/*
* Compute elapsed time since ru0 usage snapshot, and format into
* a displayable string. Result is in a static string, which is
* tacky, but no one ever claimed that the Postgres backend is
* threadable...
*/
const char *
pg_rusage_show(const PGRUsage *ru0)
{
static char result[100];
PGRUsage ru1;
pg_rusage_init(&ru1);
if (ru1.tv.tv_usec < ru0->tv.tv_usec)
{
ru1.tv.tv_sec--;
ru1.tv.tv_usec += 1000000;
}
if (ru1.ru.ru_stime.tv_usec < ru0->ru.ru_stime.tv_usec)
{
ru1.ru.ru_stime.tv_sec--;
ru1.ru.ru_stime.tv_usec += 1000000;
}
if (ru1.ru.ru_utime.tv_usec < ru0->ru.ru_utime.tv_usec)
{
ru1.ru.ru_utime.tv_sec--;
ru1.ru.ru_utime.tv_usec += 1000000;
}
snprintf(result, sizeof(result),
"CPU %d.%02ds/%d.%02du sec elapsed %d.%02d sec",
(int) (ru1.ru.ru_stime.tv_sec - ru0->ru.ru_stime.tv_sec),
(int) (ru1.ru.ru_stime.tv_usec - ru0->ru.ru_stime.tv_usec) / 10000,
(int) (ru1.ru.ru_utime.tv_sec - ru0->ru.ru_utime.tv_sec),
(int) (ru1.ru.ru_utime.tv_usec - ru0->ru.ru_utime.tv_usec) / 10000,
(int) (ru1.tv.tv_sec - ru0->tv.tv_sec),
(int) (ru1.tv.tv_usec - ru0->tv.tv_usec) / 10000);
return result;
}
/*
* lazy_vacuum_index() -- vacuum one index relation.
*
* Delete all the index entries pointing to tuples listed in
* vacrelstats->dead_tuples, and update running statistics.
*/
static void
lazy_vacuum_index(Relation indrel,
IndexBulkDeleteResult **stats,
LVRelStats *vacrelstats,
List *extra_oids)
{
IndexVacuumInfo ivinfo;
PGRUsage ru0;
pg_rusage_init(&ru0);
ivinfo.index = indrel;
ivinfo.vacuum_full = false;
ivinfo.message_level = elevel;
/* We don't yet know rel_tuples, so pass -1 */
ivinfo.num_heap_tuples = -1;
ivinfo.extra_oids = extra_oids;
/* Do bulk deletion */
*stats = index_bulk_delete(&ivinfo, *stats,
lazy_tid_reaped, (void *) vacrelstats);
ereport(elevel,
(errmsg("scanned index \"%s\" to remove %d row versions",
RelationGetRelationName(indrel),
vacrelstats->num_dead_tuples),
errdetail("%s.", pg_rusage_show(&ru0))));
}
/*
* lazy_vacuum_index() -- vacuum one index relation.
*
* Delete all the index entries pointing to tuples listed in
* vacrelstats->dead_tuples, and update running statistics.
*/
static void
lazy_vacuum_index(Relation indrel,
IndexBulkDeleteResult **stats,
LVRelStats *vacrelstats)
{
IndexVacuumInfo ivinfo;
PGRUsage ru0;
pg_rusage_init(&ru0);
ivinfo.index = indrel;
ivinfo.analyze_only = false;
ivinfo.estimated_count = true;
ivinfo.message_level = elevel;
ivinfo.num_heap_tuples = vacrelstats->old_rel_tuples;
ivinfo.strategy = vac_strategy;
/* Do bulk deletion */
*stats = index_bulk_delete(&ivinfo, *stats,
lazy_tid_reaped, (void *) vacrelstats);
ereport(elevel,
(errmsg("scanned index \"%s\" to remove %d row versions",
RelationGetRelationName(indrel),
vacrelstats->num_dead_tuples),
errdetail("%s.", pg_rusage_show(&ru0))));
}
/*
* lazy_vacuum_heap() -- second pass over the heap
*
* This routine marks dead tuples as unused and compacts out free
* space on their pages. Pages not having dead tuples recorded from
* lazy_scan_heap are not visited at all.
