* PostgreSQL logical replay/reorder buffer management
*
*
- * Copyright (c) 2012-2017, PostgreSQL Global Development Group
+ * Copyright (c) 2012-2018, PostgreSQL Global Development Group
*
*
* IDENTIFICATION
* they are written to the WAL and is responsible to reassemble them into
* toplevel transaction sized pieces. When a transaction is completely
* reassembled - signalled by reading the transaction commit record - it
- * will then call the output plugin (c.f. ReorderBufferCommit()) with the
+ * will then call the output plugin (cf. ReorderBufferCommit()) with the
* individual changes. The output plugins rely on snapshots built by
* snapbuild.c which hands them to us.
*
* transaction there will be no other data carrying records between a row's
* toast chunks and the row data itself. See ReorderBufferToast* for
* details.
+ *
+ * ReorderBuffer uses two special memory context types - SlabContext for
+ * allocations of fixed-length structures (changes and transactions), and
+ * GenerationContext for the variable-length transaction data (allocated
+ * and freed in groups with similar lifespan).
+ *
* -------------------------------------------------------------------------
*/
#include "postgres.h"
*/
static const Size max_changes_in_memory = 4096;
-/*
- * We use a very simple form of a slab allocator for frequently allocated
- * objects, simply keeping a fixed number in a linked list when unused,
- * instead pfree()ing them. Without that in many workloads aset.c becomes a
- * major bottleneck, especially when spilling to disk while decoding batch
- * workloads.
- */
-static const Size max_cached_tuplebufs = 4096 * 2; /* ~8MB */
-
/* ---------------------------------------
* primary reorderbuffer support routines
* ---------------------------------------
static ReorderBufferTXN *ReorderBufferTXNByXid(ReorderBuffer *rb,
TransactionId xid, bool create, bool *is_new,
XLogRecPtr lsn, bool create_as_top);
+static void ReorderBufferTransferSnapToParent(ReorderBufferTXN *txn,
+ ReorderBufferTXN *subtxn);
static void AssertTXNLsnOrder(ReorderBuffer *rb);
static void ReorderBufferRestoreChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
char *change);
static void ReorderBufferRestoreCleanup(ReorderBuffer *rb, ReorderBufferTXN *txn);
+static void ReorderBufferCleanupSerializedTXNs(const char *slotname);
+static void ReorderBufferSerializedPath(char *path, ReplicationSlot *slot,
+ TransactionId xid, XLogSegNo segno);
static void ReorderBufferFreeSnap(ReorderBuffer *rb, Snapshot snap);
static Snapshot ReorderBufferCopySnap(ReorderBuffer *rb, Snapshot orig_snap,
/*
- * Allocate a new ReorderBuffer
+ * Allocate a new ReorderBuffer and clean out any old serialized state from
+ * prior ReorderBuffer instances for the same slot.
*/
ReorderBuffer *
ReorderBufferAllocate(void)
HASHCTL hash_ctl;
MemoryContext new_ctx;
+ Assert(MyReplicationSlot != NULL);
+
/* allocate memory in own context, to have better accountability */
new_ctx = AllocSetContextCreate(CurrentMemoryContext,
"ReorderBuffer",
SLAB_DEFAULT_BLOCK_SIZE,
sizeof(ReorderBufferTXN));
+ buffer->tup_context = GenerationContextCreate(new_ctx,
+ "Tuples",
+ SLAB_LARGE_BLOCK_SIZE);
+
hash_ctl.keysize = sizeof(TransactionId);
hash_ctl.entrysize = sizeof(ReorderBufferTXNByIdEnt);
hash_ctl.hcxt = buffer->context;
buffer->by_txn_last_xid = InvalidTransactionId;
buffer->by_txn_last_txn = NULL;
- buffer->nr_cached_tuplebufs = 0;
-
buffer->outbuf = NULL;
buffer->outbufsize = 0;
buffer->current_restart_decoding_lsn = InvalidXLogRecPtr;
dlist_init(&buffer->toplevel_by_lsn);
- slist_init(&buffer->cached_tuplebufs);
+ dlist_init(&buffer->txns_by_base_snapshot_lsn);
+
+ /*
+ * Ensure there's no stale data from prior uses of this slot, in case some
+ * prior exit avoided calling ReorderBufferFree. Failure to do this can
+ * produce duplicated txns, and it's very cheap if there's nothing there.
+ */
+ ReorderBufferCleanupSerializedTXNs(NameStr(MyReplicationSlot->data.name));
return buffer;
}
* memory context.
*/
MemoryContextDelete(context);
+
+ /* Free disk space used by unconsumed reorder buffers */
+ ReorderBufferCleanupSerializedTXNs(NameStr(MyReplicationSlot->data.name));
}
/*
/*
* Free a ReorderBufferTXN.
