[one-users] Procedure to Create VMWare Images -- Need Help

Fri Apr 26 08:00:00 PDT 2013

Hi,

I am now able to run images downloaded from the marketplace with ESXi
nodes. However, I am still not able to deploy my own images. Probably
there is something wrong with the way I created my images.

This is the way I do to create my own images:

(1) Install a fresh new OS (such as CentOS 6.3) via vSphere Client. The
location of the VMWare images on that ESXi node is
/vmfs/volumes/datastore1/CentOS-6.3, and inside that directory there are a
bunch of files. When creating the VM I used thin-provisioning.

(2) Create a tar.gz containing the CentOS-6.3 directory, and download it
to the front end.

(3) On the front end, untar the tar.gz. Then make a temp directory, copy
the two vmdk files into the temp directory.

(4) From SunStone, I have a datastore with DS_MAD="vmware". I create a new
image inside that datastore by specifying the path of the temp directory
mentioned above. The image creation process is successful. I look into the
directory holding that image there was two files, one is
CentOS-6.3-flat.vmdk and the other is disk.vmdk.

(5) Then from SunStone I create a template with the image, and initiated
an instance. The PROLOG work fine, but the VM won't boot.  Below is the VM
log:

Fri Apr 26 22:03:14 2013 [DiM][I]: New VM state is ACTIVE.
Fri Apr 26 22:03:14 2013 [LCM][I]: New VM state is PROLOG.
Fri Apr 26 22:03:14 2013 [VM][I]: Virtual Machine has no context
Fri Apr 26 22:06:28 2013 [TM][I]: clone: Cloning
opennebula:/srv/cloud/one/var/datastores/100/8597ded816959ceff6dfffd6ec3e61c2
in /vmfs/volumes/101/152/disk.0
Fri Apr 26 22:06:28 2013 [TM][I]: ExitCode: 0
Fri Apr 26 22:06:29 2013 [LCM][I]: New VM state is BOOT
Fri Apr 26 22:06:29 2013 [VMM][I]: Generating deployment file:
/srv/cloud/one/var/vms/152/deployment.0
Fri Apr 26 22:06:29 2013 [VMM][I]: ExitCode: 0
Fri Apr 26 22:06:29 2013 [VMM][I]: Successfully execute network driver
operation: pre.
Fri Apr 26 22:06:40 2013 [VMM][I]: Command execution fail:
/srv/cloud/one/var/remotes/vmm/vmware/deploy
/srv/cloud/one/var/vms/152/deployment.0 vmware02 152 vmware02
Fri Apr 26 22:06:40 2013 [VMM][D]: deploy: Successfully defined domain
one-152.
Fri Apr 26 22:06:40 2013 [VMM][E]: deploy: Error executing: virsh -c
'esx://vmware02/?no_verify=1&auto_answer=1' start one-152 err: ExitCode: 1
Fri Apr 26 22:06:40 2013 [VMM][I]: out:
Fri Apr 26 22:06:40 2013 [VMM][I]: error: Failed to start domain one-152
Fri Apr 26 22:06:40 2013 [VMM][I]: error: internal error Could not start
domain: GenericVmConfigFault - Module DevicePowerOn power on failed.
Fri Apr 26 22:06:40 2013 [VMM][I]:
Fri Apr 26 22:06:40 2013 [VMM][I]:
Fri Apr 26 22:06:40 2013 [VMM][I]: ExitCode: 1
Fri Apr 26 22:06:40 2013 [VMM][I]: Failed to execute virtualization driver
operation: deploy.
Fri Apr 26 22:06:40 2013 [VMM][E]: Error deploying virtual machine
Fri Apr 26 22:06:40 2013 [DiM][I]: New VM state is FAILED

(6) On the ESXi node, I take a look at the vmware.log, and found this:

2013-04-26T14:06:32.709Z| vmx| Msg_Post: Error
2013-04-26T14:06:32.709Z| vmx| [msg.buslogic.notOn64bit] The BusLogic SCSI
adapter is not supported for 64-bit guests. See the documentation for the
appropriate type of SCSI adapter to use with 64-bit guests.
2013-04-26T14:06:32.709Z| vmx| [msg.moduletable.powerOnFailed] Module
DevicePowerOn power on failed.
2013-04-26T14:06:32.709Z| vmx| ----------------------------------------
2013-04-26T14:06:32.712Z| vmx| VMX_PowerOn: ModuleTable_PowerOn = 0
2013-04-26T14:06:32.712Z| vmx| MKS REMOTE stopping VNC server at 0.0.0.0:6052
2013-04-26T14:06:32.712Z| vmx| MKS local poweroff
2013-04-26T14:06:32.713Z| vmx| scsi0:0: numIOs = 0 numMergedIOs = 0
numSplitIOs = 0 ( 0.0%)
2013-04-26T14:06:32.713Z| vmx| Closing disk scsi0:0
2013-04-26T14:06:32.713Z| vmx| DISKLIB-VMFS  :
"/vmfs/volumes/5170f0fb-541ed8d5-abcf-80ee733ae51c/152/disk.0/CentOS-6.3-2-flat.vmdk"
: closed.

(7) So that look like that I am using BusLogic SCSI for my disk, but here
is what I have in disk.vmdk, LSI:

/vmfs/volumes/5170f0fb-541ed8d5-abcf-80ee733ae51c/152/disk.0 # more disk.vmdk
# Disk DescriptorFile
version=1
encoding="UTF-8"
CID=19bfd712
parentCID=ffffffff
isNativeSnapshot="no"
createType="vmfs"

# Extent description
RW 4194304 VMFS "CentOS-6.3-2-flat.vmdk"

# The Disk Data Base
#DDB

ddb.adapterType = "lsilogic"
ddb.thinProvisioned = "1"
ddb.geometry.sectors = "63"
ddb.geometry.heads = "255"
ddb.geometry.cylinders = "261"
ddb.uuid = "60 00 C2 95 9d 67 8c 34-e1 3a 25 4d 3d b4 90 48"
ddb.longContentID = "6bf0fc703c8f6b61e0bca38519bfd712"
ddb.virtualHWVersion = "8"

(8) Look further in deployment.0:

<domain type='vmware'>
	<name>one-152</name>
	<memory>1048576</memory>
	<os>
		<type arch='x86_64'>hvm</type>
	</os>
	<devices>
		<disk type='file' device='disk'>
			<source file='[101] 152/disk.0/disk.vmdk'/>
			<target dev='sda'/>
		</disk>
		<graphics type='vnc' listen='0.0.0.0' port='6052'/>
	</devices>
</domain>

(9) That's all I have. What could be wrong? The VM images was created from
the same ESXi node, how come it can not run on the same ESXi node from
OpenNebula? Is there a procedure that I can follow to create VMWare
images?

Thanks a lot.

Chen Xiang.