Commit graph

251 commits

Author SHA1 Message Date
jaredj 8eb477a635 Merge branch 'master' into develop 2021-11-08 12:04:57 +11:00
Neil Alexander b9a575919a
Try to reduce re-allocations a bit in resolveConflictsV2 2021-11-04 10:55:07 +00:00
PiotrKozimor dec05c3347
Run gofmt on dendrite - apply go 1.17 preferred build tags (#2021) 2021-11-02 16:48:48 +00:00
Neil Alexander b99f594a93
Fix #2028 (#2036) 2021-11-02 16:47:39 +00:00
alexfca 60e11f88b8
- Remove PerformQueryAllPartitions as it does not support aggreates (#24)
- Update queries to all use PartitionKeys
- Remove the _sid from queries as the PK contains the Tenant
- Fix some bugs around empty values and ordering

Co-authored-by: alexf@example.com <alexf@example.com>
2021-10-08 11:17:22 +11:00
alexfca 49f8c7fe38
- Add $$NULL$$ string to be used when a part of the DocId is known to be NULL (#23)
- Use the $$NULL$$ for the known nullable use cases

Co-authored-by: alexf@example.com <alexf@example.com>
2021-10-06 10:07:40 +11:00
alexfca db34c0950e
- Make the CosmosDocId use commas as separators instead of underscore to match Messaging (#22)
- Make the DocId for StateBlock to be bas64 and not hex

Co-authored-by: alexf@example.com <alexf@example.com>
2021-10-01 10:02:23 +10:00
alexfca 1a5d7f2bb2
- Add SetUpdateTime() to all Update statements (#20)
- Refactor Updates to use a shared method
- Remove private setXX() methods

Co-authored-by: alexf@example.com <alexf@example.com>
2021-09-24 10:56:06 +10:00
alexfca 3088238419
Add UniqueId to PartitionKey for some Dendrite tables (where possible) (#19)
* - Make all PartitionKeys include the tablename
- Update specific PKs to be item specific
- Add validation to the PerformQueryXX methods
- Fix queries that fail validation

* - Revert the PK back to CollectionName as it already includes the TableName

Co-authored-by: alexf@example.com <alexf@example.com>
2021-09-23 14:48:32 +10:00
alexfca 927238a686
Use a common way to generate CollectionName and PartitionKey (#18)
* - Create CosmosDocument as a base class
- Add CT and UT
- Refactor all tables to use the CosmosDocument

* - Add UpsertDocument method to perform updates in a generic way
- Add SetUpdateTime() to update the UT for updates
- Refactor it all

* - Add Performquery method
- Refactor appservice_events_table

* - Update naffka Topics and Messages to use the common pattern

* - Update keyserver to use the common pattern for collection and PK

* - Update mediaapi to use the common pattern for collection and pk

* - Update roomserver to use the common pattern for collectionname and pk

* - Update signingkeyserver to use the common pattern for collectionname and pk

* - Update userapi touse the common pattern for collectionname and pk

* - Update partitionOffset to use the common collectionname and pk
- Remove generic GetPartitionKey() method

Co-authored-by: alexf@example.com <alexf@example.com>
2021-09-23 09:02:37 +10:00
alexfca acf63daf79
Add CT and UT to all documents and refactor (#17)
* - Create CosmosDocument as a base class
- Add CT and UT
- Refactor all tables to use the CosmosDocument

* - Add UpsertDocument method to perform updates in a generic way
- Add SetUpdateTime() to update the UT for updates
- Refactor it all

Co-authored-by: alexf@example.com <alexf@example.com>
2021-09-20 17:41:04 +10:00
alexfca 032ad5b052
Fix Forget Room bug that prevented Room Deletion (#16)
- Fix membership Insert to ensure the OnConflict performs an Update if exists
- Fix type in the rooms bulkSelectRoomNIDsSQL

Co-authored-by: alexf@example.com <alexf@example.com>
2021-09-13 09:52:55 +10:00
alexfca fd7f25479b
Upgrade Dendrite 0.5.0 support for CosmosDB (#15)
* - Add CosmosDB back
- Add missing methods to blacklist_table.go
- Add missing methods to device_keys_table.go
- Add missing methods to events_table.go
- Add missing methods to membership_table.go
- Update state_block_table.go (due to reafctor SQL)
- Update state_snapshot_table.go (due to reafctor SQL)
- Add new key_backup_table.go
- Add new key_backup_version_table.go
- Code compiles but has runtime errors