- *
- * Deallocation might be delayed for efficiency purposes, for details check
- * the comments above max_cached_changes's definition.
*/
static void
ReorderBufferReturnTXN(ReorderBuffer *rb, ReorderBufferTXN *txn)
}
/*
- * Get an unused, possibly preallocated, ReorderBufferChange.
+ * Get an fresh ReorderBufferChange.
*/
ReorderBufferChange *
ReorderBufferGetChange(ReorderBuffer *rb)
/*
* Free an ReorderBufferChange.
- *
- * Deallocation might be delayed for efficiency purposes, for details check
- * the comments above max_cached_changes's definition.
*/
void
ReorderBufferReturnChange(ReorderBuffer *rb, ReorderBufferChange *change)
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_TRUNCATE:
break;
}
}
/*
- * Get an unused, possibly preallocated, ReorderBufferTupleBuf fitting at
- * least a tuple of size tuple_len (excluding header overhead).
+ * Get a fresh ReorderBufferTupleBuf fitting at least a tuple of size
+ * tuple_len (excluding header overhead).
*/
ReorderBufferTupleBuf *
ReorderBufferGetTupleBuf(ReorderBuffer *rb, Size tuple_len)
alloc_len = tuple_len + SizeofHeapTupleHeader;
- /*
- * Most tuples are below MaxHeapTupleSize, so we use a slab allocator for
- * those. Thus always allocate at least MaxHeapTupleSize. Note that tuples
- * generated for oldtuples can be bigger, as they don't have out-of-line
- * toast columns.
- */
- if (alloc_len < MaxHeapTupleSize)
- alloc_len = MaxHeapTupleSize;
-
-
- /* if small enough, check the slab cache */
- if (alloc_len <= MaxHeapTupleSize && rb->nr_cached_tuplebufs)
- {
- rb->nr_cached_tuplebufs--;
- tuple = slist_container(ReorderBufferTupleBuf, node,
- slist_pop_head_node(&rb->cached_tuplebufs));
- Assert(tuple->alloc_tuple_size == MaxHeapTupleSize);
-#ifdef USE_ASSERT_CHECKING
- memset(&tuple->tuple, 0xa9, sizeof(HeapTupleData));
- VALGRIND_MAKE_MEM_UNDEFINED(&tuple->tuple, sizeof(HeapTupleData));
-#endif
- tuple->tuple.t_data = ReorderBufferTupleBufData(tuple);
-#ifdef USE_ASSERT_CHECKING
- memset(tuple->tuple.t_data, 0xa8, tuple->alloc_tuple_size);
- VALGRIND_MAKE_MEM_UNDEFINED(tuple->tuple.t_data, tuple->alloc_tuple_size);
-#endif
- }
- else
- {
- tuple = (ReorderBufferTupleBuf *)
- MemoryContextAlloc(rb->context,
- sizeof(ReorderBufferTupleBuf) +
- MAXIMUM_ALIGNOF + alloc_len);
- tuple->alloc_tuple_size = alloc_len;
- tuple->tuple.t_data = ReorderBufferTupleBufData(tuple);
- }
+ tuple = (ReorderBufferTupleBuf *)
+ MemoryContextAlloc(rb->tup_context,
+ sizeof(ReorderBufferTupleBuf) +
+ MAXIMUM_ALIGNOF + alloc_len);
+ tuple->alloc_tuple_size = alloc_len;
+ tuple->tuple.t_data = ReorderBufferTupleBufData(tuple);
return tuple;
}
/*
* Free an ReorderBufferTupleBuf.
- *
- * Deallocation might be delayed for efficiency purposes, for details check
- * the comments above max_cached_changes's definition.
*/
void
ReorderBufferReturnTupleBuf(ReorderBuffer *rb, ReorderBufferTupleBuf *tuple)
{
- /* check whether to put into the slab cache, oversized tuples never are */
- if (tuple->alloc_tuple_size == MaxHeapTupleSize &&
- rb->nr_cached_tuplebufs < max_cached_tuplebufs)
- {
- rb->nr_cached_tuplebufs++;
- slist_push_head(&rb->cached_tuplebufs, &tuple->node);
- VALGRIND_MAKE_MEM_UNDEFINED(tuple->tuple.t_data, tuple->alloc_tuple_size);
- VALGRIND_MAKE_MEM_UNDEFINED(tuple, sizeof(ReorderBufferTupleBuf));
- VALGRIND_MAKE_MEM_DEFINED(&tuple->node, sizeof(tuple->node));
- VALGRIND_MAKE_MEM_DEFINED(&tuple->alloc_tuple_size, sizeof(tuple->alloc_tuple_size));
- }
- else
- {
- pfree(tuple);
- }
+ pfree(tuple);
}
/*
bool found;
Assert(TransactionIdIsValid(xid));
- Assert(!create || lsn != InvalidXLogRecPtr);
/*
* Check the one-entry lookup cache first
{
/* initialize the new entry, if creation was requested */
Assert(ent != NULL);
+ Assert(lsn != InvalidXLogRecPtr);
ent->txn = ReorderBufferGetTXN(rb);
ent->txn->xid = xid;
}
}
-
+/*
+ * AssertTXNLsnOrder
+ * Verify LSN ordering of transaction lists in the reorderbuffer
+ *
+ * Other LSN-related invariants are checked too.
