-
-
Notifications
You must be signed in to change notification settings - Fork 14.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
matrix-synapse: fix test errors #371815
matrix-synapse: fix test errors #371815
Conversation
Still seeing one other error:
|
This comment was marked as duplicate.
This comment was marked as duplicate.
Shall we wait until everything is fixed? I mean this patch is already part of 1.122.0rc1 so we would be dropping it very soon. |
for reference element-hq/synapse#17878 (comment) |
My bad for posting the duplicate. I guess I did not see the previous comment. Anyways, thanks again for all your hard work to get this resolved! |
This comment was marked as outdated.
This comment was marked as outdated.
This now works for me. |
I've also been able to deploy with this patch and the one from #369859, but it seems like a dangerous solution to me. The test seems to be related to database recovery. If the actual database recovery functionality is broken as well, I don't know whether it would be a good idea to ship this. |
@chvp for clarity: the test is about recovering from a database outage, not recovering the database itself. Mainly related to fetching events when the db is down. Test case (third party repo because it was what showed up looking for the test with google): https://mau.dev/leytilera/synapse/-/blob/v1.66.0/tests/storage/databases/main/test_events_worker.py#L197 Edit: just double checked, seems this test has not changed upstream since: https://github.com/element-hq/synapse/blob/HEAD/tests/storage/databases/main/test_events_worker.py#L345 |
I just ran your flake @Mic92 and it works for me too. Thank you for adding in this patch. |
@naturallaw777 I have beat my head against this for hours today trying to manually modify my own flake with overrides/overlays incorporating the changes from @Mic92 and have had absolutely no effect whatsoever. Where is the flake you are referencing? |
It has now been merged to the |
I've got a "twisted" patch that actually fixes the test: --- a/tests/storage/databases/main/test_events_worker.py
+++ b/tests/storage/databases/main/test_events_worker.py
@@ -372,7 +372,7 @@ class DatabaseOutageTestCase(unittest.HomeserverTestCase):
)
self.event_ids: List[str] = []
- for idx in range(1, 21): # Stream ordering starts at 1.
+ for idx in range(1, 11): # Stream ordering starts at 1.
event_json = {
"type": f"test {idx}",
"room_id": self.room_id, My assumption was that the failed tasks created a backtrace that's too long to be serialized with the twisted internal amp protocol, thus reducing the number of events, thus tasks, brings it back into the limit. |
@naturallaw777 Thank you. |
Closes #369303
Things done
nix.conf
? (See Nix manual)sandbox = relaxed
sandbox = true
nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD"
. Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/
)Add a 👍 reaction to pull requests you find important.