summaryrefslogtreecommitdiff
path: root/src/backend/catalog/toasting.c
blob: 29756eb14eb43f2a4f15ad8a7eda00bbdbd3158d (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
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
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
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
439
440
441
442
443
/*-------------------------------------------------------------------------
 *
 * toasting.c
 *	  This file contains routines to support creation of toast tables
 *
 *
 * Portions Copyright (c) 1996-2017, PostgreSQL Global Development Group
 * Portions Copyright (c) 1994, Regents of the University of California
 *
 * IDENTIFICATION
 *	  src/backend/catalog/toasting.c
 *
 *-------------------------------------------------------------------------
 */
#include "postgres.h"

#include "access/tuptoaster.h"
#include "access/xact.h"
#include "catalog/binary_upgrade.h"
#include "catalog/dependency.h"
#include "catalog/heap.h"
#include "catalog/index.h"
#include "catalog/namespace.h"
#include "catalog/pg_am.h"
#include "catalog/pg_namespace.h"
#include "catalog/pg_opclass.h"
#include "catalog/pg_type.h"
#include "catalog/toasting.h"
#include "miscadmin.h"
#include "nodes/makefuncs.h"
#include "storage/lock.h"
#include "utils/builtins.h"
#include "utils/rel.h"
#include "utils/syscache.h"

/* Potentially set by pg_upgrade_support functions */
Oid			binary_upgrade_next_toast_pg_type_oid = InvalidOid;

static void CheckAndCreateToastTable(Oid relOid, Datum reloptions,
						 LOCKMODE lockmode, bool check);
static bool create_toast_table(Relation rel, Oid toastOid, Oid toastIndexOid,
				   Datum reloptions, LOCKMODE lockmode, bool check);
static bool needs_toast_table(Relation rel);


/*
 * CreateToastTable variants
 *		If the table needs a toast table, and doesn't already have one,
 *		then create a toast table for it.
 *
 * reloptions for the toast table can be passed, too.  Pass (Datum) 0
 * for default reloptions.
 *
 * We expect the caller to have verified that the relation is a table and have
 * already done any necessary permission checks.  Callers expect this function
 * to end with CommandCounterIncrement if it makes any changes.
 */
void
AlterTableCreateToastTable(Oid relOid, Datum reloptions, LOCKMODE lockmode)
{
	CheckAndCreateToastTable(relOid, reloptions, lockmode, true);
}

void
NewHeapCreateToastTable(Oid relOid, Datum reloptions, LOCKMODE lockmode)
{
	CheckAndCreateToastTable(relOid, reloptions, lockmode, false);
}

void
NewRelationCreateToastTable(Oid relOid, Datum reloptions)
{
	CheckAndCreateToastTable(relOid, reloptions, AccessExclusiveLock, false);
}

static void
CheckAndCreateToastTable(Oid relOid, Datum reloptions, LOCKMODE lockmode, bool check)
{
	Relation	rel;

	rel = heap_open(relOid, lockmode);

	/* create_toast_table does all the work */
	(void) create_toast_table(rel, InvalidOid, InvalidOid, reloptions, lockmode, check);

	heap_close(rel, NoLock);
}

/*
 * Create a toast table during bootstrap
 *
 * Here we need to prespecify the OIDs of the toast table and its index
 */
void
BootstrapToastTable(char *relName, Oid toastOid, Oid toastIndexOid)
{
	Relation	rel;

	rel = heap_openrv(makeRangeVar(NULL, relName, -1), AccessExclusiveLock);

	if (rel->rd_rel->relkind != RELKIND_RELATION &&
		rel->rd_rel->relkind != RELKIND_MATVIEW)
		ereport(ERROR,
				(errcode(ERRCODE_WRONG_OBJECT_TYPE),
				 errmsg("\"%s\" is not a table or materialized view",
						relName)));

