Memory Leak Detected in Netty 4.1.100.Final about reactor-netty

[violetagg]

Spring Boot Version: 5.3.3

Please clarify whether this is Spring Framework or Spring Boot version

Please also provide Reactor Netty version

[Ruchika2203]

@violetagg
spring boot 2.7.17
reactor netty 1.0.38

[violetagg]

@Ruchika2203 Please provide code samples and/or reproducible example.

[Ruchika2203]

This happens on "our" code base and we do not have a reproducible example for this .
This was found with paranoid settings and is there an alternate way to get any more details.
Both applications are spring boot and right side application connects to database.

[violetagg]

@Ruchika2203 Check this FAQ, may be it will help. https://projectreactor.io/docs/netty/release/reference/index.html#faq.memory-leaks

The stack above shows that the incoming data is kept in Reactor Netty, ready to be consumed by the application, I need to see your code in order to understand why your application does not consume the incoming data.

[Ruchika2203]

Hi @violetagg ,

Here is one of the code snippet which is calling the right side app which is actually connecting to the database.
Also we see the memory leak in that datastore app only.

code.docx

[violetagg]

@Ruchika2203 I checked the code but it does not seem to have issues (although there are structures that I don't know). Try to enable Reactor Netty/Spring logs and to trace the request/response what happens. You can correlate with the connection id (https://projectreactor.io/docs/netty/release/reference/index.html#faq.logging-prefix)
In your case the connection id is 613b1879-146 from the log Hint: [613b1879-146, L:/192.168.185.101:41270 - Buffered ByteBufHolder in the inbound buffer queue

[Ruchika2203]

@violetagg Can we schedule a quick zoom connect to discuss on this issue.

[violetagg]

@Ruchika2203 It is our policy to collaborate asynchronously through GitHub issues when there is a problem that needs addressed. While we try to help anyone who raises an issue, it is often difficult to properly troubleshoot without an example. For this reason, our New Issue template asks for code to reproduce the issue and our contribution guidelines for Reactor do the same.
In order to progress on the issue we ask that a minimal example is provided and ideally checked into a new GitHub repository. If that is not possible then we will need to close the issue.

[Ruchika2203]

hi @violetagg
One more info
we are using exchange() in our code and saw an article on stack overflow where its ,mentioned that the exchange() has been deprecated due to potential memory leak.

Stacktrace link: https://stackoverflow.com/questions/58410352/spring-boot-webclients-retrieve-vs-exchange

Can it be the probable cause for netty OOM.

Code snippet :

Mono<ResponseEntity> responseEntity = webClient.patch()
.uri(dataStoreUrlWithTTLMutate, version, sessionBucketName, sessionId, timeToLive).bodyValue(data)
.exchange().flatMap(response -> response.toEntity(JsonNode.class));
PhoenixCustomPerfLogger.logExecutionTime(stopWatch, StringUtils.join(dataStoreUrlWithTTLMutate, "/", version,
"/", sessionBucketName, "/", sessionId, "/", timeToLive), HttpMethod.PATCH.name());

[violetagg]

@Ruchika2203 Yes you will experience memory leaks. You handle only the success path. Read more here:
https://docs.spring.io/spring-framework/docs/current/javadoc-api/org/springframework/web/reactive/function/client/WebClient.RequestHeadersSpec.html#exchange()

[github-actions]

If you would like us to look at this issue, please provide the requested information. If the information is not provided within the next 7 days this issue will be closed.

[github-actions]

Closing due to lack of requested feedback. If you would like us to look at this issue, please provide the requested information and we will re-open.