/*
* Given a BRIN index page, initialize it if necessary, and record it into the
* FSM if necessary. Return value is true if the FSM itself needs "vacuuming".
* The main use for this is when, during vacuuming, an uninitialized page is
* found, which could be the result of relation extension followed by a crash
* before the page can be used.
*/
bool
brin_page_cleanup(Relation idxrel, Buffer buf)
{
Page page = BufferGetPage(buf);
Size freespace;
/*
* If a page was left uninitialized, initialize it now; also record it in
* FSM.
*
* Somebody else might be extending the relation concurrently. To avoid
* re-initializing the page before they can grab the buffer lock, we
* acquire the extension lock momentarily. Since they hold the extension
* lock from before getting the page and after its been initialized, we're
* sure to see their initialization.
*/
if (PageIsNew(page))
{
LockRelationForExtension(idxrel, ShareLock);
UnlockRelationForExtension(idxrel, ShareLock);
LockBuffer(buf, BUFFER_LOCK_EXCLUSIVE);
if (PageIsNew(page))
{
brin_initialize_empty_new_buffer(idxrel, buf);
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
return true;
}
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
}
/* Nothing to be done for non-regular index pages */
if (BRIN_IS_META_PAGE(BufferGetPage(buf)) ||
BRIN_IS_REVMAP_PAGE(BufferGetPage(buf)))
return false;
/* Measure free space and record it */
freespace = br_page_get_freespace(page);
if (freespace > GetRecordedFreeSpace(idxrel, BufferGetBlockNumber(buf)))
{
RecordPageWithFreeSpace(idxrel, BufferGetBlockNumber(buf), freespace);
return true;
}
return false;
}
Datum
pg_freespace(PG_FUNCTION_ARGS)
{
Oid relid = PG_GETARG_OID(0);
int64 blkno = PG_GETARG_INT64(1);
int16 freespace;
Relation rel;
rel = relation_open(relid, AccessShareLock);
if (blkno < 0 || blkno > MaxBlockNumber)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("invalid block number")));
freespace = GetRecordedFreeSpace(rel, blkno);
relation_close(rel, AccessShareLock);
PG_RETURN_INT16(freespace);
}
/*
* This function takes an already open relation and scans its pages,
* skipping those that have the corresponding visibility map bit set.
* For pages we skip, we find the free space from the free space map
* and approximate tuple_len on that basis. For the others, we count
* the exact number of dead tuples etc.
*
* This scan is loosely based on vacuumlazy.c:lazy_scan_heap(), but
* we do not try to avoid skipping single pages.
*/
static void
statapprox_heap(Relation rel, output_type *stat)
{
BlockNumber scanned,
nblocks,
blkno;
Buffer vmbuffer = InvalidBuffer;
BufferAccessStrategy bstrategy;
TransactionId OldestXmin;
uint64 misc_count = 0;
OldestXmin = GetOldestXmin(rel, PROCARRAY_FLAGS_VACUUM);
bstrategy = GetAccessStrategy(BAS_BULKREAD);
nblocks = RelationGetNumberOfBlocks(rel);
scanned = 0;
for (blkno = 0; blkno < nblocks; blkno++)
{
Buffer buf;
Page page;
OffsetNumber offnum,
maxoff;
Size freespace;
CHECK_FOR_INTERRUPTS();
/*
* If the page has only visible tuples, then we can find out the free
* space from the FSM and move on.
*/
if (VM_ALL_VISIBLE(rel, blkno, &vmbuffer))
{
freespace = GetRecordedFreeSpace(rel, blkno);
stat->tuple_len += BLCKSZ - freespace;
stat->free_space += freespace;
continue;
}
buf = ReadBufferExtended(rel, MAIN_FORKNUM, blkno,
RBM_NORMAL, bstrategy);
LockBuffer(buf, BUFFER_LOCK_SHARE);
page = BufferGetPage(buf);
/*
* It's not safe to call PageGetHeapFreeSpace() on new pages, so we
* treat them as being free space for our purposes.
*/
if (!PageIsNew(page))
stat->free_space += PageGetHeapFreeSpace(page);
else
stat->free_space += BLCKSZ - SizeOfPageHeaderData;
if (PageIsNew(page) || PageIsEmpty(page))
{
UnlockReleaseBuffer(buf);
continue;
}
scanned++;
/*
* Look at each tuple on the page and decide whether it's live or
* dead, then count it and its size. Unlike lazy_scan_heap, we can
* afford to ignore problems and special cases.
*/
maxoff = PageGetMaxOffsetNumber(page);
for (offnum = FirstOffsetNumber;
offnum <= maxoff;
offnum = OffsetNumberNext(offnum))
{
ItemId itemid;
HeapTupleData tuple;
itemid = PageGetItemId(page, offnum);
if (!ItemIdIsUsed(itemid) || ItemIdIsRedirected(itemid) ||
ItemIdIsDead(itemid))
{
continue;
}
Assert(ItemIdIsNormal(itemid));
ItemPointerSet(&(tuple.t_self), blkno, offnum);
tuple.t_data = (HeapTupleHeader) PageGetItem(page, itemid);
tuple.t_len = ItemIdGetLength(itemid);
tuple.t_tableOid = RelationGetRelid(rel);
/*
* We count live and dead tuples, but we also need to add up
* others in order to feed vac_estimate_reltuples.
*/
switch (HeapTupleSatisfiesVacuum(&tuple, OldestXmin, buf))
{
case HEAPTUPLE_RECENTLY_DEAD:
misc_count++;
/* Fall through */
case HEAPTUPLE_DEAD:
stat->dead_tuple_len += tuple.t_len;
stat->dead_tuple_count++;
break;
case HEAPTUPLE_LIVE:
stat->tuple_len += tuple.t_len;
stat->tuple_count++;
break;
case HEAPTUPLE_INSERT_IN_PROGRESS:
case HEAPTUPLE_DELETE_IN_PROGRESS:
misc_count++;
break;
default:
elog(ERROR, "unexpected HeapTupleSatisfiesVacuum result");
break;
}
}
UnlockReleaseBuffer(buf);
}
stat->table_len = (uint64) nblocks *BLCKSZ;
stat->tuple_count = vac_estimate_reltuples(rel, false, nblocks, scanned,
stat->tuple_count + misc_count);
/*
* Calculate percentages if the relation has one or more pages.
*/
if (nblocks != 0)
{
stat->scanned_percent = 100 * scanned / nblocks;
stat->tuple_percent = 100.0 * stat->tuple_len / stat->table_len;
stat->dead_tuple_percent = 100.0 * stat->dead_tuple_len / stat->table_len;
stat->free_percent = 100.0 * stat->free_space / stat->table_len;
}
if (BufferIsValid(vmbuffer))
{
ReleaseBuffer(vmbuffer);
vmbuffer = InvalidBuffer;
}
}
