NSubstitute mocking fails when run in parallel but works fine when run manually or with [NotInParallel]

[awitton]

I have a fairly simple NSubstitute mock set up for a unit test that fails when the test is run as part of the full suite of tests. However, when run manually or with [NotInParallel], it runs as expected. As a C# testing n00b, is there something that I can do to fix this problem, or do I get over myself and let the test run [NotInParallel]

The test mock is shown below. I have tried specifying actual test strings but it didn't change the outcome.
As this is supposed to be a 'passing' test, valid string values are provided for both parameters

mockParamChecker.TestStringNotNullOrEmpty(Arg.Any<string>(), Arg.Any<string>()).Returns(true);

For testing efficiency, it is setup and injected as part of a Before(Class) method

        [Before(Class)]
        public static void HelpMasterUnitTestsSetup()
        {
            hmApiKey = "Mock API Key";
            string hmFqdn = "acme.com.au";
            string hmIpAddress = "192.168.1.1";
            int hmUserPkId = 999;

            mockParamChecker = Substitute.For<IParamChecker>();

            _sut = new HelpMaster(
                hmApiKey, 
                hmFqdn, 
                hmIpAddress, 
                hmUserPkId, 
                mockParamChecker
            );

            _hmToken = "Mock API Token";
            _sut.SetHMToken(_hmToken);

            Console.WriteLine(@"HelpMasterUnitTestsSetup: Proof that we ran the Before(Class) Setup method");
        }

The test itself is

        [Test]
        [NotInParallel]
        [Arguments("7-Zip", "25.01")]
        public async Task GetHitAppDeployChecklist_ReturnsChecklist_WhenParamsNotNullOrEmpty(string displayName, string version)
        {
            // Arrange
            mockParamChecker.TestStringNotNullOrEmpty(Arg.Any<string>(), Arg.Any<string>()).Returns(true);

            // Act
            string? result = _sut.GetHitAppDeployChecklist(displayName, version);

            // Assert
            await Assert.That(result).IsNotNull();
            await Assert.That(result).Contains("Application Update Checklist");
        }
        

[thomhurst]

I would remove the static and just set them up for each test.

By having them static, they'll be shared/leaking across tests and race conditions and other problems will occur .

Best to give each test it's own instance so it's not affected by other running tests

[awitton]

Thanks so much for the assist. That has done the trick!