-
Notifications
You must be signed in to change notification settings - Fork 16
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
Scheduler asks more than 1 worker to upload its result #428
Comments
IMO it is the root cause of iExecBlockchainComputing/iexec-result-proxy#21 (the result proxy might be smarter though). |
I agree. |
In a bunch of 100 local runs, 5 runs have failed: 2 because of a timeout and 3 because of the scheduler asking both workers to upload. During an usual run, the method |
In our 3 failing tests we can see that the workers haven't been able to upload their result onto IPFS because of an internal server error:
This issue seems to arise because of a non-unique
That's the issue explained in iExecBlockchainComputing/iexec-result-proxy#21. As pointed out there, both issues are probably linked. |
Cause:After having added some logs to ``, we can see that we have a clear race-condition.
Logs:
What happens there:
How to fix it?The first patch to do would be to revert the tests order. Then, the race condition would be reverted so that a worker can be counted for both tests. That's no more an issue as we would only use the |
Fixed in #454. |
Sometimes, the scheduler asks more than 1 worker to upload their results. This could be seen in the following logs:
Note that the last line should not happen as a task can't be transitioned from a status to the same status.
The text was updated successfully, but these errors were encountered: