Re: [one-users] help!

[Daniel Moldovan]

Hi,

Try to check if opennebula has access rights on "/var/lib/one/", that 
there is a folder "/var/lib/one/remotes", and check if you can manually 
SSH to the host you are deploying a virtual machine to, and then make 
sure you run opennebula from the same user, or check that the user from 
which you run opennnebula can ssh automatically (without any prompt) to 
the target host.


Cheers.

On 08/14/2013 10:04 AM, Lenin Abadié wrote:

hi guys!

i'm trying deploy a vm in a host but return the following erro log:

Wed Aug 14 01:34:45 2013 [DiM][I]: New VM state is ACTIVE.
Wed Aug 14 01:34:45 2013 [LCM][I]: New VM state is PROLOG.
Wed Aug 14 01:34:45 2013 [VM][I]: Virtual Machine has no context
Wed Aug 14 01:34:46 2013 [TM][I]: Command execution fail: 
/var/lib/one/remotes/tm/shared/clone 
localhost.localdomain:/var/lib/one/datastores/1/4db7457fbfefc1bbea96380d787d874e 
localhost:/var/lib/one//datastores/0/1/disk.0 1 1
Wed Aug 14 01:34:46 2013 [TM][I]: clone: Cloning 
/var/lib/one/datastores/1/4db7457fbfefc1bbea96380d787d874e in 
localhost:/var/lib/one//datastores/0/1/disk.0
Wed Aug 14 01:34:46 2013 [TM][E]: clone: Command "cd 
/var/lib/one/datastores/0/1; cp -r 
/var/lib/one/datastores/1/4db7457fbfefc1bbea96380d787d874e 
/var/lib/one/datastores/0/1/disk.0" failed: Warning: Permanently added 
'localhost' (RSA) to the list of known hosts.

Wed Aug 14 01:34:46 2013 [TM][I]: Permission denied, please try again.
Wed Aug 14 01:34:46 2013 [TM][I]: Permission denied, please try again.
Wed Aug 14 01:34:46 2013 [TM][I]: Permission denied 
(publickey,gssapi-keyex,gssapi-with-mic,password).
Wed Aug 14 01:34:46 2013 [TM][E]: Error copying 
localhost.localdomain:/var/lib/one/datastores/1/4db7457fbfefc1bbea96380d787d874e 
to localhost:/var/lib/one//datastores/0/1/disk.0

Wed Aug 14 01:34:46 2013 [TM][I]: ExitCode: 255
Wed Aug 14 01:34:46 2013 [TM][E]: Error executing image transfer 
script: Error copying 
localhost.localdomain:/var/lib/one/datastores/1/4db7457fbfefc1bbea96380d787d874e 
to localhost:/var/lib/one//datastores/0/1/disk.0

Wed Aug 14 01:34:46 2013 [DiM][I]: New VM state is FAILED

some suggestion?

regards,

--
*LENIN ERNESTO ABADIÉ OTERO
Bsc. in Computer Science - UNEMAT
M.Sc. Candidate of Computer Science
Informatic Center - UFPE
GFADS RESEARCH GROUP*
*  
 
 
 **lenin.abadie*



___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org


___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

Re: [one-users] Network Configuration - IP assignment

[Daniel . MOLDOVAN]

Only if the IP of the virtual machine is to be set statically from the VM
network configuration.