*
* Note: the reason for doing this as a second pass is we cannot remove
* the tuples until we've removed their index entries, and we want to
* process index entry removal in batches as large as possible.
*/
static void
lazy_vacuum_heap(Relation onerel, LVRelStats *vacrelstats)
{
MIRROREDLOCK_BUFMGR_DECLARE;
int tupindex;
int npages;
PGRUsage ru0;
pg_rusage_init(&ru0);
npages = 0;
tupindex = 0;
/* Fetch gp_persistent_relation_node information that will be added to XLOG record. */
RelationFetchGpRelationNodeForXLog(onerel);
while (tupindex < vacrelstats->num_dead_tuples)
{
BlockNumber tblk;
Buffer buf;
Page page;
vacuum_delay_point();
tblk = ItemPointerGetBlockNumber(&vacrelstats->dead_tuples[tupindex]);
/* -------- MirroredLock ---------- */
MIRROREDLOCK_BUFMGR_LOCK;
buf = ReadBufferWithStrategy(onerel, tblk, vac_strategy);
LockBufferForCleanup(buf);
tupindex = lazy_vacuum_page(onerel, tblk, buf, tupindex, vacrelstats);
/* Now that we've compacted the page, record its available space */
page = BufferGetPage(buf);
lazy_record_free_space(vacrelstats, tblk,
PageGetHeapFreeSpace(page));
UnlockReleaseBuffer(buf);
MIRROREDLOCK_BUFMGR_UNLOCK;
/* -------- MirroredLock ---------- */
npages++;
}
ereport(elevel,
(errmsg("\"%s\": removed %d row versions in %d pages",
RelationGetRelationName(onerel),
tupindex, npages),
errdetail("%s.",
pg_rusage_show(&ru0))));
}
/*
* lazy_cleanup_index() -- do post-vacuum cleanup for one index relation.
*/
static void
lazy_cleanup_index(Relation indrel,
IndexBulkDeleteResult *stats,
LVRelStats *vacrelstats)
{
IndexVacuumInfo ivinfo;
PGRUsage ru0;
pg_rusage_init(&ru0);
ivinfo.index = indrel;
ivinfo.analyze_only = false;
ivinfo.estimated_count = (vacrelstats->scanned_pages < vacrelstats->rel_pages);
ivinfo.message_level = elevel;
ivinfo.num_heap_tuples = vacrelstats->new_rel_tuples;
ivinfo.strategy = vac_strategy;
stats = index_vacuum_cleanup(&ivinfo, stats);
if (!stats)
return;
/*
* Now update statistics in pg_class, but only if the index says the count
* is accurate.
*/
if (!stats->estimated_count)
vac_update_relstats(indrel,
stats->num_pages,
stats->num_index_tuples,
0,
false,
InvalidTransactionId);
ereport(elevel,
(errmsg("index \"%s\" now contains %.0f row versions in %u pages",
RelationGetRelationName(indrel),
stats->num_index_tuples,
stats->num_pages),
errdetail("%.0f index row versions were removed.\n"
"%u index pages have been deleted, %u are currently reusable.\n"
"%s.",
stats->tuples_removed,
stats->pages_deleted, stats->pages_free,
pg_rusage_show(&ru0))));
pfree(stats);
}
/*
* lazy_vacuum_heap() -- second pass over the heap
*
* This routine marks dead tuples as unused and compacts out free
* space on their pages. Pages not having dead tuples recorded from
* lazy_scan_heap are not visited at all.
*
* Note: the reason for doing this as a second pass is we cannot remove
* the tuples until we've removed their index entries, and we want to
* process index entry removal in batches as large as possible.