+ *
+ * No-op if assertions are not in use.
+ */
static void
AssertTXNLsnOrder(ReorderBuffer *rb)
{
#ifdef USE_ASSERT_CHECKING
dlist_iter iter;
XLogRecPtr prev_first_lsn = InvalidXLogRecPtr;
+ XLogRecPtr prev_base_snap_lsn = InvalidXLogRecPtr;
dlist_foreach(iter, &rb->toplevel_by_lsn)
{
- ReorderBufferTXN *cur_txn;
+ ReorderBufferTXN *cur_txn = dlist_container(ReorderBufferTXN, node,
+ iter.cur);
- cur_txn = dlist_container(ReorderBufferTXN, node, iter.cur);
+ /* start LSN must be set */
Assert(cur_txn->first_lsn != InvalidXLogRecPtr);
+ /* If there is an end LSN, it must be higher than start LSN */
if (cur_txn->end_lsn != InvalidXLogRecPtr)
Assert(cur_txn->first_lsn <= cur_txn->end_lsn);
+ /* Current initial LSN must be strictly higher than previous */
if (prev_first_lsn != InvalidXLogRecPtr)
Assert(prev_first_lsn < cur_txn->first_lsn);
+ /* known-as-subtxn txns must not be listed */
Assert(!cur_txn->is_known_as_subxact);
+
prev_first_lsn = cur_txn->first_lsn;
}
+
+ dlist_foreach(iter, &rb->txns_by_base_snapshot_lsn)
+ {
+ ReorderBufferTXN *cur_txn = dlist_container(ReorderBufferTXN,
+ base_snapshot_node,
+ iter.cur);
+
+ /* base snapshot (and its LSN) must be set */
+ Assert(cur_txn->base_snapshot != NULL);
+ Assert(cur_txn->base_snapshot_lsn != InvalidXLogRecPtr);
+
+ /* current LSN must be strictly higher than previous */
+ if (prev_base_snap_lsn != InvalidXLogRecPtr)
+ Assert(prev_base_snap_lsn < cur_txn->base_snapshot_lsn);
+
+ /* known-as-subtxn txns must not be listed */
+ Assert(!cur_txn->is_known_as_subxact);
+
+ prev_base_snap_lsn = cur_txn->base_snapshot_lsn;
+ }
#endif
}
+/*
+ * ReorderBufferGetOldestTXN
+ * Return oldest transaction in reorderbuffer
+ */
ReorderBufferTXN *
ReorderBufferGetOldestTXN(ReorderBuffer *rb)
{
ReorderBufferTXN *txn;
+ AssertTXNLsnOrder(rb);
+
if (dlist_is_empty(&rb->toplevel_by_lsn))
return NULL;
- AssertTXNLsnOrder(rb);
-
txn = dlist_head_element(ReorderBufferTXN, node, &rb->toplevel_by_lsn);
Assert(!txn->is_known_as_subxact);
return txn;
}
+/*
+ * ReorderBufferGetOldestXmin
+ * Return oldest Xmin in reorderbuffer
+ *
+ * Returns oldest possibly running Xid from the point of view of snapshots
+ * used in the transactions kept by reorderbuffer, or InvalidTransactionId if
+ * there are none.
+ *
+ * Since snapshots are assigned monotonically, this equals the Xmin of the
+ * base snapshot with minimal base_snapshot_lsn.
+ */
+TransactionId
+ReorderBufferGetOldestXmin(ReorderBuffer *rb)
+{
+ ReorderBufferTXN *txn;
+
+ AssertTXNLsnOrder(rb);
+
+ if (dlist_is_empty(&rb->txns_by_base_snapshot_lsn))
+ return InvalidTransactionId;
+
+ txn = dlist_head_element(ReorderBufferTXN, base_snapshot_node,
+ &rb->txns_by_base_snapshot_lsn);
+ return txn->base_snapshot->xmin;
+}
+
void
ReorderBufferSetRestartPoint(ReorderBuffer *rb, XLogRecPtr ptr)
{
rb->current_restart_decoding_lsn = ptr;
}
+/*
+ * ReorderBufferAssignChild
+ *
+ * Make note that we know that subxid is a subtransaction of xid, seen as of
+ * the given lsn.
