rclcpp
rclcpp copied to clipboard
ros2
Add support for timers on reset callback
Based on https://github.com/ros2/rcl/pull/995, this PR adds support for the "on reset" callback for timers.
The issue which originated this need, is to wake up an events based executor when a timer is reset (which could also be caused by a time jump).
@alsora
The main issue here is the Windows CI that is failing. However, I see that the same failure is also present in a completely unrelated PR https://github.com/ros2/rclcpp/pull/1982
When I looked at the Win CI output it seemed that something related to the parameters interface in tf2_ros was failing. This seems to be kind of unrelated. Can we maybe try to tun the CI against some documentation PR to see if this is an issue with the current master of tf2?
The main issue here is the Windows CI that is failing. However, I see that the same failure is also present in a completely unrelated PR #1982
I think that was fixed in https://github.com/ros2/rclcpp/pull/2018, the error is not happening in nightlies anymore. It would be nice to rebase and rerun to confirm.
There seems to be another Windows build issue:
19:58:26 Running handlers complete
19:58:26 [2022-11-28T18:58:25+00:00] ERROR: Exception handlers complete
19:58:26 Cinc Client failed. 65 resources updated in 12 minutes 01 seconds
19:58:26 [2022-11-28T18:58:26+00:00] FATAL: LoadError: cannot load such file -- inspec/utils/deprecation/global_method
19:58:26 The command 'cmd /S /C c:\cinc-project\cinc\bin\cinc-solo.bat -c C:\TEMP\solo.rb -Eros2ci -j C:\TEMP\install_ros2_%ROS_DISTRO%.json' returned a non-zero code: 1
Seems like a random CI network issue, running it again:
The Windows CI still seems to have Network issues:
Relevant Jenkins Traceback
17:31:39 FATAL: command execution failed
17:33:44 java.nio.channels.ClosedChannelException
17:33:44 at org.jenkinsci.remoting.protocol.NetworkLayer.onRecvClosed(NetworkLayer.java:154)
17:33:44 at org.jenkinsci.remoting.protocol.impl.NIONetworkLayer.ready(NIONetworkLayer.java:179)
17:33:44 at org.jenkinsci.remoting.protocol.IOHub$OnReady.run(IOHub.java:790)
17:33:44 at jenkins.util.ContextResettingExecutorService$1.run(ContextResettingExecutorService.java:28)
17:33:44 at jenkins.security.ImpersonatingExecutorService$1.run(ImpersonatingExecutorService.java:68)
17:33:44 at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
17:33:44 at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
17:33:44 at java.lang.Thread.run(Thread.java:748)
17:33:44 Caused: java.io.IOException: Backing channel 'JNLP4-connect connection from ip-10-0-1-137.ec2.internal/10.0.1.137:49694' is disconnected.
17:33:44 at hudson.remoting.RemoteInvocationHandler.channelOrFail(RemoteInvocationHandler.java:216)
17:33:44 at hudson.remoting.RemoteInvocationHandler.invoke(RemoteInvocationHandler.java:286)
17:33:44 at com.sun.proxy.$Proxy76.isAlive(Unknown Source)
17:33:44 at hudson.Launcher$RemoteLauncher$ProcImpl.isAlive(Launcher.java:1213)
17:33:44 at hudson.Launcher$RemoteLauncher$ProcImpl.join(Launcher.java:1205)
17:33:44 at hudson.tasks.CommandInterpreter.join(CommandInterpreter.java:194)
17:33:44 at hudson.tasks.CommandInterpreter.perform(CommandInterpreter.java:144)
17:33:44 at hudson.tasks.CommandInterpreter.perform(CommandInterpreter.java:91)
17:33:44 at hudson.tasks.BuildStepMonitor$1.perform(BuildStepMonitor.java:20)
17:33:44 at hudson.model.AbstractBuild$AbstractBuildExecution.perform(AbstractBuild.java:806)
17:33:44 at hudson.model.Build$BuildExecution.build(Build.java:198)
17:33:44 at hudson.model.Build$BuildExecution.doRun(Build.java:163)
17:33:44 at hudson.model.AbstractBuild$AbstractBuildExecution.run(AbstractBuild.java:514)
17:33:44 at hudson.model.Run.execute(Run.java:1888)
17:33:44 at hudson.model.FreeStyleBuild.run(FreeStyleBuild.java:43)
17:33:44 at hudson.model.ResourceController.execute(ResourceController.java:99)
17:33:44 at hudson.model.Executor.run(Executor.java:432)
17:33:44 FATAL: Unable to delete script file C:\Windows\TEMP\jenkins469429339840863473.bat
17:33:44 java.nio.channels.ClosedChannelException
17:33:44 at org.jenkinsci.remoting.protocol.NetworkLayer.onRecvClosed(NetworkLayer.java:154)
17:33:44 at org.jenkinsci.remoting.protocol.impl.NIONetworkLayer.ready(NIONetworkLayer.java:179)
17:33:44 at org.jenkinsci.remoting.protocol.IOHub$OnReady.run(IOHub.java:790)