*/
static void
lazy_vacuum_heap(Relation onerel, LVRelStats *vacrelstats)
{
int tupindex;
int npages;
PGRUsage ru0;
pg_rusage_init(&ru0);
npages = 0;
tupindex = 0;
while (tupindex < vacrelstats->num_dead_tuples)
{
BlockNumber tblk;
Buffer buf;
Page page;
Size freespace;
vacuum_delay_point();
tblk = ItemPointerGetBlockNumber(&vacrelstats->dead_tuples[tupindex]);
buf = ReadBufferExtended(onerel, MAIN_FORKNUM, tblk, RBM_NORMAL,
vac_strategy);
if (!ConditionalLockBufferForCleanup(buf))
{
ReleaseBuffer(buf);
++tupindex;
continue;
}
tupindex = lazy_vacuum_page(onerel, tblk, buf, tupindex, vacrelstats);
/* Now that we've compacted the page, record its available space */
page = BufferGetPage(buf);
freespace = PageGetHeapFreeSpace(page);
UnlockReleaseBuffer(buf);
RecordPageWithFreeSpace(onerel, tblk, freespace);
npages++;
}
ereport(elevel,
(errmsg("\"%s\": removed %d row versions in %d pages",
RelationGetRelationName(onerel),
tupindex, npages),
errdetail("%s.",
pg_rusage_show(&ru0))));
}
/*
* lazy_cleanup_index() -- do post-vacuum cleanup for one index relation.
*/
static void
lazy_cleanup_index(Relation indrel,
IndexBulkDeleteResult *stats,
LVRelStats *vacrelstats,
List *updated_stats)
{
IndexVacuumInfo ivinfo;
PGRUsage ru0;
pg_rusage_init(&ru0);
ivinfo.index = indrel;
ivinfo.vacuum_full = false;
ivinfo.message_level = elevel;
ivinfo.num_heap_tuples = vacrelstats->rel_tuples;
ivinfo.strategy = vac_strategy;
ivinfo.extra_oids = NIL;
stats = index_vacuum_cleanup(&ivinfo, stats);
if (!stats)
return;
/* now update statistics in pg_class */
vac_update_relstats_from_list(indrel,
stats->num_pages,
stats->num_index_tuples,
false,
InvalidTransactionId,
updated_stats);
ereport(elevel,
(errmsg("index \"%s\" now contains %.0f row versions in %u pages",
RelationGetRelationName(indrel),
stats->num_index_tuples,
stats->num_pages),
errdetail("%.0f index row versions were removed.\n"
"%u index pages have been deleted, %u are currently reusable.\n"
"%s.",
stats->tuples_removed,
stats->pages_deleted, stats->pages_free,
pg_rusage_show(&ru0))));
pfree(stats);
}
/*
* 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_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))));
}
}
/*
* 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;
pg_rusage_init(&ru0);
/*
* 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).
*/
if (!ConditionalLockRelation(onerel, AccessExclusiveLock))
return;
/*
* 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->rel_pages = new_rel_pages;
vacrelstats->pages_removed = old_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))));
}
/*
* 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))));
}
}
}
/**
* @brief Entry point of the user-defined function for pg_bulkload.
* @return Returns number of loaded tuples. If the case of errors, -1 will be
* returned.
*/
Datum
pg_bulkload(PG_FUNCTION_ARGS)
{
Reader *rd = NULL;
Writer *wt = NULL;
Datum options;
MemoryContext ctx;
MemoryContext ccxt;
PGRUsage ru0;
PGRUsage ru1;
int64 count;
int64 parse_errors;
int64 skip;
WriterResult ret;
char *start;
char *end;
float8 system;
float8 user;
float8 duration;
TupleDesc tupdesc;
Datum values[PG_BULKLOAD_COLS];
bool nulls[PG_BULKLOAD_COLS];
HeapTuple result;
/* Build a tuple descriptor for our result type */
if (get_call_result_type(fcinfo, NULL, &tupdesc) != TYPEFUNC_COMPOSITE)
elog(ERROR, "return type must be a row type");
BULKLOAD_PROFILE_PUSH();
pg_rusage_init(&ru0);
/* must be the super user */
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to use pg_bulkload")));
options = PG_GETARG_DATUM(0);
ccxt = CurrentMemoryContext;
/*
* STEP 1: Initialization
*/
/* parse options and create reader and writer */
ParseOptions(options, &rd, &wt, ru0.tv.tv_sec);
/* initialize reader */
ReaderInit(rd);
/*
* We need to split PG_TRY block because gcc optimizes if-branches with
* longjmp codes too much. Local variables initialized in either branch
* cannot be handled another branch.