+ */
void
ReorderBufferAssignChild(ReorderBuffer *rb, TransactionId xid,
TransactionId subxid, XLogRecPtr lsn)
txn = ReorderBufferTXNByXid(rb, xid, true, &new_top, lsn, true);
subtxn = ReorderBufferTXNByXid(rb, subxid, true, &new_sub, lsn, false);
- if (new_sub)
+ if (new_top && !new_sub)
+ elog(ERROR, "subtransaction logged without previous top-level txn record");
+
+ if (!new_sub)
{
- /*
- * we assign subtransactions to top level transaction even if we don't
- * have data for it yet, assignment records frequently reference xids
- * that have not yet produced any records. Knowing those aren't top
- * level xids allows us to make processing cheaper in some places.
- */
- dlist_push_tail(&txn->subtxns, &subtxn->node);
- txn->nsubtxns++;
+ if (subtxn->is_known_as_subxact)
+ {
+ /* already associated, nothing to do */
+ return;
+ }
+ else
+ {
+ /*
+ * We already saw this transaction, but initially added it to the
+ * list of top-level txns. Now that we know it's not top-level,
+ * remove it from there.
+ */
+ dlist_delete(&subtxn->node);
+ }
}
- else if (!subtxn->is_known_as_subxact)
- {
- subtxn->is_known_as_subxact = true;
- Assert(subtxn->nsubtxns == 0);
- /* remove from lsn order list of top-level transactions */
- dlist_delete(&subtxn->node);
+ subtxn->is_known_as_subxact = true;
+ subtxn->toplevel_xid = xid;
+ Assert(subtxn->nsubtxns == 0);
- /* add to toplevel transaction */
- dlist_push_tail(&txn->subtxns, &subtxn->node);
- txn->nsubtxns++;
- }
- else if (new_top)
+ /* add to subtransaction list */
+ dlist_push_tail(&txn->subtxns, &subtxn->node);
+ txn->nsubtxns++;
+
+ /* Possibly transfer the subtxn's snapshot to its top-level txn. */
+ ReorderBufferTransferSnapToParent(txn, subtxn);
+
+ /* Verify LSN-ordering invariant */
+ AssertTXNLsnOrder(rb);
+}
+
+/*
+ * ReorderBufferTransferSnapToParent
+ * Transfer base snapshot from subtxn to top-level txn, if needed
+ *
+ * This is done if the top-level txn doesn't have a base snapshot, or if the
+ * subtxn's base snapshot has an earlier LSN than the top-level txn's base
+ * snapshot's LSN. This can happen if there are no changes in the toplevel
+ * txn but there are some in the subtxn, or the first change in subtxn has
+ * earlier LSN than first change in the top-level txn and we learned about
+ * their kinship only now.
+ *
+ * The subtransaction's snapshot is cleared regardless of the transfer
+ * happening, since it's not needed anymore in either case.
+ *
+ * We do this as soon as we become aware of their kinship, to avoid queueing
+ * extra snapshots to txns known-as-subtxns -- only top-level txns will
+ * receive further snapshots.
+ */
+static void
+ReorderBufferTransferSnapToParent(ReorderBufferTXN *txn,
+ ReorderBufferTXN *subtxn)
+{
+ Assert(subtxn->toplevel_xid == txn->xid);
+
+ if (subtxn->base_snapshot != NULL)
{
- elog(ERROR, "existing subxact assigned to unknown toplevel xact");
+ if (txn->base_snapshot == NULL ||
+ subtxn->base_snapshot_lsn < txn->base_snapshot_lsn)
+ {
+ /*
+ * If the toplevel transaction already has a base snapshot but
+ * it's newer than the subxact's, purge it.
+ */
+ if (txn->base_snapshot != NULL)
+ {
+ SnapBuildSnapDecRefcount(txn->base_snapshot);
+ dlist_delete(&txn->base_snapshot_node);
+ }
+
+ /*
+ * The snapshot is now the top transaction's; transfer it, and
+ * adjust the list position of the top transaction in the list by
+ * moving it to where the subtransaction is.
+ */
+ txn->base_snapshot = subtxn->base_snapshot;
+ txn->base_snapshot_lsn = subtxn->base_snapshot_lsn;
+ dlist_insert_before(&subtxn->base_snapshot_node,
+ &txn->base_snapshot_node);
+
+ /*
+ * The subtransaction doesn't have a snapshot anymore (so it
+ * mustn't be in the list.)
