Skip to content

Commit

Permalink
De-flake MSC3030 backfill test with Synapse worker replication (#492)
Browse files Browse the repository at this point in the history
Spawning from matrix-org/synapse#14028 (comment)

CI failure: https://github.com/matrix-org/synapse/actions/runs/3182998161/jobs/5189731097#step:6:15343 (from [discussion](matrix-org/synapse#14028 (comment)))
```
WORKERS=1 POSTGRES=1 COMPLEMENT_ALWAYS_PRINT_SERVER_LOGS=1 COMPLEMENT_DIR=../complement ./scripts-dev/complement.sh -run TestJumpToDateEndpoint/parallel/federation/can_paginate_after_getting_remote_event_from_timestamp_to_event_endpoint
```

### Why did this test fail with workers before?

The reason they failed before was because we're fighting against stale caches across workers racing while waiting for invalidation, see matrix-org/synapse#13185 (comment)

> Here is what happens:
>
>  1. `serverB` has `event1` stored as an `outlier` from previous requests (specifically from MSC3030 jump to date pulling in a missing `prev_event` after backfilling)
>  1. Client on `serverB` calls `/messages?dir=b`
>  1. `serverB:client_reader1` accepts the request and drives things
>  1. `serverB:client_reader1` has some backward extremities in range and requests `/backfill` from `serverA`
>  1. `serverB:client_reader1` processes the events from backfill including `event1` and puts them in the `_event_persist_queue`
>  1. `serverB:master` picks up the events from the `_event_persist_queue` and persists them to the database, de-outliers `event1` and invalidates its own cache and sends them over replication
>  1. `serverB:client_reader1` starts assembling the `/messages` response and gets `event1` out of the stale cache still as an `outlier`
>  1. `serverB:client_reader1` responds to the `/messages` request without `event1` because `outliers` are filtered out
>  1. `serverB:client_reader1` finally gets the replication data and invalidates its own cache for `event1` (too late, we already got the events from the stale cache and responded)
  • Loading branch information
MadLittleMods authored Oct 13, 2022
1 parent 1068501 commit e91a70f
Showing 1 changed file with 7 additions and 0 deletions.
7 changes: 7 additions & 0 deletions tests/msc3030_test.go
Original file line number Diff line number Diff line change
Expand Up @@ -203,6 +203,13 @@ func TestJumpToDateEndpoint(t *testing.T) {
contextResResBody := client.ParseJSON(t, contextRes)
paginationToken := client.GetJSONFieldStr(t, contextResResBody, "end")

// Hit `/messages` until `eventA` has been backfilled and replicated across
// workers (the worker persisting events isn't necessarily the same as the worker
// serving `/messages`)
fetchUntilMessagesResponseHas(t, remoteCharlie, roomID, func(ev gjson.Result) bool {
return ev.Get("event_id").Str == eventA.EventID
})

// Paginate backwards from eventB
messagesRes := remoteCharlie.MustDoFunc(t, "GET", []string{"_matrix", "client", "r0", "rooms", roomID, "messages"}, client.WithContentType("application/json"), client.WithQueries(url.Values{
"dir": []string{"b"},
Expand Down

0 comments on commit e91a70f

Please sign in to comment.