	/* create_toast_table does all the work */
	if (!create_toast_table(rel, toastOid, toastIndexOid, (Datum) 0,
							AccessExclusiveLock, false))
		elog(ERROR, "\"%s\" does not require a toast table",
			 relName);

	heap_close(rel, NoLock);
}


/*
 * create_toast_table --- internal workhorse
 *
 * rel is already opened and locked
 * toastOid and toastIndexOid are normally InvalidOid, but during
 * bootstrap they can be nonzero to specify hand-assigned OIDs
 */
static bool
create_toast_table(Relation rel, Oid toastOid, Oid toastIndexOid,
				   Datum reloptions, LOCKMODE lockmode, bool check)
{
	Oid			relOid = RelationGetRelid(rel);
	HeapTuple	reltup;
	TupleDesc	tupdesc;
	bool		shared_relation;
	bool		mapped_relation;
	Relation	toast_rel;
	Relation	class_rel;
	Oid			toast_relid;
	Oid			toast_typid = InvalidOid;
	Oid			namespaceid;
	char		toast_relname[NAMEDATALEN];
	char		toast_idxname[NAMEDATALEN];
	IndexInfo  *indexInfo;
	Oid			collationObjectId[2];
	Oid			classObjectId[2];
	int16		coloptions[2];
	ObjectAddress baseobject,
				toastobject;

	/*
	 * Toast table is shared if and only if its parent is.
	 *
	 * We cannot allow toasting a shared relation after initdb (because
	 * there's no way to mark it toasted in other databases' pg_class).
	 */
	shared_relation = rel->rd_rel->relisshared;
	if (shared_relation && !IsBootstrapProcessingMode())
		ereport(ERROR,
				(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
				 errmsg("shared tables cannot be toasted after initdb")));

	/* It's mapped if and only if its parent is, too */
	mapped_relation = RelationIsMapped(rel);

	/*
	 * Is it already toasted?
	 */
	if (rel->rd_rel->reltoastrelid != InvalidOid)
		return false;

	/*
	 * Check to see whether the table actually needs a TOAST table.
	 */
	if (!IsBinaryUpgrade)
	{
		/* Normal mode, normal check */
		if (!needs_toast_table(rel))
			return false;
	}
	else
	{
		/*
		 * In binary-upgrade mode, create a TOAST table if and only if
		 * pg_upgrade told us to (ie, a TOAST table OID has been provided).
		 *
		 * This indicates that the old cluster had a TOAST table for the
		 * current table.  We must create a TOAST table to receive the old
		 * TOAST file, even if the table seems not to need one.
		 *
		 * Contrariwise, if the old cluster did not have a TOAST table, we
		 * should be able to get along without one even if the new version's
		 * needs_toast_table rules suggest we should have one.  There is a lot
		 * of daylight between where we will create a TOAST table and where
		 * one is really necessary to avoid failures, so small cross-version
		 * differences in the when-to-create heuristic shouldn't be a problem.
		 * If we tried to create a TOAST table anyway, we would have the
		 * problem that it might take up an OID that will conflict with some
		 * old-cluster table we haven't seen yet.
		 */
		if (!OidIsValid(binary_upgrade_next_toast_pg_class_oid) ||
			!OidIsValid(binary_upgrade_next_toast_pg_type_oid))
			return false;
	}

	/*
	 * If requested check lockmode is sufficient. This is a cross check in
	 * case of errors or conflicting decisions in earlier code.
	 */
	if (check && lockmode != AccessExclusiveLock)
		elog(ERROR, "AccessExclusiveLock required to add toast table.");

	/*
	 * Create the toast table and its index
	 */
	snprintf(toast_relname, sizeof(toast_relname),
			 "pg_toast_%u", relOid);
	snprintf(toast_idxname, sizeof(toast_idxname),
			 "pg_toast_%u_index", relOid);