În Mar, Ianuarie 17, 2012 8:20 pm, Antonio Carlos Salzvedel Furtado Junior
a scris:
> What about contextualization?
>
>
> I noticed that the VM template does not have a CONTEXT definition, isn't
> it necessary?
>
> On Tue, Jan 17, 2012 at 2:50 PM, Javier Fontan
> wrote:
>
>
>> I was not aware of kvm assigning MAC addresses. If that setup suits
>> your needs then go for it, You should have no problem with it.
>>
>> On Tue, Jan 17, 2012 at 5:24 PM, Shantanu Pavgi  wrote:
>>
>>>
>>> Thanks for the reply Javier.
>>>
>>>
>>> I was able to create a VM by specifying only network interface name
>>> in
>> the NIC section. I left MAC address assignment over to the hypervisor
>> (KVM)
>> which assigns MAC address to a VM if it's not specified in the VM
>> definition file. So in this case OpenNebula didn't have any knowledge
>> of the MAC and IP address assigned to this VM. I am not sure whether
>> this will impact OpenNebula's VM management functionality though. Any
>> thoughts?
>>>
>>> Following is an example VM definition. OpenNebula version: 3.0.
>>>
>>>
>>> {{{
>>> # Capacity section
>>> NAME = ttylinux-01
>>> MEMORY = 128
>>> CPU = 1
>>>
>>>
>>> # OS and Boot section
>>> OS = [
>>> ARCH = i686,
>>> BOOT = hd ]
>>>
>>>
>>> FEATURES = [ acpi = "no" ]
>>>
>>>
>>> # Graphics I/O
>>> GRAPHICS = [
>>> TYPE = "vnc",
>>> PORT = "-1",
>>> keymap = "en-us" ]
>>>
>>> # Disk devices
>>> DISK = [
>>> IMAGE_ID = "0" ]
>>>
>>>
>>> # Network interface
>>> NIC = [
>>> BRIDGE = br0 ]
>>> }}}
>>>
>>>
>>>
>>> --
>>> Shantanu
>>>
>>>
>>> On Jan 16, 2012, at 5:05 AM, Javier Fontan wrote:
>>>
>>>
 Hello,


 Even if the VM's wont get the IP's form OpenNebula the MAC address
 must be generated. Internally MAC addresses are tied to Virtual
 Network leases so a place in a Virtual Network is needed. Just add
 a class B or C network and everything should work ok.

 You can also manually generate the mac addresses and you wont need
 the Virtual Network, in VM template:


 NIC=[MAC="00:02:01:23:45:67", IP="10.0.0.1", BRIDGE="br0"]


 But this is more cumbersome as you will need to manually generate
 different mac addresses for every interface.

 Bye


 On Mon, Jan 16, 2012 at 1:50 AM, Shantanu Pavgi 
 wrote:

>
> I am not following how OpenNebula handles network configuration
> for
>> the VMs. I am trying to create a virtual network using network config
>> file as follows:
> {{{
> NAME = "Test Network"
> TYPE = RANGED
> BRIDGE = br0
> PUBLIC=" yes"
> }}}
>
>
> In my test setup I have a DHCP server for IP address assignment.
> So I
>
>> did not put NETWORK_ADDRESS and NETWORK_SIZE values, thinking
>> OpenNebula
>> doesn't need to know about it. However I got following error upon
>> 'onvnet
>> create' command.
>
> {{{
> [VirtualNetworkAllocate] Error allocating a new virtual network.
> No
>
>> NETWORK_ADDRESS in template for Virtual Network.
>>
> }}}
>
>
> I am not sure what values should I provide in a network
> definition
>> file if IP address and MAC address assignment is not going to be
>> handled by OpenNebula. How does OpenNebula use NETWORK_ADDRESS and
>> NETWORK_SIZE values
>> during VM deployment?
>
> --
> Thanks,
> Shantanu
> ___
> Users mailing list
> Users@lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>



 --
 Javier Fontán Muiños
 Project Engineer
 OpenNebula - The Open Source Toolkit for Data Center Virtualization
 www.OpenNebula.org | jfon...@opennebula.org | @OpenNebula
 ___
 Users mailing list
 Users@lists.opennebula.org
 http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

>>>
>>
>>
>>
>> --
>> Javier Fontán Muiños
>> Project Engineer
>> OpenNebula - The Open Source Toolkit for Data Center Virtualization
>> www.OpenNebula.org | jfon...@opennebula.org | @OpenNebula
>> ___
>> Users mailing list
>> Users@lists.opennebula.org
>> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>>
>>
> ___
> Users mailing list
> Users@lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>


___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

Re: [one-users] Java OCA binding and a proxy between client and server

[Daniel . MOLDOVAN]

Hi,

Sorry for missing the previous mail regarding OCCI behind proxy.

The problem could be in the Java client or in the OCCI server.

