file for each data entry, named "<oid>.dat".
7 : * Large objects are stored in separate files named "blob_<oid>.dat",
8 : * and there's a plain-text TOC file for each BLOBS TOC entry named
9 : * "blobs_<dumpID>.toc" (or just "blobs.toc" in archive versions before 16).
10 : *
11 : * If compression is used, each data file is individually compressed and the
12 : * ".gz" suffix is added to the filenames. The TOC files are never
13 : * compressed by pg_dump, however they are accepted with the .gz suffix too,
14 : * in case the user has manually compressed them with 'gzip'.
15 : *
16 : * NOTE: This format is identical to the files written in the tar file in
17 : * the 'tar' format, except that we don't write the restore.sql file (TODO),
18 : * and the tar format doesn't support compression. Please keep the formats in
19 : * sync.
20 : *
21 : *
22 : * Portions Copyright (c) 1996-2026, PostgreSQL Global Development Group
23 : * Portions Copyright (c) 1994, Regents of the University of California
24 : * Portions Copyright (c) 2000, Philip Warner
25 : *
26 : * Rights are granted to use this software in any way so long
27 : * as this notice is not removed.
28 : *
29 : * The author is not responsible for loss or damages that may
30 : * result from its use.
31 : *
32 : * IDENTIFICATION
33 : * src/bin/pg_dump/pg_backup_directory.c
34 : *
35 : *-------------------------------------------------------------------------
36 : */
37 : #include "postgres_fe.h"
38 :
39 : #include <dirent.h>
40 : #include <sys/stat.h>
41 :
42 : #include "common/file_utils.h"
43 : #include "compress_io.h"
44 : #include "dumputils.h"
45 : #include "parallel.h"
46 : #include "pg_backup_utils.h"
47 :
48 : typedef struct
49 : {
50 : /*
51 : * Our archive location. This is basically what the user specified as his
52 : * backup file but of course here it is a directory.
53 : */
54 : char *directory;
55 :
56 : CompressFileHandle *dataFH; /* currently open data file */
57 : CompressFileHandle *LOsTocFH; /* file handle for blobs_NNN.toc */
58 : ParallelState *pstate; /* for parallel backup / restore */
59 : } lclContext;
60 :
61 : typedef struct
62 : {
63 : char *filename; /* filename excluding the directory (basename) */
64 : } lclTocEntry;
65 :
66 : /* prototypes for private functions */
67 : static void _ArchiveEntry(ArchiveHandle *AH, TocEntry *te);
68 : static void _StartData(ArchiveHandle *AH, TocEntry *te);
69 : static void _EndData(ArchiveHandle *AH, TocEntry *te);
70 : static void _WriteData(ArchiveHandle *AH, const void *data, size_t dLen);
71 : static int _WriteByte(ArchiveHandle *AH, const int i);
72 : static int _ReadByte(ArchiveHandle *AH);
73 : static void _WriteBuf(ArchiveHandle *AH, const void *buf, size_t len);
74 : static void _ReadBuf(ArchiveHandle *AH, void *buf, size_t len);
75 : static void _CloseArchive(ArchiveHandle *AH);
76 : static void _ReopenArchive(ArchiveHandle *AH);
77 : static void _PrintTocData(ArchiveHandle *AH, TocEntry *te);
78 :
79 : static void _WriteExtraToc(ArchiveHandle *AH, TocEntry *te);
80 : static void _ReadExtraToc(ArchiveHandle *AH, TocEntry *te);
81 : static void _PrintExtraToc(ArchiveHandle *AH, TocEntry *te);
82 :
83 : static void _StartLOs(ArchiveHandle *AH, TocEntry *te);
84 : static void _StartLO(ArchiveHandle *AH, TocEntry *te, Oid oid);
85 : static void _EndLO(ArchiveHandle *AH, TocEntry *te, Oid oid);
86 : static void _EndLOs(ArchiveHandle *AH, TocEntry *te);
87 : static void _LoadLOs(ArchiveHandle *AH, TocEntry *te);
88 :
89 : static void _PrepParallelRestore(ArchiveHandle *AH);
90 : static void _Clone(ArchiveHandle *AH);
91 : static void _DeClone(ArchiveHandle *AH);
92 :
93 : static int _WorkerJobRestoreDirectory(ArchiveHandle *AH, TocEntry *te);
94 : static int _WorkerJobDumpDirectory(ArchiveHandle *AH, TocEntry *te);
95 :
96 : static void setFilePath(ArchiveHandle *AH, char *buf,
97 : const char *relativeFilename);
98 :
99 : /*
100 : * Init routine required by ALL formats. This is a global routine
101 : * and should be declared in pg_backup_archiver.h
102 : *
103 : * Its task is to create any extra archive context (using AH->formatData),
104 : * and to initialize the supported function pointers.
