Lik`s forum

[lik]

VirtualServerNotHalted:Virtual server is not halted
Actions -> Stop Virtualserver
Virtualserver Halt initiated.




FluidVM FAQs
From The FluidVM Wiki
Jump to: navigation, search
Contents
[hide]

* 1 What services must be run for FluidVM?
* 2 Which ports must be left open for FluidVM to work properly?
* 3 My management server hardware crashed! How can I move FluidVM to another server?
* 4 Can the FluidVM Management server run from a VM?
* 5 Full VMs fail to start
* 6 See Also

[edit]
What services must be run for FluidVM?

The physical machine hosting the FluidVM Management server typically runs the management server and the services for the FluidVM browser based interface. It does not have to be this way, depending on the load, it is possible to move the interface services to another system.

On the FluidVM Management server:

* /etc/init.d/vkarma

This is the main back-end server

* /etc/init.d/webkarma

This is the front-end core server

* /etc/init.d/nginx

This is the front-end web server

On the compute nodes:

* /etc/init.d/fluidvm-agent-updater

This is the compute node agent service updater. Agents are kept on the management server. They are passed on to the updater and the updater in-turn starts them. The Agent Updater is also responsible for running FluidVM software updates.

* /etc/init.d/vkservices start

This command actually stops the agent. Please note that it is important to stop the agent updater before you stop the agent. If you stop the agent without stopping the agent updater, the agent updater will re-start the agent in a few seconds.


By default the installer, will turn on these services to start automatically during boot-up.
[edit]
Which ports must be left open for FluidVM to work properly?

Management server API port - 8080

Web interface core port - 8282

Web interface web server port - 8086

FluidVM Agents (Running on the Compute Nodes) - 8181

VNC ports - 5900+
[edit]
My management server hardware crashed! How can I move FluidVM to another server?

The steps are simple:

1. Install the base OS

2. Install the latest version of FluidVM

3. Reboot

4. Update your copy by running "update_fluidvm.py" from the management node.

5. Stop management server (/etc/init.d/vkarma stop)

3. Copy the database from the old system to the new one. The file in question is /var/cache/vkarma/fluidvm_development.db

4. Copy all additional OS templates (OpenVZ and Xen) over to the new system from (/openvz-templates and /xen-templates)

5. Start your new FluidVM and your old setup on new hardware is ready (/etc/init.d/vkarma start)
[edit]
Can the FluidVM Management server run from a VM?

Sure it can, but this is not recommended for production. Management of this VM in turn and resource issues are a problem.

What you can do however, is run FluidVM on an OpenVZ compute node. This way, you can provide FluidVM the resources that it needs without having to setup a VM manually.
[edit]
Full VMs fail to start

You need VT or SVM technology to be supported by the CPU, Chipset and the BIOS. You can see if the CPU has support for VT/SVM technology by looking at the value against the "VT Technology" field in the Compute Node Summary page. You can find a "Yes" or "No" there. Seeing a "Yes" here does not guarantee that your FullVMs will start. It simply means that your CPU supports VT Technology. If you see a "No" here, you can forget about running Full VMs on that server. If you see a "Yes" here and you can't run your VMs, that means VT is either disabled from the BIOS(very common) or your chipset does not support it(very rare). It is not possible to document how to enable VT from the BIOS, since it is highly vendor dependent.

If you see the following error, mostly VT has been disabled from the BIOS:

"Unable to boot up:Error: HVM guest support is unavailable: is VT/AMD-V supported by your CPU and enabled in your BIOS?"

[lik]

FluidVM has poor/limited support yet. Free support/advice/bug reports may be discussed at official community forum http://community.fluidvm.com/.