From the Java side, maybe along the specified options, also use
-Dhttp.proxySet=true. Or try setting the proxy options from the Java
client  by placing the following lines just above the OCA invocation:

System.getProperties().put("http.proxyHost", "someProxyURL");
System.getProperties().put("http.proxyPort", "someProxyPort");
System.getProperties().put("http.proxyUser", "someUserName");
System.getProperties().put("http.proxyPassword", "somePassword");
System.getProperties().put("http.proxySet", "true");


From the OCCI server side, from what I see in the OpenNebula source files,
there is a file called opennebula-3.2.0\src\cloud\occi\lib\OCCIClient.rb,
written in ruby which uses "Net::HTTP::Get" and "Net::HTTP::Post"  to
communicate with the OCCI server. Although I am no RUby programmer, from
what I know, in Ruby you need to call Net::HTTP::Proxy if wanting to
access something behind a proxy, using
Net::HTTP::Proxy(proxy.host,proxy.port,proxy.user,proxy.password) instead
of Net::HTTP. Maybe a couple of modifications to the OCCIClient.rb file
could enable OpenNebula OCA to be used behind a proxy.



În Mar, Ianuarie 17, 2012 1:50 pm, Gian Uberto Lauri a scris:
>> "CC" == Gian Uberto Lauri < > writes:
>>
>
> CC> Hello gentlemen, What should I do when, using java binding for
> CC> OCA, there is a proxy between the client and the server? It seems
> CC> that starting the tomcatwith the
>
>
> CC> -Dhttp.proxyHost=$HOST\ -Dhttp.proxyPort=$PORT\
> CC> -Dhttp.proxyUser=$USER\ -Dhttp.proxyPassword=$PASS
>
>
> Am I the only one that needs to use the OCA bindings for Java throught
> a proxy?
>
> As I said, the use of these symbol definitions:
>
>
> -Dhttp.proxyHost=$HOST -Dhttp.proxyPort=$PORT\
> -Dhttp.proxyUser=$USER -Dhttp.proxyPassword=$PASS
>
>
> does not work (the OCA client fails to contact the remote ONED).
>
> --
> ing. Gian Uberto Lauri Ricercatore / Reasearcher
> Laboratorio Ricerca e Sviluppo / Research & Development Lab.
> Area Calcolo Distribuito / Distributed Computation Area
>
>
> gianuberto.la...@eng.it
>
>
> Engineering Ingegneria Informatica spa
> Corso Stati Uniti 23/C, 35127 Padova (PD)
>
>
> Tel. +39-049.8283.571 | main(){printf(&unix["\021%six\012\0"],
> Fax  +39-049.8283.569   |(unix)["have"]+"fun"-0x60);}
> Skype: gian.uberto.lauri|  David Korn, AT&T Bell Labs
> http://www.eng.it   |  ioccc best One Liner, 1987
> ___
> Users mailing list
> Users@lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>
>


___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

Re: [one-users] application integration (service publishing) in OpenNebula?

[Daniel . MOLDOVAN]

Answers inline.

>>What I mean is: suppose I, as a user
> of OpenNebula, using my created VM's, create a Web Service, which I
> publish on the Internet. Can anyone access this (someone who has no idea
> about the private cloud, someone who is simply accessing the URL), and by
> this way uses my Web Service (created on the VM's by the means of
> OpenNebula), so, basically, uses OpenNebula "remotely" (without knowing
> it)? Or this just doesn't make sense, since the whole idea of a private
> cloud is not to provide/publish information and services to the outside
> world, and this is not even possible since the virtual context?

As Fabian said, consider each VM as a normal machine. Anything you
expose on it through a network will be visible, just as with a normal
machine. The idea of a "private cloud" is to hide the physical
location of the virtual machines (on what computer does each reside)
and hide the fact that the VM is not a physical machine.

>>Are the
> most important reasons for installing OpenNebula the performance needs? Is
> there any other reason because of which I may want to install it, besides
> the fact that I might need multiple VM's (that I can manage) to perform a
> task (and to achieve platform interoperability)?

   OpenNebula is a "hypervisor for hypervisors". The reason you want to
install it is to provide a means of managing a cluster of servers which
provide virtualization. OpenNebula most importantly allows migration of
VMs between physical machines and resource usage monitoring.


>>When the load reaches its maximum (on a task
> which a user tries to perform on OpenNebula VM's), are new VM's created
> automatically (it the physical resources allow this) to support the
> performace needs?

  I think you miss the point of what a "cloud manager" or a
"virtualization middleware" does. See OpenNebula as a middleware over
simple virtualization mechanisms such as KVN, Xen, HyperV, etc.
OpenNebula only provides a means of managing heterogeneous virtualized
environments. Automatic scaling of virtual machines is out of the scope
of this middleware.

