gitlord
/
synapse
mirror of https://github.com/matrix-org/synapse.git
1
0
Fork 0
Code Issues 0 Releases 671 Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
23339 Commits
842 Branches
288 MiB
 
 
 
 
 
 
Tree: f2f2c7c1f0
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'f2f2c7c1f0'
${ noResults }
synapse/synapse/storage/schema/__init__.py

140 lines
6.0 KiB
Raw Blame History 

  1. # Copyright 2021 The Matrix.org Foundation C.I.C.

  2. #

  3. # Licensed under the Apache License, Version 2.0 (the "License");

  4. # you may not use this file except in compliance with the License.

  5. # You may obtain a copy of the License at

  6. #

  7. #     http://www.apache.org/licenses/LICENSE-2.0

  8. #

  9. # Unless required by applicable law or agreed to in writing, software

  10. # distributed under the License is distributed on an "AS IS" BASIS,

  11. # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.

  12. # See the License for the specific language governing permissions and

  13. # limitations under the License.

  14. 

  15. SCHEMA_VERSION = 83  # remember to update the list below when updating

  16. """Represents the expectations made by the codebase about the database schema

  17. 

  18. This should be incremented whenever the codebase changes its requirements on the

  19. shape of the database schema (even if those requirements are backwards-compatible with

  20. older versions of Synapse).

  21. 

  22. See https://matrix-org.github.io/synapse/develop/development/database_schema.html

  23. for more information on how this works.

  24. 

  25. Changes in SCHEMA_VERSION = 61:

  26.     - The `user_stats_historical` and `room_stats_historical` tables are not written and

  27.       are not read (previously, they were written but not read).

  28.     - MSC2716: Add `insertion_events` and `insertion_event_edges` tables to keep track

  29.       of insertion events in order to navigate historical chunks of messages.

  30.     - MSC2716: Add `chunk_events` table to track how the chunk is labeled and

  31.       determines which insertion event it points to.

  32. 

  33. Changes in SCHEMA_VERSION = 62:

  34.     - MSC2716: Add `insertion_event_extremities` table that keeps track of which

  35.       insertion events need to be backfilled.

  36. 

  37. Changes in SCHEMA_VERSION = 63:

  38.     - The `public_room_list_stream` table is not written nor read to

  39.       (previously, it was written and read to, but not for any significant purpose).

  40.       https://github.com/matrix-org/synapse/pull/10565

  41. 

  42. Changes in SCHEMA_VERSION = 64:

  43.     - MSC2716: Rename related tables and columns from "chunks" to "batches".

  44. 

  45. Changes in SCHEMA_VERSION = 65:

  46.     - MSC2716: Remove unique event_id constraint from insertion_event_edges

  47.       because an insertion event can have multiple edges.

  48.     - Remove unused tables `user_stats_historical` and `room_stats_historical`.

  49. 

  50. Changes in SCHEMA_VERSION = 66:

  51.     - Queries on state_key columns are now disambiguated (ie, the codebase can handle

  52.       the `events` table having a `state_key` column).

  53. 

  54. Changes in SCHEMA_VERSION = 67:

  55.     - state_events.prev_state is no longer written to.

  56. 

  57. Changes in SCHEMA_VERSION = 68:

  58.     - event_reference_hashes is no longer read.

  59.     - `events` has `state_key` and `rejection_reason` columns, which are populated for

  60.       new events.

  61. 

  62. Changes in SCHEMA_VERSION = 69:

  63.     - We now write to `device_lists_changes_in_room` table.

  64.     - We now use a PostgreSQL sequence to generate future txn_ids for

  65.       `application_services_txns`. `application_services_state.last_txn` is no longer

  66.       updated.

  67. 

  68. Changes in SCHEMA_VERSION = 70:

  69.     - event_reference_hashes is no longer written to.

  70. 

  71. Changes in SCHEMA_VERSION = 71:

  72.     - event_edges.room_id is no longer read from.

  73.     - Tables related to groups are no longer accessed.

  74. 

  75. Changes in SCHEMA_VERSION = 72:

  76.     - event_edges.(room_id, is_state) are no longer written to.

  77.     - Tables related to groups are dropped.

  78.     - Unused column application_services_state.last_txn is dropped

  79.     - Cache invalidation stream id sequence now begins at 2 to match code expectation.

  80. 

  81. Changes in SCHEMA_VERSION = 73:

  82.     - thread_id column is added to event_push_actions, event_push_actions_staging

  83.       event_push_summary, receipts_linearized, and receipts_graph.

  84.     - Add table `event_failed_pull_attempts` to keep track when we fail to pull

  85.       events over federation.

  86.     - Add indexes to various tables (`event_failed_pull_attempts`, `insertion_events`,

  87.       `batch_events`) to make it easy to delete all associated rows when purging a room.

  88.     - `inserted_ts` column is added to `event_push_actions_staging` table.

  89. 

  90. Changes in SCHEMA_VERSION = 74:

  91.     - A query on `event_stream_ordering` column has now been disambiguated (i.e. the

  92.       codebase can handle the `current_state_events`, `local_current_memberships` and

  93.       `room_memberships` tables having an `event_stream_ordering` column).

  94. 

  95. Changes in SCHEMA_VERSION = 75:

  96.     - The `event_stream_ordering` column in membership tables (`current_state_events`,

  97.       `local_current_membership` & `room_memberships`) is now being populated for new

  98.       rows. When the background job to populate historical rows lands this will

  99.       become the compat schema version.

  100. 

  101. Changes in SCHEMA_VERSION = 76:

  102.     - Adds a full_user_id column to tables profiles and user_filters.

  103. 

  104. Changes in SCHEMA_VERSION = 77

  105.     - (Postgres) Add NOT VALID CHECK (full_user_id IS NOT NULL) to tables profiles and user_filters

  106. 

  107. Changes in SCHEMA_VERSION = 78

  108.     - Validate check (full_user_id IS NOT NULL) on tables profiles and user_filters

  109. 

  110. Changes in SCHEMA_VERSION = 79

  111.     - Add tables to handle in DB read-write locks.

  112.     - Add some mitigations for a painful race between foreground and background updates, cf

  113.       https://github.com/matrix-org/synapse/issues/15677.

  114. 

  115. Changes in SCHEMA_VERSION = 80

  116.     - The event_txn_id_device_id is always written to for new events.

  117.     - Add tables for the task scheduler.

  118. 

  119. Changes in SCHEMA_VERSION = 81

  120.     - The event_txn_id is no longer written to for new events.

  121. 

  122. Changes in SCHEMA_VERSION = 82

  123.     - The insertion_events, insertion_event_extremities, insertion_event_edges, and

  124.       batch_events tables are no longer purged in preparation for their removal.

  125. 

  126. Changes in SCHEMA_VERSION = 83

  127.     - The event_txn_id is no longer used.

  128. """

  129. 

  130. 

  131. SCHEMA_COMPAT_VERSION = (

  132.     # The `event_txn_id_device_id` must be written to for new events.

  133.     80

  134. )

  135. """Limit on how far the synapse codebase can be rolled back without breaking db compat

  136. 

  137. This value is stored in the database, and checked on startup. If the value in the

  138. database is greater than SCHEMA_VERSION, then Synapse will refuse to start.

  139. """