planemo tests timing out and ASGI error #13136
Comments
|
Can you please check if galaxy at the commit before that also exhibit the same issue? #13135 updated starlette and fastapi, and could be the cause for BrokenResourceError. |
|
Did a
|
|
For a5b1452 the changes in |
|
Can you check if #13137 fixes it for you ? I could see the test hang if the metadata is in error ... . If that is true we need to find a way to error out quickly. |
If no check or skip is given, self.check and self.skip need to be initialised to None. Otherwise validation with `value.missing_meta(check=self.check, skip=self.skip)` is called with `value.missing_meta(check=[""], skip=[""])`. This led to the stalling observed in galaxyproject#13136 Explanation: each data input parameter automatically gets a MetadataValidator. Since this validator was dysfunctional the job was scheduled anyway, but did not execute.
|
@davelopez do we still need to downgrade fastAPI ? I see a lot of |
|
Yeah, I also thought it was fixed but I guess it is still an issue tiangolo/fastapi#4041 so probably we should downgrade again. |
|
Is it a real issue of just polluting the logs? |
|
That's a good question, some people claim it is fixed and some are silencing it but I'm afraid I don't know if that's a good idea or what are the consequences. |
|
Yeah, all of these are kind of "logging issues". It happens when the client goes away and we write to a closed socket. You can see that the fix in https://github.com/encode/starlette/pull/1262/files is to pick up the right exception from the coroutine, I suppose that something similar should be done for the WSGI middleware we're using (which doesn't seem to be based on BaseHTTPMiddleware that got fixed). A good start might be to write a small client test |
|
It might not be the closed socket issue .... import time
import uvicorn
from fastapi import FastAPI
from fastapi.middleware.wsgi import WSGIMiddleware
def wsgi_application(env, start_response):
print('start')
time.sleep(0.3)
start_response("200 OK", {})
print('done')
return [b'bla']
app = FastAPI()
app.mount('/', WSGIMiddleware(wsgi_application))
if __name__ == "__main__":
uvicorn.run(app="timeoutapp:app", port=8001)And then |
Describe the bug
When running
planemo teston a few (two) tools of the IUC repo I see the following error after Galaxy started and the tool test execution stalls reproducibly.Tool tests stall at a line similar to (only ids differ):
Galaxy Version and/or server at which you observed the bug
Galaxy Version: dev
Commit: 060ab1a
To Reproduce
Steps to reproduce the behavior:
Run either
planemo test --biocontainers --no_dependency_resolution --no_conda_auto_init --galaxy_python_version 3.7 --galaxy_source https://github.com/galaxyproject/galaxy --galaxy_branch dev tools/macs2/macs2_callpeak.xmlorplanemo test --biocontainers --no_dependency_resolution --no_conda_auto_init --galaxy_python_version 3.7 --galaxy_source https://github.com/galaxyproject/galaxy --galaxy_branch dev tools/pygenometracks/pyGenomeTracks.xmlmacs2 stalls at the first test and pygenoetracs at test 17
Expected behavior
Screenshots
Additional context
Might be the cause of the tests timing out in the tests of #13039 here galaxyproject/tools-iuc#4318
The text was updated successfully, but these errors were encountered: