-
Bug
-
Resolution: Cannot Reproduce
-
Major
-
None
-
Platform: All, OS: Linux
If we define a Hudson Job with two Virtual Machines (different Images on the
same Vmware Host) and the job tries to power off the second virtual machine, the
Tomcat Java VM crashed.
vmware plugin version: private build based on haed version (we need the last not
released fix)
hudson version: 1.82
java jdk version: 1.6.0_04
tomcat version: 5.5.17-2.3
Here is the wrapper.log:
INFO | jvm 6 | 2008/03/06 16:36:33 | 06.03.2008 16:36:33
hudson.plugins.vmware.VirtualMachine powerOff
INFO | jvm 6 | 2008/03/06 16:36:33 | INFO: Powering off virtual machine
INFO | jvm 6 | 2008/03/06 16:36:33 | 06.03.2008 16:36:33
hudson.plugins.vmware.VirtualMachine powerOff
INFO | jvm 6 | 2008/03/06 16:36:33 | INFO: Powered down
INFO | jvm 6 | 2008/03/06 16:36:33 | 06.03.2008 16:36:33
hudson.plugins.vmware.VirtualMachine powerOff
INFO | jvm 6 | 2008/03/06 16:36:33 | INFO: Powering off virtual machine
INFO | jvm 6 | 2008/03/06 16:36:33 | #
INFO | jvm 6 | 2008/03/06 16:36:33 | # An unexpected error has been
detected by Java Runtime Environment:
INFO | jvm 6 | 2008/03/06 16:36:33 | #
INFO | jvm 6 | 2008/03/06 16:36:33 | # SIGSEGV (0xb) at pc=0x6257ee78,
pid=14022, tid=1644858288
INFO | jvm 6 | 2008/03/06 16:36:33 | #
INFO | jvm 6 | 2008/03/06 16:36:33 | # Java VM: Java HotSpot(TM) Server VM
(10.0-b19 mixed mode linux-x86)
INFO | jvm 6 | 2008/03/06 16:36:33 | # Problematic frame:
INFO | jvm 6 | 2008/03/06 16:36:33 | # C [libvix.so+0x67e78]
VixJob_WaitVargs+0x95
INFO | jvm 6 | 2008/03/06 16:36:33 | #
INFO | jvm 6 | 2008/03/06 16:36:33 | # An error report file with more
information is saved as:
INFO | jvm 6 | 2008/03/06 16:36:33 | #
/netsite/ns/http/tomcat/2.6.0/a10d40t11/hs_err_pid14022.log
INFO | jvm 6 | 2008/03/06 16:36:33 | #
INFO | jvm 6 | 2008/03/06 16:36:33 | # If you would like to submit a bug
report, please visit:
INFO | jvm 6 | 2008/03/06 16:36:33 | #
http://java.sun.com/webapps/bugreport/crash.jsp
INFO | jvm 6 | 2008/03/06 16:36:33 | # The crash happened outside the Java
Virtual Machine in native code.
INFO | jvm 6 | 2008/03/06 16:36:33 | # See problematic frame for where to
report the bug.
INFO | jvm 6 | 2008/03/06 16:36:33 | #
ERROR | wrapper | 2008/03/06 16:36:33 | JVM exited unexpectedly.
STATUS | wrapper | 2008/03/06 16:36:38 | Launching a JVM...
[JENKINS-1419] Vmware plugin crashed tomcat VM
Can you confirm if this happens again?
It will be very hard to nail down without a reproducible test case
Yes, I can reproduce this behavior with every build of this job. The Tomcat VM
of Hudson shut down every time.
Can I give you more information?
The VMware Images are both Windows XP Images.
If we are using two different jobs for every Image, the jobs runs proper.
Here the job entries:
<hudson.plugins.vmware.VMwareActivationWrapper>
<machines>
<hudson.plugins.vmware.VMwareActivationWrapper_-VMActivationConfig>
<vmxFilePath>/var/lib/vmware/Virtual_Machines/WinXPIE7Selenium/WinXPIE7.vmx</vmxFilePath>
<host>lxdevvmsrv01</host>
<powerUpMode>NORMAL_WAIT</powerUpMode>
<powerDownMode>NORMAL</powerDownMode>
<waitTimeout>300</waitTimeout>
<waitForFeedback>true</waitForFeedback>
<feedbackKey>WinXPIE7Selenium</feedbackKey>
<feedbackTimeout>180</feedbackTimeout>
</hudson.plugins.vmware.VMwareActivationWrapper_-VMActivationConfig>
<hudson.plugins.vmware.VMwareActivationWrapper_-VMActivationConfig>
<vmxFilePath>/var/lib/vmware/Virtual_Machines/WinXpIE6Selenium/WinXPIE6.vmx</vmxFilePath>
<host>lxdevvmsrv01</host>
<powerUpMode
reference="../../hudson.plugins.vmware.VMwareActivationWrapper$VMActivationConfig/powerUpMode"/>
<powerDownMode
reference="../../hudson.plugins.vmware.VMwareActivationWrapper$VMActivationConfig/powerDownMode"/>
<waitTimeout>300</waitTimeout>
<waitForFeedback>true</waitForFeedback>
<feedbackKey>WinXpIE6Selenium</feedbackKey>
<feedbackTimeout>180</feedbackTimeout>
</hudson.plugins.vmware.VMwareActivationWrapper_-VMActivationConfig>
</machines>
</hudson.plugins.vmware.VMwareActivationWrapper>
The Hudson Master configurtion:
<?xml version='1.0' encoding='UTF-8'?>
<hudson.plugins.vmware.VMwareActivationWrapper_-DescriptorImpl>
<hosts>
<hudson.plugins.vmware.VMwareActivationWrapper_-VMwareHostConfig>
<name>lxdevvmsrv01</name>
<vixLibraryPath>/home/hudson/vmware-vix/vmware-vix-distrib/vmware-vix/lib/server-1/32bit/</vixLibraryPath>
<hostName>lxdevvmsrv01</hostName>
<portNumber>904</portNumber>
<username>hudson</username>
<password>******</password>
</hudson.plugins.vmware.VMwareActivationWrapper_-VMwareHostConfig>
</hosts>
</hudson.plugins.vmware.VMwareActivationWrapper_-DescriptorImpl>
To confirm:
This issue arises when you have two VMs controlled by the one job?
Additionally, what OS are you running the VMware host on? (is this x86 or x86_64)?
What version of VMware are you using (incl build number)?
Correct, this issue arises only when we have two VMs controlled by one job
OS: CentOS release 5, x86_64
VMware Version: VMware Server 1.0.3 build-44356
Thanks,
stefan
Partial luck so!
Myself and a colleague are both running CentOS 5
Unfortunately I am using VMware Server 2 beta, but he is seeing about installing
1.0.3 to see if I can reproduce the issue.
Then I'll check if I can get it working on both!
Oh and these are Intel Core 2 Quad machines so we are also running x86_64
| Status | Original: Open [ 1 ] | New: In Progress [ 3 ] |
A-ha!
I think I know what is going wrong.
The power-down is disconnecting from the host too early.
Created an attachment (id=190)
error log file