The ClassLoadingStrategy service is advertised by the ConfigManagerActivator (a bundle activator). I don't see any errors logged for it. Maybe that bundle didn't start for some reason? Or maybe the mdsal-binding-generator-api bundle that exports the ClassLoadingStrategy class re-installed/restarted and the previous Class instance was gone? I do see that bundle outputted twice but not sure if that means it was re-installed:
2017-05-01 10:39:11,968 | INFO | pool-2-thread-1 | FeaturesServiceImpl | 8 - org.apache.karaf.features.core - 4.0.9 | mvn:org.opendaylight.mdsal/mdsal-binding-generator-api/0.10.0-Carbon mvn:org.opendaylight.mdsal/mdsal-binding-generator-api/0.10.0-Carbon 2017-05-01 10:39:12,339 | INFO | pool-2-thread-1 | FeaturesServiceImpl | 8 - org.apache.karaf.features.core - 4.0.9 | mvn:org.opendaylight.mdsal/mdsal-binding-generator-api/0.10.0-Carbon mvn:org.opendaylight.mdsal/mdsal-binding-generator-api/0.10.0-Carbon On Mon, May 1, 2017 at 4:15 PM, Colin Dixon <co...@colindixon.com> wrote: > That seems to be the root of the issues with our SFT hangs. Between SFT > hangs and the BGP test that hangs, those account for 12 of the last 16 > failures in autorelease that we don't have fixes for. > > --Colin > > > On Mon, May 1, 2017 at 4:12 PM, Tom Pantelis <tompante...@gmail.com> > wrote: > >> yeah - saw that too: >> >> 2017-05-01 10:44:13,976 | ERROR | rint Extender: 1 | BlueprintContainerImpl >> | 12 - org.apache.aries.blueprint.core - 1.7.1 | Unable to start >> blueprint container for bundle >> org.opendaylight.controller.sal-binding-broker-impl/1.5.0.Carbon due to >> unresolved dependencies >> [(objectClass=org.opendaylight.mdsal.binding.generator.api.ClassLoadingStrategy)] >> java.util.concurrent.TimeoutException >> at >> org.apache.aries.blueprint.container.BlueprintContainerImpl$1.run(BlueprintContainerImpl.java:371)[12:org.apache.aries.blueprint.core:1.7.1] >> >> >> On Mon, May 1, 2017 at 4:11 PM, Colin Dixon <co...@colindixon.com> wrote: >> >>> It seems like there's something wrong with who sal-binding-broker-impl >>> comes up and doesn't get the ClassLoadingStrategy and then SFT hangs. So >>> far, thats' the first exception that happens in every SFT hang I've looked >>> into. >>> >>> Robert, TomP, any ideas what would cause this. It looks like the AAA >>> issue we were having before, but it's not AAA. >>> >>> --Colin >>> >>> >>> On Mon, May 1, 2017 at 4:03 PM, Colin Dixon <co...@colindixon.com> >>> wrote: >>> >>>> So, Luis told me to go look at the surefire results for some of the SFT >>>> failures and I'm finding things like this: >>>> >>>> = failnever job #4 = >>>> relevant logfile: https://logs.opendaylight.org/ >>>> releng/jenkins092/autorelease-release-failnever-carbon/4/arc >>>> hives/nemo/nemo-features/odl-nemo-openflow-renderer/target/s >>>> urefire-reports/org.opendaylight.odlparent.featuretest.Singl >>>> eFeatureTest-output.txt.gz >>>> Job: https://jenkins.opendaylight.org/releng/view/autorelease/job >>>> /autorelease-release-failnever-carbon/4/ >>>> == Key lines == >>>> 2017-05-01 10:44:22,047 | ERROR | rint Extender: 2 | >>>> BlueprintContainerImpl | 12 - org.apache.aries.blueprint.core >>>> - 1.7.1 | Unable to start blueprint container for bundle >>>> org.opendaylight.aaa.shiro/0.5.0.Carbon due to unresolved dependencies >>>> [(objectClass=org.opendaylight.controller.md.sal.binding.api.DataBroker), >>>> (objectClass=org.opendaylight.aaa.cert.api.ICertificateManager)] >>>> java.util.concurrent.TimeoutException >>>> >>>> = autorelease-carbon job #271 = >>>> relevant logfile: https://logs.opendaylight.org/releng/jenkins092/aut >>>> orelease-release-carbon/285/archives/bgpcep/programming/impl >>>> /target/surefire-reports/ >>>> job: https://jenkins.opendaylight.org/releng/view/autoreleas >>>> e/job/autorelease-release-carbon/271/ >>>> 2017-04-25 16:04:47,169 | ERROR | rint Extender: 2 | >>>> BlueprintContainerImpl | 12 - org.apache.aries.blueprint.core >>>> - 1.7.1 | Unable to start blueprint container for bundle >>>> org.opendaylight.controller.sal-binding-broker-impl/1.5.0.Carbon due >>>> to unresolved dependencies [(objectClass=org.opendaylight >>>> .mdsal.binding.generator.api.ClassLoadingStrategy)] >>>> java.util.concurrent.TimeoutException >>>> >>>> This looks like the AAA issue that we supposedly fixed. However it's >>>> still cropping up in runs as recently as today. >>>> >>>> --Colin >>>> >>>> >>>> >>>> >>>> On Mon, May 1, 2017 at 12:45 PM, Colin Dixon <co...@colindixon.com> >>>> wrote: >>>> >>>>> As a random idea, to counteract these failures, would it be a truly >>>>> terrible idea to add all the SingleFeaturesTests to the excludes list for >>>>> autorelease at least for now? >>>>> >>>>> --Colin >>>>> >>>>> >>>>> On Mon, May 1, 2017 at 11:18 AM, Colin Dixon <co...@colindixon.com> >>>>> wrote: >>>>> >>>>>> add: >>>>>> * SFT for OpenDaylight :: NEMO :: OpenFlow Renderer in failnever job >>>>>> #4 >>>>>> >>>>>> --Colin >>>>>> >>>>>> >>>>>> On Mon, May 1, 2017 at 9:58 AM, Colin Dixon <co...@colindixon.com> >>>>>> wrote: >>>>>> >>>>>>> As I'm tracking autorelease failures here: >>>>>>> https://docs.google.com/spreadsheets/d/1VcB12FBiFV4GAEHZSspH >>>>>>> BNxKI_9XugJp-6Qbbw20Omk/edit#gid=1455372448 >>>>>>> >>>>>>> I'm noticing a reasonable number of hangs in SFT. Is that to be >>>>>>> expected? Is there something we can do about it? >>>>>>> >>>>>>> Examples >>>>>>> * SFT for integration features-test in autorelease-carbon job #271 >>>>>>> * SFT for netconf SSH server in the failnever job #1 >>>>>>> * SFT for features4-alto in the failnever job #3 >>>>>>> >>>>>>> Thanks, >>>>>>> --Colin >>>>>>> >>>>>>> >>>>>> >>>>> >>>> >>> >> >
_______________________________________________ controller-dev mailing list controller-dev@lists.opendaylight.org https://lists.opendaylight.org/mailman/listinfo/controller-dev