+ */
+ subtxn->base_snapshot = NULL;
+ subtxn->base_snapshot_lsn = InvalidXLogRecPtr;
+ dlist_delete(&subtxn->base_snapshot_node);
+ }
+ else
+ {
+ /* Base snap of toplevel is fine, so subxact's is not needed */
+ SnapBuildSnapDecRefcount(subtxn->base_snapshot);
+ dlist_delete(&subtxn->base_snapshot_node);
+ subtxn->base_snapshot = NULL;
+ subtxn->base_snapshot_lsn = InvalidXLogRecPtr;
+ }
}
}
TransactionId subxid, XLogRecPtr commit_lsn,
XLogRecPtr end_lsn)
{
- ReorderBufferTXN *txn;
ReorderBufferTXN *subtxn;
subtxn = ReorderBufferTXNByXid(rb, subxid, false, NULL,
if (!subtxn)
return;
- txn = ReorderBufferTXNByXid(rb, xid, false, NULL, commit_lsn, true);
-
- if (txn == NULL)
- elog(ERROR, "subxact logged without previous toplevel record");
-
- /*
- * Pass our base snapshot to the parent transaction if it doesn't have
- * one, or ours is older. That can happen if there are no changes in the
- * toplevel transaction but in one of the child transactions. This allows
- * the parent to simply use its base snapshot initially.
- */
- if (subtxn->base_snapshot != NULL &&
- (txn->base_snapshot == NULL ||
- txn->base_snapshot_lsn > subtxn->base_snapshot_lsn))
- {
- txn->base_snapshot = subtxn->base_snapshot;
- txn->base_snapshot_lsn = subtxn->base_snapshot_lsn;
- subtxn->base_snapshot = NULL;
- subtxn->base_snapshot_lsn = InvalidXLogRecPtr;
- }
-
subtxn->final_lsn = commit_lsn;
subtxn->end_lsn = end_lsn;
- if (!subtxn->is_known_as_subxact)
- {
- subtxn->is_known_as_subxact = true;
- Assert(subtxn->nsubtxns == 0);
-
- /* remove from lsn order list of top-level transactions */
- dlist_delete(&subtxn->node);
-
- /* add to subtransaction list */
- dlist_push_tail(&txn->subtxns, &subtxn->node);
- txn->nsubtxns++;
- }
+ /*
+ * Assign this subxact as a child of the toplevel xact (no-op if already
+ * done.)
+ */
+ ReorderBufferAssignChild(rb, xid, subxid, InvalidXLogRecPtr);
}
ReorderBufferReturnChange(rb, change);
}
+ /*
+ * Cleanup the base snapshot, if set.
+ */
if (txn->base_snapshot != NULL)
{
SnapBuildSnapDecRefcount(txn->base_snapshot);
- txn->base_snapshot = NULL;
- txn->base_snapshot_lsn = InvalidXLogRecPtr;
+ dlist_delete(&txn->base_snapshot_node);
}
/*
}
/*
- * Perform the replay of a transaction and it's non-aborted subtransactions.
+ * Perform the replay of a transaction and its non-aborted subtransactions.
*
* Subtransactions previously have to be processed by
* ReorderBufferCommitChild(), even if previously assigned to the toplevel
* transaction with ReorderBufferAssignChild.
*
- * We currently can only decode a transaction's contents in when their commit
- * record is read because that's currently the only place where we know about
- * cache invalidations. Thus, once a toplevel commit is read, we iterate over
- * the top and subtransactions (using a k-way merge) and replay the changes in
- * lsn order.
+ * We currently can only decode a transaction's contents when its commit
+ * record is read because that's the only place where we know about cache
+ * invalidations. Thus, once a toplevel commit is read, we iterate over the top
+ * and subtransactions (using a k-way merge) and replay the changes in lsn
+ * order.
*/
void
ReorderBufferCommit(ReorderBuffer *rb, TransactionId xid,
txn->origin_lsn = origin_lsn;
/*
- * If this transaction didn't have any real changes in our database, it's
- * OK not to have a snapshot. Note that ReorderBufferCommitChild will have
- * transferred its snapshot to this transaction if it had one and the
- * toplevel tx didn't.
+ * If this transaction has no snapshot, it didn't make any changes to the
+ * database, so there's nothing to decode. Note that
+ * ReorderBufferCommitChild will have transferred any snapshots from
+ * subtransactions if there were any.
*/
if (txn->base_snapshot == NULL)
{
* use as a normal record. It'll be cleaned up at the end
* of INSERT processing.