* Message sending + receiving working
Rooms and DMs working
- Add CrossSigningKeys table
- Add CrossSigningSigs table
- Refactor DeviceKeys yable
- Fix OneTimeKeys
- Update the KeyServer storage.go to use a PartitionStorer instead of a specific SQL PartitionOffsetStatements
- Fix small issues from the previous commit
- Implement DeleteSendToDeviceMessages

Co-authored-by: alexf@example.com <alexf@example.com>
2021-09-10 16:04:17 +10:00
Ryan W a624eab309
- Removed double imports (#1989)
- Lower cased error messages

Signed-off-by: Ryan Whittington <twentybitdev@gmail.com>

Co-authored-by: kegsay <kegan@matrix.org>
2021-09-08 17:31:03 +01:00
kegsay 7dc8fb1fe7
Add more logs (#2005)
* Add more logs

To help debug the migration issue in #1924 along with manual data-loss-inducing fixes.
Also log the origin server on processed txns to help debug buggy server origins.

* Fix query
2021-09-07 15:07:14 +01:00
Jared Jacobson 84c8cb052b Merge branch 'master' into develop
# Conflicts:
#	go.mod
#	go.sum
2021-09-03 13:54:05 +10:00
Neil Alexander 51b119107c
Don't return nonsense canonical room aliases in the public rooms responses (#1992) 2021-08-27 16:50:30 +01:00
kegsay 4cc8b28b7f
Ensure all create events have a snapshot NID of 0 (#1961)
Fixes #1924 for postgres users, though the underlying cause of why
they aren't 0 in the first place is unresolved.
2021-08-04 17:48:23 +01:00
kegsay ed04eed441
Fix sqlite migration issues (#1960)
* Do not store 'null' in the database for empty JSON arrays

This can cause issues, though it should be noted that the majority
of the time this will marshal/unmarshal just fine, see
https://play.golang.org/p/Doe2NZUgv7Q

* bugfix: sqlite migration should handle create events as having no 'before' snapshot

The state snapshot for any given event in the roomserver represents the state _before_
the event. For the create event, this is nothing, so the state snapshot nid should be 0.

In some cases this wasn't happening, resulting in a nice mix of possible options including:
 - A state snapshot without any state blocks `[]` or `null`.
 - A state snapshot with a single state block with a single event, the create event, causing
   a circular loop. This is incorrect as it represents the state before the event, not after.

* Add state key check
2021-08-04 17:08:17 +01:00
Kegan Dougal ed4097825b Factor out StatementList to sqlutil and use it in userapi
It helps with the boilerplate.
2021-07-28 18:30:04 +01:00
jahmai-ca 199c4c3cd1
Add tenant-id to all docs and cross-tenant queries (#14)
* Add tenant-id to all docs and cross-tenant queries

* Use TenantName in doc ids too

* Space

* - Rename variables and ensure code compiles

* - Fix error handling for GroupBy query

* - Fix typo

Co-authored-by: Alex Flatow <alexf@criticalarc.com>
2021-07-27 14:10:35 +10:00
kegsay 16bf94f239
Not finding the snapshot is not fatal (#1940) 2021-07-26 12:30:44 +01:00
Neil Alexander 39e8d1cc6f
Track knocking in membership updater (#1935)
* Topologically sort outliers in SendEventWithState

* Knock in membership updater

* Update gomatrixserverlib

* Update gomatrixserverlib

* Get the NID of the knock event properly for the membership updater
2021-07-22 12:26:58 +01:00
Neil Alexander c1447a58e5
Various alias fixes (#1934)
* Generate m.room.canonical_alias instead of legacy m.room.aliases

* Add omitempty tags

* Add aliases endpoint to client API

* Check power levels when setting aliases

* Don't return null on /aliases

* Don't return error if the state event fails

* Update sytest-whitelist

* Don't send updated m.room.canonical_alias events

* Don't check PLs after all because for local aliases they are apparently irrelevant

* Fix some bugs

* Allow deleting a local alias with enough PL

* Fix some more bugs

* Update sytest-whitelist

* Fix copyright notices

* Review comments
2021-07-21 16:53:50 +01:00
Neil Alexander f0f8c7f055
Optimise QueryServerJoinedToRoom (#1933)
* Optimise checking if a server is in a room

