Merge branch 'release-v1.73' into matrix-org-hotfixes

2022-12-01 10:03:23 +00:00 · 2022-12-01 10:03:23 +00:00 · 34fa1276a4
parent 9d943ba8a6 6a41e5022e
commit 34fa1276a4
7 changed files with 34 additions and 99 deletions
--- a/CHANGES.md
+++ b/CHANGES.md
@ -1,23 +1,39 @@
+Synapse 1.73.0rc2 (2022-12-01)
+==============================
+
+Please note that legacy Prometheus metric names have been removed in this release; see [the upgrade notes](https://github.com/matrix-org/synapse/blob/release-v1.73/docs/upgrade.md#legacy-prometheus-metric-names-have-now-been-removed) for more details.
+
+Bugfixes
+--------
+
+- Fix a regression in Synapse 1.73.0rc1 where Synapse's main process would stop responding to HTTP requests when a user with a large number of devices logs in. ([\#14582](https://github.com/matrix-org/synapse/issues/14582))
+
+
 Synapse 1.73.0rc1 (2022-11-29)
 ==============================

 Features
 --------

+- Speed-up `/messages` with `filter_events_for_client` optimizations. ([\#14527](https://github.com/matrix-org/synapse/issues/14527))
+- Improve DB performance by reducing amount of data that gets read in `device_lists_changes_in_room`. ([\#14534](https://github.com/matrix-org/synapse/issues/14534))
 - Adds support for handling avatar in SSO login. Contributed by @ashfame. ([\#13917](https://github.com/matrix-org/synapse/issues/13917))
 - Move MSC3030 `/timestamp_to_event` endpoints to stable `v1` location (`/_matrix/client/v1/rooms/<roomID>/timestamp_to_event?ts=<timestamp>&dir=<direction>`, `/_matrix/federation/v1/timestamp_to_event/<roomID>?ts=<timestamp>&dir=<direction>`). ([\#14471](https://github.com/matrix-org/synapse/issues/14471))
- Reduce database load of [Client-Server endpoints](https://spec.matrix.org/v1.4/client-server-api/#aggregations) which return bundled aggregations. ([\#14491](https://github.com/matrix-org/synapse/issues/14491), [\#14508](https://github.com/matrix-org/synapse/issues/14508), [\#14510](https://github.com/matrix-org/synapse/issues/14510))
+- Reduce database load of [Client-Server endpoints](https://spec.matrix.org/v1.5/client-server-api/#aggregations) which return bundled aggregations. ([\#14491](https://github.com/matrix-org/synapse/issues/14491), [\#14508](https://github.com/matrix-org/synapse/issues/14508), [\#14510](https://github.com/matrix-org/synapse/issues/14510))
 - Add unstable support for an Extensible Events room version (`org.matrix.msc1767.10`) via [MSC1767](https://github.com/matrix-org/matrix-spec-proposals/pull/1767), [MSC3931](https://github.com/matrix-org/matrix-spec-proposals/pull/3931), [MSC3932](https://github.com/matrix-org/matrix-spec-proposals/pull/3932), and [MSC3933](https://github.com/matrix-org/matrix-spec-proposals/pull/3933). ([\#14520](https://github.com/matrix-org/synapse/issues/14520), [\#14521](https://github.com/matrix-org/synapse/issues/14521), [\#14524](https://github.com/matrix-org/synapse/issues/14524))
+- Prune user's old devices on login if they have too many. ([\#14038](https://github.com/matrix-org/synapse/issues/14038), [\#14580](https://github.com/matrix-org/synapse/issues/14580))