105 : *
106 : * It should also prepare whatever its input source is for reading/writing,
107 : * and in the case of a read mode connection, it should load the Header & TOC.
108 : */
109 : void
110 0 : InitArchiveFmt_Directory(ArchiveHandle *AH)
111 : {
112 0 : lclContext *ctx;
113 :
114 : /* Assuming static functions, this can be copied for each format. */
115 0 : AH->ArchiveEntryPtr = _ArchiveEntry;
116 0 : AH->StartDataPtr = _StartData;
117 0 : AH->WriteDataPtr = _WriteData;
118 0 : AH->EndDataPtr = _EndData;
119 0 : AH->WriteBytePtr = _WriteByte;
120 0 : AH->ReadBytePtr = _ReadByte;
121 0 : AH->WriteBufPtr = _WriteBuf;
122 0 : AH->ReadBufPtr = _ReadBuf;
123 0 : AH->ClosePtr = _CloseArchive;
124 0 : AH->ReopenPtr = _ReopenArchive;
125 0 : AH->PrintTocDataPtr = _PrintTocData;
126 0 : AH->ReadExtraTocPtr = _ReadExtraToc;
127 0 : AH->WriteExtraTocPtr = _WriteExtraToc;
128 0 : AH->PrintExtraTocPtr = _PrintExtraToc;
129 :
130 0 : AH->StartLOsPtr = _StartLOs;
131 0 : AH->StartLOPtr = _StartLO;
132 0 : AH->EndLOPtr = _EndLO;
133 0 : AH->EndLOsPtr = _EndLOs;
134 :
135 0 : AH->PrepParallelRestorePtr = _PrepParallelRestore;
136 0 : AH->ClonePtr = _Clone;
137 0 : AH->DeClonePtr = _DeClone;
138 :
139 0 : AH->WorkerJobRestorePtr = _WorkerJobRestoreDirectory;
140 0 : AH->WorkerJobDumpPtr = _WorkerJobDumpDirectory;
141 :
142 : /* Set up our private context */
143 0 : ctx = (lclContext *) pg_malloc0(sizeof(lclContext));
144 0 : AH->formatData = ctx;
145 :
146 0 : ctx->dataFH = NULL;
147 0 : ctx->LOsTocFH = NULL;
148 :
149 : /*
150 : * Now open the TOC file
151 : */
152 :
153 0 : if (!AH->fSpec || strcmp(AH->fSpec, "") == 0)
154 0 : pg_fatal("no output directory specified");
155 :
156 0 : ctx->directory = AH->fSpec;
157 :
158 0 : if (AH->mode == archModeWrite)
159 : {
160 : /* we accept an empty existing directory */
161 0 : create_or_open_dir(ctx->directory);
162 0 : }
163 : else
164 : { /* Read Mode */
165 0 : char fname[MAXPGPATH];
166 0 : CompressFileHandle *tocFH;
167 :
168 0 : setFilePath(AH, fname, "toc.dat");
169 :
170 0 : tocFH = InitDiscoverCompressFileHandle(fname, PG_BINARY_R);
171 0 : if (tocFH == NULL)
172 0 : pg_fatal("could not open input file \"%s\": %m", fname);
173 :
174 0 : ctx->dataFH = tocFH;
175 :
176 : /*
177 : * The TOC of a directory format dump shares the format code of the
178 : * tar format.
