/*
* IsSystemRelation
* True iff the relation is a system catalog relation.
*
* NB: TOAST relations are considered system relations by this test
* for compatibility with the old IsSystemRelationName function.
* This is appropriate in many places but not all. Where it's not,
* also check IsToastRelation.
*
* We now just test if the relation is in the system catalog namespace;
* so it's no longer necessary to forbid user relations from having
* names starting with pg_.
*/
bool
IsSystemRelation(Relation relation)
{
return IsSystemNamespace(RelationGetNamespace(relation)) ||
IsToastNamespace(RelationGetNamespace(relation)) ||
IsAoSegmentNamespace(RelationGetNamespace(relation));
}
/*
* DropErrorTable
*
* Drop the error table from the database. This function will be called from
* destroyCdbSreh when an autogenerated error table was not used in the COPY
* operation granted KEEP wasn't specified.
*
*/
static
void DropErrorTable(CdbSreh *cdbsreh)
{
StringInfoData dropstmt;
RangeVar *errtbl_rv;
Insist(Gp_role == GP_ROLE_DISPATCH);
ereport(NOTICE,
(errcode(ERRCODE_SUCCESSFUL_COMPLETION),
errmsg("Dropping the auto-generated unused error table"),
errhint("Use KEEP in LOG INTO clause to force keeping the error table alive")));
initStringInfo(&dropstmt);
appendStringInfo(&dropstmt, "DROP TABLE %s.%s",
quote_identifier(get_namespace_name(RelationGetNamespace(cdbsreh->errtbl))),
quote_identifier(RelationGetRelationName(cdbsreh->errtbl)));
errtbl_rv = makeRangeVar(get_namespace_name(RelationGetNamespace(cdbsreh->errtbl)),
RelationGetRelationName(cdbsreh->errtbl), -1);
/* DROP the relation on the QD */
RemoveRelation(errtbl_rv,DROP_RESTRICT, NULL, RELKIND_RELATION);
/* dispatch the DROP to the QEs */
CdbDoCommand(dropstmt.data, false, /*no txn */ false);
pfree(dropstmt.data);
}
/*
* AppendOnlySegmentFileTruncateToEOF()
*
* Assumes that the segment file lock is already held.
*
* For the segment file is truncates to the eof.
*/
static void
AppendOnlySegmentFileTruncateToEOF(Relation aorel,
FileSegInfo *fsinfo)
{
const char* relname = RelationGetRelationName(aorel);
MirroredAppendOnlyOpen mirroredOpened;
int32 fileSegNo;
char filenamepath[MAXPGPATH];
int segno;
int64 segeof;
Assert(fsinfo);
Assert(RelationIsAoRows(aorel));
segno = fsinfo->segno;
relname = RelationGetRelationName(aorel);
segeof = (int64)fsinfo->eof;
/* Open and truncate the relation segfile beyond its eof */
MakeAOSegmentFileName(aorel, segno, -1, &fileSegNo, filenamepath);
elogif(Debug_appendonly_print_compaction, LOG,
"Opening AO relation \"%s.%s\", relation id %u, relfilenode %u (physical segment file #%d, logical EOF " INT64_FORMAT ")",
get_namespace_name(RelationGetNamespace(aorel)),
relname,
aorel->rd_id,
aorel->rd_node.relNode,
segno,
segeof);
if (OpenAOSegmentFile(aorel, filenamepath, fileSegNo, segeof, &mirroredOpened))
{
TruncateAOSegmentFile(&mirroredOpened, aorel, segeof, ERROR);
CloseAOSegmentFile(&mirroredOpened);
elogif(Debug_appendonly_print_compaction, LOG,
"Successfully truncated AO ROL relation \"%s.%s\", relation id %u, relfilenode %u (physical segment file #%d, logical EOF " INT64_FORMAT ")",
get_namespace_name(RelationGetNamespace(aorel)),
relname,
aorel->rd_id,
aorel->rd_node.relNode,
segno,
segeof);
}
else
{
elogif(Debug_appendonly_print_compaction, LOG,
"No gp_relation_node entry for AO ROW relation \"%s.%s\", relation id %u, relfilenode %u (physical segment file #%d, logical EOF " INT64_FORMAT ")",
get_namespace_name(RelationGetNamespace(aorel)),
relname,
aorel->rd_id,
aorel->rd_node.relNode,
segno,
segeof);
}
}
/*
* Get the max size of the relation across segments
*/
int64
cdbRelMaxSegSize(Relation rel)
{
int64 size = 0;
int i;
CdbPgResults cdb_pgresults = {NULL, 0};
StringInfoData buffer;
char *schemaName;
char *relName;
/*
* Let's ask the QEs for the size of the relation
*/
initStringInfo(&buffer);
schemaName = get_namespace_name(RelationGetNamespace(rel));
if (schemaName == NULL)
elog(ERROR, "cache lookup failed for namespace %d",
RelationGetNamespace(rel));
relName = RelationGetRelationName(rel);
/*
* Safer to pass names than oids, just in case they get out of sync between QD and QE,
* which might happen with a toast table or index, I think (but maybe not)
*/
appendStringInfo(&buffer, "select pg_relation_size('%s.%s')",
quote_identifier(schemaName), quote_identifier(relName));
CdbDispatchCommand(buffer.data, DF_WITH_SNAPSHOT, &cdb_pgresults);
for (i = 0; i < cdb_pgresults.numResults; i++)
{
struct pg_result * pgresult = cdb_pgresults.pg_results[i];
if (PQresultStatus(pgresult) != PGRES_TUPLES_OK)
{
cdbdisp_clearCdbPgResults(&cdb_pgresults);
elog(ERROR,"cdbRelMaxSegSize: resultStatus not tuples_Ok: %s %s",
PQresStatus(PQresultStatus(pgresult)),PQresultErrorMessage(pgresult));
}
else
{
Assert(PQntuples(pgresult) == 1);
int64 tempsize = 0;
(void) scanint8(PQgetvalue(pgresult, 0, 0), false, &tempsize);
if (tempsize > size)
size = tempsize;
}
}
pfree(buffer.data);
cdbdisp_clearCdbPgResults(&cdb_pgresults);
return size;
}
/*
* OpenErrorTable
*
* Open the error table for this operation, and perform all necessary checks.
*/
void OpenErrorTable(CdbSreh *cdbsreh, RangeVar *errortable)
{
AclResult aclresult;
AclMode required_access = ACL_INSERT;
Oid relOid;
/* Open and lock the error relation, using the appropriate lock type. */
cdbsreh->errtbl = heap_openrv(errortable, RowExclusiveLock);
relOid = RelationGetRelid(cdbsreh->errtbl);
/* Check relation permissions. */
aclresult = pg_class_aclcheck(relOid,
GetUserId(),
required_access);
if (aclresult != ACLCHECK_OK)
aclcheck_error(aclresult, ACL_KIND_CLASS,
RelationGetRelationName(cdbsreh->errtbl));
/* check read-only transaction */
if (XactReadOnly &&
!isTempNamespace(RelationGetNamespace(cdbsreh->errtbl)))
ereport(ERROR,
(errcode(ERRCODE_READ_ONLY_SQL_TRANSACTION),
errmsg("transaction is read-only")));
/* make sure this is a regular relation */
if (cdbsreh->errtbl->rd_rel->relkind != RELKIND_RELATION)
ereport(ERROR,
(errcode(ERRCODE_WRONG_OBJECT_TYPE),
errmsg("\"%s\" exists in the database but is a non table relation",
RelationGetRelationName(cdbsreh->errtbl))));
}
/*
* btinsert() -- insert an index tuple into a btree.
*
* Descend the tree recursively, find the appropriate location for our
* new tuple, and put it there.
*/
Datum
btinsert(PG_FUNCTION_ARGS)
{
MIRROREDLOCK_BUFMGR_VERIFY_NO_LOCK_LEAK_DECLARE;
Relation rel = (Relation) PG_GETARG_POINTER(0);
Datum *values = (Datum *) PG_GETARG_POINTER(1);
bool *isnull = (bool *) PG_GETARG_POINTER(2);
ItemPointer ht_ctid = (ItemPointer) PG_GETARG_POINTER(3);
Relation heapRel = (Relation) PG_GETARG_POINTER(4);
bool checkUnique = PG_GETARG_BOOL(5);
IndexTuple itup;
MIRROREDLOCK_BUFMGR_VERIFY_NO_LOCK_LEAK_ENTER;
if (checkUnique && (
(gp_indexcheck_insert == INDEX_CHECK_ALL && RelationIsHeap(heapRel)) ||
(gp_indexcheck_insert == INDEX_CHECK_SYSTEM &&
PG_CATALOG_NAMESPACE == RelationGetNamespace(heapRel))))
{
_bt_validate_tid(rel, ht_ctid);
}
/* generate an index tuple */
itup = index_form_tuple(RelationGetDescr(rel), values, isnull);
itup->t_tid = *ht_ctid;
_bt_doinsert(rel, itup, checkUnique, heapRel);
pfree(itup);
MIRROREDLOCK_BUFMGR_VERIFY_NO_LOCK_LEAK_EXIT;
PG_RETURN_BOOL(true);
}
/*---------------------------------------------------------------------------
* This cluster code allows for clustering multiple tables at once. Because
* of this, we cannot just run everything on a single transaction, or we
* would be forced to acquire exclusive locks on all the tables being
* clustered, simultaneously --- very likely leading to deadlock.
*
* To solve this we follow a similar strategy to VACUUM code,
* clustering each relation in a separate transaction. For this to work,
* we need to:
* - provide a separate memory context so that we can pass information in
* a way that survives across transactions
* - start a new transaction every time a new relation is clustered
* - check for validity of the information on to-be-clustered relations,
* as someone might have deleted a relation behind our back, or
* clustered one on a different index
* - end the transaction
*
* The single-relation case does not have any such overhead.
*
* We also allow a relation to be specified without index. In that case,
* the indisclustered bit will be looked up, and an ERROR will be thrown
* if there is no index with the bit set.
*---------------------------------------------------------------------------
*/
void
cluster(ClusterStmt *stmt, bool isTopLevel)
{
if (stmt->relation != NULL)
{
/* This is the single-relation case. */
Oid tableOid,
indexOid = InvalidOid;
Relation rel;
RelToCluster rvtc;
/* Find and lock the table */
rel = heap_openrv(stmt->relation, AccessExclusiveLock);
tableOid = RelationGetRelid(rel);
/* Check permissions */
if (!pg_class_ownercheck(tableOid, GetUserId()))
aclcheck_error(ACLCHECK_NOT_OWNER, ACL_KIND_CLASS,
RelationGetRelationName(rel));
/*
* Reject clustering a remote temp table ... their local buffer
* manager is not going to cope.
*/
if (isOtherTempNamespace(RelationGetNamespace(rel)))
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("cannot cluster temporary tables of other sessions")));
if (stmt->indexname == NULL)
{
ListCell *index;
/* We need to find the index that has indisclustered set. */
foreach(index, RelationGetIndexList(rel))
{
HeapTuple idxtuple;
Form_pg_index indexForm;
indexOid = lfirst_oid(index);
idxtuple = SearchSysCache(INDEXRELID,
ObjectIdGetDatum(indexOid),
0, 0, 0);
if (!HeapTupleIsValid(idxtuple))
elog(ERROR, "cache lookup failed for index %u", indexOid);
indexForm = (Form_pg_index) GETSTRUCT(idxtuple);
if (indexForm->indisclustered)
{
ReleaseSysCache(idxtuple);
break;
}
ReleaseSysCache(idxtuple);
indexOid = InvalidOid;
}
/*
* GetNewOid
* Generate a new OID that is unique within the given relation.
*
* Caller must have a suitable lock on the relation.
*
* Uniqueness is promised only if the relation has a unique index on OID.
* This is true for all system catalogs that have OIDs, but might not be
* true for user tables. Note that we are effectively assuming that the
* table has a relatively small number of entries (much less than 2^32)
* and there aren't very long runs of consecutive existing OIDs. Again,
* this is reasonable for system catalogs but less so for user tables.
*
* Since the OID is not immediately inserted into the table, there is a
* race condition here; but a problem could occur only if someone else
* managed to cycle through 2^32 OIDs and generate the same OID before we
* finish inserting our row. This seems unlikely to be a problem. Note
* that if we had to *commit* the row to end the race condition, the risk
* would be rather higher; therefore we use SnapshotDirty in the test,
* so that we will see uncommitted rows.
*/
Oid
GetNewOid(Relation relation)
{
Oid newOid;
Oid oidIndex;
Relation indexrel;
/* If relation doesn't have OIDs at all, caller is confused */
Assert(relation->rd_rel->relhasoids);
/* In bootstrap mode, we don't have any indexes to use */
if (IsBootstrapProcessingMode())
return GetNewObjectId();
/* The relcache will cache the identity of the OID index for us */
oidIndex = RelationGetOidIndex(relation);
/* If no OID index, just hand back the next OID counter value */
if (!OidIsValid(oidIndex))
{
Oid result;
/*
* System catalogs that have OIDs should *always* have a unique OID
* index; we should only take this path for user tables. Give a
* warning if it looks like somebody forgot an index.
*/
if (IsSystemRelation(relation))
elog(WARNING, "generating possibly-non-unique OID for \"%s\"",
RelationGetRelationName(relation));
result= GetNewObjectId();
if (IsSystemNamespace(RelationGetNamespace(relation)))
{
if (Gp_role == GP_ROLE_EXECUTE)
{
elog(DEBUG1,"Allocating Oid %u on relid %u %s in EXECUTE mode",result,relation->rd_id,RelationGetRelationName(relation));
}
if (Gp_role == GP_ROLE_DISPATCH)
{
elog(DEBUG5,"Allocating Oid %u on relid %u %s in DISPATCH mode",result,relation->rd_id,RelationGetRelationName(relation));
}
}
return result;
}
/* Otherwise, use the index to find a nonconflicting OID */
indexrel = index_open(oidIndex, AccessShareLock);
newOid = GetNewOidWithIndex(relation, indexrel);
index_close(indexrel, AccessShareLock);
return newOid;
}
/*
* GetNewOidWithIndex
* Guts of GetNewOid: use the supplied index
*
* This is exported separately because there are cases where we want to use
* an index that will not be recognized by RelationGetOidIndex: TOAST tables
* and pg_largeobject have indexes that are usable, but have multiple columns
* and are on ordinary columns rather than a true OID column. This code
* will work anyway, so long as the OID is the index's first column.
*
* Caller must have a suitable lock on the relation.
