I see 2 issues:

1) The MAC addresses don't appear to be set in the VCL database for centos1. I can see how this may happen. There doesn't appear to be a MAC address field when you add or edit an individual computer. The MAC address can be set when adding multiple computers. I would suggest deleting the centos1 computer from VCL and then following the "Adding Multiple Virtual Machines" section in the INSTALLATION file. This will allow you to specify all of the required information. You will enter a beginning MAC address when adding multiple computers. For VMware VMs, the MAC addresses must be in the range:
00:50:56:00:00:00-00:50:56:3F:FF:FF

You can pick anything in this range.

Since the MAC address isn't populated, DHCP must not be configured either. This would have caused problems later on when the computer is loaded. DHCP must be configured on the management node so that the VMs are automatically assigned their private IP addresses when they boot. After you add multiple computers, the contents to add to /etc/dhcpd.conf are generated and displayed on the web page. Copy and paste this into /etc/dhcpd.conf on the management node and restart the dhcpd service. After restarting dhcpd, look at the end of /var/log/messages to make sure dhcp started correctly.

I would suggest using Firefox to add multiple computers. In the past I have had issues where the dhcp info wasn't displayed with IE.

2) It doesn't look like centos1's private IP address was added to /etc/hosts. You need to add lines for all of the computers you add. I would like to automate this for the next release, but as of now it has to be done manually.

-Andy


John Ma wrote:
Hello,

I successfully captured the image from a linux vm in ESXi 4.1, thank you Andy. But I couldn't find a way to reload/deploy it to other vm guests.

I have found that the image is captured and stored at two locations:
1) at management node: /images/vmware_images/vmwarelinux-centos_image1011-v0/vmwarelinux-centos_image1011-v0-flat.vmdk 2) at ESXi datastore: /vmfs/volumes/datastore1/vmwarelinux-centos_image1011-v0/vmwarelinux-centos_image1011-v0.vmdk VCL web interface also shows the new image 'centos_image10' under Image_Manager => Image_Profiles. Strangely the original vm guest was deleted from ESXi datastore by VCL. I don't worry about that too much at this moment.

I tried different settings when deploying the new image and they all failed somehow. Attached is a cut from /var/log/vcld.log, hopefully I can get some quick help from more experienced eyes.



Thanks,
John Ma
sys admin
Marist College







From:   Andy Kurth <andy_ku...@ncsu.edu>
To:     vcl-user@incubator.apache.org
Date:   10/06/2010 04:27 PM
Subject:        Re: VCL2.2 + ESXi4.1



It looks like you need to add the entry for your VM to the /etc/hosts file on the management node:
10.10.4.61 centos

Once this is added, manually run the nmap command that failed on the management node to make sure it will work before reattempting the image capture:
/usr/bin/nmap centos -P0 -p 24 -T Aggressive

Also verify that you can ssh from the management node to this VM using its hostname:
ssh -i <path to identity key> centos

The 'vcld -setup' utility has already added all of the necessary entries to the database for the new image and imaging reservation. When the capture process failed, the VCL request (request.id = 14) was left in the database and its state was changed to 'maintenance' (request.stateid = 10). To restart the image capture process, all you need to do is change request.stateid back to the 'image' state (request.stateid = 16) and your running vcld daemon should restart the process. You don't need to run 'vcld -setup' again.

There is currently no web interface or utility to change the request state so it will have to be done manually to the database. You'll probably want to install phpmyadmin if you haven't done so already. The statement to update the request state:
update request set stateid = 16 where id = 14;

The VCL daemon will begin processing the imaging request again within a few seconds. Run 'tail -f /var/log/vcld.log' before executing the SQL statement to watch the progress.

Hope this helps,
Andy

John Ma wrote:
Hello,

I am new to VCL. Having spent two days trying to get VCL2.2 working with

ESXi4.1 but failed, I decided to ask for help.

Does anyone have a success story with ESXi4? I am using the licensed
ESXi
and have vSphere SDK for perl 4.1 installed. What I am trying to achieve

is 1)to capture a generic linux image from one of the vm guests and 2)
to
deploye/load it to multiple guests, all the guest running on the same ESXi4.1 host. I think I don't need xCAT, so it's not installed.

Attached is the script log from one of my attempts to capture image by 'vlcd -setup'.




Thanks,
John Ma
sys admin
Marist College



--
Andy Kurth
Virtual Computing Lab
Office of Information Technology
North Carolina State University
andy_ku...@ncsu.edu
919.513.4090

Reply via email to