179 : */
180 0 : AH->format = archTar;
181 0 : ReadHead(AH);
182 0 : AH->format = archDirectory;
183 0 : ReadToc(AH);
184 :
185 : /* Nothing else in the file, so close it again... */
186 0 : if (!EndCompressFileHandle(tocFH))
187 0 : pg_fatal("could not close TOC file: %m");
188 0 : ctx->dataFH = NULL;
189 0 : }
190 0 : }
191 :
192 : /*
193 : * Called by the Archiver when the dumper creates a new TOC entry.
194 : *
195 : * We determine the filename for this entry.
196 : */
197 : static void
198 0 : _ArchiveEntry(ArchiveHandle *AH, TocEntry *te)
199 : {
200 0 : lclTocEntry *tctx;
201 0 : char fn[MAXPGPATH];
202 :
203 0 : tctx = (lclTocEntry *) pg_malloc0(sizeof(lclTocEntry));
204 0 : if (strcmp(te->desc, "BLOBS") == 0)
205 : {
206 0 : snprintf(fn, MAXPGPATH, "blobs_%d.toc", te->dumpId);
207 0 : tctx->filename = pg_strdup(fn);
208 0 : }
209 0 : else if (te->dataDumper)
210 : {
211 0 : snprintf(fn, MAXPGPATH, "%d.dat", te->dumpId);
212 0 : tctx->filename = pg_strdup(fn);
213 0 : }
214 : else
215 0 : tctx->filename = NULL;
216 :
217 0 : te->formatData = tctx;
218 0 : }
219 :
220 : /*
221 : * Called by the Archiver to save any extra format-related TOC entry
222 : * data.
223 : *
224 : * Use the Archiver routines to write data - they are non-endian, and
225 : * maintain other important file information.
226 : */
227 : static void
228 0 : _WriteExtraToc(ArchiveHandle *AH, TocEntry *te)
229 : {
230 0 : lclTocEntry *tctx = (lclTocEntry *) te->formatData;
231 :
232 : /*
233 : * A dumpable object has set tctx->filename, any other object has not.
234 : * (see _ArchiveEntry).
235 : */
236 0 : if (tctx->filename)
237 0 : WriteStr(AH, tctx->filename);
238 : else
239 0 : WriteStr(AH, "");
240 0 : }
241 :
242 : /*
243 : * Called by the Archiver to read any extra format-related TOC data.
244 : *
245 : * Needs to match the order defined in _WriteExtraToc, and should also
246 : * use the Archiver input routines.
247 : */
248 : static void
249 0 : _ReadExtraToc(ArchiveHandle *AH, TocEntry *te)
250 : {
251 0 : lclTocEntry *tctx = (lclTocEntry *) te->formatData;
252 :
253 0 : if (tctx == NULL)
254 : {
255 0 : tctx = (lclTocEntry *) pg_malloc0(sizeof(lclTocEntry));
256 0 : te->formatData = tctx;
257 0 : }
258 :
259 0 : tctx->filename = ReadStr(AH);
260 0 : if (strlen(tctx->filename) == 0)
261 : {
262 0 : free(tctx->filename);
263 0 : tctx->filename = NULL;
264 0 : }
265 0 : }
266 :
267 : /*
268 : * Called by the Archiver when restoring an archive to output a comment
269 : * that includes useful information about the TOC entry.
270 : */
271 : static void
272 0 : _PrintExtraToc(ArchiveHandle *AH, TocEntry *te)
273 : {
274 0 : lclTocEntry *tctx = (lclTocEntry *) te->formatData;
275 :
276 0 : if (AH->public.verbose && tctx->filename)
277 0 : ahprintf(AH, "-- File: %s\n", tctx->filename);
278 0 : }
279 :
280 : /*
281 : * Called by the archiver when saving TABLE DATA (not schema). This routine
282 : * should save whatever format-specific information is needed to read
283 : * the archive back.
284 : *
285 : * It is called just prior to the dumper's 'DataDumper' routine being called.
286 : *
287 : * We create the data file for writing.