*/
Oid
GetNewOidWithIndex(Relation relation, Relation indexrel)
{
Oid newOid;
IndexScanDesc scan;
ScanKeyData key;
bool collides;
/* Generate new OIDs until we find one not in the table */
do
{
CHECK_FOR_INTERRUPTS();
newOid = GetNewObjectId();
ScanKeyInit(&key,
(AttrNumber) 1,
BTEqualStrategyNumber, F_OIDEQ,
ObjectIdGetDatum(newOid));
/* see notes above about using SnapshotDirty */
scan = index_beginscan(relation, indexrel,
SnapshotDirty, 1, &key);
collides = HeapTupleIsValid(index_getnext(scan, ForwardScanDirection));
index_endscan(scan);
} while (collides);
if (IsSystemNamespace(RelationGetNamespace(relation)))
{
if (Gp_role == GP_ROLE_EXECUTE)
{
if (relation->rd_id != 2604 /* pg_attrdef */ && relation->rd_id != 2606 /* pg_constraint */ && relation->rd_id != 2615 /* pg_namespace */)
elog(DEBUG1,"Allocating Oid %u with index on relid %u %s in EXECUTE mode",newOid,relation->rd_id, RelationGetRelationName(relation));
else
elog(DEBUG4,"Allocating Oid %u with index on relid %u %s in EXECUTE mode",newOid,relation->rd_id, RelationGetRelationName(relation));
}
if (Gp_role == GP_ROLE_DISPATCH)
{
elog(DEBUG5,"Allocating Oid %u with index on relid %u %s in DISPATCH mode",newOid,relation->rd_id, RelationGetRelationName(relation));
}
}
return newOid;
}
/*
* Write relation description to the output stream.
*/
void
logicalrep_write_rel(StringInfo out, Relation rel)
{
char *relname;
pq_sendbyte(out, 'R'); /* sending RELATION */
/* use Oid as relation identifier */
pq_sendint(out, RelationGetRelid(rel), 4);
/* send qualified relation name */
logicalrep_write_namespace(out, RelationGetNamespace(rel));
relname = RelationGetRelationName(rel);
pq_sendstring(out, relname);
/* send replica identity */
pq_sendbyte(out, rel->rd_rel->relreplident);
/* send the attribute info */
logicalrep_write_attrs(out, rel);
}
/*
* OIDs for catalog object are normally allocated in the master, and
* executor nodes should just use the OIDs passed by the master. But
* there are some exceptions.
*/
static bool
RelationNeedsSynchronizedOIDs(Relation relation)
{
if (IsSystemNamespace(RelationGetNamespace(relation)))
{
switch(RelationGetRelid(relation))
{
/*
* pg_largeobject is more like a user table, and has
* different contents in each segment and master.
*/
case LargeObjectRelationId:
return false;
/*
* We don't currently synchronize the OIDs of these catalogs.
* It's a bit sketchy that we don't, but we get away with it
* because these OIDs don't appear in any of the Node structs
* that are dispatched from master to segments. (Except for the
* OIDs, the contents of these tables should be in sync.)
*/
case RewriteRelationId:
case TriggerRelationId:
case AccessMethodOperatorRelationId:
case AccessMethodProcedureRelationId:
return false;
}
/*
* All other system catalogs are assumed to need synchronized
* OIDs.
*/
return true;
}
return false;
}
Datum
bt_page_items(PG_FUNCTION_ARGS)
{
text *relname = PG_GETARG_TEXT_P(0);
uint32 blkno = PG_GETARG_UINT32(1);
RangeVar *relrv;
Datum result;
char *values[BTPAGEITEMS_NCOLUMNS];
BTPageOpaque opaque;
HeapTuple tuple;
ItemId id;
FuncCallContext *fctx;
MemoryContext mctx;
struct user_args *uargs = NULL;
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
(errmsg("must be superuser to use pgstattuple functions"))));
if (SRF_IS_FIRSTCALL())
{
fctx = SRF_FIRSTCALL_INIT();
mctx = MemoryContextSwitchTo(fctx->multi_call_memory_ctx);
uargs = palloc(sizeof(struct user_args));
uargs->tupd = RelationNameGetTupleDesc(BTPAGEITEMS_TYPE);
uargs->offset = FirstOffsetNumber;
relrv = makeRangeVarFromNameList(textToQualifiedNameList(relname));
uargs->rel = relation_openrv(relrv, AccessShareLock);
if (!IS_INDEX(uargs->rel) || !IS_BTREE(uargs->rel))
elog(ERROR, "bt_page_items() can be used only on b-tree index.");
/*
* Reject attempts to read non-local temporary relations; we would
* be likely to get wrong data since we have no visibility into the
* owning session's local buffers.
*/
if (isOtherTempNamespace(RelationGetNamespace(uargs->rel)))
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("cannot access temporary tables of other sessions")));
if (blkno == 0)
elog(ERROR, "Block 0 is a meta page.");
CHECK_RELATION_BLOCK_RANGE(uargs->rel, blkno);
uargs->buffer = ReadBuffer(uargs->rel, blkno);
uargs->page = BufferGetPage(uargs->buffer);
opaque = (BTPageOpaque) PageGetSpecialPointer(uargs->page);
if (P_ISDELETED(opaque))
elog(NOTICE, "bt_page_items(): this page is deleted.");
fctx->max_calls = PageGetMaxOffsetNumber(uargs->page);
fctx->user_fctx = uargs;
MemoryContextSwitchTo(mctx);
}
fctx = SRF_PERCALL_SETUP();
uargs = fctx->user_fctx;
if (fctx->call_cntr < fctx->max_calls)
{
IndexTuple itup;
id = PageGetItemId(uargs->page, uargs->offset);
if (!ItemIdIsValid(id))
elog(ERROR, "Invalid ItemId.");
itup = (IndexTuple) PageGetItem(uargs->page, id);
{
int j = 0;
BlockNumber blkno = BlockIdGetBlockNumber(&(itup->t_tid.ip_blkid));
values[j] = palloc(32);
snprintf(values[j++], 32, "%d", uargs->offset);
values[j] = palloc(32);
snprintf(values[j++], 32, "(%u,%u)", blkno, itup->t_tid.ip_posid);
values[j] = palloc(32);
snprintf(values[j++], 32, "%d", (int) IndexTupleSize(itup));
values[j] = palloc(32);
snprintf(values[j++], 32, "%c", IndexTupleHasNulls(itup) ? 't' : 'f');
values[j] = palloc(32);
snprintf(values[j++], 32, "%c", IndexTupleHasVarwidths(itup) ? 't' : 'f');
{
int off;
char *dump;
char *ptr = (char *) itup + IndexInfoFindDataOffset(itup->t_info);
dump = palloc(IndexTupleSize(itup) * 3);
memset(dump, 0, IndexTupleSize(itup) * 3);
for (off = 0;
off < IndexTupleSize(itup) - IndexInfoFindDataOffset(itup->t_info);
off++)
{
if (dump[0] == '\0')
sprintf(dump, "%02x", *(ptr + off) & 0xff);
else
{
char buf[4];
sprintf(buf, " %02x", *(ptr + off) & 0xff);
strcat(dump, buf);
}
}
values[j] = dump;
}
tuple = BuildTupleFromCStrings(TupleDescGetAttInMetadata(uargs->tupd), values);
result = TupleGetDatum(TupleDescGetSlot(uargs->tupd), tuple);
}
uargs->offset = uargs->offset + 1;
SRF_RETURN_NEXT(fctx, result);
}
else
{
ReleaseBuffer(uargs->buffer);
relation_close(uargs->rel, AccessShareLock);
SRF_RETURN_DONE(fctx);
}
}
/*
* ExecRefreshMatView -- execute a REFRESH MATERIALIZED VIEW command
*
* This refreshes the materialized view by creating a new table and swapping
* the relfilenodes of the new table and the old materialized view, so the OID
* of the original materialized view is preserved. Thus we do not lose GRANT
* nor references to this materialized view.
*
* If WITH NO DATA was specified, this is effectively like a TRUNCATE;
* otherwise it is like a TRUNCATE followed by an INSERT using the SELECT
* statement associated with the materialized view. The statement node's
* skipData field shows whether the clause was used.
*
* Indexes are rebuilt too, via REINDEX. Since we are effectively bulk-loading
* the new heap, it's better to create the indexes afterwards than to fill them
* incrementally while we load.
*
* The matview's "populated" state is changed based on whether the contents
* reflect the result set of the materialized view's query.
*/
ObjectAddress
ExecRefreshMatView(RefreshMatViewStmt *stmt, const char *queryString,
ParamListInfo params, char *completionTag)
{
Oid matviewOid;
Relation matviewRel;
RewriteRule *rule;
List *actions;
Query *dataQuery;
Oid tableSpace;
Oid relowner;
Oid OIDNewHeap;
DestReceiver *dest;
uint64 processed = 0;
bool concurrent;
LOCKMODE lockmode;
char relpersistence;
Oid save_userid;
int save_sec_context;
int save_nestlevel;
ObjectAddress address;
/* Determine strength of lock needed. */
concurrent = stmt->concurrent;
lockmode = concurrent ? ExclusiveLock : AccessExclusiveLock;
/*
* Get a lock until end of transaction.
*/
matviewOid = RangeVarGetRelidExtended(stmt->relation,
lockmode, 0,
RangeVarCallbackOwnsTable, NULL);
matviewRel = table_open(matviewOid, NoLock);
/* Make sure it is a materialized view. */
if (matviewRel->rd_rel->relkind != RELKIND_MATVIEW)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("\"%s\" is not a materialized view",
RelationGetRelationName(matviewRel))));
/* Check that CONCURRENTLY is not specified if not populated. */
if (concurrent && !RelationIsPopulated(matviewRel))
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("CONCURRENTLY cannot be used when the materialized view is not populated")));
/* Check that conflicting options have not been specified. */
if (concurrent && stmt->skipData)
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
errmsg("CONCURRENTLY and WITH NO DATA options cannot be used together")));
/*
* Check that everything is correct for a refresh. Problems at this point
* are internal errors, so elog is sufficient.
*/
if (matviewRel->rd_rel->relhasrules == false ||
matviewRel->rd_rules->numLocks < 1)
elog(ERROR,
"materialized view \"%s\" is missing rewrite information",
RelationGetRelationName(matviewRel));
if (matviewRel->rd_rules->numLocks > 1)
elog(ERROR,
"materialized view \"%s\" has too many rules",
RelationGetRelationName(matviewRel));
rule = matviewRel->rd_rules->rules[0];
if (rule->event != CMD_SELECT || !(rule->isInstead))
elog(ERROR,
"the rule for materialized view \"%s\" is not a SELECT INSTEAD OF rule",
RelationGetRelationName(matviewRel));
actions = rule->actions;
if (list_length(actions) != 1)
elog(ERROR,
"the rule for materialized view \"%s\" is not a single action",
RelationGetRelationName(matviewRel));
/*
* Check that there is a unique index with no WHERE clause on one or more
* columns of the materialized view if CONCURRENTLY is specified.
*/
if (concurrent)
{
List *indexoidlist = RelationGetIndexList(matviewRel);
ListCell *indexoidscan;
bool hasUniqueIndex = false;
foreach(indexoidscan, indexoidlist)
{
Oid indexoid = lfirst_oid(indexoidscan);
Relation indexRel;
indexRel = index_open(indexoid, AccessShareLock);
hasUniqueIndex = is_usable_unique_index(indexRel);
index_close(indexRel, AccessShareLock);
if (hasUniqueIndex)
break;
}
list_free(indexoidlist);
if (!hasUniqueIndex)
ereport(ERROR,
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
errmsg("cannot refresh materialized view \"%s\" concurrently",
quote_qualified_identifier(get_namespace_name(RelationGetNamespace(matviewRel)),
RelationGetRelationName(matviewRel))),
errhint("Create a unique index with no WHERE clause on one or more columns of the materialized view.")));
}
/*
* DefineIndex
* Creates a new index.
*
* 'heapRelation': the relation the index will apply to.
* 'indexRelationName': the name for the new index, or NULL to indicate
* that a nonconflicting default name should be picked.
* 'indexRelationId': normally InvalidOid, but during bootstrap can be
* nonzero to specify a preselected OID for the index.
* 'accessMethodName': name of the AM to use.
* 'tableSpaceName': name of the tablespace to create the index in.
* NULL specifies using the appropriate default.
* 'attributeList': a list of IndexElem specifying columns and expressions
* to index on.
* 'predicate': the partial-index condition, or NULL if none.
* 'rangetable': needed to interpret the predicate.
* 'options': reloptions from WITH (in list-of-DefElem form).
* 'unique': make the index enforce uniqueness.
* 'primary': mark the index as a primary key in the catalogs.
* 'isconstraint': index is for a PRIMARY KEY or UNIQUE constraint,
* so build a pg_constraint entry for it.
* 'is_alter_table': this is due to an ALTER rather than a CREATE operation.
* 'check_rights': check for CREATE rights in the namespace. (This should
* be true except when ALTER is deleting/recreating an index.)
* 'skip_build': make the catalog entries but leave the index file empty;
* it will be filled later.
* 'quiet': suppress the NOTICE chatter ordinarily provided for constraints.
* 'concurrent': avoid blocking writers to the table while building.
*/
void
DefineIndex(RangeVar *heapRelation,
char *indexRelationName,
Oid indexRelationId,
char *accessMethodName,
char *tableSpaceName,
List *attributeList,
Expr *predicate,
List *rangetable,
List *options,
bool unique,
bool primary,
bool isconstraint,
bool is_alter_table,
bool check_rights,
bool skip_build,
bool quiet,
bool concurrent)
{
Oid *classObjectId;
Oid accessMethodId;
Oid relationId;
Oid namespaceId;
Oid tablespaceId;
Relation rel;
HeapTuple tuple;
Form_pg_am accessMethodForm;
RegProcedure amoptions;
Datum reloptions;
IndexInfo *indexInfo;
int numberOfAttributes;
List *old_xact_list;
ListCell *lc;
uint32 ixcnt;
LockRelId heaprelid;
LOCKTAG heaplocktag;
Snapshot snapshot;
Relation pg_index;
HeapTuple indexTuple;
Form_pg_index indexForm;
/*
* count attributes in index
*/
numberOfAttributes = list_length(attributeList);
if (numberOfAttributes <= 0)
ereport(ERROR,
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
errmsg("must specify at least one column")));
if (numberOfAttributes > INDEX_MAX_KEYS)
ereport(ERROR,
(errcode(ERRCODE_TOO_MANY_COLUMNS),
errmsg("cannot use more than %d columns in an index",
INDEX_MAX_KEYS)));
/*
* Open heap relation, acquire a suitable lock on it, remember its OID
*
* Only SELECT ... FOR UPDATE/SHARE are allowed while doing a standard
* index build; but for concurrent builds we allow INSERT/UPDATE/DELETE
* (but not VACUUM).
