aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorPeter Eisentraut <peter@eisentraut.org>2024-12-03 09:04:20 +0100
committerPeter Eisentraut <peter@eisentraut.org>2024-12-03 09:04:20 +0100
commit1acf10549e64c6a52ced570d712fcba1a2f5d1ec (patch)
tree27d398dde27f323f29ff6f1963c2d331aede72b9
parent1ba0782ce90cb4261098de59b49ae5cb2326566b (diff)
downloadpostgresql-1acf10549e64c6a52ced570d712fcba1a2f5d1ec.tar.gz
postgresql-1acf10549e64c6a52ced570d712fcba1a2f5d1ec.zip
Fix temporary memory leak in system table index scans
Commit 811af9786b introduced palloc() calls into systable_beginscan() and systable_beginscan_ordered(). But there was no pfree(), as is the usual style. It turns out that an ANALYZE of a partitioned table can invoke many thousand system table index scans, and this memory is not cleaned up until the end of the command, so this can temporarily leak quite a bit of memory. Maybe there are improvements to be made at a higher level about this, but for now, insert a couple of corresponding pfree() calls to fix this particular issue. Reported-by: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://www.postgresql.org/message-id/Z0XTfIq5xUtbkiIh@pryzbyj2023
-rw-r--r--src/backend/access/index/genam.c4
1 files changed, 4 insertions, 0 deletions
diff --git a/src/backend/access/index/genam.c b/src/backend/access/index/genam.c
index 60c61039d66..4b4ebff6a17 100644
--- a/src/backend/access/index/genam.c
+++ b/src/backend/access/index/genam.c
@@ -449,6 +449,8 @@ systable_beginscan(Relation heapRelation,
snapshot, nkeys, 0);
index_rescan(sysscan->iscan, idxkey, nkeys, NULL, 0);
sysscan->scan = NULL;
+
+ pfree(idxkey);
}
else
{
@@ -713,6 +715,8 @@ systable_beginscan_ordered(Relation heapRelation,
index_rescan(sysscan->iscan, idxkey, nkeys, NULL, 0);
sysscan->scan = NULL;
+ pfree(idxkey);
+
/*
* If CheckXidAlive is set then set a flag to indicate that system table
* scan is in-progress. See detailed comments in xact.c where these