*/
+ if (specinsert == NULL)
+ elog(ERROR, "invalid ordering of speculative insertion changes");
Assert(specinsert->data.tp.oldtuple == NULL);
change = specinsert;
change->action = REORDER_BUFFER_CHANGE_INSERT;
if (!RelationIsLogicallyLogged(relation))
goto change_done;
+ /*
+ * Ignore temporary heaps created during DDL unless the
+ * plugin has asked for them.
+ */
+ if (relation->rd_rel->relrewrite && !rb->output_rewrites)
+ goto change_done;
+
/*
* For now ignore sequence changes entirely. Most of the
* time they don't log changes using records we
specinsert = change;
break;
+ case REORDER_BUFFER_CHANGE_TRUNCATE:
+ {
+ int i;
+ int nrelids = change->data.truncate.nrelids;
+ int nrelations = 0;
+ Relation *relations;
+
+ relations = palloc0(nrelids * sizeof(Relation));
+ for (i = 0; i < nrelids; i++)
+ {
+ Oid relid = change->data.truncate.relids[i];
+ Relation relation;
+
+ relation = RelationIdGetRelation(relid);
+
+ if (relation == NULL)
+ elog(ERROR, "could not open relation with OID %u", relid);
+
+ if (!RelationIsLogicallyLogged(relation))
+ continue;
+
+ relations[nrelations++] = relation;
+ }
+
+ rb->apply_truncate(rb, txn, nrelations, relations, change);
+
+ for (i = 0; i < nrelations; i++)
+ RelationClose(relations[i]);
+
+ break;
+ }
+
case REORDER_BUFFER_CHANGE_MESSAGE:
rb->message(rb, txn, change->lsn, true,
change->data.msg.prefix,
* Iterate through all (potential) toplevel TXNs and abort all that are
* older than what possibly can be running. Once we've found the first
* that is alive we stop, there might be some that acquired an xid earlier
- * but started writing later, but it's unlikely and they will cleaned up
- * in a later call to ReorderBufferAbortOld().
+ * but started writing later, but it's unlikely and they will be cleaned
+ * up in a later call to this function.
*/
dlist_foreach_modify(it, &rb->toplevel_by_lsn)
{
if (TransactionIdPrecedes(txn->xid, oldestRunningXid))
{
+ /*
+ * We set final_lsn on a transaction when we decode its commit or
+ * abort record, but we never see those records for crashed
+ * transactions. To ensure cleanup of these transactions, set
+ * final_lsn to that of their last change; this causes
+ * ReorderBufferRestoreCleanup to do the right thing.
+ */
+ if (txn->serialized && txn->final_lsn == 0)
+ {
+ ReorderBufferChange *last =
+ dlist_tail_element(ReorderBufferChange, node, &txn->changes);
+
+ txn->final_lsn = last->lsn;
+ }
+
elog(DEBUG2, "aborting old transaction %u", txn->xid);
/* remove potential on-disk data, and deallocate this tx */
/*
* Execute invalidations happening outside the context of a decoded
* transaction. That currently happens either for xid-less commits
- * (c.f. RecordTransactionCommit()) or for invalidations in uninteresting
+ * (cf. RecordTransactionCommit()) or for invalidations in uninteresting
* transactions (via ReorderBufferForget()).
*/
void
}
/*
- * Setup the base snapshot of a transaction. The base snapshot is the snapshot
- * that is used to decode all changes until either this transaction modifies
- * the catalog or another catalog modifying transaction commits.
+ * Set up the transaction's base snapshot.
*
- * Needs to be called before any changes are added with
- * ReorderBufferQueueChange().
+ * If we know that xid is a subtransaction, set the base snapshot on the
+ * top-level transaction instead.
*/
void
ReorderBufferSetBaseSnapshot(ReorderBuffer *rb, TransactionId xid,
ReorderBufferTXN *txn;
bool is_new;
+ AssertArg(snap != NULL);
+
+ /*
+ * Fetch the transaction to operate on. If we know it's a subtransaction,
+ * operate on its top-level transaction instead.
+ */
txn = ReorderBufferTXNByXid(rb, xid, true, &is_new, lsn, true);
+ if (txn->is_known_as_subxact)
+ txn = ReorderBufferTXNByXid(rb, txn->toplevel_xid, false,
+ NULL, InvalidXLogRecPtr, false);
Assert(txn->base_snapshot == NULL);
- Assert(snap != NULL);
txn->base_snapshot = snap;
txn->base_snapshot_lsn = lsn;
+ dlist_push_tail(&rb->txns_by_base_snapshot_lsn, &txn->base_snapshot_node);
+
+ AssertTXNLsnOrder(rb);
}
/*
}
/*
- * Have we already added the first snapshot?
+ * ReorderBufferXidHasBaseSnapshot
+ * Have we already set the base snapshot for the given txn/subtxn?