*/
rel = heap_openrv(heapRelation,
(concurrent ? ShareUpdateExclusiveLock : ShareLock));
relationId = RelationGetRelid(rel);
namespaceId = RelationGetNamespace(rel);
/* Note: during bootstrap may see uncataloged relation */
if (rel->rd_rel->relkind != RELKIND_RELATION &&
rel->rd_rel->relkind != RELKIND_UNCATALOGED)
ereport(ERROR,
(errcode(ERRCODE_WRONG_OBJECT_TYPE),
errmsg("\"%s\" is not a table",
heapRelation->relname)));
/*
* Don't try to CREATE INDEX on temp tables of other backends.
*/
if (isOtherTempNamespace(namespaceId))
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("cannot create indexes on temporary tables of other sessions")));
/*
* Verify we (still) have CREATE rights in the rel's namespace.
* (Presumably we did when the rel was created, but maybe not anymore.)
* Skip check if caller doesn't want it. Also skip check if
* bootstrapping, since permissions machinery may not be working yet.
*/
if (check_rights && !IsBootstrapProcessingMode())
{
AclResult aclresult;
aclresult = pg_namespace_aclcheck(namespaceId, GetUserId(),
ACL_CREATE);
if (aclresult != ACLCHECK_OK)
aclcheck_error(aclresult, ACL_KIND_NAMESPACE,
get_namespace_name(namespaceId));
}
/*
* Select tablespace to use. If not specified, use default_tablespace
* (which may in turn default to database's default).
*/
if (tableSpaceName)
{
tablespaceId = get_tablespace_oid(tableSpaceName);
if (!OidIsValid(tablespaceId))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
errmsg("tablespace \"%s\" does not exist",
tableSpaceName)));
}
else
{
tablespaceId = GetDefaultTablespace();
/* note InvalidOid is OK in this case */
}
/* Check permissions except when using database's default */
if (OidIsValid(tablespaceId))
{
AclResult aclresult;
aclresult = pg_tablespace_aclcheck(tablespaceId, GetUserId(),
ACL_CREATE);
if (aclresult != ACLCHECK_OK)
aclcheck_error(aclresult, ACL_KIND_TABLESPACE,
get_tablespace_name(tablespaceId));
}
/*
* Force shared indexes into the pg_global tablespace. This is a bit of a
* hack but seems simpler than marking them in the BKI commands.
*/
if (rel->rd_rel->relisshared)
tablespaceId = GLOBALTABLESPACE_OID;
/*
* Select name for index if caller didn't specify
*/
if (indexRelationName == NULL)
{
if (primary)
indexRelationName = ChooseRelationName(RelationGetRelationName(rel),
NULL,
"pkey",
namespaceId);
else
{
IndexElem *iparam = (IndexElem *) linitial(attributeList);
indexRelationName = ChooseRelationName(RelationGetRelationName(rel),
iparam->name,
"key",
namespaceId);
}
}
/*
* look up the access method, verify it can handle the requested features
*/
tuple = SearchSysCache(AMNAME,
PointerGetDatum(accessMethodName),
0, 0, 0);
if (!HeapTupleIsValid(tuple))
{
/*
* Hack to provide more-or-less-transparent updating of old RTREE
* indexes to GIST: if RTREE is requested and not found, use GIST.
*/
if (strcmp(accessMethodName, "rtree") == 0)
{
ereport(NOTICE,
(errmsg("substituting access method \"gist\" for obsolete method \"rtree\"")));
accessMethodName = "gist";
tuple = SearchSysCache(AMNAME,
PointerGetDatum(accessMethodName),
0, 0, 0);
}
if (!HeapTupleIsValid(tuple))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
errmsg("access method \"%s\" does not exist",
accessMethodName)));
}
accessMethodId = HeapTupleGetOid(tuple);
accessMethodForm = (Form_pg_am) GETSTRUCT(tuple);
if (unique && !accessMethodForm->amcanunique)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("access method \"%s\" does not support unique indexes",
accessMethodName)));
if (numberOfAttributes > 1 && !accessMethodForm->amcanmulticol)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("access method \"%s\" does not support multicolumn indexes",
accessMethodName)));
amoptions = accessMethodForm->amoptions;
ReleaseSysCache(tuple);
/*
* If a range table was created then check that only the base rel is
* mentioned.
*/
if (rangetable != NIL)
{
if (list_length(rangetable) != 1 || getrelid(1, rangetable) != relationId)
ereport(ERROR,
(errcode(ERRCODE_INVALID_COLUMN_REFERENCE),
errmsg("index expressions and predicates may refer only to the table being indexed")));
}
/*
* Validate predicate, if given
*/
if (predicate)
CheckPredicate(predicate);
/*
* Extra checks when creating a PRIMARY KEY index.
*/
if (primary)
{
List *cmds;
ListCell *keys;
/*
* If ALTER TABLE, check that there isn't already a PRIMARY KEY. In
* CREATE TABLE, we have faith that the parser rejected multiple pkey
* clauses; and CREATE INDEX doesn't have a way to say PRIMARY KEY, so
* it's no problem either.
*/
if (is_alter_table &&
relationHasPrimaryKey(rel))
{
ereport(ERROR,
(errcode(ERRCODE_INVALID_TABLE_DEFINITION),
errmsg("multiple primary keys for table \"%s\" are not allowed",
RelationGetRelationName(rel))));
}
/*
* Check that all of the attributes in a primary key are marked as not
* null, otherwise attempt to ALTER TABLE .. SET NOT NULL
*/
cmds = NIL;
foreach(keys, attributeList)
{
IndexElem *key = (IndexElem *) lfirst(keys);
HeapTuple atttuple;
if (!key->name)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("primary keys cannot be expressions")));
/* System attributes are never null, so no problem */
if (SystemAttributeByName(key->name, rel->rd_rel->relhasoids))
continue;
atttuple = SearchSysCacheAttName(relationId, key->name);
if (HeapTupleIsValid(atttuple))
{
if (!((Form_pg_attribute) GETSTRUCT(atttuple))->attnotnull)
{
/* Add a subcommand to make this one NOT NULL */
AlterTableCmd *cmd = makeNode(AlterTableCmd);
cmd->subtype = AT_SetNotNull;
cmd->name = key->name;
cmds = lappend(cmds, cmd);
}
ReleaseSysCache(atttuple);
}
else
{
/*
* This shouldn't happen during CREATE TABLE, but can happen
* during ALTER TABLE. Keep message in sync with
* transformIndexConstraints() in parser/analyze.c.
*/
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_COLUMN),
errmsg("column \"%s\" named in key does not exist",
key->name)));
}
}
/*
* XXX: Shouldn't the ALTER TABLE .. SET NOT NULL cascade to child
* tables? Currently, since the PRIMARY KEY itself doesn't cascade,
* we don't cascade the notnull constraint(s) either; but this is
* pretty debatable.
*
* XXX: possible future improvement: when being called from ALTER
* TABLE, it would be more efficient to merge this with the outer
* ALTER TABLE, so as to avoid two scans. But that seems to
* complicate DefineIndex's API unduly.
*/
if (cmds)
AlterTableInternal(relationId, cmds, false);
}
/*
* Post vacuum, iterate over all entries in index, check if the h_tid
* of each entry exists and is not dead. For specific system tables,
* also ensure that the key in index entry matches the corresponding
* attribute in the heap tuple.
*/
void
_bt_validate_vacuum(Relation irel, Relation hrel, TransactionId oldest_xmin)
{
MIRROREDLOCK_BUFMGR_DECLARE;
BlockNumber blkno;
BlockNumber num_pages;
Buffer ibuf = InvalidBuffer;
Buffer hbuf = InvalidBuffer;
Page ipage;
BTPageOpaque opaque;
IndexTuple itup;
HeapTupleData htup;
OffsetNumber maxoff,
minoff,
offnum;
Oid ioid,
hoid;
bool isnull;
blkno = BTREE_METAPAGE + 1;
num_pages = RelationGetNumberOfBlocks(irel);
elog(LOG, "btvalidatevacuum: index %s, heap %s",
RelationGetRelationName(irel), RelationGetRelationName(hrel));
MIRROREDLOCK_BUFMGR_LOCK;
for (; blkno < num_pages; blkno++)
{
ibuf = ReadBuffer(irel, blkno);
ipage = BufferGetPage(ibuf);
opaque = (BTPageOpaque) PageGetSpecialPointer(ipage);
if (!PageIsNew(ipage))
_bt_checkpage(irel, ibuf);
if (P_ISLEAF(opaque))
{
minoff = P_FIRSTDATAKEY(opaque);
maxoff = PageGetMaxOffsetNumber(ipage);
for (offnum = minoff;
offnum <= maxoff;
offnum = OffsetNumberNext(offnum))
{
itup = (IndexTuple) PageGetItem(ipage,
PageGetItemId(ipage, offnum));
ItemPointerCopy(&itup->t_tid, &htup.t_self);
/*
* TODO: construct a tid bitmap based on index tids
* and fetch heap tids in order afterwards. That will
* also allow validating if a heap tid appears twice
* in a unique index.
*/
if (!heap_release_fetch(hrel, SnapshotAny, &htup,
&hbuf, true, NULL))
{
elog(ERROR, "btvalidatevacuum: tid (%d,%d) from index %s "
"not found in heap %s",
ItemPointerGetBlockNumber(&itup->t_tid),
ItemPointerGetOffsetNumber(&itup->t_tid),
RelationGetRelationName(irel),
RelationGetRelationName(hrel));
}
switch (HeapTupleSatisfiesVacuum(hrel, htup.t_data, oldest_xmin, hbuf))
{
case HEAPTUPLE_RECENTLY_DEAD:
case HEAPTUPLE_LIVE:
case HEAPTUPLE_INSERT_IN_PROGRESS:
case HEAPTUPLE_DELETE_IN_PROGRESS:
/* these tuples are considered alive by vacuum */
break;
case HEAPTUPLE_DEAD:
elog(ERROR, "btvalidatevacuum: vacuum did not remove "
"dead tuple (%d,%d) from heap %s and index %s",
ItemPointerGetBlockNumber(&itup->t_tid),
ItemPointerGetOffsetNumber(&itup->t_tid),
RelationGetRelationName(hrel),
RelationGetRelationName(irel));
break;
default:
elog(ERROR, "btvalidatevacuum: invalid visibility");
break;
}
switch(RelationGetRelid(irel))
{
case DatabaseOidIndexId:
case TypeOidIndexId:
case ClassOidIndexId:
case ConstraintOidIndexId:
hoid = HeapTupleGetOid(&htup);
ioid = index_getattr(itup, 1, RelationGetDescr(irel), &isnull);
if (hoid != ioid)
{
elog(ERROR,
"btvalidatevacuum: index oid(%d) != heap oid(%d)"
" tuple (%d,%d) index %s", ioid, hoid,
ItemPointerGetBlockNumber(&itup->t_tid),
ItemPointerGetOffsetNumber(&itup->t_tid),
RelationGetRelationName(irel));
}
break;
case GpRelationNodeOidIndexId:
hoid = heap_getattr(&htup, 1, RelationGetDescr(hrel), &isnull);
ioid = index_getattr(itup, 1, RelationGetDescr(irel), &isnull);
if (hoid != ioid)
{
elog(ERROR,
"btvalidatevacuum: index oid(%d) != heap oid(%d)"
" tuple (%d,%d) index %s", ioid, hoid,
ItemPointerGetBlockNumber(&itup->t_tid),
ItemPointerGetOffsetNumber(&itup->t_tid),
RelationGetRelationName(irel));
}
int4 hsegno = heap_getattr(&htup, 2, RelationGetDescr(hrel), &isnull);
int4 isegno = index_getattr(itup, 2, RelationGetDescr(irel), &isnull);
if (isegno != hsegno)
{
elog(ERROR,
"btvalidatevacuum: index segno(%d) != heap segno(%d)"
" tuple (%d,%d) index %s", isegno, hsegno,
ItemPointerGetBlockNumber(&itup->t_tid),
ItemPointerGetOffsetNumber(&itup->t_tid),
RelationGetRelationName(irel));
}
break;
default:
break;
}
if (RelationGetNamespace(irel) == PG_AOSEGMENT_NAMESPACE)
{
int4 isegno = index_getattr(itup, 1, RelationGetDescr(irel), &isnull);
int4 hsegno = heap_getattr(&htup, 1, RelationGetDescr(hrel), &isnull);
if (isegno != hsegno)
{
elog(ERROR,
"btvalidatevacuum: index segno(%d) != heap segno(%d)"
" tuple (%d,%d) index %s", isegno, hsegno,
ItemPointerGetBlockNumber(&itup->t_tid),
ItemPointerGetOffsetNumber(&itup->t_tid),
RelationGetRelationName(irel));
}
}
}
}
if (BufferIsValid(ibuf))
ReleaseBuffer(ibuf);
}
if (BufferIsValid(hbuf))
ReleaseBuffer(hbuf);
MIRROREDLOCK_BUFMGR_UNLOCK;
}
/*
* lazy_vacuum_rel() -- perform LAZY VACUUM for one heap relation
*
* This routine vacuums a single heap, cleans out its indexes, and
* updates its relpages and reltuples statistics.
*
* At entry, we have already established a transaction and opened
* and locked the relation.
*/
void
lazy_vacuum_rel(Relation onerel, VacuumStmt *vacstmt,
BufferAccessStrategy bstrategy)
{
LVRelStats *vacrelstats;
Relation *Irel;
int nindexes;
BlockNumber possibly_freeable;
PGRUsage ru0;
TimestampTz starttime = 0;
long secs;
int usecs;
double read_rate,
write_rate;
bool scan_all;
TransactionId freezeTableLimit;
BlockNumber new_rel_pages;
double new_rel_tuples;
BlockNumber new_rel_allvisible;
TransactionId new_frozen_xid;
/* measure elapsed time iff autovacuum logging requires it */
if (IsAutoVacuumWorkerProcess() && Log_autovacuum_min_duration >= 0)
{
pg_rusage_init(&ru0);
starttime = GetCurrentTimestamp();
}
if (vacstmt->options & VACOPT_VERBOSE)
elevel = INFO;
else
elevel = DEBUG2;
vac_strategy = bstrategy;
vacuum_set_xid_limits(vacstmt->freeze_min_age, vacstmt->freeze_table_age,
onerel->rd_rel->relisshared,
&OldestXmin, &FreezeLimit, &freezeTableLimit);
scan_all = TransactionIdPrecedesOrEquals(onerel->rd_rel->relfrozenxid,
freezeTableLimit);
vacrelstats = (LVRelStats *) palloc0(sizeof(LVRelStats));
vacrelstats->old_rel_pages = onerel->rd_rel->relpages;
vacrelstats->old_rel_tuples = onerel->rd_rel->reltuples;
vacrelstats->num_index_scans = 0;
/* Open all indexes of the relation */
vac_open_indexes(onerel, RowExclusiveLock, &nindexes, &Irel);
vacrelstats->hasindex = (nindexes > 0);
/* Do the vacuuming */
lazy_scan_heap(onerel, vacrelstats, Irel, nindexes, scan_all);
/* Done with indexes */
vac_close_indexes(nindexes, Irel, NoLock);
/*
* Optionally truncate the relation.