288 : */
289 : static void
290 0 : _StartData(ArchiveHandle *AH, TocEntry *te)
291 : {
292 0 : lclTocEntry *tctx = (lclTocEntry *) te->formatData;
293 0 : lclContext *ctx = (lclContext *) AH->formatData;
294 0 : char fname[MAXPGPATH];
295 :
296 0 : setFilePath(AH, fname, tctx->filename);
297 :
298 0 : ctx->dataFH = InitCompressFileHandle(AH->compression_spec);
299 :
300 0 : if (!ctx->dataFH->open_write_func(fname, PG_BINARY_W, ctx->dataFH))
301 0 : pg_fatal("could not open output file \"%s\": %m", fname);
302 0 : }
303 :
304 : /*
305 : * Called by archiver when dumper calls WriteData. This routine is
306 : * called for both LO and table data; it is the responsibility of
307 : * the format to manage each kind of data using StartLO/StartData.
308 : *
309 : * It should only be called from within a DataDumper routine.
310 : *
311 : * We write the data to the open data file.
312 : */
313 : static void
314 0 : _WriteData(ArchiveHandle *AH, const void *data, size_t dLen)
315 : {
316 0 : lclContext *ctx = (lclContext *) AH->formatData;
317 0 : CompressFileHandle *CFH = ctx->dataFH;
318 :
319 0 : if (dLen <= 0)
320 0 : return;
321 0 : CFH->write_func(data, dLen, CFH);
322 0 : }
323 :
324 : /*
325 : * Called by the archiver when a dumper's 'DataDumper' routine has
326 : * finished.
327 : *
328 : * We close the data file.
329 : */
330 : static void
331 0 : _EndData(ArchiveHandle *AH, TocEntry *te)
332 : {
333 0 : lclContext *ctx = (lclContext *) AH->formatData;
334 :
335 : /* Close the file */
336 0 : if (!EndCompressFileHandle(ctx->dataFH))
337 0 : pg_fatal("could not close data file: %m");
338 :
339 0 : ctx->dataFH = NULL;
340 0 : }
341 :
342 : /*
343 : * Print data for a given file (can be a LO as well)
344 : */
345 : static void
346 0 : _PrintFileData(ArchiveHandle *AH, char *filename)
347 : {
348 0 : size_t cnt;
349 0 : char *buf;
350 0 : size_t buflen;
351 0 : CompressFileHandle *CFH;
352 :
353 0 : if (!filename)
354 0 : return;
355 :
356 0 : CFH = InitDiscoverCompressFileHandle(filename, PG_BINARY_R);
357 0 : if (!CFH)
358 0 : pg_fatal("could not open input file \"%s\": %m", filename);
359 :
360 0 : buflen = DEFAULT_IO_BUFFER_SIZE;
361 0 : buf = pg_malloc(buflen);
362 :
363 0 : while ((cnt = CFH->read_func(buf, buflen, CFH)) > 0)
364 : {
365 0 : ahwrite(buf, 1, cnt, AH);
366 : }
367 :
368 0 : free(buf);
369 0 : if (!EndCompressFileHandle(CFH))
370 0 : pg_fatal("could not close data file \"%s\": %m", filename);
371 0 : }
372 :
373 : /*
374 : * Print data for a given TOC entry
375 : */
376 : static void
377 0 : _PrintTocData(ArchiveHandle *AH, TocEntry *te)
378 : {
379 0 : lclTocEntry *tctx = (lclTocEntry *) te->formatData;
380 :
381 0 : if (!tctx->filename)
382 0 : return;
383 :
384 0 : if (strcmp(te->desc, "BLOBS") == 0)
385 0 : _LoadLOs(AH, te);
386 : else
387 : {
388 0 : char fname[MAXPGPATH];
389 :
390 0 : setFilePath(AH, fname, tctx->filename);
391 0 : _PrintFileData(AH, fname);
392 0 : }
393 0 : }
394 :
395 : static void
396 0 : _LoadLOs(ArchiveHandle *AH, TocEntry *te)
397 : {
398 0 : Oid oid;
399 0 : lclContext *ctx = (lclContext *) AH->formatData;
400 0 : lclTocEntry *tctx = (lclTocEntry *) te->formatData;
401 0 : CompressFileHandle *CFH;
402 0 : char tocfname[MAXPGPATH];
403 0 : char line[MAXPGPATH];
404 :
405 0 : StartRestoreLOs(AH);
406 :
407 : /*
408 : * Note: before archive v16, there was always only one BLOBS TOC entry,
409 : * now there can be multiple. Furthermore, although the actual filename
410 : * was always "blobs.toc" before v16, the value of tctx->filename did not
411 : * match that before commit 548e50976 fixed it. For simplicity we assume
412 : * it must be "blobs.toc" in all archives before v16.