*/
bool
ReorderBufferXidHasBaseSnapshot(ReorderBuffer *rb, TransactionId xid)
{
ReorderBufferTXN *txn;
- txn = ReorderBufferTXNByXid(rb, xid, false, NULL, InvalidXLogRecPtr,
- false);
+ txn = ReorderBufferTXNByXid(rb, xid, false,
+ NULL, InvalidXLogRecPtr, false);
/* transaction isn't known yet, ergo no snapshot */
if (txn == NULL)
return false;
- /*
- * TODO: It would be a nice improvement if we would check the toplevel
- * transaction in subtransactions, but we'd need to keep track of a bit
- * more state.
- */
+ /* a known subtxn? operate on top-level txn instead */
+ if (txn->is_known_as_subxact)
+ txn = ReorderBufferTXNByXid(rb, txn->toplevel_xid, false,
+ NULL, InvalidXLogRecPtr, false);
+
return txn->base_snapshot != NULL;
}
int fd = -1;
XLogSegNo curOpenSegNo = 0;
Size spilled = 0;
- char path[MAXPGPATH];
elog(DEBUG2, "spill %u changes in XID %u to disk",
(uint32) txn->nentries_mem, txn->xid);
if (fd == -1 ||
!XLByteInSeg(change->lsn, curOpenSegNo, wal_segment_size))
{
- XLogRecPtr recptr;
+ char path[MAXPGPATH];
if (fd != -1)
CloseTransientFile(fd);
XLByteToSeg(change->lsn, curOpenSegNo, wal_segment_size);
- XLogSegNoOffsetToRecPtr(curOpenSegNo, 0, recptr, wal_segment_size);
/*
* No need to care about TLIs here, only used during a single run,
* so each LSN only maps to a specific WAL record.
*/
- sprintf(path, "pg_replslot/%s/xid-%u-lsn-%X-%X.snap",
- NameStr(MyReplicationSlot->data.name), txn->xid,
- (uint32) (recptr >> 32), (uint32) recptr);
+ ReorderBufferSerializedPath(path, MyReplicationSlot, txn->xid,
+ curOpenSegNo);
/* open segment, create it if necessary */
fd = OpenTransientFile(path,
if (fd < 0)
ereport(ERROR,
(errcode_for_file_access(),
- errmsg("could not open file \"%s\": %m",
- path)));
+ errmsg("could not open file \"%s\": %m", path)));
}
ReorderBufferSerializeChange(rb, txn, fd, change);
}
break;
}
+ case REORDER_BUFFER_CHANGE_TRUNCATE:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
ondisk->size = sz;
+ errno = 0;
pgstat_report_wait_start(WAIT_EVENT_REORDER_BUFFER_WRITE);
if (write(fd, rb->outbuf, ondisk->size) != ondisk->size)
{
int save_errno = errno;
CloseTransientFile(fd);
- errno = save_errno;
+
+ /* if write didn't set errno, assume problem is no disk space */
+ errno = save_errno ? save_errno : ENOSPC;
ereport(ERROR,
(errcode_for_file_access(),
errmsg("could not write to data file for XID %u: %m",
if (*fd == -1)
{
- XLogRecPtr recptr;
char path[MAXPGPATH];
/* first time in */
if (*segno == 0)
- {
XLByteToSeg(txn->first_lsn, *segno, wal_segment_size);
- }
Assert(*segno != 0 || dlist_is_empty(&txn->changes));
- XLogSegNoOffsetToRecPtr(*segno, 0, recptr, wal_segment_size);
/*
* No need to care about TLIs here, only used during a single run,
* so each LSN only maps to a specific WAL record.
*/
- sprintf(path, "pg_replslot/%s/xid-%u-lsn-%X-%X.snap",
- NameStr(MyReplicationSlot->data.name), txn->xid,
- (uint32) (recptr >> 32), (uint32) recptr);
+ ReorderBufferSerializedPath(path, MyReplicationSlot, txn->xid,
+ *segno);
*fd = OpenTransientFile(path, O_RDONLY | PG_BINARY);
if (*fd < 0 && errno == ENOENT)
(errcode_for_file_access(),
errmsg("could not open file \"%s\": %m",
path)));
-
}
/*
break;
}
/* the base struct contains all the data, easy peasy */
+ case REORDER_BUFFER_CHANGE_TRUNCATE:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
for (cur = first; cur <= last; cur++)
{
char path[MAXPGPATH];
- XLogRecPtr recptr;
- XLogSegNoOffsetToRecPtr(cur, 0, recptr, wal_segment_size);
-
- sprintf(path, "pg_replslot/%s/xid-%u-lsn-%X-%X.snap",
- NameStr(MyReplicationSlot->data.name), txn->xid,
- (uint32) (recptr >> 32), (uint32) recptr);
+ ReorderBufferSerializedPath(path, MyReplicationSlot, txn->xid, cur);
if (unlink(path) != 0 && errno != ENOENT)
ereport(ERROR,
(errcode_for_file_access(),
}
}
+/*
+ * Remove any leftover serialized reorder buffers from a slot directory after a
+ * prior crash or decoding session exit.