17:33:44 at jenkins.util.ContextResettingExecutorService$1.run(ContextResettingExecutorService.java:28)
17:33:44 at jenkins.security.ImpersonatingExecutorService$1.run(ImpersonatingExecutorService.java:68)
17:33:44 at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
17:33:44 at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
17:33:44 at java.lang.Thread.run(Thread.java:748)
17:33:44 Caused: hudson.remoting.ChannelClosedException: Channel "hudson.remoting.Channel@30159803:JNLP4-connect connection from ip-10-0-1-137.ec2.internal/10.0.1.137:49694": Remote call on JNLP4-connect connection from ip-10-0-1-137.ec2.internal/10.0.1.137:49694 failed. The channel is closing down or has closed down
17:33:44 at hudson.remoting.Channel.call(Channel.java:994)
17:33:44 at hudson.FilePath.act(FilePath.java:1165)
17:33:44 at hudson.FilePath.act(FilePath.java:1154)
17:33:44 at hudson.FilePath.delete(FilePath.java:1681)
17:33:44 at hudson.tasks.CommandInterpreter.perform(CommandInterpreter.java:162)
17:33:44 at hudson.tasks.CommandInterpreter.perform(CommandInterpreter.java:91)
17:33:44 at hudson.tasks.BuildStepMonitor$1.perform(BuildStepMonitor.java:20)
17:33:44 at hudson.model.AbstractBuild$AbstractBuildExecution.perform(AbstractBuild.java:806)
17:33:44 at hudson.model.Build$BuildExecution.build(Build.java:198)
17:33:44 at hudson.model.Build$BuildExecution.doRun(Build.java:163)
17:33:44 at hudson.model.AbstractBuild$AbstractBuildExecution.run(AbstractBuild.java:514)
17:33:44 at hudson.model.Run.execute(Run.java:1888)
17:33:44 at hudson.model.FreeStyleBuild.run(FreeStyleBuild.java:43)
17:33:44 at hudson.model.ResourceController.execute(ResourceController.java:99)
17:33:44 at hudson.model.Executor.run(Executor.java:432)
17:33:44 Build step 'Execute Windows batch command' marked build as failure
17:33:44 ERROR: Step ‘Record compiler warnings and static analysis results’ aborted due to exception:
17:33:44 java.io.IOException: No workspace found for ci_windows #18344
17:33:44 at io.jenkins.plugins.analysis.core.steps.IssuesRecorder.perform(IssuesRecorder.java:577)
17:33:44 at hudson.tasks.BuildStepMonitor$1.perform(BuildStepMonitor.java:20)
17:33:44 at hudson.model.AbstractBuild$AbstractBuildExecution.perform(AbstractBuild.java:806)
17:33:44 at hudson.model.AbstractBuild$AbstractBuildExecution.performAllBuildSteps(AbstractBuild.java:755)
17:33:44 at hudson.model.Build$BuildExecution.post2(Build.java:178)
17:33:44 at hudson.model.AbstractBuild$AbstractBuildExecution.post(AbstractBuild.java:699)
17:33:44 at hudson.model.Run.execute(Run.java:1913)
17:33:44 at hudson.model.FreeStyleBuild.run(FreeStyleBuild.java:43)
17:33:44 at hudson.model.ResourceController.execute(ResourceController.java:99)
17:33:44 at hudson.model.Executor.run(Executor.java:432)
17:33:44 ERROR: Step ‘Publish Cobertura Coverage Report’ failed: no workspace for ci_windows #18344
17:33:44 ERROR: Step ‘Publish xUnit test result report’ failed: no workspace for ci_windows #18344
17:33:44 Finished: FAILURE
The Windows CI still seems to have Network issues:
Yes, I will run it again in some hours, after checking with the infra team if there are any windows CI issues.
Thanks for everyone's patience on these infrastructure issues. We use low-bid spot instances in order to keep costs down but these cause us trouble every year around the shopping season. The ROS infra team is currently working to add some on-demand capacity to allow jobs to continue.
In order to run that transition I am unfortunately going to abort but I will requeue it again and link it here once the transition is complete.
Alright. Because is running on a relatively fresh agent, there may still be issues related to other reliability issues caused by the holiday activity but this agent should at least stick around.
The run still fails :/, but it at least failed relatively early with an infrastructure related error.
@fujitatomoya @ivanpauno @hidmic it looks like all comments have been addressed; can you have a final look before we merge this PR and its related rcl PR?
I am good to go with this, but requesting another review from maintainers. CC: @ivanpauno @hidmic @wjwwood
be advised that this must be merged with https://github.com/ros2/rcl/pull/995
Ping; it looks like the rcl PR (https://github.com/ros2/rcl/pull/995) has been approved, can we get an additional review here?