413 : */
414 0 : if (AH->version < K_VERS_1_16)
415 0 : setFilePath(AH, tocfname, "blobs.toc");
416 : else
417 0 : setFilePath(AH, tocfname, tctx->filename);
418 :
419 0 : CFH = ctx->LOsTocFH = InitDiscoverCompressFileHandle(tocfname, PG_BINARY_R);
420 :
421 0 : if (ctx->LOsTocFH == NULL)
422 0 : pg_fatal("could not open large object TOC file \"%s\" for input: %m",
423 : tocfname);
424 :
425 : /* Read the LOs TOC file line-by-line, and process each LO */
426 0 : while ((CFH->gets_func(line, MAXPGPATH, CFH)) != NULL)
427 : {
428 0 : char lofname[MAXPGPATH + 1];
429 0 : char path[MAXPGPATH];
430 :
431 : /* Can't overflow because line and lofname are the same length */
432 0 : if (sscanf(line, "%u %" CppAsString2(MAXPGPATH) "s\n", &oid, lofname) != 2)
433 0 : pg_fatal("invalid line in large object TOC file \"%s\": \"%s\"",
434 : tocfname, line);
435 :
436 0 : StartRestoreLO(AH, oid, AH->public.ropt->dropSchema);
437 0 : snprintf(path, MAXPGPATH, "%s/%s", ctx->directory, lofname);
438 0 : _PrintFileData(AH, path);
439 0 : EndRestoreLO(AH, oid);
440 0 : }
441 0 : if (!CFH->eof_func(CFH))
442 0 : pg_fatal("error reading large object TOC file \"%s\"",
443 : tocfname);
444 :
445 0 : if (!EndCompressFileHandle(ctx->LOsTocFH))
446 0 : pg_fatal("could not close large object TOC file \"%s\": %m",
447 : tocfname);
448 :
449 0 : ctx->LOsTocFH = NULL;
450 :
451 0 : EndRestoreLOs(AH);
452 0 : }
453 :
454 :
455 : /*
456 : * Write a byte of data to the archive.
457 : * Called by the archiver to do integer & byte output to the archive.
458 : * These routines are only used to read & write the headers & TOC.
459 : */
460 : static int
461 0 : _WriteByte(ArchiveHandle *AH, const int i)
462 : {
463 0 : unsigned char c = (unsigned char) i;
464 0 : lclContext *ctx = (lclContext *) AH->formatData;
465 0 : CompressFileHandle *CFH = ctx->dataFH;
466 :
467 0 : CFH->write_func(&c, 1, CFH);
468 0 : return 1;
469 0 : }
470 :
471 : /*
472 : * Read a byte of data from the archive.
473 : * Called by the archiver to read bytes & integers from the archive.
474 : * These routines are only used to read & write headers & TOC.
475 : * EOF should be treated as a fatal error.
476 : */
477 : static int
478 0 : _ReadByte(ArchiveHandle *AH)
479 : {
480 0 : lclContext *ctx = (lclContext *) AH->formatData;
481 0 : CompressFileHandle *CFH = ctx->dataFH;
482 :
483 0 : return CFH->getc_func(CFH);
484 0 : }
485 :
486 : /*
487 : * Write a buffer of data to the archive.
488 : * Called by the archiver to write a block of bytes to the TOC or a data file.
489 : */
490 : static void
491 0 : _WriteBuf(ArchiveHandle *AH, const void *buf, size_t len)
492 : {
493 0 : lclContext *ctx = (lclContext *) AH->formatData;
494 0 : CompressFileHandle *CFH = ctx->dataFH;
495 :
496 0 : CFH->write_func(buf, len, CFH);
497 0 : }
498 :
499 : /*
500 : * Read a block of bytes from the archive.
501 : *
502 : * Called by the archiver to read a block of bytes from the archive
503 : */
504 : static void
505 0 : _ReadBuf(ArchiveHandle *AH, void *buf, size_t len)
506 : {
507 0 : lclContext *ctx = (lclContext *) AH->formatData;
508 0 : CompressFileHandle *CFH = ctx->dataFH;
509 :
510 : /*
511 : * We do not expect a short read, so fail if we get one. The read_func
512 : * already dealt with any outright I/O error.
