[🐛 Bug]: Memory leak by creating multiple HttpClient-*-SelectorManager threads when using jdk11 http client in loop with selenium grid

[pogrebnyaksa]

What happened?

I experience memory leak in my app - selenium spawns multiple threads.

How can we reproduce the issue?

<properties>
        <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
        <selenium.version>4.6.0</selenium.version>
        <lombok.version>1.18.24</lombok.version>
        <spring-web.version>5.3.23</spring-web.version>
        <jackson-databind.version>2.14.0</jackson-databind.version>
    </properties>

    <dependencies>
        <dependency>
            <groupId>org.seleniumhq.selenium</groupId>
            <artifactId>selenium-chrome-driver</artifactId>
            <version>${selenium.version}</version>
        </dependency>
        <dependency>
            <groupId>org.seleniumhq.selenium</groupId>
            <artifactId>selenium-devtools-v107</artifactId>
            <version>${selenium.version}</version>
        </dependency>
        <dependency>
            <groupId>org.seleniumhq.selenium</groupId>
            <artifactId>selenium-support</artifactId>
            <version>${selenium.version}</version>
        </dependency>
        <dependency>
            <groupId>org.seleniumhq.selenium</groupId>
            <artifactId>selenium-http-jdk-client</artifactId>
            <version>${selenium.version}</version>
        </dependency>

        <!-- https://mvnrepository.com/artifact/org.springframework/spring-web -->
        <dependency>
            <groupId>org.springframework</groupId>
            <artifactId>spring-web</artifactId>
            <version>${spring-web.version}</version>
        </dependency>
        <dependency>
            <groupId>com.fasterxml.jackson.core</groupId>
            <artifactId>jackson-databind</artifactId>
            <version>${jackson-databind.version}</version>
        </dependency>



        <dependency>
            <groupId>org.projectlombok</groupId>
            <artifactId>lombok</artifactId>
            <version>${lombok.version}</version>
            <scope>provided</scope>
        </dependency>
    </dependencies>



    
    @SneakyThrows
    public static void main(String[] args) {
        System.setProperty("webdriver.http.factory", "jdk-http-client");

        while (true) {
            WebDriver driver = null;
            try {
                ChromeOptions chromeOptions = new ChromeOptions();
                chromeOptions.setHeadless(false);
                chromeOptions.addArguments(
                        "--ignore-certificate-errors",
                        "--auto-open-devtools-for-tabs"
                );
                driver = RemoteWebDriver.builder()
                        .augmentUsing(new Augmenter())
                        .oneOf(chromeOptions)
                        .address(GRID_URL)
                        .build();
                driver.get("http://example.com/");
            } catch (Exception e) {
                log.info("Exception ", e);
            } finally{
                if (driver != null) {
                    driver.quit();
                }
            }

            TimeUnit.SECONDS.sleep(5);
        }
    }

If you make thread dumps periodically - you can see that number of HttpClient-*-SelectorManager threads is increasing rather quickly (in 5 minutes I see 90 threads).

Relevant log output

no log output required

Operating System

Debian 9, Ubuntu 20.04.3

Selenium version

Java 4.6.0

What are the browser(s) and version(s) where you see this issue?

chrome 107.0

What are the browser driver(s) and version(s) where you see this issue?

ChromeDriver 107.0.5304.62

Are you using Selenium Grid?

Yes, 4.6.0

[github-actions]

@pogrebnyaksa, thank you for creating this issue. We will troubleshoot it as soon as we can.

---

Info for maintainers

Triage this issue by using labels.

If information is missing, add a helpful comment and then I-issue-template label.

If the issue is a question, add the I-question label.

If the issue is valid but there is no time to troubleshoot it, consider adding the help wanted label.

If the issue requires changes or fixes from an external project (e.g., ChromeDriver, GeckoDriver, MSEdgeDriver, W3C), add the applicable G-* label, and it will provide the correct link and auto-close the issue.

After troubleshooting the issue, please add the R-awaiting answer label.

Thank you!

[github-actions]

Hi, @pogrebnyaksa.
Please follow the issue template, we need more information to reproduce the issue.

Either a complete code snippet and URL/HTML (if more than one file is needed, provide a GitHub repo and instructions to run the code), the specific versions used, or a more detailed description to help us understand the issue.

Note: If you cannot share your code and URL/HTML, any complete code snippet and URL/HTML that reproduces the issue is good enough.

Reply to this issue when all information is provided, thank you.

[diemol]

Also, some evidence and graphs showing the leak would be very helpful to troubleshoot this.

[joerg1985]

@pogrebnyaksa
could you replace the TimeUnit.SECONDS.sleep(5); with a System.gc(); call?
The selector threads should stop when GC is triggered.

In the real world the GC should automatically run after some time.
The loop does not generate enought trash to make it run automatically ;)

[pogrebnyaksa]

Also, some evidence and graphs showing the leak would be very helpful to troubleshoot this.

Please, look into repo https://github.com/pogrebnyaksa/selenium-repo.
There you can find sources, threaddumps and screenshoot with threads increasing.

[pogrebnyaksa]

@pogrebnyaksa could you replace the TimeUnit.SECONDS.sleep(5); with a System.gc(); call? The selector threads should stop when GC is triggered.

In the real world the GC should automatically run after some time. The loop does not generate enought trash to make it run automatically ;)

You are right, manual call GC cleans all hanging threads from jdk11 http client, but it is not normal to spawn such huge amount of threads until GC clean them all. I have reached 700 of threads and that's not the end - it pollutes thread dumps and makes monitoring of service more complicated. Also I think it is a good practice to shutdown threads when they are not needed any more

[joerg1985]

@pogrebnyaksa I agree a graceful shutdown is good practice and Selenium has a mechanism to clear unused clients.
But there is no way to stop a single java HttpClient, this is usually done automatically.

I created a PR to call System.gc() when a client is not in use any more.
This should help in your case, but i am not sure if they will merge it.

[titusfortner]

Is this a Java issue or a Rust issue? The Java code for Selenium Manager does not do anything with http commands directly, that's all in the selenium-manager executable.