*/
PG_TRY();
{
/* truncate heap */
if (wt->truncate)
TruncateTable(wt->relid);
/* initialize writer */
WriterInit(wt);
/* initialize checker */
CheckerInit(&rd->checker, wt->rel, wt->tchecker);
/* initialize parser */
ParserInit(rd->parser, &rd->checker, rd->infile, wt->desc,
wt->multi_process, PG_GET_COLLATION());
}
PG_CATCH();
{
if (rd)
ReaderClose(rd, true);
if (wt)
WriterClose(wt, true);
PG_RE_THROW();
}
PG_END_TRY();
/* No throwable codes here! */
PG_TRY();
{
/* create logger */
CreateLogger(rd->logfile, wt->verbose, rd->infile[0] == ':');
start = timeval_to_cstring(ru0.tv);
LoggerLog(INFO, "\npg_bulkload %s on %s\n\n",
PG_BULKLOAD_VERSION, start);
ReaderDumpParams(rd);
WriterDumpParams(wt);
LoggerLog(INFO, "\n");
BULKLOAD_PROFILE(&prof_init);
/*
* STEP 2: Build heap
*/
/* Switch into its memory context */
Assert(wt->context);
ctx = MemoryContextSwitchTo(wt->context);
/* Loop for each input file record. */
while (wt->count < rd->limit)
{
HeapTuple tuple;
CHECK_FOR_INTERRUPTS();
/* read tuple */
BULKLOAD_PROFILE_PUSH();
tuple = ReaderNext(rd);
BULKLOAD_PROFILE_POP();
BULKLOAD_PROFILE(&prof_reader);
if (tuple == NULL)
break;
/* write tuple */
BULKLOAD_PROFILE_PUSH();
WriterInsert(wt, tuple);
wt->count += 1;
BULKLOAD_PROFILE_POP();
BULKLOAD_PROFILE(&prof_writer);
MemoryContextReset(wt->context);
BULKLOAD_PROFILE(&prof_reset);
}
MemoryContextSwitchTo(ctx);
/*
* STEP 3: Finalize heap and merge indexes
*/
count = wt->count;
parse_errors = rd->parse_errors;
/*
* close writer first and reader second because shmem_exit callback
* is managed by a simple stack.
*/
ret = WriterClose(wt, false);
wt = NULL;
skip = ReaderClose(rd, false);
rd = NULL;
}
PG_CATCH();
{
ErrorData *errdata;
MemoryContext ecxt;
ecxt = MemoryContextSwitchTo(ccxt);
errdata = CopyErrorData();
LoggerLog(INFO, "%s\n", errdata->message);
FreeErrorData(errdata);
/* close writer first, and reader second */
if (wt)
WriterClose(wt, true);
if (rd)
ReaderClose(rd, true);
MemoryContextSwitchTo(ecxt);
PG_RE_THROW();
}
PG_END_TRY();
count -= ret.num_dup_new;
LoggerLog(INFO, "\n"
" " int64_FMT " Rows skipped.\n"
" " int64_FMT " Rows successfully loaded.\n"
" " int64_FMT " Rows not loaded due to parse errors.\n"
" " int64_FMT " Rows not loaded due to duplicate errors.\n"
" " int64_FMT " Rows replaced with new rows.\n\n",
skip, count, parse_errors, ret.num_dup_new, ret.num_dup_old);
pg_rusage_init(&ru1);
system = diffTime(ru1.ru.ru_stime, ru0.ru.ru_stime);
user = diffTime(ru1.ru.ru_utime, ru0.ru.ru_utime);
duration = diffTime(ru1.tv, ru0.tv);
end = timeval_to_cstring(ru1.tv);
memset(nulls, 0, sizeof(nulls));
values[0] = Int64GetDatum(skip);
values[1] = Int64GetDatum(count);
values[2] = Int64GetDatum(parse_errors);
values[3] = Int64GetDatum(ret.num_dup_new);
values[4] = Int64GetDatum(ret.num_dup_old);
values[5] = Float8GetDatumFast(system);
values[6] = Float8GetDatumFast(user);
values[7] = Float8GetDatumFast(duration);
LoggerLog(INFO,
"Run began on %s\n"
"Run ended on %s\n\n"
"CPU %.2fs/%.2fu sec elapsed %.2f sec\n",
start, end, system, user, duration);
LoggerClose();
result = heap_form_tuple(tupdesc, values, nulls);
BULKLOAD_PROFILE(&prof_fini);
BULKLOAD_PROFILE_POP();
BULKLOAD_PROFILE_PRINT();
PG_RETURN_DATUM(HeapTupleGetDatum(result));
}
/*
* 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;
PageFreeSpaceInfo *pageSpaces;
int n;
int i,
j;
PGRUsage ru0;
pg_rusage_init(&ru0);
/*
* 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).
