Some Questions about opennebula

Ou_ian-June 8th, 2011 AM
Would you like to ask how to set the vm ip address to DHCP?

All our VMS need to connect to the company's network (Internet) through eth0 (obtained through the company's DHCP ), all VMS must use the registered MAC address to obtain the correct IP address. How can I set the IP address in the network template of opennebula?

I have tested a single KVM instance. I first establish br0 through bridge and then bridge eth0 and tap so that the VM can correctly obtain the IP address and connect to the Internet, but I cannot figure out how to configure the network in opennebula.

Vpsee-June 9th, 2011 AM
@ Ou_ian
You need to use onevnet to create and configure a Public Virtual Network. The configuration file is as follows. For details, refer to the official documentation.
-----------------------
Name = "public"
Type = Fixed

Bridge = vbr1

Leases = [IP = 172.16.39.2, MAC = 50: 20: 20: 20: 22]
Leases = [IP = 172.16.39.3, MAC = 50: 20: 20: 20: 23]
-----------------------

Ou_ian-June 10th, 2011 AM
Thanks to vpsee, but my problem is that the DHCP server is not something I can control, so I cannot define a combination of MAC addresses and IP addresses in fixed.

The opennebula documentation is hopeless. Can I define a virtual switch as in ESX?

Vpsee-June 13th, 2011 pm
@ Ou_ian
Do you understand what you mean? Does your DHCP server need to register your MAC address? For example, some companies need to report the MAC address to the Administrator to restrict which computers can connect to the company's network, and then the Administrator binds the MAC address and IP address to the DHCP server. If the DHCP server does not have this restriction and is not bound to an IP address or a MAC address, you can specify the IP address of the opennebula VM at will, and set the IP address, mask, and gateway like the common kvm vm.

You can also disable opennebula from managing the virtual network, so that each opennebula VM gets an IP address from DHCP like a common kvm vm.

Ou_ian-June 15th, 2011 :53 AM
Thanks to vpsee, this is what I mean. In this case, the leases defined in fixed is not a VLAN, and then the VLAN is connected to eth0 through the br0 bridge?

If so, the networking of opennebula is also limited. Can I use virtual switch-All VMS to bridge through virtual switch and eth0 like ESX?

In addition, if opennebula is not allowed to manage the virtual network, I tried to set the target to tap in the NIC of the Virtual Machine template, and the IP address still cannot be obtained from DHCP, however, if I directly use qemu-KVM to run the VM image, I only need to add the-net tap parameter, so there is no problem in obtaining the IP address... Opennebula is really a headache... No serious replies have been received from countless emails...

Maybe I should try openstack?

[Email protected]-June 15th, 2011 AM
I would like to ask an image creation question, how can I create a centos-based VMware Virtual Machine image. And then use opennebula. Could you please help me with the steps?
Thank you!

[Email protected]-June 16th, 2011 pm
@ Vpsee
I also want to ask a question. If I configure opennebula to use KVM Virtual Technology and try again with Vmware after the test is complete, I just need to change the corresponding im, Vm, and TM to VMware.

 

Q: How is the tcpu FCPU acpu displayed in onehost calculated? For example, create two virtual machines (Virtual Machine parameter 1 cpu mb memory ...) Then, his acpu may have 200 left.
# Onehost create node01 im_xen vmm_xen tm_ssh

# Onehost list
ID name cluster rvm tcpu FCPU acpu tmem fmem stat
1 node01 default 0 400 400 400 3.8g 383 m on
I want to verify how many virtual machines can be created on the opennebula host node by using web applications and Virtual Machine creation parameters (CPU memory size. But I don't quite understand how his acpu is calculated? I have tested that if acpu is 0 and there are still many fcpus, you cannot create virtual machines on this node.

 

The IMG file can be used directly, depending on how your IMG works. Does the IMG built-in kernel boot system? Or do I need to start with the external kernel? Generally, you can start the external kernel (the kernel is on the parent machine,
OS = [
Kernel = "/boot/vmlinuz-2.6.18-4-xen-amd64 ",
Initrd = "/boot/initrd. img-2.6.18-4-xen-amd64 ",
Root = "sda1"]