Optionally crash benchmark QPS client worker if it fails to connect to the server target. #36511
+38
−7
Conversation
This file contains 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
apolcyn
reviewed
May 2, 2024
| grpc_core::Crash("Client failed to connect to all channels"); | ||
| } else { | ||
| client->AwaitThreadsCompletion(); | ||
| ctx->TryCancel(); |
There was a problem hiding this comment.
I don't think TryCancel is necessary?
Simply terminating the RPC with the status code and message below seems ideal
There was a problem hiding this comment.
TryCancel ensures that the stream passed in from the QPS JSON driver is closed, so that the driver will terminate if the client fails to connect. If the TryCancel isn't there, the driver will hang forever waiting for the client to write the initial status.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
By default, the QPS worker will crash if it cannot connect to the server target while operating as a client.
When explicitly set to false, the
die_on_connection_failureflag (true by default) will keep the QPS worker process running if it fails to connect to the server target. Instead of dying, the QPS worker will finish the client threads, cancel the context to end the benchmark run and terminate the QPS JSON driver, and revert to the idle state of waiting for aRunClientorRunServerinvocation.This is useful in benchmark setups where runs are invoked continuously on long-lived QPS worker processes, reducing spurious worker deaths and the need for manual restarts due to transient connectivity failures.
CC: @apolcyn