Install cnMaestro-1.6.0-r5 on OpenStack

Looking for help solving a problem trying to install the .ova for cnMaestro onto our company's OpenStack environment.

Openstack seems to not like .ova files, so I extracted the .VMDK from the .OVA and uploaded that image to our image library.  The virtual machine seemed to create OK with 4 CPU Cores, 8GB RAM, and 100GB disk, but the initial boot of the server fails as shown:

"Target filesystem doesn't have requested /sbin/init"

"No init found.  Try passing init= bootarg."

and we're stuck with the (initramfs) prompt.

Any ideas?

We will check and let you know if we can support openstack.

We dont have any immediate plans to support openstack, but we will see if in future we can add it.

Thanks for your feedback.

Rupam

I did find a solution.  It took some time, so I will leave the notes here for anyone else that might find this in the future.

TLDR:  Convert the OVA into QCOW2 and launch that from Openstack.

Place the file "cnmaestro-on-premises_1.6.0-r5_amd.ova" on a linux server..  I used /tmp/cambium/ in this example.  Adjust for your system as needed.

You will need root, or sudo capability to install the "qemu-img" package if it is not already on the machine.  For Debian or Ubuntu: "apt-get install qemu-img".  I used CentOS, so it was "yum install qemu-img"

Now, you can convert the file.

cd /tmp/cambium
tar -xvf cnmaestro-on-premises_1.6.0-r5_amd64.ova
qemu-img convert -O qcow2 cnmaestro-on-premises_1.6.0-r5_amd64-disk1.vmdk cnmaestro-on-premises_1.6.0-r5_amd64.qcow2

Now you have a .QCOW2 file that you can send up to your Openstack image repository, and launch an instance from that.

If you send it locally using the linux "glance" command from the Openstack python tools, the command is along the lines of:

glance image-create --name "cnMaestro-1.6.0-r5" --disk-format qcow2 --file cnmaestro-on-premises_1.6.0-r5_amd64.qcow2 --container-format bare

Pick your own "name" of course.

1 Like

Thanks great to hear you found something.. we will pass this info to our SIT team.

Thanks for sharing the solution you found for others searching this thread in the future!