 The basic idea is that a VM is created, deployed, and the life cycle of
the VM is monitored using OpenNebula. The VM can be
started/stopped/migrated but automatic scaling is not possible, due to
different factors, among which the unavailability of such a mechanism on
the underlying hypervisors (only XEN allows changing the VM allocated
memory without recreating the machine).

  Any scaling mechanism should be built on top of OpenNebula.


>>Suppose I have an intranet in a company
> environment, with more interconnected physical machine that I'm actually
> going to use, does this change anything? Do I have to set up a separate
> physical network for the physical machines that I'm planning to include
> in the private cloud?

  If you read the OpenNebula documentation, the only requirements are that
a bridge exists on each OpenNebula node to connect the virtual machines.
To what that bridge connects, it is up to the infrastructure provider,
but no special requirements are needed in terms of network for the VMs.

 The "live migration" mechanism requires a storage NAS and that can
influence the network, but mainly any network will do fine with
OpenNebula.

În Joi, Decembrie 15, 2011 10:54 am, biro lehel a scris:
> Hello Fabian. Thanks again for your reply. I really appreciate you for
> taking the time.
>
> I read what you wrote a couple of times, and (I think) it helped me to
> clarify some things. But still, I have a few questions and issues for
> which I am looking for a clear answer. I put them in bullets:
>
> As I understand so far, OpenNebula has two types of users: the
> administrator, who basically has control over "everything", and the
> users, who can authenticate securely, instantiate some VM's, and do the
> work necessary for them. My question: can OpenNebula have another "layer"
> of users, some kind of "end-users"? What I mean is: suppose I, as a user
> of OpenNebula, using my created VM's, create a Web Service, which I
> publish on the Internet. Can anyone access this (someone who has no idea
> about the private cloud, someone who is simply accessing the URL), and by
> this way uses my Web Service (created on the VM's by the means of
> OpenNebula), so, basically, uses OpenNebula "remotely" (without knowing
> it)? Or this just doesn't make sense, since the whole idea of a private
> cloud is not to provide/publish information and services to the outside
> world, and this is not even possible since the virtual context?Are the
> most important reasons for installing OpenNebula the performance needs? Is
> there any other reason because of which I may want to install it, besides
> the fact that I might need multiple VM's (that I can manage) to perform a
> task (and to achieve platform interoperability)? I mean this has be the
> main point of it, right?When the load reaches its maximum (on a task
> which a user tries to perform on OpenNebu

Re: [one-users] managing images

[Daniel . MOLDOVAN]

The valid values for target are : hda, sda, sdb, and so on.
1,2,3 are partitions on that disk. And yes, the disk in my case is a blank
disk created on the fly with fs disk type on which i have installed an OS
by booting the VM from a cdrom image and now I use it for my virtual
machines.