 Bugfixes
 --------

- Fix #12383: paginate room messages from the start if no from is given. Contributed by @gnunicorn . ([\#14149](https://github.com/matrix-org/synapse/issues/14149))
- Fix a bug introduced in 1.58.0 where a user with presence state 'org.matrix.msc3026.busy' would mistakenly be set to 'online' when calling `/sync` or `/events` on a worker process. ([\#14393](https://github.com/matrix-org/synapse/issues/14393))
+- Fix a long-standing bug where paginating from the start of a room did not work. Contributed by @gnunicorn. ([\#14149](https://github.com/matrix-org/synapse/issues/14149))
+- Fix a bug introduced in Synapse 1.58.0 where a user with presence state `org.matrix.msc3026.busy` would mistakenly be set to `online` when calling `/sync` or `/events` on a worker process. ([\#14393](https://github.com/matrix-org/synapse/issues/14393))
 - Fix a bug introduced in Synapse 1.70.0 where a receipt's thread ID was not sent over federation. ([\#14466](https://github.com/matrix-org/synapse/issues/14466))
 - Fix a long-standing bug where the [List media admin API](https://matrix-org.github.io/synapse/latest/admin_api/media_admin_api.html#list-all-media-in-a-room) would fail when processing an image with broken thumbnail information. ([\#14537](https://github.com/matrix-org/synapse/issues/14537))
 - Fix a bug introduced in Synapse 1.67.0 where two logging context warnings would be logged on startup. ([\#14574](https://github.com/matrix-org/synapse/issues/14574))
+- In application service transactions that include the experimental `org.matrix.msc3202.device_one_time_key_counts` key, include a duplicate key of `org.matrix.msc3202.device_one_time_keys_count` to match the name proposed by [MSC3202](https://github.com/matrix-org/matrix-spec-proposals/pull/3202). ([\#14565](https://github.com/matrix-org/synapse/issues/14565))
+- Fix a bug introduced in Synapse 0.9 where Synapse would fail to fetch server keys whose IDs contain a forward slash. ([\#14490](https://github.com/matrix-org/synapse/issues/14490))


 Improved Documentation
@ -35,30 +51,22 @@ Deprecations and Removals
 Internal Changes
 ----------------

- Prune user's old devices on login if they have too many. ([\#14038](https://github.com/matrix-org/synapse/issues/14038))
- Add missing type hints to `HomeServer`. ([\#14055](https://github.com/matrix-org/synapse/issues/14055))
+- Improve type hinting throughout Synapse. ([\#14055](https://github.com/matrix-org/synapse/issues/14055), [\#14412](https://github.com/matrix-org/synapse/issues/14412), [\#14529](https://github.com/matrix-org/synapse/issues/14529), [\#14452](https://github.com/matrix-org/synapse/issues/14452)).
 - Remove old stream ID tracking code. Contributed by Nick @Beeper (@fizzadar). ([\#14376](https://github.com/matrix-org/synapse/issues/14376), [\#14468](https://github.com/matrix-org/synapse/issues/14468))
 - Remove the `worker_main_http_uri` configuration setting. This is now handled via internal replication. ([\#14400](https://github.com/matrix-org/synapse/issues/14400), [\#14476](https://github.com/matrix-org/synapse/issues/14476))
+- Refactor `federation_sender` and `pusher` configuration loading. ([\#14496](https://github.com/matrix-org/synapse/issues/14496))
+([\#14509](https://github.com/matrix-org/synapse/issues/14509), [\#14573](https://github.com/matrix-org/synapse/issues/14573))
 - Faster joins: do not wait for full state when creating events to send. ([\#14403](https://github.com/matrix-org/synapse/issues/14403))
 - Faster joins: filter out non local events when a room doesn't have its full state. ([\#14404](https://github.com/matrix-org/synapse/issues/14404))
 - Faster joins: send events to initial list of servers if we don't have the full state yet. ([\#14408](https://github.com/matrix-org/synapse/issues/14408))
- Remove duplicated type information from type hints. ([\#14412](https://github.com/matrix-org/synapse/issues/14412))
+- Faster joins: use servers list approximation received during `send_join` (potentially updated with received membership events) in `assert_host_in_room`. ([\#14515](https://github.com/matrix-org/synapse/issues/14515))