*
* Don't even think about it unless we have a shot at releasing a goodly
* number of pages. Otherwise, the time taken isn't worth it.
*/
possibly_freeable = vacrelstats->rel_pages - vacrelstats->nonempty_pages;
if (possibly_freeable > 0 &&
(possibly_freeable >= REL_TRUNCATE_MINIMUM ||
possibly_freeable >= vacrelstats->rel_pages / REL_TRUNCATE_FRACTION))
lazy_truncate_heap(onerel, vacrelstats);
/* Vacuum the Free Space Map */
FreeSpaceMapVacuum(onerel);
/*
* Update statistics in pg_class.
*
* A corner case here is that if we scanned no pages at all because every
* page is all-visible, we should not update relpages/reltuples, because
* we have no new information to contribute. In particular this keeps
* us from replacing relpages=reltuples=0 (which means "unknown tuple
* density") with nonzero relpages and reltuples=0 (which means "zero
* tuple density") unless there's some actual evidence for the latter.
*
* We do update relallvisible even in the corner case, since if the
* table is all-visible we'd definitely like to know that. But clamp
* the value to be not more than what we're setting relpages to.
*
* Also, don't change relfrozenxid if we skipped any pages, since then
* we don't know for certain that all tuples have a newer xmin.
*/
new_rel_pages = vacrelstats->rel_pages;
new_rel_tuples = vacrelstats->new_rel_tuples;
if (vacrelstats->scanned_pages == 0 && new_rel_pages > 0)
{
new_rel_pages = vacrelstats->old_rel_pages;
new_rel_tuples = vacrelstats->old_rel_tuples;
}
new_rel_allvisible = visibilitymap_count(onerel);
if (new_rel_allvisible > new_rel_pages)
new_rel_allvisible = new_rel_pages;
new_frozen_xid = FreezeLimit;
if (vacrelstats->scanned_pages < vacrelstats->rel_pages)
new_frozen_xid = InvalidTransactionId;
vac_update_relstats(onerel,
new_rel_pages,
new_rel_tuples,
new_rel_allvisible,
vacrelstats->hasindex,
new_frozen_xid);
/* report results to the stats collector, too */
pgstat_report_vacuum(RelationGetRelid(onerel),
onerel->rd_rel->relisshared,
new_rel_tuples);
/* and log the action if appropriate */
if (IsAutoVacuumWorkerProcess() && Log_autovacuum_min_duration >= 0)
{
TimestampTz endtime = GetCurrentTimestamp();
if (Log_autovacuum_min_duration == 0 ||
TimestampDifferenceExceeds(starttime, endtime,
Log_autovacuum_min_duration))
{
TimestampDifference(starttime, endtime, &secs, &usecs);
read_rate = 0;
write_rate = 0;
if ((secs > 0) || (usecs > 0))
{
read_rate = (double) BLCKSZ * VacuumPageMiss / (1024 * 1024) /
(secs + usecs / 1000000.0);
write_rate = (double) BLCKSZ * VacuumPageDirty / (1024 * 1024) /
(secs + usecs / 1000000.0);
}
ereport(LOG,
(errmsg("automatic vacuum of table \"%s.%s.%s\": index scans: %d\n"
"pages: %d removed, %d remain\n"
"tuples: %.0f removed, %.0f remain\n"
"buffer usage: %d hits, %d misses, %d dirtied\n"
"avg read rate: %.3f MiB/s, avg write rate: %.3f MiB/s\n"
"system usage: %s",
get_database_name(MyDatabaseId),
get_namespace_name(RelationGetNamespace(onerel)),
RelationGetRelationName(onerel),
vacrelstats->num_index_scans,
vacrelstats->pages_removed,
vacrelstats->rel_pages,
vacrelstats->tuples_deleted,
vacrelstats->new_rel_tuples,
VacuumPageHit,
VacuumPageMiss,
VacuumPageDirty,
read_rate,write_rate,
pg_rusage_show(&ru0))));
}
}
}
/*
* lazy_vacuum_rel() -- perform LAZY VACUUM for one heap relation
*
* This routine vacuums a single heap, cleans out its indexes, and
* updates its relpages and reltuples statistics.
*
* At entry, we have already established a transaction and opened
* and locked the relation.
*/
void
lazy_vacuum_rel(Relation onerel, VacuumStmt *vacstmt,
BufferAccessStrategy bstrategy)
{
LVRelStats *vacrelstats;
Relation *Irel;
int nindexes;
BlockNumber possibly_freeable;
PGRUsage ru0;
TimestampTz starttime = 0;
bool scan_all;
TransactionId freezeTableLimit;
pg_rusage_init(&ru0);
/* measure elapsed time iff autovacuum logging requires it */
if (IsAutoVacuumWorkerProcess() && Log_autovacuum_min_duration > 0)
starttime = GetCurrentTimestamp();
if (vacstmt->options & VACOPT_VERBOSE)
elevel = INFO;
else
elevel = DEBUG2;
vac_strategy = bstrategy;
vacuum_set_xid_limits(vacstmt->freeze_min_age, vacstmt->freeze_table_age,
onerel->rd_rel->relisshared,
&OldestXmin, &FreezeLimit, &freezeTableLimit);
scan_all = TransactionIdPrecedesOrEquals(onerel->rd_rel->relfrozenxid,
freezeTableLimit);
vacrelstats = (LVRelStats *) palloc0(sizeof(LVRelStats));
vacrelstats->old_rel_tuples = onerel->rd_rel->reltuples;
vacrelstats->num_index_scans = 0;
/* Open all indexes of the relation */
vac_open_indexes(onerel, RowExclusiveLock, &nindexes, &Irel);
vacrelstats->hasindex = (nindexes > 0);
/* Do the vacuuming */
lazy_scan_heap(onerel, vacrelstats, Irel, nindexes, scan_all);
/* Done with indexes */
vac_close_indexes(nindexes, Irel, NoLock);
/*
* Optionally truncate the relation.
*
* Don't even think about it unless we have a shot at releasing a goodly
* number of pages. Otherwise, the time taken isn't worth it.
*/
possibly_freeable = vacrelstats->rel_pages - vacrelstats->nonempty_pages;
if (possibly_freeable > 0 &&
(possibly_freeable >= REL_TRUNCATE_MINIMUM ||
possibly_freeable >= vacrelstats->rel_pages / REL_TRUNCATE_FRACTION))
lazy_truncate_heap(onerel, vacrelstats);
/* Vacuum the Free Space Map */
FreeSpaceMapVacuum(onerel);
/*
* Update statistics in pg_class. But don't change relfrozenxid if we
* skipped any pages.
*/
vac_update_relstats(onerel,
vacrelstats->rel_pages, vacrelstats->new_rel_tuples,
vacrelstats->hasindex,
(vacrelstats->scanned_pages < vacrelstats->rel_pages) ?
InvalidTransactionId :
FreezeLimit);
/* report results to the stats collector, too */
pgstat_report_vacuum(RelationGetRelid(onerel),
onerel->rd_rel->relisshared,
vacrelstats->new_rel_tuples);
/* and log the action if appropriate */
if (IsAutoVacuumWorkerProcess() && Log_autovacuum_min_duration >= 0)
{
if (Log_autovacuum_min_duration == 0 ||
TimestampDifferenceExceeds(starttime, GetCurrentTimestamp(),
Log_autovacuum_min_duration))
ereport(LOG,
(errmsg("automatic vacuum of table \"%s.%s.%s\": index scans: %d\n"
"pages: %d removed, %d remain\n"
"tuples: %.0f removed, %.0f remain\n"
"system usage: %s",
get_database_name(MyDatabaseId),
get_namespace_name(RelationGetNamespace(onerel)),
RelationGetRelationName(onerel),
vacrelstats->num_index_scans,
vacrelstats->pages_removed,
vacrelstats->rel_pages,
vacrelstats->tuples_deleted,
vacrelstats->new_rel_tuples,
pg_rusage_show(&ru0))));
}
}
/*
* vacuum_appendonly_rel() -- vaccum an append-only relation
*
* This procedure will be what gets executed both for VACUUM
* and VACUUM FULL (and also ANALYZE or any other thing that
* needs the pg_class stats updated).
*
* The function can compact append-only segment files or just
* truncating the segment file to its existing eof.
*
* Afterwards, the reltuples and relpages information in pg_class
* are updated. reltuples is the same as "pg_aoseg_<oid>:tupcount"
* column and we simulate relpages by subdividing the eof value
* ("pg_aoseg_<oid>:eof") over the defined page size.
*
*
* There are txn ids, hint bits, free space, dead tuples,
* etc. these are all irrelevant in the append only relation context.
*
*/
void
vacuum_appendonly_rel(Relation aorel, VacuumStmt *vacstmt)
{
char *relname;
PGRUsage ru0;
Assert(RelationIsAoRows(aorel) || RelationIsAoCols(aorel));
Assert(!vacummStatement_IsInAppendOnlyCleanupPhase(vacstmt));
pg_rusage_init(&ru0);
relname = RelationGetRelationName(aorel);
ereport(elevel,
(errmsg("vacuuming \"%s.%s\"",
get_namespace_name(RelationGetNamespace(aorel)),
relname)));
if (Gp_role == GP_ROLE_DISPATCH)
{
return;
}
Assert(list_length(vacstmt->appendonly_compaction_insert_segno) <= 1);
if (vacstmt->appendonly_compaction_insert_segno == NULL)
{
elogif(Debug_appendonly_print_compaction, LOG,
"Vacuum drop phase %s", RelationGetRelationName(aorel));
if (RelationIsAoRows(aorel))
{
AppendOnlyDrop(aorel, vacstmt->appendonly_compaction_segno);
}
else
{
Assert(RelationIsAoCols(aorel));
AOCSDrop(aorel, vacstmt->appendonly_compaction_segno);
}
}
else
{
int insert_segno = linitial_int(vacstmt->appendonly_compaction_insert_segno);
if (insert_segno == APPENDONLY_COMPACTION_SEGNO_INVALID)
{
elogif(Debug_appendonly_print_compaction, LOG,
"Vacuum pseudo-compaction phase %s", RelationGetRelationName(aorel));
}
else
{
elogif(Debug_appendonly_print_compaction, LOG,
"Vacuum compaction phase %s", RelationGetRelationName(aorel));
if (RelationIsAoRows(aorel))
{
AppendOnlyCompact(aorel,
vacstmt->appendonly_compaction_segno,
insert_segno, vacstmt->full);
}
else
{
Assert(RelationIsAoCols(aorel));
AOCSCompact(aorel,
vacstmt->appendonly_compaction_segno,
insert_segno, vacstmt->full);
}
}
}
}
desc = lfirst(lc);
if (desc->indexRelid == indexRel->rd_id)
return desc;
}
heapRel = relation_open(indexRel->rd_index->indrelid, AccessShareLock);
desc = palloc0(sizeof(ZDBIndexDescriptor));
/* these all come from the actual index */
desc->indexRelid = RelationGetRelid(indexRel);
desc->heapRelid = RelationGetRelid(heapRel);
desc->isShadow = ZDBIndexOptionsGetShadow(indexRel) != NULL;
desc->logit = false;
desc->databaseName = pstrdup(get_database_name(MyDatabaseId));
desc->schemaName = pstrdup(get_namespace_name(RelationGetNamespace(heapRel)));
desc->tableName = pstrdup(RelationGetRelationName(heapRel));
desc->options = ZDBIndexOptionsGetOptions(indexRel) == NULL ? NULL : pstrdup(ZDBIndexOptionsGetOptions(indexRel));
desc->searchPreference = ZDBIndexOptionsGetSearchPreference(indexRel);
desc->refreshInterval = ZDBIndexOptionsGetRefreshInterval(indexRel);
desc->bulk_concurrency = ZDBIndexOptionsGetBulkConcurrency(indexRel);
desc->batch_size = ZDBIndexOptionsGetBatchSize(indexRel);
desc->ignoreVisibility = ZDBIndexOptionsGetIgnoreVisibility(indexRel);
desc->fieldLists = ZDBIndexOptionsGetFieldLists(indexRel);
desc->alwaysResolveJoins = ZDBIndexOptionsAlwaysResolveJoins(indexRel);
heapTupDesc = RelationGetDescr(heapRel);
for (i = 0; i < heapTupDesc->natts; i++) {
if (heapTupDesc->attrs[i]->atttypid == JSONOID) {
desc->hasJson = true;
/*
* refresh_by_match_merge
*
* Refresh a materialized view with transactional semantics, while allowing
* concurrent reads.
*
* This is called after a new version of the data has been created in a
* temporary table. It performs a full outer join against the old version of
* the data, producing "diff" results. This join cannot work if there are any
* duplicated rows in either the old or new versions, in the sense that every
* column would compare as equal between the two rows. It does work correctly
* in the face of rows which have at least one NULL value, with all non-NULL
* columns equal. The behavior of NULLs on equality tests and on UNIQUE
* indexes turns out to be quite convenient here; the tests we need to make
* are consistent with default behavior. If there is at least one UNIQUE
* index on the materialized view, we have exactly the guarantee we need.
*
* The temporary table used to hold the diff results contains just the TID of
* the old record (if matched) and the ROW from the new table as a single
* column of complex record type (if matched).
*
* Once we have the diff table, we perform set-based DELETE and INSERT
* operations against the materialized view, and discard both temporary
* tables.
*
* Everything from the generation of the new data to applying the differences
* takes place under cover of an ExclusiveLock, since it seems as though we
* would want to prohibit not only concurrent REFRESH operations, but also
* incremental maintenance. It also doesn't seem reasonable or safe to allow
* SELECT FOR UPDATE or SELECT FOR SHARE on rows being updated or deleted by
* this command.
