如果添加参数/选项不能解决问题，那么可能是/tmp目录的权限问题。

确保执行Chrome的用户有权限在/tmp文件夹下创建文件夹/files。这就是我案例中的解

我的问题是php-webdriver中的一个bug。我的代码是:

$chromeOptions = new ChromeOptions();
$chromeOptions->addArguments([
    '--headless',
    '--no-sandbox',
    '--disable-gpu',
    '--disable-dev-shm-usage',
    '--no-proxy-server'
]);
$chromeOptions->setExperimentalOption('detach', true);
$capabilities = DesiredCapabilities::chrome();
$capabilities->setCapability(
    ChromeOptions::CAPABILITY,
    $chromeOptions
);

如你所见，一切都符合其他可能的原因。

但实际上这些功能并没有传递给chromedriver。我不得不改变设置chrome选项功能:

$capabilities->setCapability(
    ChromeOptions::CAPABILITY_W3C, // <<< Have to use W3C capabilities with recent versions of Chromedriver.
    $chromeOptions->toArray() // <<<<< bug in php-webdriver 1.9, object not getting serialized automatically like with the deprecated capability ChromeOptions::CAPABILITY
);

我的设置是:

$ chromedriver --version
ChromeDriver 79.0.3945.36 (3582db32b33893869b8c1339e8f4d9ed1816f143-refs/branch-heads/3945@{#614})
$ java -jar selenium-server-standalone-3.141.59.jar --version
Selenium server version: 3.141.59, revision: e82be7d358

做了一个bug报告和一个PR来修复php-webdriver中的这个bug

https://github.com/php-webdriver/php-webdriver/issues/849

我最近也面临着同样的问题，经过一些试验和错误，它也适用于我。

必须在最上面:

options.addArguments("--no-sandbox"); //has to be the very first option

BaseSeleniumTests.java

public abstract class BaseSeleniumTests {

    private static final String CHROMEDRIVER_EXE = "chromedriver.exe";
    private static final String IEDRIVER_EXE = "IEDriverServer.exe";
    private static final String FFDRIVER_EXE = "geckodriver.exe";
    protected WebDriver driver;

    @Before
    public void setUp() {
        loadChromeDriver();
    }

    @After
    public void tearDown() {
        if (driver != null) {
            driver.close();
            driver.quit();
        }
    }

    private void loadChromeDriver() {
        ClassLoader classLoader = getClass().getClassLoader();
        String filePath = classLoader.getResource(CHROMEDRIVER_EXE).getFile();
        DesiredCapabilities capabilities = DesiredCapabilities.chrome();
        ChromeDriverService service = new ChromeDriverService.Builder()
                .usingDriverExecutable(new File(filePath))
                .build();
        ChromeOptions options = new ChromeOptions();
        options.addArguments("--no-sandbox"); // Bypass OS security model, MUST BE THE VERY FIRST OPTION
        options.addArguments("--headless");
        options.setExperimentalOption("useAutomationExtension", false);
        options.addArguments("start-maximized"); // open Browser in maximized mode
        options.addArguments("disable-infobars"); // disabling infobars
        options.addArguments("--disable-extensions"); // disabling extensions
        options.addArguments("--disable-gpu"); // applicable to windows os only
        options.addArguments("--disable-dev-shm-usage"); // overcome limited resource problems
        options.merge(capabilities);
        this.driver = new ChromeDriver(service, options);
    }

}

GoogleSearchPageTraditionalSeleniumTests.java

@RunWith(SpringRunner.class)
@SpringBootTest
public class GoogleSearchPageTraditionalSeleniumTests extends BaseSeleniumTests {

    @Test
    public void getSearchPage() {
        this.driver.get("https://www.google.com");
        WebElement element = this.driver.findElement(By.name("q"));
        assertNotNull(element);
    }

}

pom.xml

<dependencies>
        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter</artifactId>
        </dependency>
        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-test</artifactId>
            <scope>test</scope>
        </dependency>
        <dependency>
            <groupId>org.seleniumhq.selenium</groupId>
            <artifactId>selenium-java</artifactId>
            <scope>test</scope>
        </dependency>
</dependencies>

我在Ubuntu 20和Python Selenium上遇到了这个问题，我首先单独下载了chromedriver，然后使用sudo apt install chromium-browser，尽管它们是同一个版本，但这种情况一直发生。

我的修复是使用提供的chrome驱动程序，附带的回购包位于

/snap/bin/chromium.chromedriver

driver = webdriver.Chrome(chrome_options=options, executable_path='/snap/bin/chromium.chromedriver')

我也经历过这个问题，提出的解决方案似乎都不起作用。然后我发现问题是我在WSL版本1上运行，似乎chromedriver与windows浏览器一起工作，而不是与aptitude安装的浏览器。

为了使它与WSL版本1和版本2兼容，我发现(并在debian WSLv2和ubuntu WSLv1中进行了测试，它可以工作)，该平台的版本显示的单词Microsoft的第一个大写字母。

所以解决方案是这样的:

    import platform
    from selenium.webdriver.chrome.service import Service as ChromeService
    from seleniumwire import webdriver

    chromedriver = 'chromedriver.exe'
    
    # According to:
    # https://stackoverflow.com/a/71879688/7019069
    # When using WSL v1 the chromedriver.exe of local chrome of windows is used
    # In WSL v2 (updated version) it does work using the installation shown in the README.md)
    # And according to
    # https://github.com/microsoft/WSL/issues/4555
    # It is possible to differentiate the version of WSL by the first uppercase of the platform 
    # version of Microsoft. Therefore only the linux chromedriver is used if that word is matched 
    # in the platform of the driver. 
    if platform.system() == 'Linux' and not re.search(re.escape('Microsoft'), platform.platform()):
        chromedriver = 'chromedriver'

    driver_path = os.path.join(drivers_path, chromedriver)
  service = ChromeService(driver_path)

    driver = webdriver.Chrome(
        service=service,
        options=__get_chrome_options(headless)
    )

我被这个问题困扰了很长一段时间，最后通过添加一个额外的选项来解决它:

options.addArguments(”——crash-dumps-dir = / tmp”)