	/* this is pretty painful...  need a tuple descriptor */
	tupdesc = CreateTemplateTupleDesc(3, false);
	TupleDescInitEntry(tupdesc, (AttrNumber) 1,
					   "chunk_id",
					   OIDOID,
					   -1, 0);
	TupleDescInitEntry(tupdesc, (AttrNumber) 2,
					   "chunk_seq",
					   INT4OID,
					   -1, 0);
	TupleDescInitEntry(tupdesc, (AttrNumber) 3,
					   "chunk_data",
					   BYTEAOID,
					   -1, 0);

	/*
	 * Ensure that the toast table doesn't itself get toasted, or we'll be
	 * toast :-(.  This is essential for chunk_data because type bytea is
	 * toastable; hit the other two just to be sure.
	 */
	tupdesc->attrs[0]->attstorage = 'p';
	tupdesc->attrs[1]->attstorage = 'p';
	tupdesc->attrs[2]->attstorage = 'p';

	/*
	 * Toast tables for regular relations go in pg_toast; those for temp
	 * relations go into the per-backend temp-toast-table namespace.
	 */
	if (isTempOrTempToastNamespace(rel->rd_rel->relnamespace))
		namespaceid = GetTempToastNamespace();
	else
		namespaceid = PG_TOAST_NAMESPACE;

	/*
	 * Use binary-upgrade override for pg_type.oid, if supplied.  We might be
	 * in the post-schema-restore phase where we are doing ALTER TABLE to
	 * create TOAST tables that didn't exist in the old cluster.
	 */
	if (IsBinaryUpgrade && OidIsValid(binary_upgrade_next_toast_pg_type_oid))
	{
		toast_typid = binary_upgrade_next_toast_pg_type_oid;
		binary_upgrade_next_toast_pg_type_oid = InvalidOid;
	}

	toast_relid = heap_create_with_catalog(toast_relname,
										   namespaceid,
										   rel->rd_rel->reltablespace,
										   toastOid,
										   toast_typid,
										   InvalidOid,
										   rel->rd_rel->relowner,
										   tupdesc,
										   NIL,
										   RELKIND_TOASTVALUE,
										   rel->rd_rel->relpersistence,
										   shared_relation,
										   mapped_relation,
										   true,
										   0,
										   ONCOMMIT_NOOP,
										   reloptions,
										   false,
										   true,
										   true,
										   NULL);
	Assert(toast_relid != InvalidOid);

	/* make the toast relation visible, else heap_open will fail */
	CommandCounterIncrement();

	/* ShareLock is not really needed here, but take it anyway */
	toast_rel = heap_open(toast_relid, ShareLock);

	/*
	 * Create unique index on chunk_id, chunk_seq.
	 *
	 * NOTE: the normal TOAST access routines could actually function with a
	 * single-column index on chunk_id only. However, the slice access
	 * routines use both columns for faster access to an individual chunk. In
	 * addition, we want it to be unique as a check against the possibility of
	 * duplicate TOAST chunk OIDs. The index might also be a little more
	 * efficient this way, since btree isn't all that happy with large numbers
	 * of equal keys.
	 */

	indexInfo = makeNode(IndexInfo);
	indexInfo->ii_NumIndexAttrs = 2;
	indexInfo->ii_KeyAttrNumbers[0] = 1;
	indexInfo->ii_KeyAttrNumbers[1] = 2;
	indexInfo->ii_Expressions = NIL;
	indexInfo->ii_ExpressionsState = NIL;
	indexInfo->ii_Predicate = NIL;
	indexInfo->ii_PredicateState = NULL;
	indexInfo->ii_ExclusionOps = NULL;
	indexInfo->ii_ExclusionProcs = NULL;
	indexInfo->ii_ExclusionStrats = NULL;
	indexInfo->ii_Unique = true;
	indexInfo->ii_ReadyForInserts = true;
	indexInfo->ii_Concurrent = false;
	indexInfo->ii_BrokenHotChain = false;
	indexInfo->ii_AmCache = NULL;
	indexInfo->ii_Context = CurrentMemoryContext;