*/
static void
refresh_by_match_merge(Oid matviewOid, Oid tempOid, Oid relowner,
int save_sec_context)
{
StringInfoData querybuf;
Relation matviewRel;
Relation tempRel;
char *matviewname;
char *tempname;
char *diffname;
TupleDesc tupdesc;
bool foundUniqueIndex;
List *indexoidlist;
ListCell *indexoidscan;
int16 relnatts;
bool *usedForQual;
initStringInfo(&querybuf);
matviewRel = heap_open(matviewOid, NoLock);
matviewname = quote_qualified_identifier(get_namespace_name(RelationGetNamespace(matviewRel)),
RelationGetRelationName(matviewRel));
tempRel = heap_open(tempOid, NoLock);
tempname = quote_qualified_identifier(get_namespace_name(RelationGetNamespace(tempRel)),
RelationGetRelationName(tempRel));
diffname = make_temptable_name_n(tempname, 2);
relnatts = matviewRel->rd_rel->relnatts;
usedForQual = (bool *) palloc0(sizeof(bool) * relnatts);
/* Open SPI context. */
if (SPI_connect() != SPI_OK_CONNECT)
elog(ERROR, "SPI_connect failed");
/* Analyze the temp table with the new contents. */
appendStringInfo(&querybuf, "ANALYZE %s", tempname);
if (SPI_exec(querybuf.data, 0) != SPI_OK_UTILITY)
elog(ERROR, "SPI_exec failed: %s", querybuf.data);
/*
* We need to ensure that there are not duplicate rows without NULLs in
* the new data set before we can count on the "diff" results. Check for
* that in a way that allows showing the first duplicated row found. Even
* after we pass this test, a unique index on the materialized view may
* find a duplicate key problem.
*/
resetStringInfo(&querybuf);
appendStringInfo(&querybuf,
"SELECT newdata FROM %s newdata "
"WHERE newdata IS NOT NULL AND EXISTS "
"(SELECT * FROM %s newdata2 WHERE newdata2 IS NOT NULL "
"AND newdata2 OPERATOR(pg_catalog.*=) newdata "
"AND newdata2.ctid OPERATOR(pg_catalog.<>) "
"newdata.ctid) LIMIT 1",
tempname, tempname);
if (SPI_execute(querybuf.data, false, 1) != SPI_OK_SELECT)
elog(ERROR, "SPI_exec failed: %s", querybuf.data);
if (SPI_processed > 0)
{
ereport(ERROR,
(errcode(ERRCODE_CARDINALITY_VIOLATION),
errmsg("new data for \"%s\" contains duplicate rows without any null columns",
RelationGetRelationName(matviewRel)),
errdetail("Row: %s",
SPI_getvalue(SPI_tuptable->vals[0], SPI_tuptable->tupdesc, 1))));
}
SetUserIdAndSecContext(relowner,
save_sec_context | SECURITY_LOCAL_USERID_CHANGE);
/* Start building the query for creating the diff table. */
resetStringInfo(&querybuf);
appendStringInfo(&querybuf,
"CREATE TEMP TABLE %s AS "
"SELECT mv.ctid AS tid, newdata "
"FROM %s mv FULL JOIN %s newdata ON (",
diffname, matviewname, tempname);
/*
* Get the list of index OIDs for the table from the relcache, and look up
* each one in the pg_index syscache. We will test for equality on all
* columns present in all unique indexes which only reference columns and
* include all rows.
*/
tupdesc = matviewRel->rd_att;
foundUniqueIndex = false;
indexoidlist = RelationGetIndexList(matviewRel);
foreach(indexoidscan, indexoidlist)
{
Oid indexoid = lfirst_oid(indexoidscan);
Relation indexRel;
Form_pg_index indexStruct;
indexRel = index_open(indexoid, RowExclusiveLock);
indexStruct = indexRel->rd_index;
/*
* We're only interested if it is unique, valid, contains no
* expressions, and is not partial.
*/
if (indexStruct->indisunique &&
IndexIsValid(indexStruct) &&
RelationGetIndexExpressions(indexRel) == NIL &&
RelationGetIndexPredicate(indexRel) == NIL)
{
int numatts = indexStruct->indnatts;
int i;
/* Add quals for all columns from this index. */
for (i = 0; i < numatts; i++)
{
int attnum = indexStruct->indkey.values[i];
Oid type;
Oid op;
const char *colname;
/*
* Only include the column once regardless of how many times
* it shows up in how many indexes.
*/
if (usedForQual[attnum - 1])
continue;
usedForQual[attnum - 1] = true;
/*
* Actually add the qual, ANDed with any others.
*/
if (foundUniqueIndex)
appendStringInfoString(&querybuf, " AND ");
colname = quote_identifier(NameStr((tupdesc->attrs[attnum - 1])->attname));
appendStringInfo(&querybuf, "newdata.%s ", colname);
type = attnumTypeId(matviewRel, attnum);
op = lookup_type_cache(type, TYPECACHE_EQ_OPR)->eq_opr;
mv_GenerateOper(&querybuf, op);
appendStringInfo(&querybuf, " mv.%s", colname);
foundUniqueIndex = true;
}
}
/* Keep the locks, since we're about to run DML which needs them. */
index_close(indexRel, NoLock);
}
/*
* lazy_truncate_heap - try to truncate off any empty pages at the end
*/
static void
lazy_truncate_heap(Relation onerel, LVRelStats *vacrelstats)
{
BlockNumber old_rel_pages = vacrelstats->rel_pages;
BlockNumber new_rel_pages;
PGRUsage ru0;
int lock_retry;
pg_rusage_init(&ru0);
/*
* Loop until no more truncating can be done.
*/
do
{
/*
* We need full exclusive lock on the relation in order to do
* truncation. If we can't get it, give up rather than waiting --- we
* don't want to block other backends, and we don't want to deadlock
* (which is quite possible considering we already hold a lower-grade
* lock).
*/
vacrelstats->lock_waiter_detected = false;
lock_retry = 0;
while (true)
{
if (ConditionalLockRelation(onerel, AccessExclusiveLock))
break;
/*
* Check for interrupts while trying to (re-)acquire the exclusive
* lock.
*/
CHECK_FOR_INTERRUPTS();
if (++lock_retry > (AUTOVACUUM_TRUNCATE_LOCK_TIMEOUT /
AUTOVACUUM_TRUNCATE_LOCK_WAIT_INTERVAL))
{
/*
* We failed to establish the lock in the specified number of
* retries. This means we give up truncating. Suppress the
* ANALYZE step. Doing an ANALYZE at this point will reset the
* dead_tuple_count in the stats collector, so we will not get
* called by the autovacuum launcher again to do the truncate.
*/
vacrelstats->lock_waiter_detected = true;
ereport(LOG,
(errmsg("automatic vacuum of table \"%s.%s.%s\": "
"could not (re)acquire exclusive "
"lock for truncate scan",
get_database_name(MyDatabaseId),
get_namespace_name(RelationGetNamespace(onerel)),
RelationGetRelationName(onerel))));
return;
}
pg_usleep(AUTOVACUUM_TRUNCATE_LOCK_WAIT_INTERVAL);
}
/*
* Now that we have exclusive lock, look to see if the rel has grown
* whilst we were vacuuming with non-exclusive lock. If so, give up;
* the newly added pages presumably contain non-deletable tuples.
*/
new_rel_pages = RelationGetNumberOfBlocks(onerel);
if (new_rel_pages != old_rel_pages)
{
/*
* Note: we intentionally don't update vacrelstats->rel_pages with
* the new rel size here. If we did, it would amount to assuming
* that the new pages are empty, which is unlikely. Leaving the
* numbers alone amounts to assuming that the new pages have the
* same tuple density as existing ones, which is less unlikely.
*/
UnlockRelation(onerel, AccessExclusiveLock);
return;
}
/*
* Scan backwards from the end to verify that the end pages actually
* contain no tuples. This is *necessary*, not optional, because
* other backends could have added tuples to these pages whilst we
* were vacuuming.
*/
new_rel_pages = count_nondeletable_pages(onerel, vacrelstats);
if (new_rel_pages >= old_rel_pages)
{
/* can't do anything after all */
UnlockRelation(onerel, AccessExclusiveLock);
return;
}
/*
* Okay to truncate.
*/
RelationTruncate(onerel, new_rel_pages);
/*
* We can release the exclusive lock as soon as we have truncated.
* Other backends can't safely access the relation until they have
* processed the smgr invalidation that smgrtruncate sent out ... but
* that should happen as part of standard invalidation processing once
* they acquire lock on the relation.
*/
UnlockRelation(onerel, AccessExclusiveLock);
/*
* Update statistics. Here, it *is* correct to adjust rel_pages
* without also touching reltuples, since the tuple count wasn't
* changed by the truncation.
*/
vacrelstats->pages_removed += old_rel_pages - new_rel_pages;
vacrelstats->rel_pages = new_rel_pages;
ereport(elevel,
(errmsg("\"%s\": truncated %u to %u pages",
RelationGetRelationName(onerel),
old_rel_pages, new_rel_pages),
errdetail("%s.",
pg_rusage_show(&ru0))));
old_rel_pages = new_rel_pages;
} while (new_rel_pages > vacrelstats->nonempty_pages &&
vacrelstats->lock_waiter_detected);
}
/*
* IsToastRelation
* True iff relation is a TOAST support relation (or index).
*/
bool
IsToastRelation(Relation relation)
{
return IsToastNamespace(RelationGetNamespace(relation));
}
Datum
spgstat(PG_FUNCTION_ARGS)
{
text *name=PG_GETARG_TEXT_P(0);
char *relname=text_to_cstring(name);
RangeVar *relvar;
Relation index;
List *relname_list;
Oid relOid;
BlockNumber blkno = SPGIST_HEAD_BLKNO;
BlockNumber totalPages = 0,
innerPages = 0,
emptyPages = 0;
double usedSpace = 0.0;
char res[1024];
int bufferSize = -1;
int64 innerTuples = 0,
leafTuples = 0;
relname_list = stringToQualifiedNameList(relname);
relvar = makeRangeVarFromNameList(relname_list);
relOid = RangeVarGetRelid(relvar, false);
index = index_open(relOid, AccessExclusiveLock);
if ( index->rd_am == NULL )
elog(ERROR, "Relation %s.%s is not an index",
get_namespace_name(RelationGetNamespace(index)),
RelationGetRelationName(index) );
totalPages = RelationGetNumberOfBlocks(index);
for(blkno=SPGIST_HEAD_BLKNO; blkno<totalPages; blkno++)
{
Buffer buffer;
Page page;
buffer = ReadBuffer(index, blkno);
LockBuffer(buffer, BUFFER_LOCK_SHARE);
page = BufferGetPage(buffer);
if (SpGistPageIsLeaf(page))
{
leafTuples += SpGistPageGetMaxOffset(page);
}
else
{
innerPages++;
innerTuples += SpGistPageGetMaxOffset(page);
}
if (bufferSize < 0)
bufferSize = BufferGetPageSize(buffer) - MAXALIGN(sizeof(SpGistPageOpaqueData)) -
SizeOfPageHeaderData;
usedSpace += bufferSize - (PageGetFreeSpace(page) + sizeof(ItemIdData));
if (PageGetFreeSpace(page) + sizeof(ItemIdData) == bufferSize)
emptyPages++;
UnlockReleaseBuffer(buffer);
}
index_close(index, AccessExclusiveLock);
totalPages--; /* metapage */
snprintf(res, sizeof(res),
"totalPages: %u\n"
"innerPages: %u\n"
"leafPages: %u\n"
"emptyPages: %u\n"
"usedSpace: %.2f kbytes\n"
"freeSpace: %.2f kbytes\n"
"fillRatio: %.2f%c\n"
"leafTuples: %lld\n"
"innerTuples: %lld",
totalPages, innerPages, totalPages - innerPages, emptyPages,
usedSpace / 1024.0,
(( (double) bufferSize ) * ( (double) totalPages ) - usedSpace) / 1024,
100.0 * ( usedSpace / (( (double) bufferSize ) * ( (double) totalPages )) ),
'%',
leafTuples, innerTuples
);
PG_RETURN_TEXT_P(CStringGetTextDatum(res));
}
/* ------------------------------------------------
* bt_metap()
*
* Get a btree meta-page information
*
* Usage: SELECT * FROM bt_metap('t1_pkey')
* ------------------------------------------------
*/
Datum
bt_metap(PG_FUNCTION_ARGS)
{
text *relname = PG_GETARG_TEXT_P(0);
Buffer buffer;
Relation rel;
RangeVar *relrv;
Datum result;
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
(errmsg("must be superuser to use pgstattuple functions"))));
relrv = makeRangeVarFromNameList(textToQualifiedNameList(relname));
rel = relation_openrv(relrv, AccessShareLock);
if (!IS_INDEX(rel) || !IS_BTREE(rel))
elog(ERROR, "bt_metap() can be used only on b-tree index.");
/*
* Reject attempts to read non-local temporary relations; we would
* be likely to get wrong data since we have no visibility into the
* owning session's local buffers.
*/
if (isOtherTempNamespace(RelationGetNamespace(rel)))
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("cannot access temporary tables of other sessions")));
buffer = ReadBuffer(rel, 0);
{
BTMetaPageData *metad;
TupleDesc tupleDesc;
int j;
char *values[BTMETAP_NCOLUMNS];
HeapTuple tuple;
Page page = BufferGetPage(buffer);
metad = BTPageGetMeta(page);
tupleDesc = RelationNameGetTupleDesc(BTMETAP_TYPE);
j = 0;
values[j] = palloc(32);
snprintf(values[j++], 32, "%d", metad->btm_magic);
values[j] = palloc(32);
snprintf(values[j++], 32, "%d", metad->btm_version);
values[j] = palloc(32);
snprintf(values[j++], 32, "%d", metad->btm_root);
values[j] = palloc(32);
snprintf(values[j++], 32, "%d", metad->btm_level);
values[j] = palloc(32);
snprintf(values[j++], 32, "%d", metad->btm_fastroot);
values[j] = palloc(32);
snprintf(values[j++], 32, "%d", metad->btm_fastlevel);
tuple = BuildTupleFromCStrings(TupleDescGetAttInMetadata(tupleDesc),
values);
result = TupleGetDatum(TupleDescGetSlot(tupleDesc), tuple);
}
ReleaseBuffer(buffer);
relation_close(rel, AccessShareLock);
PG_RETURN_DATUM(result);
}
char *
SPI_getnspname(Relation rel)
{
return get_namespace_name(RelationGetNamespace(rel));
}
/*
* IsAoSegmentRelation
* True iff relation is an AO segment support relation (or index).
*/
bool
IsAoSegmentRelation(Relation relation)
{
return IsAoSegmentNamespace(RelationGetNamespace(relation));
}
/*
* lazy_vacuum_rel() -- perform LAZY VACUUM for one heap relation
*
* This routine vacuums a single heap, cleans out its indexes, and
* updates its relpages and reltuples statistics.
