Fix get destinations to catch up query. (#9114)

t was doing a sequential scan on `destination_rooms`, which took
minutes.

2 files changed, 12 insertions, 13 deletions

diff --git a/changelog.d/9114.bugfix b/changelog.d/9114.bugfix
new file mode 100644

index 0000000000..77112abd5c
--- /dev/null
+++ b/changelog.d/9114.bugfix
@@ -0,0 +1 @@
+Fix bug in federation catchup logic that caused outbound federation to be delayed for large servers after start up. Introduced in v1.21.0.
diff --git a/synapse/storage/databases/main/transactions.py b/synapse/storage/databases/main/transactions.py

index 59207cadd4..cea595ff19 100644
--- a/synapse/storage/databases/main/transactions.py
+++ b/synapse/storage/databases/main/transactions.py
@@ -464,19 +464,17 @@ class TransactionStore(TransactionWorkerStore):
         txn: LoggingTransaction, now_time_ms: int, after_destination: Optional[str]
     ) -> List[str]:
         q = """
-            SELECT destination FROM destinations
-                WHERE destination IN (
-                    SELECT destination FROM destination_rooms
-                        WHERE destination_rooms.stream_ordering >
-                            destinations.last_successful_stream_ordering
-                )
-                AND destination > ?
-                AND (
-                    retry_last_ts IS NULL OR
-                    retry_last_ts + retry_interval < ?
-                )
-                ORDER BY destination
-                LIMIT 25
+            SELECT DISTINCT destination FROM destinations
+            INNER JOIN destination_rooms USING (destination)
+                WHERE
+                    stream_ordering > last_successful_stream_ordering
+                    AND destination > ?
+                    AND (
+                        retry_last_ts IS NULL OR
+                        retry_last_ts + retry_interval < ?
+                    )
+                    ORDER BY destination
+                    LIMIT 25
         """
         txn.execute(
             q,