Closed Rvh91 closed 3 months ago
Hi @Rvh91 could you please help me with following
Hi @nbhoski, We are using version 2.13.1, on Jenkins Version 2.440.3. Honestly it's a bit difficult to backtrack when the issues started, before I never looked at the jenkins logs of the controller, only the console output of the individual builds, so I never noticed when we started to get the 'writePath / Serialization / CpsFlowExecution' and related errors. We have had the issues with the random failures for a while, but never suspected the Matlab Plugin before (we had an 'unclean' kill of a running job, that in combination with the flowNode issues resulted in us suspecting issues in that direction initially). Because of that, I had manually cleaned out all old builds on both the controller and agent, related to that project.
I'm going through the commits on our develop branch, but it is almost impossible to backtrack, as the majority of our build history is now marked as failed. The last build that is still marked as successful ( 19th of march ) also made use of the "runMATLABBuild" command.
I have forked our repo, and started with an empty jenkinsfile to troubleshoot. The issues started popping up there when I started to use the "runMATLABBuild" command. After changing those lines to the aforementioned runMATLABCommand(command: 'buildtool testReport')
instead of runMATLABBuild(tasks: 'testReport')
the issues went away. Reverting back to the runMATLABBuild also results in those errors returning.
Hope this helps.
_I'm going through the commits on our develop branch, but it is almost impossible to backtrack, as the majority of our build history is now marked as failed. The last build that is still marked as successful ( 19th of march ) also made use of the "runMATLABBuild" command
For above on 19 the march when build was successful were you still using the same plugin version?
I have forked our repo, and started with an empty jenkinsfile to troubleshoot. The issues started popping up there when I started to use the "runMATLABBuild" command. After changing those lines to the aforementioned runMATLABCommand(command: 'buildtool testReport') instead of runMATLABBuild(tasks: 'testReport') the issues went away. Reverting back to the runMATLABBuild also results in those errors returning.
Can you please try above by using plugin version prior to 2.13.0 and see if this is reproducible?
@nbhoski, I don't know to be honest, at the time I did not have admin privileges on our Jenkins controller. Is there any way for me to find our when plugins have been updated? I will rollback the plugin to 2.12.0 and try to reproduce the issue.
@nbhoski, I don't know to be honest, at the time I did not have admin privileges on our Jenkins controller. Is there any way for me to find our when plugins have been updated? I will rollback the plugin to 2.12.0 and try to reproduce the issue.
Sure , and also could you share buildfile.m that you are using ? so that I could reproduce it locally with latest plugin version
I have tried it with the 2.12.0, and I dont see the same errors in the jenkins log (FlowNode persistSafe, CpsFlowExecution related errors, and FilePath _getChannelForSerialization), which leads me to believe that this version of the plugin does not have the error.
I dont think I'm allowed to share the buildfile though.. Also not sure if that would be of any use, as its calling many functions that you wont have?
Hi @Rvh91 since ths is warnning and I can reproduce it without any extra efforts. This should not cause the build failure anyways and this set of warning does not get displayed in the console log. I will work on fixing this warning message however the newer version of plugin should not break your build except for these warning messages.
Marked this as bug will work on this
This is fixed
We have had multiple builds break that we could not explain, where in the Jenkins logs we would end up with the errors listed below. Builds that passed all stages succesfull, could be marked as failed a few hours later, which completely broke our build history. We have setup the Jenkins job as a multibranch pipeline.
Changing the step in the jenkinsFile from
to
seems to have resolved the issue for us.
In the console output:
But it can also happen on a successful pipeline:
NOTE: that the snipped above is taken from a different build, whereas the rest of the errors listed here are all from the same build, hence the different flownode numbers
In the Jenkins logs (truncated with the dots to fit the allowed message size, full logs can be found as attached file) fullJenkinsLog.txt :