 - Fix type logic in TCP replication code that prevented correctly ignoring blank commands. ([\#14449](https://github.com/matrix-org/synapse/issues/14449))
- Enable mypy's [`strict_equality` check](https://mypy.readthedocs.io/en/stable/command_line.html#cmdoption-mypy-strict-equality) by default. ([\#14452](https://github.com/matrix-org/synapse/issues/14452))
 - Remove option to skip locking of tables when performing emulated upserts, to avoid a class of bugs in future. ([\#14469](https://github.com/matrix-org/synapse/issues/14469))
 - `scripts-dev/federation_client`: Fix routing on servers with `.well-known` files. ([\#14479](https://github.com/matrix-org/synapse/issues/14479))
 - Reduce default third party invite rate limit to 216 invites per day. ([\#14487](https://github.com/matrix-org/synapse/issues/14487))
- Fix a bug introduced in Synapse 0.9 where it would fail to fetch server keys whose IDs contain a forward slash. ([\#14490](https://github.com/matrix-org/synapse/issues/14490))
- Refactor `federation_sender` and `pusher` configuration loading. ([\#14496](https://github.com/matrix-org/synapse/issues/14496))
-([\#14509](https://github.com/matrix-org/synapse/issues/14509), [\#14573](https://github.com/matrix-org/synapse/issues/14573))
- Faster joins: use servers list approximation received during `send_join` (potentially updated with received membership events) in `assert_host_in_room`. ([\#14515](https://github.com/matrix-org/synapse/issues/14515))
 - Refactor conversion of device list changes in room to outbound pokes to track unconverted rows using a `(stream ID, room ID)` position instead of updating the `converted_to_destinations` flag on every row. ([\#14516](https://github.com/matrix-org/synapse/issues/14516))
 - Add more prompts to the bug report form. ([\#14522](https://github.com/matrix-org/synapse/issues/14522))
 - Extend editorconfig rules on indent and line length to `.pyi` files. ([\#14526](https://github.com/matrix-org/synapse/issues/14526))
- Speed-up `/messages` with `filter_events_for_client` optimizations. ([\#14527](https://github.com/matrix-org/synapse/issues/14527))
- Add missing type hints. ([\#14529](https://github.com/matrix-org/synapse/issues/14529))
- Improve DB performance by reducing amount of data that gets read in `device_lists_changes_in_room`. ([\#14534](https://github.com/matrix-org/synapse/issues/14534))
- In application service transactions that include the experimental `org.matrix.msc3202.device_one_time_key_counts` key, include a duplicate key of `org.matrix.msc3202.device_one_time_keys_count` to match the name proposed by [MSC3202](https://github.com/matrix-org/matrix-spec-proposals/blob/travis/msc/otk-dl-appservice/proposals/3202-encrypted-appservices.md). ([\#14565](https://github.com/matrix-org/synapse/issues/14565))
 - Run Rust CI when `Cargo.lock` changes. This is particularly useful for dependabot updates. ([\#14571](https://github.com/matrix-org/synapse/issues/14571))
 - Fix a possible variable shadow in `create_new_client_event`. ([\#14575](https://github.com/matrix-org/synapse/issues/14575))
 - Bump various dependencies in the `poetry.lock` file and in CI scripts. ([\#14557](https://github.com/matrix-org/synapse/issues/14557), [\#14559](https://github.com/matrix-org/synapse/issues/14559), [\#14560](https://github.com/matrix-org/synapse/issues/14560), [\#14500](https://github.com/matrix-org/synapse/issues/14500), [\#14501](https://github.com/matrix-org/synapse/issues/14501), [\#14502](https://github.com/matrix-org/synapse/issues/14502), [\#14503](https://github.com/matrix-org/synapse/issues/14503), [\#14504](https://github.com/matrix-org/synapse/issues/14504), [\#14505](https://github.com/matrix-org/synapse/issues/14505)).
--- a/debian/changelog
+++ b/debian/changelog
@ -1,3 +1,9 @@
+matrix-synapse-py3 (1.73.0~rc2) stable; urgency=medium
+
+  * New Synapse release 1.73.0rc2.
+
+ -- Synapse Packaging team <packages@matrix.org>  Thu, 01 Dec 2022 10:02:19 +0000
+
 matrix-synapse-py3 (1.73.0~rc1) stable; urgency=medium

