Fix graceful shutdown race condition #362
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Currently, when `Topology` is shutting down, it calls `Terminator.trap_exit/1` to allow it to handle `terminate/2` callback and shutdown producers gracefully. However, `Terminator.trap_exit/1` is using `GenServer.cast/2` which makes it fully asynchronous, making it possible for `Topology`'s terminate callback to return before `Terminator` starts trapping exits, allowing `Terminator` to be shut down without putting producers into "draining" state. In my system when we're running ~50-100 Broadway instances we're seeing ~50% of them shutting down properly and the rest of them get stuck with producers going full blast unaware of the shutdown. This change simply changes `Terminator.trap_exit/1` to use `GenServer.call/2` instead of `GenServer.cast/2` to make it fully synchronous, which fixed the issue. Since I don't know Broadway internals at all, please let me know if there is a better way to fix this.
Member
|
💚 💙 💜 💛 ❤️ |
Pull Request Test Coverage Report for Build 0160e33666b8ae8fdfb245bb0ac3b8d5d0a49817-PR-362Details
💛 - Coveralls |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Currently, when
Topologyis shutting down, it callsTerminator.trap_exit/1to allow it to handleterminate/2callback and shutdown producers gracefully. However,Terminator.trap_exit/1is usingGenServer.cast/2which makes it fully asynchronous, making it possible forTopology's terminate callback to return beforeTerminatorstarts trapping exits, allowingTerminatorto be shut down without putting producers into "draining" state.In my system when we're running ~50-100 Broadway instances we're seeing ~50% of them shutting down properly and the rest of them get stuck with producers going full blast unaware of the shutdown.
This change simply changes
Terminator.trap_exit/1to useGenServer.call/2instead ofGenServer.cast/2to make it fully synchronous, which fixed the issue. Since I don't know Broadway internals at all, please let me know if there is a better way to fix this.