513 : */
514 0 : if (CFH->read_func(buf, len, CFH) != len)
515 0 : pg_fatal("could not read from input file: end of file");
516 0 : }
517 :
518 : /*
519 : * Close the archive.
520 : *
521 : * When writing the archive, this is the routine that actually starts
522 : * the process of saving it to files. No data should be written prior
523 : * to this point, since the user could sort the TOC after creating it.
524 : *
525 : * If an archive is to be written, this routine must call:
526 : * WriteHead to save the archive header
527 : * WriteToc to save the TOC entries
528 : * WriteDataChunks to save all data & LOs.
529 : */
530 : static void
531 0 : _CloseArchive(ArchiveHandle *AH)
532 : {
533 0 : lclContext *ctx = (lclContext *) AH->formatData;
534 :
535 0 : if (AH->mode == archModeWrite)
536 : {
537 0 : CompressFileHandle *tocFH;
538 0 : pg_compress_specification compression_spec = {0};
539 0 : char fname[MAXPGPATH];
540 :
541 0 : setFilePath(AH, fname, "toc.dat");
542 :
543 : /* this will actually fork the processes for a parallel backup */
544 0 : ctx->pstate = ParallelBackupStart(AH);
545 :
546 : /* The TOC is always created uncompressed */
547 0 : compression_spec.algorithm = PG_COMPRESSION_NONE;
548 0 : tocFH = InitCompressFileHandle(compression_spec);
549 0 : if (!tocFH->open_write_func(fname, PG_BINARY_W, tocFH))
550 0 : pg_fatal("could not open output file \"%s\": %m", fname);
551 0 : ctx->dataFH = tocFH;
552 :
553 : /*
554 : * Write 'tar' in the format field of the toc.dat file. The directory
555 : * is compatible with 'tar', so there's no point having a different
556 : * format code for it.
557 : */
558 0 : AH->format = archTar;
559 0 : WriteHead(AH);
560 0 : AH->format = archDirectory;
561 0 : WriteToc(AH);
562 0 : if (!EndCompressFileHandle(tocFH))
563 0 : pg_fatal("could not close TOC file: %m");
564 0 : WriteDataChunks(AH, ctx->pstate);
565 :
566 0 : ParallelBackupEnd(AH, ctx->pstate);
567 :
568 : /*
569 : * In directory mode, there is no need to sync all the entries
570 : * individually. Just recurse once through all the files generated.
571 : */
572 0 : if (AH->dosync)
573 0 : sync_dir_recurse(ctx->directory, AH->sync_method);
574 0 : }
575 0 : AH->FH = NULL;
576 0 : }
577 :
578 : /*
579 : * Reopen the archive's file handle.
580 : */
581 : static void
582 0 : _ReopenArchive(ArchiveHandle *AH)
583 : {
584 : /*
585 : * Our TOC is in memory, our data files are opened by each child anyway as
586 : * they are separate. We support reopening the archive by just doing
587 : * nothing.
588 : */
589 0 : }
590 :
591 : /*
592 : * LO support
593 : */
594 :
595 : /*
596 : * Called by the archiver when starting to save BLOB DATA (not schema).
597 : * It is called just prior to the dumper's DataDumper routine.
598 : *
599 : * We open the large object TOC file here, so that we can append a line to
600 : * it for each LO.
601 : */
602 : static void
603 0 : _StartLOs(ArchiveHandle *AH, TocEntry *te)
604 : {
605 0 : lclContext *ctx = (lclContext *) AH->formatData;
606 0 : lclTocEntry *tctx = (lclTocEntry *) te->formatData;
607 0 : pg_compress_specification compression_spec = {0};
608 0 : char fname[MAXPGPATH];
609 :
610 0 : setFilePath(AH, fname, tctx->filename);
611 :
612 : /* The LO TOC file is never compressed */
613 0 : compression_spec.algorithm = PG_COMPRESSION_NONE;
614 0 : ctx->LOsTocFH = InitCompressFileHandle(compression_spec);
615 0 : if (!ctx->LOsTocFH->open_write_func(fname, "ab", ctx->LOsTocFH))
616 0 : pg_fatal("could not open output file \"%s\": %m", fname);
617 0 : }
618 :
619 : /*
620 : * Called by the archiver when we're about to start dumping a LO.
