Merge branch 'develop' into avoid_duplicate_filters

1 files changed, 26 insertions, 7 deletions

diff --git a/synapse/storage/events.py b/synapse/storage/events.py
index dbd63078c6..ea6879c619 100644
--- a/synapse/storage/events.py
+++ b/synapse/storage/events.py
@@ -207,6 +207,18 @@ class EventsStore(SQLBaseStore):
             where_clause="contains_url = true AND outlier = false",
         )
 
+        # an event_id index on event_search is useful for the purge_history
+        # api. Plus it means we get to enforce some integrity with a UNIQUE
+        # clause
+        self.register_background_index_update(
+            "event_search_event_id_idx",
+            index_name="event_search_event_id_idx",
+            table="event_search",
+            columns=["event_id"],
+            unique=True,
+            psql_only=True,
+        )
+
         self._event_persist_queue = _EventPeristenceQueue()
 
     def persist_events(self, events_and_contexts, backfilled=False):
@@ -2022,6 +2034,8 @@ class EventsStore(SQLBaseStore):
                 400, "topological_ordering is greater than forward extremeties"
             )
 
+        logger.debug("[purge] looking for events to delete")
+
         txn.execute(
             "SELECT event_id, state_key FROM events"
             " LEFT JOIN state_events USING (room_id, event_id)"
@@ -2030,6 +2044,14 @@ class EventsStore(SQLBaseStore):
         )
         event_rows = txn.fetchall()
 
+        to_delete = [
+            (event_id,) for event_id, state_key in event_rows
+            if state_key is None and not self.hs.is_mine_id(event_id)
+        ]
+        logger.info(
+            "[purge] found %i events before cutoff, of which %i are remote"
+            " non-state events to delete", len(event_rows), len(to_delete))
+
         for event_id, state_key in event_rows:
             txn.call_after(self._get_state_group_for_event.invalidate, (event_id,))
 
@@ -2080,6 +2102,7 @@ class EventsStore(SQLBaseStore):
         )
 
         state_rows = txn.fetchall()
+        logger.debug("[purge] found %i redundant state groups", len(state_rows))
 
         # make a set of the redundant state groups, so that we can look them up
         # efficiently
@@ -2173,10 +2196,6 @@ class EventsStore(SQLBaseStore):
         )
 
         # Delete all remote non-state events
-        to_delete = [
-            (event_id,) for event_id, state_key in event_rows
-            if state_key is None and not self.hs.is_mine_id(event_id)
-        ]
         for table in (
             "events",
             "event_json",
@@ -2192,7 +2211,7 @@ class EventsStore(SQLBaseStore):
             "event_signatures",
             "rejections",
         ):
-            logger.debug("[purge] removing non-state events from %s", table)
+            logger.debug("[purge] removing remote non-state events from %s", table)
 
             txn.executemany(
                 "DELETE FROM %s WHERE event_id = ?" % (table,),
@@ -2200,7 +2219,7 @@ class EventsStore(SQLBaseStore):
             )
 
         # Mark all state and own events as outliers
-        logger.debug("[purge] marking events as outliers")
+        logger.debug("[purge] marking remaining events as outliers")
         txn.executemany(
             "UPDATE events SET outlier = ?"
             " WHERE event_id = ?",
@@ -2210,7 +2229,7 @@ class EventsStore(SQLBaseStore):
             ]
         )
 
-        logger.debug("[purge] done")
+        logger.info("[purge] done")
 
     @defer.inlineCallbacks
     def is_event_after(self, event_id1, event_id2):