/ Check-in [7bd44794]
Login
SQLite training in Houston TX on 2019-11-05 (details)
Part of the 2019 Tcl Conference

Many hyperlinks are disabled.
Use anonymous login to enable hyperlinks.

Overview
Comment:Prohibit backup if the destination is using WAL and has a different page size from the source.
Downloads: Tarball | ZIP archive | SQL archive
Timelines: family | ancestors | descendants | both | trunk
Files: files | file ages | folders
SHA1: 7bd44794c482beee16c684712545275e2bf63dfa
User & Date: drh 2010-05-05 16:23:27
References
2010-05-05
18:46
Do not compare page sizes on source and destination of backup until transactions are started and the page sizes are locked. This is a fix to check-in [7bd44794c4]. check-in: ec715778 user: drh tags: trunk
Context
2010-05-05
18:20
Simplifications to the SHM locking implementation in os_unix.c. check-in: 9de05bfb user: drh tags: trunk
16:23
Prohibit backup if the destination is using WAL and has a different page size from the source. check-in: 7bd44794 user: drh tags: trunk
15:33
When closing a WAL database, if the exclusive lock on the database file is obtained and the database successfully checkpointed, delete the wal-index file from the file system. check-in: 2ac5d96c user: dan tags: trunk
Changes
Hide Diffs Unified Diffs Ignore Whitespace Patch

Changes to src/backup.c.

213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
...
283
284
285
286
287
288
289



290
291
292
293
294
295
296








297


298
299
300
301
302
303
304
...
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
...
379
380
381
382
383
384
385
386
387


388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
  assert( !isFatalError(p->rc) );
  assert( iSrcPg!=PENDING_BYTE_PAGE(p->pSrc->pBt) );
  assert( zSrcData );

  /* Catch the case where the destination is an in-memory database and the
  ** page sizes of the source and destination differ. 
  */
  if( nSrcPgsz!=nDestPgsz && sqlite3PagerIsMemdb(sqlite3BtreePager(p->pDest)) ){
    rc = SQLITE_READONLY;
  }

  /* This loop runs once for each destination page spanned by the source 
  ** page. For each iteration, variable iOff is set to the byte offset
  ** of the destination page.
  */
................................................................................
}

/*
** Copy nPage pages from the source b-tree to the destination.
*/
int sqlite3_backup_step(sqlite3_backup *p, int nPage){
  int rc;




  sqlite3_mutex_enter(p->pSrcDb->mutex);
  sqlite3BtreeEnter(p->pSrc);
  if( p->pDestDb ){
    sqlite3_mutex_enter(p->pDestDb->mutex);
  }









  rc = p->rc;


  if( !isFatalError(rc) ){
    Pager * const pSrcPager = sqlite3BtreePager(p->pSrc);     /* Source pager */
    Pager * const pDestPager = sqlite3BtreePager(p->pDest);   /* Dest pager */
    int ii;                            /* Iterator variable */
    int nSrcPage = -1;                 /* Size of source db in pages */
    int bCloseTrans = 0;               /* True if src db requires unlocking */

................................................................................
    ** is to make sure that the schema-version really does change in
    ** the case where the source and destination databases have the
    ** same schema version.
    */
    if( rc==SQLITE_DONE 
     && (rc = sqlite3BtreeUpdateMeta(p->pDest,1,p->iDestSchema+1))==SQLITE_OK
    ){
      const int nSrcPagesize = sqlite3BtreeGetPageSize(p->pSrc);
      const int nDestPagesize = sqlite3BtreeGetPageSize(p->pDest);
      int nDestTruncate;
  
      if( p->pDestDb ){
        sqlite3ResetInternalSchema(p->pDestDb, 0);
      }

      /* Set nDestTruncate to the final number of pages in the destination
................................................................................
      ** round up. In this case the call to sqlite3OsTruncate() below will
      ** fix the size of the file. However it is important to call
      ** sqlite3PagerTruncateImage() here so that any pages in the 
      ** destination file that lie beyond the nDestTruncate page mark are
      ** journalled by PagerCommitPhaseOne() before they are destroyed
      ** by the file truncation.
      */
      if( nSrcPagesize<nDestPagesize ){
        int ratio = nDestPagesize/nSrcPagesize;


        nDestTruncate = (nSrcPage+ratio-1)/ratio;
        if( nDestTruncate==(int)PENDING_BYTE_PAGE(p->pDest->pBt) ){
          nDestTruncate--;
        }
      }else{
        nDestTruncate = nSrcPage * (nSrcPagesize/nDestPagesize);
      }
      sqlite3PagerTruncateImage(pDestPager, nDestTruncate);

      if( nSrcPagesize<nDestPagesize ){
        /* If the source page-size is smaller than the destination page-size,
        ** two extra things may need to happen:
        **
        **   * The destination may need to be truncated, and
        **
        **   * Data stored on the pages immediately following the 
        **     pending-byte page in the source database may need to be
        **     copied into the destination database.
        */
        const i64 iSize = (i64)nSrcPagesize * (i64)nSrcPage;
        sqlite3_file * const pFile = sqlite3PagerFile(pDestPager);