621 : *
622 : * We create a file to write the LO to.
623 : */
624 : static void
625 0 : _StartLO(ArchiveHandle *AH, TocEntry *te, Oid oid)
626 : {
627 0 : lclContext *ctx = (lclContext *) AH->formatData;
628 0 : char fname[MAXPGPATH];
629 :
630 0 : snprintf(fname, MAXPGPATH, "%s/blob_%u.dat", ctx->directory, oid);
631 :
632 0 : ctx->dataFH = InitCompressFileHandle(AH->compression_spec);
633 0 : if (!ctx->dataFH->open_write_func(fname, PG_BINARY_W, ctx->dataFH))
634 0 : pg_fatal("could not open output file \"%s\": %m", fname);
635 0 : }
636 :
637 : /*
638 : * Called by the archiver when the dumper is finished writing a LO.
639 : *
640 : * We close the LO file and write an entry to the LO TOC file for it.
641 : */
642 : static void
643 0 : _EndLO(ArchiveHandle *AH, TocEntry *te, Oid oid)
644 : {
645 0 : lclContext *ctx = (lclContext *) AH->formatData;
646 0 : CompressFileHandle *CFH = ctx->LOsTocFH;
647 0 : char buf[50];
648 0 : int len;
649 :
650 : /* Close the BLOB data file itself */
651 0 : if (!EndCompressFileHandle(ctx->dataFH))
652 0 : pg_fatal("could not close LO data file: %m");
653 0 : ctx->dataFH = NULL;
654 :
655 : /* register the LO in blobs_NNN.toc */
656 0 : len = snprintf(buf, sizeof(buf), "%u blob_%u.dat\n", oid, oid);
657 0 : CFH->write_func(buf, len, CFH);
658 0 : }
659 :
660 : /*
661 : * Called by the archiver when finishing saving BLOB DATA.
662 : *
663 : * We close the LOs TOC file.
664 : */
665 : static void
666 0 : _EndLOs(ArchiveHandle *AH, TocEntry *te)
667 : {
668 0 : lclContext *ctx = (lclContext *) AH->formatData;
669 :
670 0 : if (!EndCompressFileHandle(ctx->LOsTocFH))
671 0 : pg_fatal("could not close LOs TOC file: %m");
672 0 : ctx->LOsTocFH = NULL;
673 0 : }
674 :
675 : /*
676 : * Gets a relative file name and prepends the output directory, writing the
677 : * result to buf. The caller needs to make sure that buf is MAXPGPATH bytes
678 : * big. Can't use a static char[MAXPGPATH] inside the function because we run
679 : * multithreaded on Windows.
680 : */
681 : static void
682 0 : setFilePath(ArchiveHandle *AH, char *buf, const char *relativeFilename)
683 : {
684 0 : lclContext *ctx = (lclContext *) AH->formatData;
685 0 : char *dname;
686 :
687 0 : dname = ctx->directory;
688 :
689 0 : if (strlen(dname) + 1 + strlen(relativeFilename) + 1 > MAXPGPATH)
690 0 : pg_fatal("file name too long: \"%s\"", dname);
691 :
692 0 : strcpy(buf, dname);
693 0 : strcat(buf, "/");
694 0 : strcat(buf, relativeFilename);
695 0 : }
696 :
697 : /*
698 : * Prepare for parallel restore.
699 : *
700 : * The main thing that needs to happen here is to fill in TABLE DATA and BLOBS
701 : * TOC entries' dataLength fields with appropriate values to guide the
702 : * ordering of restore jobs. The source of said data is format-dependent,
703 : * as is the exact meaning of the values.
704 : *
705 : * A format module might also choose to do other setup here.
706 : */
707 : static void
708 0 : _PrepParallelRestore(ArchiveHandle *AH)
709 : {
710 0 : TocEntry *te;
711 :
712 0 : for (te = AH->toc->next; te != AH->toc; te = te->next)
713 : {
714 0 : lclTocEntry *tctx = (lclTocEntry *) te->formatData;
715 0 : char fname[MAXPGPATH];
716 0 : struct stat st;
717 :
718 : /*
719 : * A dumpable object has set tctx->filename, any other object has not.