În Mar, Februarie 1, 2011 3:02 pm, anoop Lekshmanan a scris:
> Hi,
>
>
> Thanks for the reply.
>
>
> So Target can be any name like hda1 sda1 etc right?
>
>
> You mean  "virtualDiskFilePath" is a blank disk created on Node using dd
> command
>
> or is it the .img created uisng virt-install.
>
> Thanks,
> Anoop
>
>
> On Tue, Feb 1, 2011 at 6:10 PM,  wrote:
>
>
>> Hy,
>> The target is the virtual target for the virtual disk. I have no used
>> the images repository until now and i am using KVM so i'm not entirely
>> sure what DISK types XEN supports . However, this might be useful so :
>> using an already existing virtual machine virtual disk i declare
>>
>> DISK   = [
>> type = disk, clone=no, size =4096, format = ext3, source=
>> "virtualDiskFilePath",
>> target   = hda, readonly = "no" ]
>> . Then, i have in the OS section
>> OS = [
>> boot = hd, ROOT = hda
>> ]
>> Note that hda is just the target inside the virtual machine.
>>
>>
>> And for booting a XEN kernel located on the Host computer, NOT in the
>> virtual machine, use OS = [
>> boot = hd, ROOT = hda,
>> kernel = "/boot/vmlinuz-2.6.35-22-server", initrd =
>> "/boot/initrd.img-2.6.35-22-server",
>> ]
>> where "boot/vmlinuz-2.6.35-22-server" should be replaced by the path to
>> your kernel and its name, as shown in
>> http://opennebula.org/documentation:rel2.0:template OS
>> section.
>>
>> Hope this helps.
>>
>>
>> Regards
>>
>>
>> În Mar, Februarie 1, 2011 1:16 pm, anoop Lekshmanan a scris:
>>
>>> Thanks Daniel.
>>>
>>>
>>>
>>> I have Xen installed and XenDom0 running on a node and I have
>>> successfully connected to the Node.
>>>
>>> ID NAME  CLUSTER  RVM   TCPU   FCPU   ACPUTMEM
>>> FMEM
>>> STAT
>>> 0 node-0 default0400  395 400 7.9G
>>> 1024K  off
>>> 1 node1  default0400  398 400
>>> 7.9G  7.2G   on
>>>
>>>
>>>
>>> I have created an image template of CentOS DVD ISO:
>>> NAME  = "CentOS 5.5 Base"
>>> PATH  =
>>> /opt/cloud/one/images/CentOS-5.5-x86_64-bin-DVD-1of2.iso
>>> PUBLIC= YES
>>> DESCRIPTION   = "CentOS 5.5 for testing"
>>>
>>>
>>>
>>> Then registered the template and oneimage lsited it:
>>>
>>>
>>>
>>> oneimage list ID USER NAME TYPE
>>> REGTIME
>>> PUB PER STAT
>>> #VMS
>>> 1 oneadmin  CentOS 5.5 Base   OS   Jan 31, 2011 11:14 Yes  No
>>> rdy 0
>>>
>>> I have created vnet using onevnet with name LAN2.
>>>
>>>
>>>
>>> Now I am trying to start a VM on the node using this image.
>>>
>>>
>>>
>>> I have created the vm template as:
>>>
>>>
>>>
>>> CPU= 1
>>> MEMORY = 2056
>>>
>>>
>>>
>>> DISK = [ image  = "CentOS 5.5 Base" ]
>>>
>>>
>>>
>>> DISK = [ type   = swap,
>>> size   = 3072 ]
>>>
>>> #swap of 3 GB
>>> DISK = [ type   = swap,
>>> size   = 3072 ] #/ root file system of 10 GB
>>>
>>>
>>> DISK = [ TYPE   = fs,
>>> SIZE   = 10240,
>>> FORMAT = ext3,
>>> SAVE   = yes,
>>> TARGET = ? ]
>>> NIC= [ NETWORK = "LAN2" ]
>>>
>>>
>>>
>>> GRAPHICS = [type="vnc",listen="0.0.0.0",port="1"]
>>>
>>>
>>>
>>> What could be the Target, I have /vm on hda3 ( not a separate
>>> partition but / of my Xen Host) I am trying to have a 10 GB / mount and
>>> 3 GB swap space. How can I set
>>>
>> the
>>> rest  like kernel etc
>>>
>>> I have no idea how to create the disk,  I have no separate partitions
>>> for Xen guest, partitions I have are / /boot and swap. I have a /vm
>>> directory for keeping Guests.  Should I re-instal the machine with
>>> separate /vm partition?
>>>
>>> I am stuck on this for some time. It would be great if some one can
>>> help me.
>>>
>>> Thanks,
>>> Anoop
>>>
>>>
>>>
>>>
>>>
>>> On Tue, Feb 1, 2011 at 12:42 PM, 
>>> wrote:
>>>
>>>
>>>
 Hello,
 Look at http://www.opennebula.org/documentation:rel2.0:template .
 And
 depending on your needs you can create both swap partitions and disk
 and so on. You need at least one boot partition which is the root
 for the OS
 in the case you are using KVM. For XEN you can use a kernel located
 on the host to boot the VM. Regarding the disk concepts,think at
 each disk as a separate storage drive, depending on the storage
 type(cdrom, disk, swap, etc).

