database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-11-22 12:22:45 +03:00
Code Activity

Remove remants of "snapshot too old"

Browse Source 
Remove the 'whenTaken' and 'lsn' fields from SnapshotData. After the
removal of the "snapshot too old" feature, they were never set to a
non-zero value.

This largely reverts commit 3e2f3c2e42, which added the
OldestActiveSnapshot tracking, and the init_toast_snapshot()
function. That was only required for setting the 'whenTaken' and 'lsn'
fields. SnapshotToast is now a constant again, like SnapshotSelf and
SnapshotAny. I kept a thin get_toast_snapshot() wrapper around
SnapshotToast though, to check that you have a registered or active
snapshot. That's still a useful sanity check.

Reviewed-by: Nathan Bossart, Andres Freund, Tom Lane
Discussion: https://www.postgresql.org/message-id/cd4b4f8c-e63a-41c0-95f6-6e6cd9b83f6d@iki.fi
This commit is contained in:
Heikki Linnakangas
2024-12-09 18:13:03 +02:00
parent f64ec81a81
commit 4d8275046c
10 changed files with 28 additions and 114 deletions
Download Patch File Download Diff File Expand all files Collapse all files

51
src/backend/access/common/toast_internals.c
View File

@@ -393,7 +393,6 @@ toast_delete_datum(Relation rel, Datum value, bool is_speculative)
HeapTuple toasttup;
int num_indexes;
int validIndex;
SnapshotData SnapshotToast;
if (!VARATT_IS_EXTERNAL_ONDISK(attr))
return;
@@ -425,9 +424,8 @@ toast_delete_datum(Relation rel, Datum value, bool is_speculative)
* sequence or not, but since we've already locked the index we might as
* well use systable_beginscan_ordered.)
*/
init_toast_snapshot(&SnapshotToast);
toastscan = systable_beginscan_ordered(toastrel, toastidxs[validIndex],
&SnapshotToast, 1, &toastkey);
get_toast_snapshot(), 1, &toastkey);
while ((toasttup = systable_getnext_ordered(toastscan, ForwardScanDirection)) != NULL)
{
/*
@@ -631,41 +629,28 @@ toast_close_indexes(Relation *toastidxs, int num_indexes, LOCKMODE lock)
}
/* ----------
* init_toast_snapshot
* get_toast_snapshot
*
* Initialize an appropriate TOAST snapshot. We must use an MVCC snapshot
* to initialize the TOAST snapshot; since we don't know which one to use,
* just use the oldest one.
* Return the TOAST snapshot. Detoasting *must* happen in the same
* transaction that originally fetched the toast pointer.
*/
void
init_toast_snapshot(Snapshot toast_snapshot)
Snapshot
get_toast_snapshot(void)
{
Snapshot snapshot = GetOldestSnapshot();
/*
* GetOldestSnapshot returns NULL if the session has no active snapshots.
* We can get that if, for example, a procedure fetches a toasted value
* into a local variable, commits, and then tries to detoast the value.
* Such coding is unsafe, because once we commit there is nothing to
* prevent the toast data from being deleted. Detoasting *must* happen in
* the same transaction that originally fetched the toast pointer. Hence,
* rather than trying to band-aid over the problem, throw an error. (This
* is not very much protection, because in many scenarios the procedure
* would have already created a new transaction snapshot, preventing us
* from detecting the problem. But it's better than nothing, and for sure
* we shouldn't expend code on masking the problem more.)
* We cannot directly check that detoasting happens in the same
* transaction that originally fetched the toast pointer, but at least
* check that the session has some active snapshots. It might not if, for
* example, a procedure fetches a toasted value into a local variable,
* commits, and then tries to detoast the value. Such coding is unsafe,
* because once we commit there is nothing to prevent the toast data from
* being deleted. (This is not very much protection, because in many
* scenarios the procedure would have already created a new transaction
* snapshot, preventing us from detecting the problem. But it's better
* than nothing.)
*/
if (snapshot == NULL)
if (!HaveRegisteredOrActiveSnapshot())
elog(ERROR, "cannot fetch toast data without an active snapshot");
/*
* Catalog snapshots can be returned by GetOldestSnapshot() even if not
* registered or active. That easily hides bugs around not having a
* snapshot set up - most of the time there is a valid catalog snapshot.
* So additionally insist that the current snapshot is registered or
* active.
*/
Assert(HaveRegisteredOrActiveSnapshot());
InitToastSnapshot(*toast_snapshot, snapshot->lsn, snapshot->whenTaken);
return &SnapshotToastData;
}

4
src/backend/access/heap/heaptoast.c
View File

@@ -639,7 +639,6 @@ heap_fetch_toast_slice(Relation toastrel, Oid valueid, int32 attrsize,
int endchunk;
int num_indexes;
int validIndex;
SnapshotData SnapshotToast;
/* Look for the valid index of toast relation */
validIndex = toast_open_indexes(toastrel,
@@ -685,9 +684,8 @@ heap_fetch_toast_slice(Relation toastrel, Oid valueid, int32 attrsize,
}
/* Prepare for scan */
init_toast_snapshot(&SnapshotToast);
toastscan = systable_beginscan_ordered(toastrel, toastidxs[validIndex],
&SnapshotToast, nscankeys, toastkey);
get_toast_snapshot(), nscankeys, toastkey);
/*
* Read the chunks by index