720 : * (see _ArchiveEntry).
721 : */
722 0 : if (tctx->filename == NULL)
723 0 : continue;
724 :
725 : /* We may ignore items not due to be restored */
726 0 : if ((te->reqs & (REQ_DATA | REQ_STATS)) == 0)
727 0 : continue;
728 :
729 : /*
730 : * Stat the file and, if successful, put its size in dataLength. When
731 : * using compression, the physical file size might not be a very good
732 : * guide to the amount of work involved in restoring the file, but we
733 : * only need an approximate indicator of that.
734 : */
735 0 : setFilePath(AH, fname, tctx->filename);
736 :
737 0 : if (stat(fname, &st) == 0)
738 0 : te->dataLength = st.st_size;
739 0 : else if (AH->compression_spec.algorithm != PG_COMPRESSION_NONE)
740 : {
741 0 : if (AH->compression_spec.algorithm == PG_COMPRESSION_GZIP)
742 0 : strlcat(fname, ".gz", sizeof(fname));
743 0 : else if (AH->compression_spec.algorithm == PG_COMPRESSION_LZ4)
744 0 : strlcat(fname, ".lz4", sizeof(fname));
745 0 : else if (AH->compression_spec.algorithm == PG_COMPRESSION_ZSTD)
746 0 : strlcat(fname, ".zst", sizeof(fname));
747 :
748 0 : if (stat(fname, &st) == 0)
749 0 : te->dataLength = st.st_size;
750 0 : }
751 :
752 : /*
753 : * If this is a BLOBS entry, what we stat'd was blobs_NNN.toc, which
754 : * most likely is a lot smaller than the actual blob data. We don't
755 : * have a cheap way to estimate how much smaller, but fortunately it
756 : * doesn't matter too much as long as we get the LOs processed
757 : * reasonably early. Arbitrarily scale up by a factor of 1K.
758 : */
759 0 : if (strcmp(te->desc, "BLOBS") == 0)
760 0 : te->dataLength *= 1024;
761 0 : }
762 0 : }
763 :
764 : /*
765 : * Clone format-specific fields during parallel restoration.
766 : */
767 : static void
768 0 : _Clone(ArchiveHandle *AH)
769 : {
770 0 : lclContext *ctx = (lclContext *) AH->formatData;
771 :
772 0 : AH->formatData = (lclContext *) pg_malloc(sizeof(lclContext));
773 0 : memcpy(AH->formatData, ctx, sizeof(lclContext));
774 0 : ctx = (lclContext *) AH->formatData;
775 :
776 : /*
777 : * TOC-entry-local state isn't an issue because any one TOC entry is
778 : * touched by just one worker child.
779 : */
780 :
781 : /*
782 : * We also don't copy the ParallelState pointer (pstate), only the leader
783 : * process ever writes to it.
784 : */
785 0 : }
786 :
787 : static void
788 0 : _DeClone(ArchiveHandle *AH)
789 : {
790 0 : lclContext *ctx = (lclContext *) AH->formatData;
791 :
792 0 : free(ctx);
793 0 : }
794 :
795 : /*
796 : * This function is executed in the child of a parallel backup for a
797 : * directory-format archive and dumps the actual data for one TOC entry.
798 : */
799 : static int
800 0 : _WorkerJobDumpDirectory(ArchiveHandle *AH, TocEntry *te)
801 : {
802 : /*
803 : * This function returns void. We either fail and die horribly or
804 : * succeed... A failure will be detected by the parent when the child dies
805 : * unexpectedly.
806 : */
807 0 : WriteDataChunksForTocEntry(AH, te);
808 :
809 0 : return 0;
810 : }
811 :
812 : /*
813 : * This function is executed in the child of a parallel restore from a
814 : * directory-format archive and restores the actual data for one TOC entry.
815 : */
816 : static int
817 0 : _WorkerJobRestoreDirectory(ArchiveHandle *AH, TocEntry *te)
818 : {
819 0 : return parallel_restore(AH, te);
820 : }
|
