Hi Mark, > On Jun 5, 2023, at 17:14, Mark Thomas <ma...@apache.org> wrote: > > On 05/06/2023 03:41, Han Li wrote: >> Mark, >> I encounter one error, as follow: >> 20 Testcase: testPerformance took 11.807 sec >> 21 FAILED >> 22 5666 >> 23 junit.framework.AssertionFailedError: 5666 >> 24 at >> org.apache.catalina.mapper.TestMapperPerformance.testPerformance(TestMapperPerformance.java:49) >> 25 at >> org.apache.catalina.mapper.TestMapperPerformance.testPerformance(TestMapperPerformance.java:32) >> 26 at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native >> Method) >> 27 at >> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:77) >> 28 at >> java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) >> But I re-tested this entry, and it did not reproduce. >> According to your comment [1], I think that need feedback to you. >> // Takes ~1s on markt's laptop. If this takes more than 5s something >> // probably needs looking at. If this fails repeatedly then we may need >> // to increase this limit. >> ENV: >> OpenJDK 17.0.2 >> macOS 12.3.1 >> NOTE: I have never encountered this problem before, the difference between >> this and before is that I change the value of ‘test.threads' property from 8 >> to 12. ; ) > > Hi Han, > > How many cores do you have available on that machine? If it is less than 12 > you might be seeing the effects of context switching. 12, I think maybe I was doing something else at the time and it affected this test. > > Looking at the date on that test and the comment, I think the laptop being > referenced is a work laptop that has since been replaced. My current personal > desktop and work laptop both run the whole test in under 2s if I run it > single threaded.
> > The ~1s time refers to an individual test and there are 9 on those in the > test overall so running the whole test in <2s indicates that the 5s per > individual test limit should be very generous. > > If the failure only happened once and you can't repeat it I think it is safe > to ignore it for now. If it happens again or we see similar issues with the > CI systems then we might need to look more closely. I have reran the whole test single thread twice, the result all under 1s. Look great. Let’s ignore it. ;) Han > > Mark > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org > <mailto:dev-unsubscr...@tomcat.apache.org> > For additional commands, e-mail: dev-h...@tomcat.apache.org > <mailto:dev-h...@tomcat.apache.org>