	collationObjectId[0] = InvalidOid;
	collationObjectId[1] = InvalidOid;

	classObjectId[0] = OID_BTREE_OPS_OID;
	classObjectId[1] = INT4_BTREE_OPS_OID;

	coloptions[0] = 0;
	coloptions[1] = 0;

	index_create(toast_rel, toast_idxname, toastIndexOid, InvalidOid,
				 indexInfo,
				 list_make2("chunk_id", "chunk_seq"),
				 BTREE_AM_OID,
				 rel->rd_rel->reltablespace,
				 collationObjectId, classObjectId, coloptions, (Datum) 0,
				 true, false, false, false,
				 true, false, false, true, false);

	heap_close(toast_rel, NoLock);

	/*
	 * Store the toast table's OID in the parent relation's pg_class row
	 */
	class_rel = heap_open(RelationRelationId, RowExclusiveLock);

	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relOid));
	if (!HeapTupleIsValid(reltup))
		elog(ERROR, "cache lookup failed for relation %u", relOid);

	((Form_pg_class) GETSTRUCT(reltup))->reltoastrelid = toast_relid;

	if (!IsBootstrapProcessingMode())
	{
		/* normal case, use a transactional update */
		CatalogTupleUpdate(class_rel, &reltup->t_self, reltup);
	}
	else
	{
		/* While bootstrapping, we cannot UPDATE, so overwrite in-place */
		heap_inplace_update(class_rel, reltup);
	}

	heap_freetuple(reltup);

	heap_close(class_rel, RowExclusiveLock);

	/*
	 * Register dependency from the toast table to the master, so that the
	 * toast table will be deleted if the master is.  Skip this in bootstrap
	 * mode.
	 */
	if (!IsBootstrapProcessingMode())
	{
		baseobject.classId = RelationRelationId;
		baseobject.objectId = relOid;
		baseobject.objectSubId = 0;
		toastobject.classId = RelationRelationId;
		toastobject.objectId = toast_relid;
		toastobject.objectSubId = 0;

		recordDependencyOn(&toastobject, &baseobject, DEPENDENCY_INTERNAL);
	}

	/*
	 * Make changes visible
	 */
	CommandCounterIncrement();

	return true;
}

/*
 * Check to see whether the table needs a TOAST table.  It does only if
 * (1) there are any toastable attributes, and (2) the maximum length
 * of a tuple could exceed TOAST_TUPLE_THRESHOLD.  (We don't want to
 * create a toast table for something like "f1 varchar(20)".)
 */
static bool
needs_toast_table(Relation rel)
{
	int32		data_length = 0;
	bool		maxlength_unknown = false;
	bool		has_toastable_attrs = false;
	TupleDesc	tupdesc;
	Form_pg_attribute *att;
	int32		tuple_length;
	int			i;

	tupdesc = rel->rd_att;
	att = tupdesc->attrs;

	for (i = 0; i < tupdesc->natts; i++)
	{
		if (att[i]->attisdropped)
			continue;
		data_length = att_align_nominal(data_length, att[i]->attalign);
		if (att[i]->attlen > 0)
		{
			/* Fixed-length types are never toastable */
			data_length += att[i]->attlen;
		}
		else
		{
			int32		maxlen = type_maximum_size(att[i]->atttypid,
												   att[i]->atttypmod);

			if (maxlen < 0)
				maxlength_unknown = true;
			else
				data_length += maxlen;
			if (att[i]->attstorage != 'p')
				has_toastable_attrs = true;
		}
	}
	if (!has_toastable_attrs)
		return false;			/* nothing to toast? */
	if (maxlength_unknown)
		return true;			/* any unlimited-length attrs? */
	tuple_length = MAXALIGN(SizeofHeapTupleHeader +
							BITMAPLEN(tupdesc->natts)) +
		MAXALIGN(data_length);
	return (tuple_length > TOAST_TUPLE_THRESHOLD);
}