New Question

Openstack Juno -Windows 2012 STD/ R2 fails after sysprep

asked 2016-06-23 16:00:21 +0300

Parth Gandhi gravatar image


I am creating Windows RAW image file for KVM Hypervisor for Openstack using the autounattend answer file which installs the OS, install cloudbase init and then do a sysprep with another unattend answer file. Once the image is created I use the image to spawn the instances. However after creating first instance all other instance fail to spawn, using the base image, with the below error message. I have tried modifying the setting in the unattend answer file but in vain.

""The computer restarted unexpectedly or encountered an unexpected error. Windows installation cannot proceed""

PS: I have 2 Openstack Juno setup and this image works fine in one and behaves as mentioned above in other.

edit retag flag offensive close merge delete

2 answers

Sort by » oldest newest most voted

answered 2016-07-27 10:01:02 +0300

Parth Gandhi gravatar image

I was able to fix the issue by modifying the unattend file. I am still investigating as to what was the issue in the original answer file which was causing the image to fail however I now have the working images of 2012 R2 and STD.

edit flag offensive delete link more

answered 2016-06-23 18:15:56 +0300

avladu gravatar image


you need to make sure that you are using kvm (libvirt_type=kvm in the compute's nova.conf). More information you can find here:

Thanks, Adrian Vladu.

edit flag offensive delete link more


Hi Adrian, I checked and both the environment has the above setting in nova.conf file.

Parth Gandhi gravatar imageParth Gandhi ( 2016-06-24 07:27:50 +0300 )edit

Moreover, you need the virtio drivers provided by fedora, to be installed if you are using virtio-scsi or virtio-net:

avladu gravatar imageavladu ( 2016-06-27 00:00:36 +0300 )edit

I have installed the virtIO drivers as well. The image works on one environment and fails on other.

Parth Gandhi gravatar imageParth Gandhi ( 2016-06-27 11:57:52 +0300 )edit

I understand. If possible, can you provide the qemu-kvm command used in either environments on a pastebin? The command can be retrieved by issuing a 'ps aux | grep kvm' on the compute nodes. I would like to see if there is any difference between the commands, in order to reproduce your issue. Thanks

avladu gravatar imageavladu ( 2016-06-27 15:30:44 +0300 )edit

Hi Please find the outout for working : and non working :

Parth Gandhi gravatar imageParth Gandhi ( 2016-06-30 12:13:03 +0300 )edit

You have an issue with the unattend file and the error is a sysprep one. There are 4 log files where you can check what the actual error is: C:\Windows\Panther\setupact.log/setuperr.log and C:\Windows\Panther\UnattendGC\setupact.log/setuperr.log

avladu gravatar imageavladu ( 2016-06-30 15:49:48 +0300 )edit

There is a high probability that the error you are getting is network or driver related. If you can provide the log files I have mentioned above, it would suffice to get to the bottom of your issue.

avladu gravatar imageavladu ( 2016-06-30 15:51:11 +0300 )edit

I will. However what surprise me is that the same image works on my other environment with the same unattend file. I will try to get the above mentioned logs.

Parth Gandhi gravatar imageParth Gandhi ( 2016-06-30 16:00:59 +0300 )edit

I was unable to copy/paste the logs or copy the file to working machine hence I have taken the screenshot. the link for the same is :

Parth Gandhi gravatar imageParth Gandhi ( 2016-06-30 16:53:29 +0300 )edit

I would suggest adding to your unattend.xml the section that disables devices drivers persistence: . As a side note, I strongly advise to use the already tested generation scripts: .

avladu gravatar imageavladu ( 2016-06-30 17:11:15 +0300 )edit

I made the changes in my unattend file as per the link and disabled the drive driver persistence and created a new image. I was able to create one instance from it and then it failed again for the subsequent instance.

Parth Gandhi gravatar imageParth Gandhi ( 2016-07-01 09:59:14 +0300 )edit

setuperr.log says unable to modify bcd boot entries - the error is now different than the previous one.

Parth Gandhi gravatar imageParth Gandhi ( 2016-07-01 10:00:29 +0300 )edit

Can you try to reproduce the issue using our qcow2 trial image? you can download it from here: .

avladu gravatar imageavladu ( 2016-07-01 11:05:25 +0300 )edit

I have already tried it and unfortunately it is also failing in the similar way after creating one instance on this environment. We took the trial image - converted it to RAW from Qcow2 - installed patches on it and then created a syspreped image.

Parth Gandhi gravatar imageParth Gandhi ( 2016-07-01 12:43:38 +0300 )edit

It might be an issue with the re-syspreped image you are creating. Does the unmodified cloudbase image also fail?

avladu gravatar imageavladu ( 2016-07-01 15:12:30 +0300 )edit

The unmodified image still works.

Parth Gandhi gravatar imageParth Gandhi ( 2016-07-11 09:14:14 +0300 )edit

Hi, I was able to make the image working for one environment after modifying the unattend file. however it again fails in the other environment after creating 4 instance. Below is the unattend file.

Parth Gandhi gravatar imageParth Gandhi ( 2016-07-19 10:46:31 +0300 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Question Tools

1 follower


Asked: 2016-06-23 16:00:21 +0300

Seen: 541 times

Last updated: Jul 27 '16