 P.S. Is ONE not ON :P




 În Lun, Ianuarie 31, 2011 10:27 pm, anoop Lekshmanan a scris:


> Thanks Nikolay.
>
>
>
>
> One more question, on deploying like this how can I
> choose/customize disk for thus deployed VMs like /, swap, boot
> etc. I had gone through ON
> docs, but I dont understand the disk concepts, should we create
 partitions
> /,swap,
> boot on the node for VM. What do

Re: [one-users] managing images

[Daniel . MOLDOVAN]

Hy,
The target is the virtual target for the virtual disk. I have no used the
images repository until now and i am using KVM so i'm not entirely sure
what DISK types XEN supports . However, this might be useful so : using an
already existing virtual machine virtual disk i declare

DISK   = [
   type = disk,
   clone=no,
   size =4096,
   format = ext3,
   source= "virtualDiskFilePath",
   target   = hda,
   readonly = "no"
   ]
. Then, i have in the OS section
OS = [
   boot = hd,
   ROOT = hda
  ]
Note that hda is just the target inside the virtual machine.

And for booting a XEN kernel located on the Host computer, NOT in the
virtual machine, use
OS = [
   boot = hd,
   ROOT = hda,
   kernel = "/boot/vmlinuz-2.6.35-22-server",
   initrd = "/boot/initrd.img-2.6.35-22-server",
  ]
where "boot/vmlinuz-2.6.35-22-server" should be replaced by the path to
your kernel and its name,
as shown in http://opennebula.org/documentation:rel2.0:template OS section.

Hope this helps.

Regards

În Mar, Februarie 1, 2011 1:16 pm, anoop Lekshmanan a scris:
> Thanks Daniel.
>
>
> I have Xen installed and XenDom0 running on a node and I have
> successfully connected to the Node.
>
> ID NAME  CLUSTER  RVM   TCPU   FCPU   ACPUTMEMFMEM
> STAT
> 0 node-0 default0400  395 400
> 7.9G  1024K  off
> 1 node1  default0400  398 400
> 7.9G  7.2G   on
>
>
> I have created an image template of CentOS DVD ISO:
> NAME  = "CentOS 5.5 Base"
> PATH  = /opt/cloud/one/images/CentOS-5.5-x86_64-bin-DVD-1of2.iso
> PUBLIC= YES
> DESCRIPTION   = "CentOS 5.5 for testing"
>
>
> Then registered the template and oneimage lsited it:
>
>
> oneimage list ID USER NAME TYPE  REGTIME
> PUB PER STAT
> #VMS
> 1 oneadmin  CentOS 5.5 Base   OS   Jan 31, 2011 11:14 Yes  No
> rdy 0
>
> I have created vnet using onevnet with name LAN2.
>
>
> Now I am trying to start a VM on the node using this image.
>
>
> I have created the vm template as:
>
>
> CPU= 1
> MEMORY = 2056
>
>
> DISK = [ image  = "CentOS 5.5 Base" ]
>
>
> DISK = [ type   = swap,
> size   = 3072 ]
>
> #swap of 3 GB
> DISK = [ type   = swap,
> size   = 3072 ] #/ root file system of 10 GB
>
>
> DISK = [ TYPE   = fs,
> SIZE   = 10240,
> FORMAT = ext3,
> SAVE   = yes,
> TARGET = ? ]
> NIC= [ NETWORK = "LAN2" ]
>
>
> GRAPHICS = [type="vnc",listen="0.0.0.0",port="1"]
>
>
> What could be the Target, I have /vm on hda3 ( not a separate partition
> but / of my Xen Host)
> I am trying to have a 10 GB / mount and 3 GB swap space. How can I set the
>  rest  like kernel etc
>
> I have no idea how to create the disk,  I have no separate partitions for
>  Xen guest, partitions I have are / /boot and swap. I have a /vm
> directory for keeping Guests.  Should I re-instal the machine with
> separate /vm partition?
>
> I am stuck on this for some time. It would be great if some one can help
> me.
>
> Thanks,
> Anoop
>
>
>
>
> On Tue, Feb 1, 2011 at 12:42 PM,  wrote:
>
>
>> Hello,
>> Look at http://www.opennebula.org/documentation:rel2.0:template . And
>> depending on your needs you can create both swap partitions and disk and
>>  so on. You need at least one boot partition which is the root for the
>> OS
>> in the case you are using KVM. For XEN you can use a kernel located on
>> the host to boot the VM. Regarding the disk concepts,think at each disk
>> as a separate storage drive, depending on the storage type(cdrom, disk,
>> swap, etc).
>>
>> P.S. Is ONE not ON :P
>>
>>
>>
>> În Lun, Ianuarie 31, 2011 10:27 pm, anoop Lekshmanan a scris:
>>
>>> Thanks Nikolay.
>>>
>>>
>>>
>>> One more question, on deploying like this how can I choose/customize
>>> disk for thus deployed VMs like /, swap, boot etc. I had gone through
>>> ON
>>> docs, but I dont understand the disk concepts, should we create
>> partitions
>>> /,swap,
>>> boot on the node for VM. What does Target means, should we create
>>> target partitions before running the deploy commands.
>>>
>>> Sorry for bothering you, I am totally confused on this part.
>>>
>>>
>>>
>>> Thanks,
>>> Anoop
>>>
>>>
>>>
>>> On Tue, Feb 1, 2011 at 1:42 AM,  wrote:
>>>
>>>
>>>
 anoop Lekshmanan wrote on 31/01/11 23:04:

