OK, with the above changes, the job is now crapping out faster using the 180s timeout. Oops, wrong units. I'll trigger another one with a longer timeout - 7200s should be better.
https://ci.eclipse.org/webtools/job/webtools-webservices_R3_10/ >=191 On Thu, May 31, 2018 at 9:58 AM, Nick Boldt <nbo...@redhat.com> wrote: > Yes, GEF legacy is being used. I asked about this recently and Carl > explained why [1]. > > [1] https://bugs.eclipse.org/bugs/show_bug.cgi?id=533942#c5 > > Side issue aside, old logs are potentially available if you see them in > Jenkins. If you can't find them under artifacts for a given build, then > they're gone. Since we haven't previously archived the .log files, they're > gone. But for future builds we can fix that. > > Setting a longer timeout? Done: I've added -Dsurefire.exitTimeout=180 > -Dsurefire.timeout=180 to the maven invocations, and upped the overall > build timeout from 150 to 300 mins. > > I've also increased the number of sets of artifacts to store in the job > from 1 to 4, and the retention rules from 7/5/2/1 to 10/10/10/4. And... > I've added **/.log to the list of files to archive for each build so those > will be retained. This might have to change again later if we chew up too > much disk, but for now it'll help with troubleshooting. > > I've also given you (using your email address above) access to configure > the job once you're logged in to ci.eclipse.org. If that's not the email > address associated with your eclipse.org userid, let me know and I can > add your other identity. > > Job is running with the above config changes: > > https://ci.eclipse.org/webtools/job/webtools-webservices_R3_10/ >=189 > > Anything else you think would help here? > > Thanks for looking into this problem! > > Nick > > On Tue, May 29, 2018 at 1:11 PM, Shane Clarke <shane_cla...@hotmail.com> > wrote: > >> Hi Nick, >> >> I had a look at these. >> >> I'm not able to reproduce the issue locally and i'm running with a build >> with the dependencies listed in http://git.eclipse.org/c/we >> btools/webtools.releng.aggregator.git/tree/wtp-parent/pom.xml#n45 >> >> Just to note the gef-repo.url and the gef-legacy-repo.url look like >> they're both pulling in GEF 3.1.0 ? (unrelated to this issue) >> >> The log file https://ci.eclipse.org/webtools/job/webtools-webservice >> s_R3_10/ws/jaxws/tests/org.eclipse.jst.ws.cxf.tests/ >> target/work/data/.metadata/.log/*view*/ is full of these type of errors: >> >> !MESSAGE No IModelProvider exists for project >> P/TestProject2_1527516764789 of version: EJB Module 3.0 >> !STACK 0 >> java.lang.NullPointerException: No IModelProvider exists for project >> P/TestProject2_1527516764789 of version: EJB Module 3.0 >> at org.eclipse.jst.j2ee.model.ModelProviderManager.getModelProv >> ider(ModelProviderManager.java:101) >> at org.eclipse.jst.j2ee.model.ModelProviderManager.getModelProv >> ider(ModelProviderManager.java:281) >> at org.eclipse.jst.jee.ui.internal.navigator.JEE5ContentProvide >> r.getCachedModelProvider(JEE5ContentProvider.java:76) >> at org.eclipse.jst.jee.ui.internal.navigator.Ejb3ContentProvide >> r.getNewContentProviderInstance(Ejb3ContentProvider.java:212) >> at org.eclipse.jst.jee.ui.internal.navigator.LoadingJeeDDJob. >> run(LoadingJeeDDJob.java:50) >> at org.eclipse.core.internal.jobs.Worker.run(Worker.java:60) >> >> !ENTRY org.eclipse.core.jobs 4 2 2018-05-28 10:14:37.987 >> !MESSAGE An internal error occurred during: "Loading descriptor for >> TestProject2_1527516764789.". >> !STACK 0 >> java.lang.NullPointerException >> at org.eclipse.jst.jee.ui.internal.navigator.JEE5ContentProvide >> r.getCachedModelProvider(JEE5ContentProvider.java:77) >> at org.eclipse.jst.jee.ui.internal.navigator.Ejb3ContentProvide >> r.getNewContentProviderInstance(Ejb3ContentProvider.java:212) >> at org.eclipse.jst.jee.ui.internal.navigator.LoadingJeeDDJob. >> run(LoadingJeeDDJob.java:50) >> at org.eclipse.core.internal.jobs.Worker.run(Worker.java:60) >> >> What looks to be happening is the tests themselves aren't failing. The >> errors are a result of the Project Explorer view (Common Navigator View) >> being open in the Java EE perspective and it's refreshing as a result of >> the tests creating EJB projects. The errors are introducing a lag in the >> test run and the number of them is adding up to push the overall test >> runtime over the 30 minute timeout. >> >> I'm not sure what's caused this to start happening now. These tests are >> part of a SAP contribution to the project around 9 years ago and they >> haven't really changed i think since that. >> >> Are the test log files for the earlier successful builds e.g. #183 >> maintained? And is it possible to extend the timeout limit for these tests >> for a test run to see if they pass? >> >> Thanks, >> Shane >> ------------------------------ >> *From:* Nick Boldt <nbo...