*
* At entry, we have already established a transaction and opened
* and locked the relation.
*
* The return value indicates whether this function has held off
* interrupts -- caller must RESUME_INTERRUPTS() after commit if true.
*/
bool
lazy_vacuum_rel(Relation onerel, VacuumStmt *vacstmt,
BufferAccessStrategy bstrategy, List *updated_stats)
{
LVRelStats *vacrelstats;
Relation *Irel;
int nindexes;
BlockNumber possibly_freeable;
PGRUsage ru0;
TimestampTz starttime = 0;
bool heldoff = false;
pg_rusage_init(&ru0);
/* measure elapsed time iff autovacuum logging requires it */
if (IsAutoVacuumWorkerProcess() && Log_autovacuum_min_duration > 0)
starttime = GetCurrentTimestamp();
if (vacstmt->verbose)
elevel = INFO;
else
elevel = DEBUG2;
if (Gp_role == GP_ROLE_DISPATCH)
elevel = DEBUG2; /* vacuum and analyze messages aren't interesting from the QD */
#ifdef FAULT_INJECTOR
if (vacuumStatement_IsInAppendOnlyDropPhase(vacstmt))
{
FaultInjector_InjectFaultIfSet(
CompactionBeforeSegmentFileDropPhase,
DDLNotSpecified,
"", // databaseName
""); // tableName
}
if (vacummStatement_IsInAppendOnlyCleanupPhase(vacstmt))
{
FaultInjector_InjectFaultIfSet(
CompactionBeforeCleanupPhase,
DDLNotSpecified,
"", // databaseName
""); // tableName
}
#endif
/*
* MPP-23647. Update xid limits for heap as well as appendonly
* relations. This allows setting relfrozenxid to correct value
* for an appendonly (AO/CO) table.
*/
vac_strategy = bstrategy;
vacuum_set_xid_limits(vacstmt->freeze_min_age, onerel->rd_rel->relisshared,
&OldestXmin, &FreezeLimit);
/*
* Execute the various vacuum operations. Appendonly tables are treated
* differently.
*/
if (RelationIsAoRows(onerel) || RelationIsAoCols(onerel))
{
lazy_vacuum_aorel(onerel, vacstmt, updated_stats);
return false;
}
vacrelstats = (LVRelStats *) palloc0(sizeof(LVRelStats));
/* heap relation */
/* Set threshold for interesting free space = average request size */
/* XXX should we scale it up or down? Adjust vacuum.c too, if so */
vacrelstats->threshold = GetAvgFSMRequestSize(&onerel->rd_node);
vacrelstats->num_index_scans = 0;
/* Open all indexes of the relation */
vac_open_indexes(onerel, RowExclusiveLock, &nindexes, &Irel);
vacrelstats->hasindex = (nindexes > 0);
/* Do the vacuuming */
lazy_scan_heap(onerel, vacrelstats, Irel, nindexes, updated_stats);
/* Done with indexes */
vac_close_indexes(nindexes, Irel, NoLock);
/*
* Optionally truncate the relation.
*
* Don't even think about it unless we have a shot at releasing a goodly
* number of pages. Otherwise, the time taken isn't worth it.
*
* Note that after we've truncated the heap, it's too late to abort the
* transaction; doing so would lose the sinval messages needed to tell
* the other backends about the table being shrunk. We prevent interrupts
* in that case; caller is responsible for re-enabling them after
* committing the transaction.
*/
possibly_freeable = vacrelstats->rel_pages - vacrelstats->nonempty_pages;
if (possibly_freeable > 0 &&
(possibly_freeable >= REL_TRUNCATE_MINIMUM ||
possibly_freeable >= vacrelstats->rel_pages / REL_TRUNCATE_FRACTION))
{
HOLD_INTERRUPTS();
heldoff = true;
lazy_truncate_heap(onerel, vacrelstats);
}
/* Update shared free space map with final free space info */
lazy_update_fsm(onerel, vacrelstats);
if (vacrelstats->tot_free_pages > MaxFSMPages)
ereport(WARNING,
(errmsg("relation \"%s.%s\" contains more than \"max_fsm_pages\" pages with useful free space",
get_namespace_name(RelationGetNamespace(onerel)),
RelationGetRelationName(onerel)),
/* Only suggest VACUUM FULL if > 20% free */
(vacrelstats->tot_free_pages > vacrelstats->rel_pages * 0.20) ?
errhint("Consider using VACUUM FULL on this relation or increasing the configuration parameter \"max_fsm_pages\".") :
errhint("Consider increasing the configuration parameter \"max_fsm_pages\".")));
/* Update statistics in pg_class */
vac_update_relstats_from_list(onerel,
vacrelstats->rel_pages,
vacrelstats->rel_tuples,
vacrelstats->hasindex,
FreezeLimit,
updated_stats);
/* report results to the stats collector, too */
pgstat_report_vacuum(RelationGetRelid(onerel), onerel->rd_rel->relisshared,
true /*vacrelstats->scanned_all*/,
vacstmt->analyze, vacrelstats->rel_tuples);
if (gp_indexcheck_vacuum == INDEX_CHECK_ALL ||
(gp_indexcheck_vacuum == INDEX_CHECK_SYSTEM &&
PG_CATALOG_NAMESPACE == RelationGetNamespace(onerel)))
{
int i;
for (i = 0; i < nindexes; i++)
{
if (Irel[i]->rd_rel->relam == BTREE_AM_OID)
_bt_validate_vacuum(Irel[i], onerel, OldestXmin);
}
}
/* and log the action if appropriate */
if (IsAutoVacuumWorkerProcess() && Log_autovacuum_min_duration >= 0)
{
if (Log_autovacuum_min_duration == 0 ||
TimestampDifferenceExceeds(starttime, GetCurrentTimestamp(),
Log_autovacuum_min_duration))
ereport(LOG,
(errmsg("automatic vacuum of table \"%s.%s.%s\": index scans: %d\n"
"pages: %d removed, %d remain\n"
"tuples: %.0f removed, %.0f remain\n"
"system usage: %s",
get_database_name(MyDatabaseId),
get_namespace_name(RelationGetNamespace(onerel)),
RelationGetRelationName(onerel),
vacrelstats->num_index_scans,
vacrelstats->pages_removed, vacrelstats->rel_pages,
vacrelstats->tuples_deleted, vacrelstats->rel_tuples,
pg_rusage_show(&ru0))));
}
return heldoff;
}
/*
table_log()
trigger function for logging table changes
parameter:
- log table name (optional)
return:
- trigger data (for Pg)
*/
Datum table_log(PG_FUNCTION_ARGS) {
TriggerData *trigdata = (TriggerData *) fcinfo->context;
int ret;
char query[250 + NAMEDATALEN]; /* for getting table infos (250 chars (+ one times the length of all names) should be enough) */
int number_columns = 0; /* counts the number columns in the table */
int number_columns_log = 0; /* counts the number columns in the table */
char *orig_schema;
char *log_schema;
char *log_table;
int use_session_user = 0; /* should we write the current (session) user to the log table? */
/*
* Some checks first...
*/
#ifdef TABLE_LOG_DEBUG
elog(NOTICE, "start table_log()");
#endif /* TABLE_LOG_DEBUG */
/* called by trigger manager? */
if (!CALLED_AS_TRIGGER(fcinfo)) {
elog(ERROR, "table_log: not fired by trigger manager");
}
/* must only be called for ROW trigger */
if (TRIGGER_FIRED_FOR_STATEMENT(trigdata->tg_event)) {
elog(ERROR, "table_log: can't process STATEMENT events");
}
/* must only be called AFTER */
if (TRIGGER_FIRED_BEFORE(trigdata->tg_event)) {
elog(ERROR, "table_log: must be fired after event");
}
/* now connect to SPI manager */
ret = SPI_connect();
if (ret != SPI_OK_CONNECT) {
elog(ERROR, "table_log: SPI_connect returned %d", ret);
}
/* get schema name for the table, in case we need it later */
orig_schema = get_namespace_name(RelationGetNamespace(trigdata->tg_relation));
#ifdef TABLE_LOG_DEBUG
elog(NOTICE, "prechecks done, now getting original table attributes");
#endif /* TABLE_LOG_DEBUG */
number_columns = count_columns(trigdata->tg_relation->rd_att);
if (number_columns < 1) {
elog(ERROR, "table_log: number of columns in table is < 1, can this happen?");
}
#ifdef TABLE_LOG_DEBUG
elog(NOTICE, "number columns in orig table: %i", number_columns);
#endif /* TABLE_LOG_DEBUG */
if (trigdata->tg_trigger->tgnargs > 3) {
elog(ERROR, "table_log: too many arguments to trigger");
}
/* name of the log schema */
if (trigdata->tg_trigger->tgnargs > 2) {
/* check if a log schema argument is given, if yes, use it */
log_schema = trigdata->tg_trigger->tgargs[2];
} else {
/* if no, use orig_schema */
log_schema = orig_schema;
}
/* should we write the current user? */
if (trigdata->tg_trigger->tgnargs > 1) {
/* check if a second argument is given */
/* if yes, use it, if it is true */
if (atoi(trigdata->tg_trigger->tgargs[1]) == 1) {
use_session_user = 1;
#ifdef TABLE_LOG_DEBUG
elog(NOTICE, "will write session user to 'trigger_user'");
#endif /* TABLE_LOG_DEBUG */
}
}
/* name of the log table */
if (trigdata->tg_trigger->tgnargs > 0) {
/* check if a logtable argument is given */
/* if yes, use it */
log_table = (char *) palloc((strlen(trigdata->tg_trigger->tgargs[0]) + 2) * sizeof(char));
sprintf(log_table, "%s", trigdata->tg_trigger->tgargs[0]);
} else {
/* if no, use 'table name' + '_log' */
log_table = (char *) palloc((strlen(do_quote_ident(SPI_getrelname(trigdata->tg_relation))) + 5) * sizeof(char));
sprintf(log_table, "%s_log", SPI_getrelname(trigdata->tg_relation));
}
#ifdef TABLE_LOG_DEBUG
elog(NOTICE, "log table: %s", log_table);
#endif /* TABLE_LOG_DEBUG */
#ifdef TABLE_LOG_DEBUG
elog(NOTICE, "now check, if log table exists");
#endif /* TABLE_LOG_DEBUG */
/* get the number columns in the table */
snprintf(query, 249, "%s.%s", do_quote_ident(log_schema), do_quote_ident(log_table));
number_columns_log = count_columns(RelationNameGetTupleDesc(query));
if (number_columns_log < 1) {
elog(ERROR, "could not get number columns in relation %s", log_table);
}
#ifdef TABLE_LOG_DEBUG
elog(NOTICE, "number columns in log table: %i", number_columns_log);
#endif /* TABLE_LOG_DEBUG */
/* check if the logtable has 3 (or now 4) columns more than our table */
/* +1 if we should write the session user */
if (use_session_user == 0) {
/* without session user */
if (number_columns_log != number_columns + 3 && number_columns_log != number_columns + 4) {
elog(ERROR, "number colums in relation %s(%d) does not match columns in %s(%d)", SPI_getrelname(trigdata->tg_relation), number_columns, log_table, number_columns_log);
}
} else {
/* with session user */
if (number_columns_log != number_columns + 3 + 1 && number_columns_log != number_columns + 4 + 1) {
elog(ERROR, "number colums in relation %s does not match columns in %s", SPI_getrelname(trigdata->tg_relation), log_table);
}
}
#ifdef TABLE_LOG_DEBUG
elog(NOTICE, "log table OK");
#endif /* TABLE_LOG_DEBUG */
/* For each column in key ... */
#ifdef TABLE_LOG_DEBUG
elog(NOTICE, "copy data ...");
#endif /* TABLE_LOG_DEBUG */
if (TRIGGER_FIRED_BY_INSERT(trigdata->tg_event)) {
/* trigger called from INSERT */
#ifdef TABLE_LOG_DEBUG
elog(NOTICE, "mode: INSERT -> new");
#endif /* TABLE_LOG_DEBUG */
__table_log(trigdata, "INSERT", "new", trigdata->tg_trigtuple, number_columns, log_table, use_session_user, log_schema);
} else if (TRIGGER_FIRED_BY_UPDATE(trigdata->tg_event)) {
/* trigger called from UPDATE */
#ifdef TABLE_LOG_DEBUG
elog(NOTICE, "mode: UPDATE -> old");
#endif /* TABLE_LOG_DEBUG */
__table_log(trigdata, "UPDATE", "old", trigdata->tg_trigtuple, number_columns, log_table, use_session_user, log_schema);
#ifdef TABLE_LOG_DEBUG
elog(NOTICE, "mode: UPDATE -> new");
#endif /* TABLE_LOG_DEBUG */
__table_log(trigdata, "UPDATE", "new", trigdata->tg_newtuple, number_columns, log_table, use_session_user, log_schema);
} else if (TRIGGER_FIRED_BY_DELETE(trigdata->tg_event)) {
/* trigger called from DELETE */
#ifdef TABLE_LOG_DEBUG
elog(NOTICE, "mode: DELETE -> old");
#endif /* TABLE_LOG_DEBUG */
__table_log(trigdata, "DELETE", "old", trigdata->tg_trigtuple, number_columns, log_table, use_session_user, log_schema);
} else {
elog(ERROR, "trigger fired by unknown event");
}
#ifdef TABLE_LOG_DEBUG
elog(NOTICE, "cleanup, trigger done");
#endif /* TABLE_LOG_DEBUG */
/* clean up */
pfree(log_table);
/* close SPI connection */
SPI_finish();
/* return trigger data */
return PointerGetDatum(trigdata->tg_trigtuple);
}
/*
* lazy_scan_heap() -- scan an open heap relation
*
* This routine sets commit status bits, builds lists of dead tuples
* and pages with free space, and calculates statistics on the number
* of live tuples in the heap. When done, or when we run low on space
* for dead-tuple TIDs, invoke vacuuming of indexes and heap.
*
* If there are no indexes then we just vacuum each dirty page as we
* process it, since there's no point in gathering many tuples.