        assert( pFile );
        assert( (i64)nDestTruncate*(i64)nDestPagesize >= iSize || (
              nDestTruncate==(int)(PENDING_BYTE_PAGE(p->pDest->pBt)-1)
           && iSize>=PENDING_BYTE && iSize<=PENDING_BYTE+nDestPagesize
        ));
        if( SQLITE_OK==(rc = sqlite3PagerCommitPhaseOne(pDestPager, 0, 1))
         && SQLITE_OK==(rc = backupTruncateFile(pFile, iSize))
         && SQLITE_OK==(rc = sqlite3PagerSync(pDestPager))
        ){
          i64 iOff;
          i64 iEnd = MIN(PENDING_BYTE + nDestPagesize, iSize);
          for(
            iOff=PENDING_BYTE+nSrcPagesize; 
            rc==SQLITE_OK && iOff<iEnd; 
            iOff+=nSrcPagesize
          ){
            PgHdr *pSrcPg = 0;
            const Pgno iSrcPg = (Pgno)((iOff/nSrcPagesize)+1);
            rc = sqlite3PagerGet(pSrcPager, iSrcPg, &pSrcPg);
            if( rc==SQLITE_OK ){
              u8 *zData = sqlite3PagerGetData(pSrcPg);
              rc = sqlite3OsWrite(pFile, zData, nSrcPagesize, iOff);
            }
            sqlite3PagerUnref(pSrcPg);
          }
        }
      }else{
        rc = sqlite3PagerCommitPhaseOne(pDestPager, 0, 0);
      }







|







 







>
>
>







>
>
>
>
>
>
>
>
|
>
>







 







<
<







 







|
|
>
>





|



|









|



|

|






|

|

|


|



|







213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
...
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
...
372
373
374
375
376
377
378


379
380
381
382
383
384
385
...
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
  assert( !isFatalError(p->rc) );
  assert( iSrcPg!=PENDING_BYTE_PAGE(p->pSrc->pBt) );
  assert( zSrcData );

  /* Catch the case where the destination is an in-memory database and the
  ** page sizes of the source and destination differ. 
  */
  if( nSrcPgsz!=nDestPgsz && sqlite3PagerIsMemdb(pDestPager) ){
    rc = SQLITE_READONLY;
  }

  /* This loop runs once for each destination page spanned by the source 
  ** page. For each iteration, variable iOff is set to the byte offset
  ** of the destination page.
  */
................................................................................
}

/*
** Copy nPage pages from the source b-tree to the destination.
*/
int sqlite3_backup_step(sqlite3_backup *p, int nPage){
  int rc;
  int destMode;       /* Destination journal mode */
  int pgszSrc;        /* Source page size */
  int pgszDest;       /* Destination page size */

  sqlite3_mutex_enter(p->pSrcDb->mutex);
  sqlite3BtreeEnter(p->pSrc);
  if( p->pDestDb ){
    sqlite3_mutex_enter(p->pDestDb->mutex);
  }

  /* Do not allow backup if the destination database is in WAL mode */
  destMode = sqlite3PagerJournalMode(sqlite3BtreePager(p->pDest),
                                     PAGER_JOURNALMODE_QUERY);
  pgszSrc = sqlite3BtreeGetPageSize(p->pSrc);
  pgszDest = sqlite3BtreeGetPageSize(p->pDest);
  if( destMode==PAGER_JOURNALMODE_WAL && pgszSrc!=pgszDest ){
    rc = SQLITE_READONLY;
  }else{
    rc = p->rc;
  }

  if( !isFatalError(rc) ){
    Pager * const pSrcPager = sqlite3BtreePager(p->pSrc);     /* Source pager */
    Pager * const pDestPager = sqlite3BtreePager(p->pDest);   /* Dest pager */
    int ii;                            /* Iterator variable */
    int nSrcPage = -1;                 /* Size of source db in pages */
    int bCloseTrans = 0;               /* True if src db requires unlocking */

................................................................................
    ** is to make sure that the schema-version really does change in
    ** the case where the source and destination databases have the
    ** same schema version.
    */
    if( rc==SQLITE_DONE 
     && (rc = sqlite3BtreeUpdateMeta(p->pDest,1,p->iDestSchema+1))==SQLITE_OK
    ){


      int nDestTruncate;
  
      if( p->pDestDb ){
        sqlite3ResetInternalSchema(p->pDestDb, 0);
      }

      /* Set nDestTruncate to the final number of pages in the destination
................................................................................
      ** round up. In this case the call to sqlite3OsTruncate() below will
      ** fix the size of the file. However it is important to call
      ** sqlite3PagerTruncateImage() here so that any pages in the 
      ** destination file that lie beyond the nDestTruncate page mark are
      ** journalled by PagerCommitPhaseOne() before they are destroyed
      ** by the file truncation.
      */
      assert( pgszSrc==sqlite3BtreeGetPageSize(p->pSrc) );
      assert( pgszDest==sqlite3BtreeGetPageSize(p->pDest) );
      if( pgszSrc<pgszDest ){
        int ratio = pgszDest/pgszSrc;
        nDestTruncate = (nSrcPage+ratio-1)/ratio;
        if( nDestTruncate==(int)PENDING_BYTE_PAGE(p->pDest->pBt) ){
          nDestTruncate--;
        }
      }else{
        nDestTruncate = nSrcPage * (pgszSrc/pgszDest);
      }
      sqlite3PagerTruncateImage(pDestPager, nDestTruncate);