*/
if (!ConditionalLockRelation(onerel, AccessExclusiveLock))
return;
/*
* 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)
{
/* might as well use the latest news when we update pg_class stats */
vacrelstats->rel_pages = new_rel_pages;
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,
/* markPersistentAsPhysicallyTruncated */ true);
/*
* Drop free-space info for removed blocks; these must not get entered
* into the FSM!
*/
pageSpaces = vacrelstats->free_pages;
n = vacrelstats->num_free_pages;
j = 0;
for (i = 0; i < n; i++)
{
if (pageSpaces[i].blkno < new_rel_pages)
{
pageSpaces[j] = pageSpaces[i];
j++;
}
}
vacrelstats->num_free_pages = j;
/*
* If tot_free_pages was more than num_free_pages, we can't tell for sure
* what its correct value is now, because we don't know which of the
* forgotten pages are getting truncated. Conservatively set it equal to
* num_free_pages.
*/
vacrelstats->tot_free_pages = j;
/* We destroyed the heap ordering, so mark array unordered */
vacrelstats->fs_is_heap = false;
/* update statistics */
vacrelstats->rel_pages = new_rel_pages;
vacrelstats->pages_removed = old_rel_pages - new_rel_pages;
/*
* We can't keep the exclusive lock until commit, since this will cause
* deadlock, see MPP-5733.
*/
UnlockRelation(onerel, AccessExclusiveLock);
ereport(elevel,
(errmsg("\"%s\": truncated %u to %u pages",
RelationGetRelationName(onerel),
old_rel_pages, new_rel_pages),
errdetail("%s.",
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))));
}
/*
* 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;
}
/*
* 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);
}
}
}
}
/*
* 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;
PageFreeSpaceInfo *pageSpaces;
int n;
int i,
j;
PGRUsage ru0;
/*
* Persistent table TIDs are stored in other locations like gp_relation_node
* and changeTracking logs, which continue to have references to CTID even
* if PT tuple is marked deleted. This TID is used to read tuple during
* crash recovery or segment resyncs. Hence need to avoid truncating
* persistent tables to avoid error / crash in heap_fetch using the TID
* on lazy vacuum.
*/
if (GpPersistent_IsPersistentRelation(RelationGetRelid(onerel)))
return;
pg_rusage_init(&ru0);
/*
* 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).
*/
if (!ConditionalLockRelation(onerel, AccessExclusiveLock))
return;
/*
* 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)
{
/* might as well use the latest news when we update pg_class stats */
vacrelstats->rel_pages = new_rel_pages;
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,
/* markPersistentAsPhysicallyTruncated */ true);
/*
* Note: once we have truncated, we *must* keep the exclusive lock until
* commit. The sinval message that will be sent at commit (as a result of
* vac_update_relstats()) must be received by other backends, to cause
* them to reset their rd_targblock values, before they can safely access
* the table again.
*/
/*
* Drop free-space info for removed blocks; these must not get entered
* into the FSM!
*/
pageSpaces = vacrelstats->free_pages;
n = vacrelstats->num_free_pages;
j = 0;
for (i = 0; i < n; i++)
{
if (pageSpaces[i].blkno < new_rel_pages)
{
pageSpaces[j] = pageSpaces[i];
j++;
}
}
vacrelstats->num_free_pages = j;
/*
* If tot_free_pages was more than num_free_pages, we can't tell for sure
* what its correct value is now, because we don't know which of the
* forgotten pages are getting truncated. Conservatively set it equal to
* num_free_pages.
*/
vacrelstats->tot_free_pages = j;
/* We destroyed the heap ordering, so mark array unordered */
vacrelstats->fs_is_heap = false;
/* update statistics */
vacrelstats->rel_pages = new_rel_pages;
vacrelstats->pages_removed = old_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))));
}
/*
* 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);
}