* Fix queries

* Fix queries
2021-07-21 13:06:32 +01:00
Neil Alexander f63068df3b
Only include go-sqlite3 on the relevant binaries (#1900)
* Only include go-sqlite3 on the relevant binaries

* The driver name is always sqlite3 now

* Update to matrix-org/go-sqlite3-js@e537baa
2021-07-20 11:18:14 +01:00
David Spenler 8d8fe485b4
Fix failing ban tests (#1884)
* Add room membership and powerlevel checks for func SendBan

* Added non-error return to func GetStateEvent when no state events with the specified state key are found

* Add passing tests to whitelist

* Fixed formatting

* Update roomserver/storage/shared/storage.go

Co-authored-by: Neil Alexander <neilalexander@users.noreply.github.com>
Co-authored-by: kegsay <kegan@matrix.org>
Co-authored-by: kegsay <kegsay@gmail.com>
2021-07-19 18:33:05 +01:00
Neil Alexander 09d3bab838
Metric fixes
Squashed commit of the following:

commit c6eb4d8bbf80320ec2b6d416c77659b0343e5e47
Author: Neil Alexander <neilalexander@users.noreply.github.com>
Date:   Mon Jul 19 16:52:57 2021 +0100

    Fix bug

commit d420966d9ac44936728960a8d38602662b58f1c3
Author: Neil Alexander <neilalexander@users.noreply.github.com>
Date:   Mon Jul 19 16:46:12 2021 +0100

    Update metric

commit 0ad6e37846e2ebbbd0e33a38274094bd15b8f11b
Author: Neil Alexander <neilalexander@users.noreply.github.com>
Date:   Mon Jul 19 16:30:14 2021 +0100

    Fix observe for calculateStateDurations
2021-07-19 17:20:55 +01:00
Neil Alexander eb2a8e4c0b
Set buckets for dendrite_roomserver_calculate_state_duration_microseconds 2021-07-19 16:07:06 +01:00
Neil Alexander b20d402f39
dendrite_roomserver_calculate_state_duration_microseconds as histogram rather than summary 2021-07-19 15:34:12 +01:00
kegsay e80098e186
bugfix: retire invites even when we cannot talk to the remote server to make/send_leave (#1918)
* bugfix: retire invites even when we cannot talk to the remote server to make/send_leave

Also modify the leave response in /sync to include a fake event as this is ultimately
what clients (and sytest) will use to determine leave-ness.

* hash the event ID

* Base64 not hex
2021-07-14 10:39:17 +01:00
kegsay f8ae391a5b
Expose more data when outputting output room events (#1916)
* Add more logging for content fields

* Fix fields
2021-07-13 11:19:21 +01:00
Neil Alexander acec6fa979
Move a couple of callers to helpers.IsServerCurrentlyInRoom over to the query API (#1912) 2021-07-09 17:49:59 +01:00
Neil Alexander c8408a6387
Add more optimised code path for checking if we're in a room (#1909)
* Add more optimised code path for checking if we're in a room

* Fix database queries

* Fix federation API test

* Fix logging

* Review comments

* Make separate API call for room membership
2021-07-09 16:36:45 +01:00
kegsay 3e50bac944
bugfix: order the state blocks so recreating state snapshots works correctly (#1908)
* Logging

* Revert "Logging"

This reverts commit 23ce334182.

* bugfix: order the state blocks so recreating state snapshots works correctly
2021-07-09 10:49:49 +01:00
Neil Alexander 816e1a402b
Fix bug when rejecting invites (#1907)
* Fix rejecting invites maybe

* Remove comment that is no longer correct

* Review comment on performFederatedRejectInvite
2021-07-08 14:54:03 +01:00
kegsay 5a09290c32
db migration: handle create events with no state blocks from v0.1.0 (#1904) 2021-07-07 17:07:33 +01:00
Neil Alexander 192a7a7923
Roomserver input backpressure metric
Squashed commit of the following:

commit 56e934ac0aeedcfb2c072010959ba49734d4e0cb
Author: Neil Alexander <neilalexander@users.noreply.github.com>
Date:   Fri Jul 2 09:39:30 2021 +0100

    Fix metric

commit 3911f3a0c17b164b012e881c085ceca30f5de408
Author: Neil Alexander <neilalexander@users.noreply.github.com>
Date:   Fri Jul 2 09:36:29 2021 +0100

    Register correct metric

commit a9ddbfaed421538a701151801e9451198a8be4f3
Author: Neil Alexander <neilalexander@users.noreply.github.com>
Date:   Fri Jul 2 09:33:33 2021 +0100