*/
static void
lazy_scan_heap(Relation onerel, LVRelStats *vacrelstats,
Relation *Irel, int nindexes, bool scan_all)
{
BlockNumber nblocks,
blkno;
HeapTupleData tuple;
char *relname;
BlockNumber empty_pages,
vacuumed_pages;
double num_tuples,
tups_vacuumed,
nkeep,
nunused;
IndexBulkDeleteResult **indstats;
int i;
PGRUsage ru0;
Buffer vmbuffer = InvalidBuffer;
BlockNumber next_not_all_visible_block;
bool skipping_all_visible_blocks;
pg_rusage_init(&ru0);
relname = RelationGetRelationName(onerel);
ereport(elevel,
(errmsg("vacuuming \"%s.%s\"",
get_namespace_name(RelationGetNamespace(onerel)),
relname)));
empty_pages = vacuumed_pages = 0;
num_tuples = tups_vacuumed = nkeep = nunused = 0;
indstats = (IndexBulkDeleteResult **)
palloc0(nindexes * sizeof(IndexBulkDeleteResult *));
nblocks = RelationGetNumberOfBlocks(onerel);
vacrelstats->rel_pages = nblocks;
vacrelstats->scanned_pages = 0;
vacrelstats->nonempty_pages = 0;
vacrelstats->latestRemovedXid = InvalidTransactionId;
lazy_space_alloc(vacrelstats, nblocks);
/*
* We want to skip pages that don't require vacuuming according to the
* visibility map, but only when we can skip at least SKIP_PAGES_THRESHOLD
* consecutive pages. Since we're reading sequentially, the OS should be
* doing readahead for us, so there's no gain in skipping a page now and
* then; that's likely to disable readahead and so be counterproductive.
* Also, skipping even a single page means that we can't update
* relfrozenxid, so we only want to do it if we can skip a goodly number
* of pages.
*
* Before entering the main loop, establish the invariant that
* next_not_all_visible_block is the next block number >= blkno that's not
* all-visible according to the visibility map, or nblocks if there's no
* such block. Also, we set up the skipping_all_visible_blocks flag,
* which is needed because we need hysteresis in the decision: once we've
* started skipping blocks, we may as well skip everything up to the next
* not-all-visible block.
*
* Note: if scan_all is true, we won't actually skip any pages; but we
* maintain next_not_all_visible_block anyway, so as to set up the
* all_visible_according_to_vm flag correctly for each page.
*/
for (next_not_all_visible_block = 0;
next_not_all_visible_block < nblocks;
next_not_all_visible_block++)
{
if (!visibilitymap_test(onerel, next_not_all_visible_block, &vmbuffer))
break;
vacuum_delay_point();
}
if (next_not_all_visible_block >= SKIP_PAGES_THRESHOLD)
skipping_all_visible_blocks = true;
else
skipping_all_visible_blocks = false;
for (blkno = 0; blkno < nblocks; blkno++)
{
Buffer buf;
Page page;
OffsetNumber offnum,
maxoff;
bool tupgone,
hastup;
int prev_dead_count;
OffsetNumber frozen[MaxOffsetNumber];
int nfrozen;
Size freespace;
bool all_visible_according_to_vm;
bool all_visible;
bool has_dead_tuples;
if (blkno == next_not_all_visible_block)
{
/* Time to advance next_not_all_visible_block */
for (next_not_all_visible_block++;
next_not_all_visible_block < nblocks;
next_not_all_visible_block++)
{
if (!visibilitymap_test(onerel, next_not_all_visible_block,
&vmbuffer))
break;
vacuum_delay_point();
}
/*
* We know we can't skip the current block. But set up
* skipping_all_visible_blocks to do the right thing at the
* following blocks.
*/
if (next_not_all_visible_block - blkno > SKIP_PAGES_THRESHOLD)
skipping_all_visible_blocks = true;
else
skipping_all_visible_blocks = false;
all_visible_according_to_vm = false;
}
else
{
/* Current block is all-visible */
if (skipping_all_visible_blocks && !scan_all)
continue;
all_visible_according_to_vm = true;
}
vacuum_delay_point();
vacrelstats->scanned_pages++;
/*
* If we are close to overrunning the available space for dead-tuple
* TIDs, pause and do a cycle of vacuuming before we tackle this page.
*/
if ((vacrelstats->max_dead_tuples - vacrelstats->num_dead_tuples) < MaxHeapTuplesPerPage &&
vacrelstats->num_dead_tuples > 0)
{
/* Log cleanup info before we touch indexes */
vacuum_log_cleanup_info(onerel, vacrelstats);
/* Remove index entries */
for (i = 0; i < nindexes; i++)
lazy_vacuum_index(Irel[i],
&indstats[i],
vacrelstats);
/* Remove tuples from heap */
lazy_vacuum_heap(onerel, vacrelstats);
/*
* Forget the now-vacuumed tuples, and press on, but be careful
* not to reset latestRemovedXid since we want that value to be
* valid.
*/
vacrelstats->num_dead_tuples = 0;
vacrelstats->num_index_scans++;
}
buf = ReadBufferExtended(onerel, MAIN_FORKNUM, blkno,
RBM_NORMAL, vac_strategy);
/* We need buffer cleanup lock so that we can prune HOT chains. */
LockBufferForCleanup(buf);
page = BufferGetPage(buf);
if (PageIsNew(page))
{
/*
* An all-zeroes page could be left over if a backend extends the
* relation but crashes before initializing the page. Reclaim such
* pages for use.
*
* We have to be careful here because we could be looking at a
* page that someone has just added to the relation and not yet
* been able to initialize (see RelationGetBufferForTuple). To
* protect against that, release the buffer lock, grab the
* relation extension lock momentarily, and re-lock the buffer. If
* the page is still uninitialized by then, it must be left over
* from a crashed backend, and we can initialize it.
*
* We don't really need the relation lock when this is a new or
* temp relation, but it's probably not worth the code space to
* check that, since this surely isn't a critical path.
*
* Note: the comparable code in vacuum.c need not worry because
* it's got exclusive lock on the whole relation.
*/
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
LockRelationForExtension(onerel, ExclusiveLock);
UnlockRelationForExtension(onerel, ExclusiveLock);
LockBufferForCleanup(buf);
if (PageIsNew(page))
{
ereport(WARNING,
(errmsg("relation \"%s\" page %u is uninitialized --- fixing",
relname, blkno)));
PageInit(page, BufferGetPageSize(buf), 0);
empty_pages++;
}
freespace = PageGetHeapFreeSpace(page);
MarkBufferDirty(buf);
UnlockReleaseBuffer(buf);
RecordPageWithFreeSpace(onerel, blkno, freespace);
continue;
}
if (PageIsEmpty(page))
{
empty_pages++;
freespace = PageGetHeapFreeSpace(page);
if (!PageIsAllVisible(page))
{
PageSetAllVisible(page);
SetBufferCommitInfoNeedsSave(buf);
}
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
/* Update the visibility map */
if (!all_visible_according_to_vm)
{
visibilitymap_pin(onerel, blkno, &vmbuffer);
LockBuffer(buf, BUFFER_LOCK_SHARE);
if (PageIsAllVisible(page))
visibilitymap_set(onerel, blkno, PageGetLSN(page), &vmbuffer);
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
}
ReleaseBuffer(buf);
RecordPageWithFreeSpace(onerel, blkno, freespace);
continue;
}
/*
* Prune all HOT-update chains in this page.
*
* We count tuples removed by the pruning step as removed by VACUUM.
*/
tups_vacuumed += heap_page_prune(onerel, buf, OldestXmin, false,
&vacrelstats->latestRemovedXid);
/*
* Now scan the page to collect vacuumable items and check for tuples
* requiring freezing.
*/
all_visible = true;
has_dead_tuples = false;
nfrozen = 0;
hastup = false;
prev_dead_count = vacrelstats->num_dead_tuples;
maxoff = PageGetMaxOffsetNumber(page);
for (offnum = FirstOffsetNumber;
offnum <= maxoff;
offnum = OffsetNumberNext(offnum))
{
ItemId itemid;
itemid = PageGetItemId(page, offnum);
/* Unused items require no processing, but we count 'em */
if (!ItemIdIsUsed(itemid))
{
nunused += 1;
continue;
}
/* Redirect items mustn't be touched */
if (ItemIdIsRedirected(itemid))
{
hastup = true; /* this page won't be truncatable */
continue;
}
ItemPointerSet(&(tuple.t_self), blkno, offnum);
/*
* DEAD item pointers are to be vacuumed normally; but we don't
* count them in tups_vacuumed, else we'd be double-counting (at
* least in the common case where heap_page_prune() just freed up
* a non-HOT tuple).
*/
if (ItemIdIsDead(itemid))
{
lazy_record_dead_tuple(vacrelstats, &(tuple.t_self));
all_visible = false;
continue;
}
Assert(ItemIdIsNormal(itemid));
tuple.t_data = (HeapTupleHeader) PageGetItem(page, itemid);
tuple.t_len = ItemIdGetLength(itemid);
tupgone = false;
switch (HeapTupleSatisfiesVacuum(tuple.t_data, OldestXmin, buf))
{
case HEAPTUPLE_DEAD:
/*
* Ordinarily, DEAD tuples would have been removed by
* heap_page_prune(), but it's possible that the tuple
* state changed since heap_page_prune() looked. In
* particular an INSERT_IN_PROGRESS tuple could have
* changed to DEAD if the inserter aborted. So this
* cannot be considered an error condition.
*
* If the tuple is HOT-updated then it must only be
* removed by a prune operation; so we keep it just as if
* it were RECENTLY_DEAD. Also, if it's a heap-only
* tuple, we choose to keep it, because it'll be a lot
* cheaper to get rid of it in the next pruning pass than
* to treat it like an indexed tuple.
*/
if (HeapTupleIsHotUpdated(&tuple) ||
HeapTupleIsHeapOnly(&tuple))
nkeep += 1;
else
tupgone = true; /* we can delete the tuple */
all_visible = false;
break;
case HEAPTUPLE_LIVE:
/* Tuple is good --- but let's do some validity checks */
if (onerel->rd_rel->relhasoids &&
!OidIsValid(HeapTupleGetOid(&tuple)))
elog(WARNING, "relation \"%s\" TID %u/%u: OID is invalid",
relname, blkno, offnum);
/*
* Is the tuple definitely visible to all transactions?
*
* NB: Like with per-tuple hint bits, we can't set the
* PD_ALL_VISIBLE flag if the inserter committed
* asynchronously. See SetHintBits for more info. Check
* that the HEAP_XMIN_COMMITTED hint bit is set because of
* that.
*/
if (all_visible)
{
TransactionId xmin;
if (!(tuple.t_data->t_infomask & HEAP_XMIN_COMMITTED))
{
all_visible = false;
break;
}
/*
* The inserter definitely committed. But is it old
* enough that everyone sees it as committed?
*/
xmin = HeapTupleHeaderGetXmin(tuple.t_data);
if (!TransactionIdPrecedes(xmin, OldestXmin))
{
all_visible = false;
break;
}
}
break;
case HEAPTUPLE_RECENTLY_DEAD:
/*
* If tuple is recently deleted then we must not remove it
* from relation.
*/
nkeep += 1;
all_visible = false;
break;
case HEAPTUPLE_INSERT_IN_PROGRESS:
/* This is an expected case during concurrent vacuum */
all_visible = false;
break;
case HEAPTUPLE_DELETE_IN_PROGRESS:
/* This is an expected case during concurrent vacuum */
all_visible = false;
break;
default:
elog(ERROR, "unexpected HeapTupleSatisfiesVacuum result");
break;
}
if (tupgone)
{
lazy_record_dead_tuple(vacrelstats, &(tuple.t_self));
HeapTupleHeaderAdvanceLatestRemovedXid(tuple.t_data,
&vacrelstats->latestRemovedXid);
tups_vacuumed += 1;
has_dead_tuples = true;
}
else
{
num_tuples += 1;
hastup = true;
/*
* Each non-removable tuple must be checked to see if it needs
* freezing. Note we already have exclusive buffer lock.
*/
if (heap_freeze_tuple(tuple.t_data, FreezeLimit,
InvalidBuffer))
frozen[nfrozen++] = offnum;
}
} /* scan along page */
/*
* If we froze any tuples, mark the buffer dirty, and write a WAL
* record recording the changes. We must log the changes to be
* crash-safe against future truncation of CLOG.
*/
if (nfrozen > 0)
{
MarkBufferDirty(buf);
if (RelationNeedsWAL(onerel))
{
XLogRecPtr recptr;
recptr = log_heap_freeze(onerel, buf, FreezeLimit,
frozen, nfrozen);
PageSetLSN(page, recptr);
PageSetTLI(page, ThisTimeLineID);
}
}
/*
* If there are no indexes then we can vacuum the page right now
* instead of doing a second scan.
*/
if (nindexes == 0 &&
vacrelstats->num_dead_tuples > 0)
{
/* Remove tuples from heap */
lazy_vacuum_page(onerel, blkno, buf, 0, vacrelstats);
/*
* Forget the now-vacuumed tuples, and press on, but be careful
* not to reset latestRemovedXid since we want that value to be
* valid.
*/
vacrelstats->num_dead_tuples = 0;
vacuumed_pages++;
}
freespace = PageGetHeapFreeSpace(page);
/* Update the all-visible flag on the page */
if (!PageIsAllVisible(page) && all_visible)
{
PageSetAllVisible(page);
SetBufferCommitInfoNeedsSave(buf);
}
/*
* It's possible for the value returned by GetOldestXmin() to move
* backwards, so it's not wrong for us to see tuples that appear to
* not be visible to everyone yet, while PD_ALL_VISIBLE is already
* set. The real safe xmin value never moves backwards, but
* GetOldestXmin() is conservative and sometimes returns a value
* that's unnecessarily small, so if we see that contradiction it just
* means that the tuples that we think are not visible to everyone yet
* actually are, and the PD_ALL_VISIBLE flag is correct.
*
* There should never be dead tuples on a page with PD_ALL_VISIBLE
* set, however.
*/
else if (PageIsAllVisible(page) && has_dead_tuples)
{
elog(WARNING, "page containing dead tuples is marked as all-visible in relation \"%s\" page %u",
relname, blkno);
PageClearAllVisible(page);
SetBufferCommitInfoNeedsSave(buf);
/*
* Normally, we would drop the lock on the heap page before
* updating the visibility map, but since this case shouldn't
* happen anyway, don't worry about that.
*/
visibilitymap_clear(onerel, blkno);
}
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
/* Update the visibility map */
if (!all_visible_according_to_vm && all_visible)
{
visibilitymap_pin(onerel, blkno, &vmbuffer);
LockBuffer(buf, BUFFER_LOCK_SHARE);
if (PageIsAllVisible(page))
visibilitymap_set(onerel, blkno, PageGetLSN(page), &vmbuffer);
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
}
ReleaseBuffer(buf);
/* Remember the location of the last page with nonremovable tuples */
if (hastup)
vacrelstats->nonempty_pages = blkno + 1;
/*
* If we remembered any tuples for deletion, then the page will be
* visited again by lazy_vacuum_heap, which will compute and record
* its post-compaction free space. If not, then we're done with this
* page, so remember its free space as-is. (This path will always be
* taken if there are no indexes.)