      if( pgszSrc<pgszDest ){
        /* If the source page-size is smaller than the destination page-size,
        ** two extra things may need to happen:
        **
        **   * The destination may need to be truncated, and
        **
        **   * Data stored on the pages immediately following the 
        **     pending-byte page in the source database may need to be
        **     copied into the destination database.
        */
        const i64 iSize = (i64)pgszSrc * (i64)nSrcPage;
        sqlite3_file * const pFile = sqlite3PagerFile(pDestPager);

        assert( pFile );
        assert( (i64)nDestTruncate*(i64)pgszDest >= iSize || (
              nDestTruncate==(int)(PENDING_BYTE_PAGE(p->pDest->pBt)-1)
           && iSize>=PENDING_BYTE && iSize<=PENDING_BYTE+pgszDest
        ));
        if( SQLITE_OK==(rc = sqlite3PagerCommitPhaseOne(pDestPager, 0, 1))
         && SQLITE_OK==(rc = backupTruncateFile(pFile, iSize))
         && SQLITE_OK==(rc = sqlite3PagerSync(pDestPager))
        ){
          i64 iOff;
          i64 iEnd = MIN(PENDING_BYTE + pgszDest, iSize);
          for(
            iOff=PENDING_BYTE+pgszSrc; 
            rc==SQLITE_OK && iOff<iEnd; 
            iOff+=pgszSrc
          ){
            PgHdr *pSrcPg = 0;
            const Pgno iSrcPg = (Pgno)((iOff/pgszSrc)+1);
            rc = sqlite3PagerGet(pSrcPager, iSrcPg, &pSrcPg);
            if( rc==SQLITE_OK ){
              u8 *zData = sqlite3PagerGetData(pSrcPg);
              rc = sqlite3OsWrite(pFile, zData, pgszSrc, iOff);
            }
            sqlite3PagerUnref(pSrcPg);
          }
        }
      }else{
        rc = sqlite3PagerCommitPhaseOne(pDestPager, 0, 0);
      }

Changes to src/sqlite.h.in.

5485
5486
5487
5488
5489
5490
5491
5492

5493
5494
5495




5496
5497
5498
5499
5500
5501
5502
** from source to destination, then it returns [SQLITE_DONE].
** ^If an error occurs while running sqlite3_backup_step(B,N),
** then an [error code] is returned. ^As well as [SQLITE_OK] and
** [SQLITE_DONE], a call to sqlite3_backup_step() may return [SQLITE_READONLY],
** [SQLITE_NOMEM], [SQLITE_BUSY], [SQLITE_LOCKED], or an
** [SQLITE_IOERR_ACCESS | SQLITE_IOERR_XXX] extended error code.
**
** ^The sqlite3_backup_step() might return [SQLITE_READONLY] if the destination

** database was opened read-only or if
** the destination is an in-memory database with a different page size
** from the source database.




**
** ^If sqlite3_backup_step() cannot obtain a required file-system lock, then
** the [sqlite3_busy_handler | busy-handler function]
** is invoked (if one is specified). ^If the 
** busy-handler returns non-zero before the lock is available, then 
** [SQLITE_BUSY] is returned to the caller. ^In this case the call to
** sqlite3_backup_step() can be retried later. ^If the source







|
>
|
<
|
>
>
>
>







5485
5486
5487
5488
5489
5490
5491
5492
5493
5494

5495
5496
5497
5498
5499
5500
5501
5502
5503
5504
5505
5506
** from source to destination, then it returns [SQLITE_DONE].
** ^If an error occurs while running sqlite3_backup_step(B,N),
** then an [error code] is returned. ^As well as [SQLITE_OK] and
** [SQLITE_DONE], a call to sqlite3_backup_step() may return [SQLITE_READONLY],
** [SQLITE_NOMEM], [SQLITE_BUSY], [SQLITE_LOCKED], or an
** [SQLITE_IOERR_ACCESS | SQLITE_IOERR_XXX] extended error code.
**
** ^(The sqlite3_backup_step() might return [SQLITE_READONLY] if
** <ol>
** <li> the destination database was opened read-only, or

** <li> the destination database is using write-ahead-log journaling
** and the destination and source page sizes differ, or
** <li> The destination database is an in-memory database and the
** destination and source page sizes differ.
** </ol>)^
**
** ^If sqlite3_backup_step() cannot obtain a required file-system lock, then
** the [sqlite3_busy_handler | busy-handler function]
** is invoked (if one is specified). ^If the 
** busy-handler returns non-zero before the lock is available, then 
** [SQLITE_BUSY] is returned to the caller. ^In this case the call to
** sqlite3_backup_step() can be retried later. ^If the source