 Hi Nikolay,


>
> Thanks!
>
>
>
> I have used the registered image with VM template but on onevm
> create template-name the vm state continues be on pending.
>
 To run VM it needs to be deployed apart from creating VM
 $ onevm deploy  



 VM_id is virtual machine id  (specified in first column on 'onevm
 list') Host_id is registered in ON host id ('onehost list') E.g.
 $ onevm deploy 9 3



 Hope it helps,
 Nikolay.



 ___
 Users mailing list
 Users@lists.opennebula.org
 http://lists.opennebula.org/listinfo.cgi/users

Re: [one-users] managing images

[Daniel . MOLDOVAN]

Hello,
Look at http://www.opennebula.org/documentation:rel2.0:template . And
depending on your needs you can create both swap partitions and disk and
so on. You need at least one boot partition which is the root for the OS
in the case you are using KVM. For XEN you can use a kernel located on the
host to boot the VM. Regarding the disk concepts,think at each disk as a
separate storage drive, depending on the storage type(cdrom, disk, swap,
etc).

P.S. Is ONE not ON :P


În Lun, Ianuarie 31, 2011 10:27 pm, anoop Lekshmanan a scris:
> Thanks Nikolay.
>
>
> One more question, on deploying like this how can I choose/customize disk
>  for thus deployed VMs like /, swap, boot etc. I had gone through ON
> docs, but I dont understand the disk concepts, should we create partitions
> /,swap,
> boot on the node for VM. What does Target means, should we create target
> partitions before running the deploy commands.
>
> Sorry for bothering you, I am totally confused on this part.
>
>
> Thanks,
> Anoop
>
>
> On Tue, Feb 1, 2011 at 1:42 AM,  wrote:
>
>
>> anoop Lekshmanan wrote on 31/01/11 23:04:
>>
>> Hi Nikolay,
>>
>>>
>>> Thanks!
>>>
>>>
>>> I have used the registered image with VM template but on onevm create
>>>  template-name the vm state continues be on pending.
>>>
>> To run VM it needs to be deployed apart from creating VM
>> $ onevm deploy  
>>
>>
>> VM_id is virtual machine id  (specified in first column on 'onevm
>> list') Host_id is registered in ON host id ('onehost list')
>> E.g.
>> $ onevm deploy 9 3
>>
>>
>> Hope it helps,
>> Nikolay.
>>
>>
>> ___
>> Users mailing list
>> Users@lists.opennebula.org
>> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>>
>>
> ___
> Users mailing list
> Users@lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>


___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

Re: [one-users] VMware Driver Addon installation

[Daniel . MOLDOVAN]

Considering that libvirt is the linux virtualization toolkit used for the
virtualization process by the hypervisors, it should be installed where
the virtualization process takes place, and that is on the cluster nodes.

În Joi, Ianuarie 27, 2011 9:59 am, antoni artigues a scris:
> Hello
>
>
> We want to install the VMware Driver Addon.
>
>
> In the documentation there is one prerequisite: Install libvirt.
>
>
> Only one question about that: Where do we have to install libvirt?
> Only in the front-end or also in each cluster node?
>
>
> Thanks in advance
>
>
> Antoni Artigues
>
>
> ___
> Users mailing list
> Users@lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>
>


___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

Re: [one-users] Unable to ping and ssh to VMs

[Daniel . MOLDOVAN]

Contextualization  must be used in order to be able to assign the desired
IP to the virtual machine. The contextualization script reads the MAC of
the VM(opennebula assigns the MAC you have specified) and based on the MAC
fields it assigns an IP. The contextualization script must be places
inside your virtual machine.