    Try to capture RS input backpressure metric
2021-07-02 09:48:55 +01:00
kegsay c849e74dfc
db migration: fix #1844 and add additional assertions (#1889)
* db migration: fix #1844 and add additional assertions

- Migration scripts will now check to see if there are any unconverted
  snapshot IDs and fail the migration if there are any. This should
  prevent people from getting a corrupt database in the event the root
  cause is still unknown.
- Add an ORDER BY clause when doing batch queries in the postgres
  migration. LIMIT and OFFSET without ORDER BY are undefined and must
  not be relied upon to produce a deterministic ordering (e.g row order).
  See https://www.postgresql.org/docs/current/queries-limit.html

* Linting

Co-authored-by: Neil Alexander <neilalexander@users.noreply.github.com>
2021-06-29 11:25:17 +01:00
Neil Alexander 7c3991ee2f
Use a custom FIFO queue for the RS input API (#1888)
* Use a FIFO queue instead of a channel to reduce backpressure

* Make sure someone wakes up

* Tweaks

* Add comments
2021-06-28 15:11:36 +01:00
Neil Alexander 5357df36c9
Fix panic in roomserver 2021-06-21 09:41:12 +01:00
alexfca ae2eea039e
- Update setXX() to use the object and not require the PartitionKey (#13)
- Remove redundant setXX() code
2021-06-01 14:52:59 +10:00
alexfca ed89984f9d
Implement Cosmos DB for the Media API (#11)
* - Update config to use CosmosDB for MediaAPI
- Implement CosmosDB for MediaAPI
- Fix bug in Public Room Creation

* - Remove file: config line
2021-05-31 13:20:09 +10:00
alexfca 5d68daef80
Implement Cosmos DB for the RoomServer Service (#5)
* - Implement Cosmos for the devices_table
- Use the ConnectionString in the YAML to include the Tenant
- Revert all other non implemented tables back to use SQLLite3

* - Change the Config to use "test.criticicalarc.com" Container
- Add generic function GetDocumentOrNil to standardize GetDocument
- Add func to return CrossPartition queries for Aggregates
- Add func GetNextSequence() as generic seq generator for AutoIncrement
- Add cosmosdbutil.ErrNoRows to return (emulate) sql.ErrNoRows
- Add a "fake" ExclusiveWriterFake
- Add standard "getXX", "setXX" and "queryXX" to all TABLE class files
- Add specific Table SEQ for the Events table
- Add specific Table SEQ for the Rooms table
- Add specific Table SEQ for the StateSnapshot table
2021-05-20 14:42:33 +10:00
alexfca b696923333
- Implement Cosmos for the devices_table (#4)
- Use the ConnectionString in the YAML to include the Tenant
- Revert all other non implemented tables back to use SQLLite3
2021-05-12 16:30:49 +10:00
Alex Flatow 5ded872da9 - Add CosmosDB as a Datasource type
- Use the SQLLite as a base for the CosmosDB package(s)
- Update the ConnString to use file: from cosmosdb: so it still works
- Add a yaml file for the config to use CosmosDB
2021-05-06 15:09:44 +10:00
Alex Flatow 4b42a0addb storage.go 2021-05-06 14:41:27 +10:00
Neil Alexander c67d8da3eb
Fix bug in SQLite migration 2021-04-26 13:45:47 +01:00
Neil Alexander 5ce1fe80de
State storage refactor (#1839)
* Hash-deduplicated state storage (and migrations) for PostgreSQL and SQLite

* Refactor droomserver database setup for migrations

* Fix conflict statements

* Update migration names

* Set a boundary for old to new block/snapshot IDs so we don't rewrite them more than once accidentally

* Create sequence if not exists

* Fix boundary queries

* Fix boundary queries

* Use Query

* Break out queries a bit

* More sequence tweaks

* Query parameters are not playing the game

* Injection escaping may not work for CREATE SEQUENCE after all

* Fix snapshot sequence name

* Use boundaried IDs in SQLite too

* Use IFNULL for SQLite

* Use COALESCE in PostgreSQL

* Review comments @Kegsay
2021-04-26 13:25:57 +01:00
Kegsay af41f6d454
Add Sentry support (#1803)
* Add Sentry support

* Use HTTP Sentry properly maybe

* Capture panics

* Log fed Sentry stuff correctly

* British english linter
2021-03-24 10:25:24 +00:00