+ */
+static void
+ReorderBufferCleanupSerializedTXNs(const char *slotname)
+{
+ DIR *spill_dir;
+ struct dirent *spill_de;
+ struct stat statbuf;
+ char path[MAXPGPATH * 2 + 12];
+
+ sprintf(path, "pg_replslot/%s", slotname);
+
+ /* we're only handling directories here, skip if it's not ours */
+ if (lstat(path, &statbuf) == 0 && !S_ISDIR(statbuf.st_mode))
+ return;
+
+ spill_dir = AllocateDir(path);
+ while ((spill_de = ReadDirExtended(spill_dir, path, INFO)) != NULL)
+ {
+ /* only look at names that can be ours */
+ if (strncmp(spill_de->d_name, "xid", 3) == 0)
+ {
+ snprintf(path, sizeof(path),
+ "pg_replslot/%s/%s", slotname,
+ spill_de->d_name);
+
+ if (unlink(path) != 0)
+ ereport(ERROR,
+ (errcode_for_file_access(),
+ errmsg("could not remove file \"%s\" during removal of pg_replslot/%s/xid*: %m",
+ path, slotname)));
+ }
+ }
+ FreeDir(spill_dir);
+}
+
+/*
+ * Given a replication slot, transaction ID and segment number, fill in the
+ * corresponding spill file into 'path', which is a caller-owned buffer of size
+ * at least MAXPGPATH.
+ */
+static void
+ReorderBufferSerializedPath(char *path, ReplicationSlot *slot, TransactionId xid,
+ XLogSegNo segno)
+{
+ XLogRecPtr recptr;
+
+ XLogSegNoOffsetToRecPtr(segno, 0, wal_segment_size, recptr);
+
+ snprintf(path, MAXPGPATH, "pg_replslot/%s/xid-%u-lsn-%X-%X.tmp",
+ NameStr(MyReplicationSlot->data.name),
+ xid,
+ (uint32) (recptr >> 32), (uint32) recptr);
+}
+
/*
* Delete all data spilled to disk after we've restarted/crashed. It will be
* recreated when the respective slots are reused.
DIR *logical_dir;
struct dirent *logical_de;
- DIR *spill_dir;
- struct dirent *spill_de;
-
logical_dir = AllocateDir("pg_replslot");
while ((logical_de = ReadDir(logical_dir, "pg_replslot")) != NULL)
{
- struct stat statbuf;
- char path[MAXPGPATH * 2 + 12];
-
if (strcmp(logical_de->d_name, ".") == 0 ||
strcmp(logical_de->d_name, "..") == 0)
continue;
* ok, has to be a surviving logical slot, iterate and delete
* everything starting with xid-*
*/
- sprintf(path, "pg_replslot/%s", logical_de->d_name);
-
- /* we're only creating directories here, skip if it's not our's */
- if (lstat(path, &statbuf) == 0 && !S_ISDIR(statbuf.st_mode))
- continue;
-
- spill_dir = AllocateDir(path);
- while ((spill_de = ReadDir(spill_dir, path)) != NULL)
- {
- if (strcmp(spill_de->d_name, ".") == 0 ||
- strcmp(spill_de->d_name, "..") == 0)
- continue;
-
- /* only look at names that can be ours */
- if (strncmp(spill_de->d_name, "xid", 3) == 0)
- {
- sprintf(path, "pg_replslot/%s/%s", logical_de->d_name,
- spill_de->d_name);
-
- if (unlink(path) != 0)
- ereport(PANIC,
- (errcode_for_file_access(),
- errmsg("could not remove file \"%s\": %m",
- path)));
- }
- }
- FreeDir(spill_dir);
+ ReorderBufferCleanupSerializedTXNs(logical_de->d_name);
}
FreeDir(logical_dir);
}
new_ent->combocid = ent->combocid;
}
}
+
+ CloseTransientFile(fd);
}
/*
- * Check whether the TransactionOId 'xid' is in the pre-sorted array 'xip'.
+ * Check whether the TransactionOid 'xid' is in the pre-sorted array 'xip'.
*/
static bool
TransactionIdInArray(TransactionId xid, TransactionId *xip, Size num)
projects / postgresql / blobdiff