Skip to content
This repository has been archived by the owner on Oct 31, 2023. It is now read-only.

fix race conditions caused by LoopingCallService running in separate threads #3660

Closed
etam opened this issue Dec 11, 2018 · 1 comment · Fixed by #3661
Closed

fix race conditions caused by LoopingCallService running in separate threads #3660

etam opened this issue Dec 11, 2018 · 1 comment · Fixed by #3661
Assignees
@etam
Labels
bug in progress

Comments

@etam
Copy link
Contributor

etam commented Dec 11, 2018

Every LoopingCallService in golem/client.py runs in a separate thread, through twisted.internet.threads.deferToThread. But there are no measures preventing race conditions.

This caused problems before: #2849
This causes problems now: #3245

There is an issue to make easy to optionally to run LoopingCallService in main thread: #2851

Possible solution: run LoopingCallService by default in main thread and optionally in a separate thread. This needs little changes in code, but requires extensive performance testing.
@etam etam self-assigned this Dec 11, 2018
This was referenced Dec 11, 2018
Closed
Merged
@ghost ghost added the in progress label Dec 11, 2018
@etam etam added the _current label Dec 12, 2018
@etam etam mentioned this issue Feb 7, 2019
Closed
@etam etam mentioned this issue Feb 3, 2020
Closed
@etam
Copy link
Contributor Author

etam commented Feb 3, 2020

This is serious now: https://github.com/golemfactory/golem/issues/5083#issuecomment-581436846

@etam etam closed this as completed in #3661 Feb 7, 2020
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Assignees

@etam etam

Labels
bug in progress
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

LoopingCallService sync by default golemfactory/clay
2 participants
@etam @shadeofblue