[JENKINS-40424] mkdir failing for folders with no existing parent on windows

When my jenkins pipeline job needs to create a folder on my windows slave (be it the workspace folder, a temporary folder to hold a batch file, whatever) , such as

C:\Jenkins\workspace\pipeline-test2@​tmp\durable-487de3d5

Then it will fail if the parent does not exist

So in the above case, if

C:\Jenkins\workspace\pipeline-test2@​tmp

already exists (e.g. I created it manually), then durable-487de3d5 will be created correctly, and the script will work.

But if the parent does not exist, then I get the exception

```
Failed to mkdirs: C:\Jenkins\workspace\pipeline-test2@​tmp\durable-487de3d5
hudson.FilePath.mkdirs(FilePath.java:1169)
org.jenkinsci.plugins.durabletask.FileMonitoringTask$FileMonitoringController.(FileMonitoringTask.java:101)
org.jenkinsci.plugins.durabletask.WindowsBatchScript$BatchController.(WindowsBatchScript.java:94)
org.jenkinsci.plugins.durabletask.WindowsBatchScript$BatchController.(WindowsBatchScript.java:92)
org.jenkinsci.plugins.durabletask.WindowsBatchScript.doLaunch(WindowsBatchScript.java:60)
org.jenkinsci.plugins.durabletask.FileMonitoringTask.launchWithCookie(FileMonitoringTask.java:66)
org.jenkinsci.plugins.durabletask.FileMonitoringTask.launch(FileMonitoringTask.java:61)
org.jenkinsci.plugins.workflow.steps.durable_task.DurableTaskStep$Execution.start(DurableTaskStep.java:158)
org.jenkinsci.plugins.workflow.cps.DSL.invokeStep(DSL.java:184)
org.jenkinsci.plugins.workflow.cps.DSL.invokeMethod(DSL.java:126)
org.jenkinsci.plugins.workflow.cps.CpsScript.invokeMethod(CpsScript.java:108)
org.codehaus.groovy.runtime.callsite.PogoMetaClassSite.call(PogoMetaClassSite.java:48)
org.codehaus.groovy.runtime.callsite.CallSiteArray.defaultCall(CallSiteArray.java:48)
org.codehaus.groovy.runtime.callsite.AbstractCallSite.call(AbstractCallSite.java:113)
com.cloudbees.groovy.cps.sandbox.DefaultInvoker.methodCall(DefaultInvoker.java:18)
WorkflowScript.batch(WorkflowScript:20)
WorkflowScript.runTests(WorkflowScript:77)

```

Please note a couple of things

1) This is not a remote drive (I've noticed a number of tickets referring to mkdir problems which involve remote drives with windows)

2) This issue did not originally exist on this machine, previously folders were created as expected, so something changed. I have no idea if that was Jenkins, some windows update, a change in windows config... unfortunately the problem didn't manifest itself straight away because I was testing with workspaces which had already been successfully created before the problem began.

3) If I run a different windows machine as a slave, it does not have this issue.

I have no idea how to investigate this further, is there a way to step through the slave code in a debugger?

---
Originally reported by jonhodgson, imported from: mkdir failing for folders with no existing parent on windows

• status: Open
  
• priority: Critical
  
• component(s): workflow-durable-task-step-plugin
  
• resolution: Unresolved
  
• votes: 0
  
• watchers: 2
  
• imported: 20251212-090250
  

Raw content of original issue

When my jenkins pipeline job needs to create a folder on my windows slave (be it the workspace folder, a temporary folder to hold a batch file, whatever) , such as


C:\Jenkins\workspace\pipeline-test2@tmp\durable-487de3d5


Then it will fail if the parent does not exist


So in the above case, if 


C:\Jenkins\workspace\pipeline-test2@tmp


already exists (e.g. I created it manually), then durable-487de3d5 will be created correctly, and the script will work.


But if the parent does not exist, then I get the exception




Failed to mkdirs: C:\Jenkins\workspace\pipeline-test2@tmp\durable-487de3d5

hudson.FilePath.mkdirs(FilePath.java:1169)

org.jenkinsci.plugins.durabletask.FileMonitoringTask$FileMonitoringController.<init>(FileMonitoringTask.java:101)

org.jenkinsci.plugins.durabletask.WindowsBatchScript$BatchController.<init>(WindowsBatchScript.java:94)

org.jenkinsci.plugins.durabletask.WindowsBatchScript$BatchController.<init>(WindowsBatchScript.java:92)

org.jenkinsci.plugins.durabletask.WindowsBatchScript.doLaunch(WindowsBatchScript.java:60)

org.jenkinsci.plugins.durabletask.FileMonitoringTask.launchWithCookie(FileMonitoringTask.java:66)

org.jenkinsci.plugins.durabletask.FileMonitoringTask.launch(FileMonitoringTask.java:61)

org.jenkinsci.plugins.workflow.steps.durable_task.DurableTaskStep$Execution.start(DurableTaskStep.java:158)

org.jenkinsci.plugins.workflow.cps.DSL.invokeStep(DSL.java:184)

org.jenkinsci.plugins.workflow.cps.DSL.invokeMethod(DSL.java:126)

org.jenkinsci.plugins.workflow.cps.CpsScript.invokeMethod(CpsScript.java:108)

org.codehaus.groovy.runtime.callsite.PogoMetaClassSite.call(PogoMetaClassSite.java:48)

org.codehaus.groovy.runtime.callsite.CallSiteArray.defaultCall(CallSiteArray.java:48)

org.codehaus.groovy.runtime.callsite.AbstractCallSite.call(AbstractCallSite.java:113)

com.cloudbees.groovy.cps.sandbox.DefaultInvoker.methodCall(DefaultInvoker.java:18)

WorkflowScript.batch(WorkflowScript:20)

WorkflowScript.runTests(WorkflowScript:77)





Please note a couple of things

1) This is not a remote drive (I've noticed a number of tickets referring to mkdir problems which involve remote drives with windows)

2) This issue did not originally exist on this machine, previously folders were created as expected, so something changed. I have no idea if that was Jenkins, some windows update, a change in windows config... unfortunately the problem didn't manifest itself straight away because I was testing with workspaces which had already been successfully created before the problem began.

3) If I run a different windows machine as a slave, it does not have this issue.


I have no idea how to investigate this further, is there a way to step through the slave code in a debugger?

• environment: Jenkins 2.35 java 1.8.0_111-b14 (slave and master), Debian master, Windows 7 slave