  * New Synapse release 1.73.0rc1.
--- a/pyproject.toml
+++ b/pyproject.toml
@ -57,7 +57,7 @@ manifest-path = "rust/Cargo.toml"

 [tool.poetry]
 name = "matrix-synapse"
-version = "1.73.0rc1"
+version = "1.73.0rc2"
 description = "Homeserver for the Matrix decentralised comms protocol"
 authors = ["Matrix.org Team and Contributors <packages@matrix.org>"]
 license = "Apache-2.0"
--- a/synapse/handlers/device.py
+++ b/synapse/handlers/device.py
@ -421,9 +421,6 @@ class DeviceHandler(DeviceWorkerHandler):

        self._check_device_name_length(initial_device_display_name)

-        # Prune the user's device list if they already have a lot of devices.
-        await self._prune_too_many_devices(user_id)
-
        if device_id is not None:
            new_device = await self.store.store_device(
                user_id=user_id,
@ -455,14 +452,6 @@ class DeviceHandler(DeviceWorkerHandler):

        raise errors.StoreError(500, "Couldn't generate a device ID.")

-    async def _prune_too_many_devices(self, user_id: str) -> None:
-        """Delete any excess old devices this user may have."""
-        device_ids = await self.store.check_too_many_devices_for_user(user_id)
-        if not device_ids:
-            return
-
-        await self.delete_devices(user_id, device_ids)
-
    async def _delete_stale_devices(self) -> None:
        """Background task that deletes devices which haven't been accessed for more than
        a configured time period.
@ -492,7 +481,7 @@ class DeviceHandler(DeviceWorkerHandler):
            device_ids = [d for d in device_ids if d != except_device_id]
        await self.delete_devices(user_id, device_ids)

-    async def delete_devices(self, user_id: str, device_ids: Collection[str]) -> None:
+    async def delete_devices(self, user_id: str, device_ids: List[str]) -> None:
        """Delete several devices

        Args:
--- a/synapse/storage/databases/main/devices.py
+++ b/synapse/storage/databases/main/devices.py
@ -1533,71 +1533,6 @@ class DeviceBackgroundUpdateStore(SQLBaseStore):

        return rows

-    async def check_too_many_devices_for_user(self, user_id: str) -> Collection[str]:
-        """Check if the user has a lot of devices, and if so return the set of
-        devices we can prune.
-
-        This does *not* return hidden devices or devices with E2E keys.
-        """
-
-        num_devices = await self.db_pool.simple_select_one_onecol(
-            table="devices",
-            keyvalues={"user_id": user_id, "hidden": False},
-            retcol="COALESCE(COUNT(*), 0)",
-            desc="count_devices",
-        )
-
-        # We let users have up to ten devices without pruning.
-        if num_devices <= 10:
-            return ()
-
-        # We prune everything older than N days.
-        max_last_seen = self._clock.time_msec() - 14 * 24 * 60 * 60 * 1000
-
-        if num_devices > 50:
-            # If the user has more than 50 devices, then we chose a last seen
-            # that ensures we keep at most 50 devices.
-            sql = """
-                SELECT last_seen FROM devices
-                LEFT JOIN e2e_device_keys_json USING (user_id, device_id)
-                WHERE
-                    user_id = ?
-                    AND NOT hidden
-                    AND last_seen IS NOT NULL
-                    AND key_json IS NULL
-                ORDER BY last_seen DESC
-                LIMIT 1
-                OFFSET 50
-            """
-
-            rows = await self.db_pool.execute(
-                "check_too_many_devices_for_user_last_seen", None, sql, (user_id,)
-            )
-            if rows:
-                max_last_seen = max(rows[0][0], max_last_seen)
-
-        # Now fetch the devices to delete.
-        sql = """
-            SELECT DISTINCT device_id FROM devices
-            LEFT JOIN e2e_device_keys_json USING (user_id, device_id)
-            WHERE
-                user_id = ?
-                AND NOT hidden
-                AND last_seen < ?
-                AND key_json IS NULL
-        """
-
-        def check_too_many_devices_for_user_txn(
-            txn: LoggingTransaction,
-        ) -> Collection[str]:
-            txn.execute(sql, (user_id, max_last_seen))
-            return {device_id for device_id, in txn}
-
-        return await self.db_pool.runInteraction(
-            "check_too_many_devices_for_user",
-            check_too_many_devices_for_user_txn,
-        )
-

 class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore):
    # Because we have write access, this will be a StreamIdGenerator
@ -1656,7 +1591,6 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore):
                values={},
                insertion_values={
                    "display_name": initial_device_display_name,
-                    "last_seen": self._clock.time_msec(),
                    "hidden": False,
                },
                desc="store_device",
@ -1702,7 +1636,7 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore):
            )
            raise StoreError(500, "Problem storing device.")

-    async def delete_devices(self, user_id: str, device_ids: Collection[str]) -> None:
+    async def delete_devices(self, user_id: str, device_ids: List[str]) -> None:
        """Deletes several devices.

        Args:
--- a/tests/handlers/test_device.py
+++ b/tests/handlers/test_device.py
@ -115,7 +115,7 @@ class DeviceTestCase(unittest.HomeserverTestCase):
                "device_id": "xyz",
                "display_name": "display 0",
                "last_seen_ip": None,
-                "last_seen_ts": 1000000,
+                "last_seen_ts": None,
            },
            device_map["xyz"],
        )
--- a/tests/storage/test_client_ips.py
+++ b/tests/storage/test_client_ips.py
@ -169,8 +169,6 @@ class ClientIpStoreTestCase(unittest.HomeserverTestCase):
            )
        )

-        last_seen = self.clock.time_msec()
-
        if after_persisting:
            # Trigger the storage loop
            self.reactor.advance(10)
@ -191,7 +189,7 @@ class ClientIpStoreTestCase(unittest.HomeserverTestCase):
                        "device_id": device_id,
                        "ip": None,
                        "user_agent": None,
-                        "last_seen": last_seen,
+                        "last_seen": None,
                    },
                ],
            )