@redhat.com> >> *Sent:* Monday 28 May 2018 20:42 >> *To:* Shane Clarke >> *Cc:* General discussion of project-wide or architectural issues.; Keith >> Chong >> *Subject:* Re: [webtools.webservices] Build issue - stuck tests in >> jst.ws.cxf >> >> No dep changes since May 9, in terms of URLs used. >> >> http://git.eclipse.org/c/webtools/webtools.releng.aggregator >> .git/log/wtp-parent/pom.xml >> >> However, since some URLs are a "latest build" style, you can see that the >> deps can change for Platform, DTP, JGit/EGit, & EMF. >> >> http://git.eclipse.org/c/webtools/webtools.releng.aggregator >> .git/tree/wtp-parent/pom.xml#n45 >> >> The rest of the requirements/depdencies are static URLs with specific >> releases in them, including the old Graphiti 0.15.0.M6 release and some >> emf.* projects from 2017/06. >> >> ---- >> >> Here's a link to the .log files generated by Eclipse while the build runs >> tests: >> >> https://ci.eclipse.org/webtools/job/webtools-webservices_R3_10/ws/**/.log >> >> And here's the log from the test that's failing. >> >> https://ci.eclipse.org/webtools/job/webtools-webservices_R3_ >> 10/ws/jaxws/tests/org.eclipse.jst.ws.cxf.tests/target/work/ >> data/.metadata/.log/*view*/ >> >> Does that shine any light on the failures? >> >> Nick >> >> >> On Mon, May 28, 2018 at 3:15 PM, Shane Clarke <shane.cla...@outlook.com> >> wrote: >> >> Hi Nick, >> >> There isn't much to go on in those logs. I do see >> >> org.eclipse.jst.ws.cxf.tests: An unexpected error occured while launching >> the test runtime (return code 143). See log for details. >> >> But the only other reference is to the test exceeding the timeout and the >> process being killed. >> >> I think those builds are running the platform I20180527-0800 build? I ran >> the tests with that and couldn't produce any issues. >> >> Were there any dependency changes between builds 183 and 184? >> >> Thanks, >> Shane >> ------------------------------ >> *From:* Nick Boldt <nbo...@redhat.com> >> *Sent:* Monday 28 May 2018 15:59 >> *To:* General discussion of project-wide or architectural issues. >> *Cc:* Shane Clarke; Keith Chong >> *Subject:* [webtools.webservices] Build issue - stuck tests in jst.ws.cxf >> >> For two builds in a row, the WTP Webservices build has been stuck running >> tests, specifically in *org.eclipse.jst.ws.cxf.tests.AllTestsSuite. *This >> test suite times out after 30 mins, suggesting either stuck UI or some >> other blocking process that's waiting for something to happen before >> proceeding. >> >> https://hudson.eclipse.org/webtools/view/webtools_R3_10/job/ >> webtools-webservices_R3_10/185/console >> https://hudson.eclipse.org/webtools/view/webtools_R3_10/job/ >> webtools-webservices_R3_10/184/console >> https://hudson.eclipse.org/webtools/view/webtools_R3_10/job/ >> webtools-webservices_R3_10/184/testReport/ (325 tests did not complete >> as expected) >> >> Shane, Keith, if you're back from PTO, can you have a look? >> >> -- >> >> Nick Boldt >> >> Principal Software Engineer, RHCSA >> >> Productization Lead :: JBoss Tools & Dev Studio >> >> IM: @nickboldt / @nboldt / http://nick.divbyzero.com >> <https://red.ht/sig> >> TRIED. TESTED. TRUSTED. <https://redhat.com/trusted> >> @ @redhatnews <https://twitter.com/redhatnews> Red Hat >> <https://www.facebook.com/RedHatInc> >> <https://www.facebook.com/RedHatInc> >> >> >> “The Only Thing That Is Constant Is Change” - Heraclitus >> >> >> >> >> -- >> >> Nick Boldt >> >> Principal Software Engineer, RHCSA >> >> Productization Lead :: JBoss Tools & Dev Studio >> >> IM: @nickboldt / @nboldt / http://nick.divbyzero.com >> <https://red.ht/sig> >> TRIED. TESTED. TRUSTED. <https://redhat.com/trusted> >> @ @redhatnews <https://twitter.com/redhatnews> Red Hat >> <https://www.facebook.com/RedHatInc> >> <https://www.facebook.com/RedHatInc> >> >> >> “The Only Thing That Is Constant Is Change” - Heraclitus >> > > > > -- > > Nick Boldt > > Principal Software Engineer, RHCSA > > Productization Lead :: JBoss Tools & Dev Studio > > IM: @nickboldt / @nboldt / http://nick.divbyzero.com > <https://red.ht/sig> > TRIED. TESTED. TRUSTED. <https://redhat.com/trusted> > @ @redhatnews <https://twitter.com/redhatnews> Red Hat > <https://www.facebook.com/RedHatInc> > <https://www.facebook.com/RedHatInc> > > > “The Only Thing That Is Constant Is Change” - Heraclitus > -- Nick Boldt Principal Software Engineer, RHCSA Productization Lead :: JBoss Tools & Dev Studio IM: @nickboldt / @nboldt / http://nick.divbyzero.com <https://red.ht/sig> TRIED. TESTED. TRUSTED. <https://redhat.com/trusted> @ @redhatnews <https://twitter.com/redhatnews> Red Hat <https://www.facebook.com/RedHatInc> <https://www.facebook.com/RedHatInc> “The Only Thing That Is Constant Is Change” - Heraclitus
_______________________________________________ wtp-dev mailing list wtp-dev@eclipse.org To change your delivery options, retrieve your password, or unsubscribe from this list, visit https://dev.eclipse.org/mailman/listinfo/wtp-dev