Clarify docs on hot standby lock release

simonat2ndQuadrant · simonat2ndQuadrant · commit 759340c8f524 · 2012-11-13T15:58:35.000-03:00
Andres Freund and Simon Riggs
diff --git a/src/backend/storage/ipc/procarray.c b/src/backend/storage/ipc/procarray.c
@@ -1394,6 +1394,11 @@ GetSnapshotData(Snapshot snapshot)
  * We don't worry about updating other counters, we want to keep this as
  * simple as possible and leave GetSnapshotData() as the primary code for
  * that bookkeeping.
+ *
+ * Note that if any transaction has overflowed its cached subtransactions
+ * then there is no real need include any subtransactions. That isn't a
+ * common enough case to worry about optimising the size of the WAL record,
+ * and we may wish to see that data for diagnostic purposes anyway.
  */
 RunningTransactions
 GetRunningTransactionData(void)
diff --git a/src/backend/storage/ipc/standby.c b/src/backend/storage/ipc/standby.c
@@ -513,6 +513,10 @@ CheckRecoveryConflictDeadlock(void)
  * RelationLockList, so we can keep track of the various entries made by
  * the Startup process's virtual xid in the shared lock table.
  *
+ * We record the lock against the top-level xid, rather than individual
+ * subtransaction xids. This means AccessExclusiveLocks held by aborted
+ * subtransactions are not released as early as possible on standbys.
+ *
  * List elements use type xl_rel_lock, since the WAL record type exactly
  * matches the information that we need to keep track of.
  *
@@ -646,8 +650,8 @@ StandbyReleaseAllLocks(void)
 
 /*
  * StandbyReleaseOldLocks
- *		Release standby locks held by XIDs that aren't running, as long
- *		as they're not prepared transactions.
+ *		Release standby locks held by top-level XIDs that aren't running,
+ *		as long as they're not prepared transactions.
  */
 void
 StandbyReleaseOldLocks(int nxids, TransactionId *xids)
