-
Bug
-
Resolution: Unresolved
-
Major
-
None
-
Hudson 1.383
Master: Ubuntu, glassfish
Slave: Windows XP, Cygwin, running via JNLP
It looks like either cygpath sometimes does not return the right information, or Hudson does not use the cygpath information to find /bin/bash. The log snippet below shows a few things:
- Timing - It found bash twice within 10 seconds, and then the 3rd time it failed to find it
- When it worked, Hudson called C:\cygwin\bin\bash. When it failed, Hudson called /bin/bash. This is what makes me think that it didn't use the cygpath value
I found one thread in the forums which suggested just using bash instead of /bin/bash in the shell environment variable. Would that take care of something like this?
The two success cases:
09:56:19 $ C:\cygwin\bin\cygpath -w /bin/bash 09:56:19 [spectrum_tap2] $ C:\cygwin\bin\bash -xe C:\DOCUME~1\OR-ENG~1\LOCALS~1\Temp\hudson3300316933213918666.sh ... 09:56:19 $ C:\cygwin\bin\cygpath -w /bin/bash 09:56:20 [spectrum_tap2] $ C:\cygwin\bin\bash -xe C:\DOCUME~1\OR-ENG~1\LOCALS~1\Temp\hudson2834466993129794225.sh 09:56:20 + '[' '' '!=' true ']' 09:56:20 + echo 'Build description: Build skipped' 09:56:20 Build description: Build skipped
And the failure, 3 seconds later:
09:56:23 $ C:\cygwin\bin\cygpath -w /bin/bash 09:56:27 [spectrum_tap2] $ /bin/bash -xe C:\DOCUME~1\OR-ENG~1\LOCALS~1\Temp\hudson5829219388252082290.sh 09:56:29 FATAL: command execution failed 09:56:29 java.io.IOException: Cannot run program "/bin/bash" (in directory "C:\Hudson\workspace\spectrum_tap2"): CreateProcess error=3, The system cannot find the path specified 09:56:29 at java.lang.ProcessBuilder.start(Unknown Source) 09:56:29 at hudson.Proc$LocalProc.<init>(Proc.java:192) 09:56:29 at hudson.Proc$LocalProc.<init>(Proc.java:164) 09:56:29 at hudson.Launcher$LocalLauncher.launch(Launcher.java:638) 09:56:29 at hudson.Launcher$ProcStarter.start(Launcher.java:273) 09:56:29 at hudson.Launcher$RemoteLaunchCallable.call(Launcher.java:793) 09:56:29 at hudson.Launcher$RemoteLaunchCallable.call(Launcher.java:767) 09:56:29 at hudson.remoting.UserRequest.perform(UserRequest.java:114) 09:56:29 at hudson.remoting.UserRequest.perform(UserRequest.java:48) 09:56:29 at hudson.remoting.Request$2.run(Request.java:270) 09:56:29 at java.util.concurrent.Executors$RunnableAdapter.call(Unknown Source) 09:56:29 at java.util.concurrent.FutureTask$Sync.innerRun(Unknown Source) 09:56:29 at java.util.concurrent.FutureTask.run(Unknown Source) 09:56:29 at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(Unknown Source) 09:56:29 at java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source) 09:56:29 at hudson.remoting.Engine$1$1.run(Engine.java:58) 09:56:29 at java.lang.Thread.run(Unknown Source) 09:56:29 Caused by: java.io.IOException: CreateProcess error=3, The system cannot find the path specified 09:56:29 at java.lang.ProcessImpl.create(Native Method) 09:56:29 at java.lang.ProcessImpl.<init>(Unknown Source) 09:56:29 at java.lang.ProcessImpl.start(Unknown Source) 09:56:29 ... 17 more
I don't know what component this should be, so I set it to 'core'.
[JENKINS-8477] Random failure to find bash on Windows/Cygwin slave
Allan Eayrs
added a comment - I am in the same team as Saurav. We see this occur on an irregular basis, but it does happen. Our solution is to run the build again. Most times this problem with cygpath goes away. This is without any other intervention, aside from starting the project again.
Allan Eayrs
added a comment - I am in the same team as Saurav. We see this occur on an irregular basis, but it does happen. Our solution is to run the build again. Most times this problem with cygpath goes away. This is without any other intervention, aside from starting the project again. 
This happens when "C:\cygwin\bin\cygpath -w /bin/bash" returns non-zero exit code. Any idea why that can happen?