*/
if (vacrelstats->num_dead_tuples == prev_dead_count)
RecordPageWithFreeSpace(onerel, blkno, freespace);
}
/* save stats for use later */
vacrelstats->scanned_tuples = num_tuples;
vacrelstats->tuples_deleted = tups_vacuumed;
/* now we can compute the new value for pg_class.reltuples */
vacrelstats->new_rel_tuples = vac_estimate_reltuples(onerel, false,
nblocks,
vacrelstats->scanned_pages,
num_tuples);
/* If any tuples need to be deleted, perform final vacuum cycle */
/* XXX put a threshold on min number of tuples here? */
if (vacrelstats->num_dead_tuples > 0)
{
/* Log cleanup info before we touch indexes */
vacuum_log_cleanup_info(onerel, vacrelstats);
/* Remove index entries */
for (i = 0; i < nindexes; i++)
lazy_vacuum_index(Irel[i],
&indstats[i],
vacrelstats);
/* Remove tuples from heap */
lazy_vacuum_heap(onerel, vacrelstats);
vacrelstats->num_index_scans++;
}
/* Release the pin on the visibility map page */
if (BufferIsValid(vmbuffer))
{
ReleaseBuffer(vmbuffer);
vmbuffer = InvalidBuffer;
}
/* Do post-vacuum cleanup and statistics update for each index */
for (i = 0; i < nindexes; i++)
lazy_cleanup_index(Irel[i], indstats[i], vacrelstats);
/* If no indexes, make log report that lazy_vacuum_heap would've made */
if (vacuumed_pages)
ereport(elevel,
(errmsg("\"%s\": removed %.0f row versions in %u pages",
RelationGetRelationName(onerel),
tups_vacuumed, vacuumed_pages)));
ereport(elevel,
(errmsg("\"%s\": found %.0f removable, %.0f nonremovable row versions in %u out of %u pages",
RelationGetRelationName(onerel),
tups_vacuumed, num_tuples,
vacrelstats->scanned_pages, nblocks),
errdetail("%.0f dead row versions cannot be removed yet.\n"
"There were %.0f unused item pointers.\n"
"%u pages are entirely empty.\n"
"%s.",
nkeep,
nunused,
empty_pages,
pg_rusage_show(&ru0))));
}
/*
* lazy_scan_heap() -- scan an open heap relation
*
* This routine sets commit status bits, builds lists of dead tuples
* and pages with free space, and calculates statistics on the number
* of live tuples in the heap. When done, or when we run low on space
* for dead-tuple TIDs, invoke vacuuming of indexes and heap.
*
* If there are no indexes then we just vacuum each dirty page as we
* process it, since there's no point in gathering many tuples.
*/
static void
lazy_scan_heap(Relation onerel, LVRelStats *vacrelstats,
Relation *Irel, int nindexes, List *updated_stats)
{
MIRROREDLOCK_BUFMGR_DECLARE;
BlockNumber nblocks,
blkno;
HeapTupleData tuple;
char *relname;
BlockNumber empty_pages,
vacuumed_pages;
double num_tuples,
tups_vacuumed,
nkeep,
nunused;
IndexBulkDeleteResult **indstats;
int i;
int reindex_count = 1;
PGRUsage ru0;
/* Fetch gp_persistent_relation_node information that will be added to XLOG record. */
RelationFetchGpRelationNodeForXLog(onerel);
pg_rusage_init(&ru0);
relname = RelationGetRelationName(onerel);
ereport(elevel,
(errmsg("vacuuming \"%s.%s\"",
get_namespace_name(RelationGetNamespace(onerel)),
relname)));
empty_pages = vacuumed_pages = 0;
num_tuples = tups_vacuumed = nkeep = nunused = 0;
indstats = (IndexBulkDeleteResult **)
palloc0(nindexes * sizeof(IndexBulkDeleteResult *));
nblocks = RelationGetNumberOfBlocks(onerel);
vacrelstats->rel_pages = nblocks;
vacrelstats->nonempty_pages = 0;
lazy_space_alloc(vacrelstats, nblocks);
for (blkno = 0; blkno < nblocks; blkno++)
{
Buffer buf;
Page page;
OffsetNumber offnum,
maxoff;
bool tupgone,
hastup;
int prev_dead_count;
OffsetNumber frozen[MaxOffsetNumber];
int nfrozen;
vacuum_delay_point();
/*
* If we are close to overrunning the available space for dead-tuple
* TIDs, pause and do a cycle of vacuuming before we tackle this page.
*/
if ((vacrelstats->max_dead_tuples - vacrelstats->num_dead_tuples) < MaxHeapTuplesPerPage &&
vacrelstats->num_dead_tuples > 0)
{
/* Remove index entries */
for (i = 0; i < nindexes; i++)
lazy_vacuum_index(Irel[i], &indstats[i], vacrelstats);
reindex_count++;
/* Remove tuples from heap */
lazy_vacuum_heap(onerel, vacrelstats);
/* Forget the now-vacuumed tuples, and press on */
vacrelstats->num_dead_tuples = 0;
vacrelstats->num_index_scans++;
}
/* -------- MirroredLock ---------- */
MIRROREDLOCK_BUFMGR_LOCK;
buf = ReadBufferWithStrategy(onerel, blkno, vac_strategy);
/* We need buffer cleanup lock so that we can prune HOT chains. */
LockBufferForCleanup(buf);
page = BufferGetPage(buf);
if (PageIsNew(page))
{
/*
* An all-zeroes page could be left over if a backend extends the
* relation but crashes before initializing the page. Reclaim such
* pages for use.
*
* We have to be careful here because we could be looking at a
* page that someone has just added to the relation and not yet
* been able to initialize (see RelationGetBufferForTuple). To
* protect against that, release the buffer lock, grab the
* relation extension lock momentarily, and re-lock the buffer. If
* the page is still uninitialized by then, it must be left over
* from a crashed backend, and we can initialize it.
*
* We don't really need the relation lock when this is a new or
* temp relation, but it's probably not worth the code space to
* check that, since this surely isn't a critical path.
*
* Note: the comparable code in vacuum.c need not worry because
* it's got exclusive lock on the whole relation.
*/
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
MIRROREDLOCK_BUFMGR_UNLOCK;
/* -------- MirroredLock ---------- */
LockRelationForExtension(onerel, ExclusiveLock);
UnlockRelationForExtension(onerel, ExclusiveLock);
/* -------- MirroredLock ---------- */
MIRROREDLOCK_BUFMGR_LOCK;
LockBufferForCleanup(buf);
if (PageIsNew(page))
{
ereport(WARNING,
(errmsg("relation \"%s\" page %u is uninitialized --- fixing",
relname, blkno)));
PageInit(page, BufferGetPageSize(buf), 0);
/* must record in xlog so that changetracking will know about this change */
log_heap_newpage(onerel, page, blkno);
empty_pages++;
lazy_record_free_space(vacrelstats, blkno,
PageGetHeapFreeSpace(page));
}
MarkBufferDirty(buf);
UnlockReleaseBuffer(buf);
MIRROREDLOCK_BUFMGR_UNLOCK;
/* -------- MirroredLock ---------- */
continue;
}
if (PageIsEmpty(page))
{
empty_pages++;
lazy_record_free_space(vacrelstats, blkno,
PageGetHeapFreeSpace(page));
UnlockReleaseBuffer(buf);
MIRROREDLOCK_BUFMGR_UNLOCK;
/* -------- MirroredLock ---------- */
continue;
}
/*
* Prune all HOT-update chains in this page.
*
* We count tuples removed by the pruning step as removed by VACUUM.
*/
tups_vacuumed += heap_page_prune(onerel, buf, OldestXmin,
false, false);
/*
* Now scan the page to collect vacuumable items and check for tuples
* requiring freezing.
*/
nfrozen = 0;
hastup = false;
prev_dead_count = vacrelstats->num_dead_tuples;
maxoff = PageGetMaxOffsetNumber(page);
for (offnum = FirstOffsetNumber;
offnum <= maxoff;
offnum = OffsetNumberNext(offnum))
{
ItemId itemid;
itemid = PageGetItemId(page, offnum);
/* Unused items require no processing, but we count 'em */
if (!ItemIdIsUsed(itemid))
{
nunused += 1;
continue;
}
/* Redirect items mustn't be touched */
if (ItemIdIsRedirected(itemid))
{
hastup = true; /* this page won't be truncatable */
continue;
}
ItemPointerSet(&(tuple.t_self), blkno, offnum);
/*
* DEAD item pointers are to be vacuumed normally; but we don't
* count them in tups_vacuumed, else we'd be double-counting (at
* least in the common case where heap_page_prune() just freed up
* a non-HOT tuple).
*/
if (ItemIdIsDead(itemid))
{
lazy_record_dead_tuple(vacrelstats, &(tuple.t_self));
continue;
}
Assert(ItemIdIsNormal(itemid));
tuple.t_data = (HeapTupleHeader) PageGetItem(page, itemid);
tuple.t_len = ItemIdGetLength(itemid);
tupgone = false;
switch (HeapTupleSatisfiesVacuum(onerel, tuple.t_data, OldestXmin, buf))
{
case HEAPTUPLE_DEAD:
/*
* Ordinarily, DEAD tuples would have been removed by
* heap_page_prune(), but it's possible that the tuple
* state changed since heap_page_prune() looked. In
* particular an INSERT_IN_PROGRESS tuple could have
* changed to DEAD if the inserter aborted. So this
* cannot be considered an error condition.
*
* If the tuple is HOT-updated then it must only be
* removed by a prune operation; so we keep it just as if
* it were RECENTLY_DEAD. Also, if it's a heap-only
* tuple, we choose to keep it, because it'll be a lot
* cheaper to get rid of it in the next pruning pass than
* to treat it like an indexed tuple.
*/
if (HeapTupleIsHotUpdated(&tuple) ||
HeapTupleIsHeapOnly(&tuple))
nkeep += 1;
else
tupgone = true; /* we can delete the tuple */
break;
case HEAPTUPLE_LIVE:
/* Tuple is good --- but let's do some validity checks */
if (onerel->rd_rel->relhasoids &&
!OidIsValid(HeapTupleGetOid(&tuple)))
elog(WARNING, "relation \"%s\" TID %u/%u: OID is invalid",
relname, blkno, offnum);
break;
case HEAPTUPLE_RECENTLY_DEAD:
/*
* If tuple is recently deleted then we must not remove it
* from relation.
*/
nkeep += 1;
break;
case HEAPTUPLE_INSERT_IN_PROGRESS:
/* This is an expected case during concurrent vacuum */
break;
case HEAPTUPLE_DELETE_IN_PROGRESS:
/* This is an expected case during concurrent vacuum */
break;
default:
elog(ERROR, "unexpected HeapTupleSatisfiesVacuum result");
break;
}
if (tupgone)
{
lazy_record_dead_tuple(vacrelstats, &(tuple.t_self));
tups_vacuumed += 1;
}
else
{
num_tuples += 1;
hastup = true;
/*
* Each non-removable tuple must be checked to see if it needs
* freezing. Note we already have exclusive buffer lock.
*/
if (heap_freeze_tuple(tuple.t_data, &FreezeLimit,
InvalidBuffer, false))
frozen[nfrozen++] = offnum;
}
} /* scan along page */
/*
* If we froze any tuples, mark the buffer dirty, and write a WAL
* record recording the changes. We must log the changes to be
* crash-safe against future truncation of CLOG.
*/
if (nfrozen > 0)
{
MarkBufferDirty(buf);
/* no XLOG for temp tables, though */
if (!onerel->rd_istemp)
{
XLogRecPtr recptr;
recptr = log_heap_freeze(onerel, buf, FreezeLimit,
frozen, nfrozen);
PageSetLSN(page, recptr);
}
}
/*
* If there are no indexes then we can vacuum the page right now
* instead of doing a second scan.
*/
if (nindexes == 0 &&
vacrelstats->num_dead_tuples > 0)
{
/* Remove tuples from heap */
lazy_vacuum_page(onerel, blkno, buf, 0, vacrelstats);
/* Forget the now-vacuumed tuples, and press on */
vacrelstats->num_dead_tuples = 0;
vacuumed_pages++;
}
/*
* If we remembered any tuples for deletion, then the page will be
* visited again by lazy_vacuum_heap, which will compute and record
* its post-compaction free space. If not, then we're done with this
* page, so remember its free space as-is. (This path will always be
* taken if there are no indexes.)
*/
if (vacrelstats->num_dead_tuples == prev_dead_count)
{
lazy_record_free_space(vacrelstats, blkno,
PageGetHeapFreeSpace(page));
}
/* Remember the location of the last page with nonremovable tuples */
if (hastup)
vacrelstats->nonempty_pages = blkno + 1;
UnlockReleaseBuffer(buf);
MIRROREDLOCK_BUFMGR_UNLOCK;
/* -------- MirroredLock ---------- */
}
/* save stats for use later */
vacrelstats->rel_tuples = num_tuples;
vacrelstats->tuples_deleted = tups_vacuumed;
/* If any tuples need to be deleted, perform final vacuum cycle */
/* XXX put a threshold on min number of tuples here? */
if (vacrelstats->num_dead_tuples > 0)
{
/* Remove index entries */
for (i = 0; i < nindexes; i++)
lazy_vacuum_index(Irel[i], &indstats[i], vacrelstats);
reindex_count++;
/* Remove tuples from heap */
lazy_vacuum_heap(onerel, vacrelstats);
vacrelstats->num_index_scans++;
}
/* Do post-vacuum cleanup and statistics update for each index */
for (i = 0; i < nindexes; i++)
lazy_cleanup_index(Irel[i], indstats[i], vacrelstats, updated_stats);
/* If no indexes, make log report that lazy_vacuum_heap would've made */
if (vacuumed_pages)
ereport(elevel,
(errmsg("\"%s\": removed %.0f row versions in %u pages",
RelationGetRelationName(onerel),
tups_vacuumed, vacuumed_pages)));
ereport(elevel,
(errmsg("\"%s\": found %.0f removable, %.0f nonremovable row versions in %u pages",
RelationGetRelationName(onerel),
tups_vacuumed, num_tuples, nblocks),
errdetail("%.0f dead row versions cannot be removed yet.\n"
"There were %.0f unused item pointers.\n"
"%u pages contain useful free space.\n"
"%u pages are entirely empty.\n"
"%s.",
nkeep,
nunused,
vacrelstats->tot_free_pages,
empty_pages,
pg_rusage_show(&ru0))));
}