Look at http://www.opennebula.org/documentation:rel2.2:cong and
http://www.opennebula.org/documentation:rel2.0:vmg

Considering you have no IP assigned( an IP should be assigned
automatically to the VM ), i assume you might have a problem with the
bridge you are connecting the VM network interface to.

Regards,
Daniel

P.S. sometimes the VM needs a reboot after deployment in order to get the
correct IP


> Hi Daniel,
>
>
> I manage to connect to the VM using vnc but there is no IP configuration
> on its interface. How should i configure in order for it to get the valid
> IP address like the
> ttylinux image does?
>
>
> On Fri, Dec 17, 2010 at 3:29 PM,  wrote:
>
>
>> You could read the hundred mails regarding this problem. Use VNC,
>> connect to the VM and see what the assigned IP is. Most probably the VM
>> has a different IP than expected. Or none at all.
>>
>> În Vin, Decembrie 17, 2010 7:52 am, Taylor Tay a scris:
>>
>>> Hi, I manage to setup OpenNebula with the ttylinux sample working
>>> within a private network. When i created my own VM and run, i cant
>>> seems to ping or ssh into it. I have already install ssh and open the
>>> port 22 on the VM itself therefore i have no idea where it go wrong.
>>> The log files did not
>>> show any error. Do i need to specify anything on the VM network
>>> interface itself beforehand?
>>> ___
>>> Users mailing list
>>> Users@lists.opennebula.org
>>> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>>>
>>>
>>>
>>
>>
>>
>


___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

[one-users] Virtual machine IP

[Daniel . MOLDOVAN]

Hello,

Does anyone now what is the mecanism behind assigning IP addresses to the
virtual machines based on their MAC? Or has someone managed to assign a
static IP to a virtual machine?

I am using KVM. I have a situation in which i atach a virtual machine to a
bridge, but the virtual machine never gets the IP i have assigned.
Opennebula reports that the machine has the IP i have assigned, but that
is not true. A manual check reveals a totally different IP.

I have noticed that libvirt allows only the definition of the MAC and not
of the IP of the virtual machine.

Also, OpenNebula generates the MAC of the virtual machine from the network
IP if provided by performing a hexadecimal translation for the last 4
bytes. For example, 192.168.1.105  = MAC 02:00:0c:a8:01:69 .

So when i specify a 192.168.1.105 IP i always get a :69 mac but i get
another IP address. For example,:
for  192.168.1.105 i get 192.168.1.102,
for  192.168.1.106 i get 192.168.1.10,
for 192.168.1.110 i get 192.168.1.114.

I can't figure out a pattern or find out the source of the IP's. Dnsmasq
seems stopped. I have tried attaching the virtual machines to a custom
defined bridge or to the virbr0 created by opennebula express install. I
have tried defining other networks using virsh and attaching the virtual
machines to them with no luck.

This may be a problem of dhcp IP assignment, but i can't manage to find
the dhcp server used.

Thank,
Daniel

___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

[one-users] [Fwd: Re: Opennebula development question]

[Daniel . MOLDOVAN]

Thanks to all for the answers.

And i have a question for opennebula developement. Why in
/one/tm_commands/ssh/tm_context.sh, the temporary directory is
TMP_DIR="$ONE_LOCATION/var/$DST_HASH" ? Should not be /tmp/$DST_HASH?
Because usually oneadmin does not have access to create folders in var in
the case opennebula is installed system wide as in my case(as a result
from express install).

Regards,
Daniel

> Hi, I'm new to this so I might be wrong, but I'm going to try to give a
> few answers below... hopefully someone will correct me if needed.
>
> El 18/11/10 17:27, daniel.moldo...@cs.utcluj.ro escribió:
>
>> Hy,
>>
>>
>> Is there any method of assigning a fixed/static IP to a virtual
>> machine? I have tried anything, even creating a virtual network with
>> only one IP as lease like below and assign this network to the virtual
>> machine in the vm template.
>>
>> NAME = "VM2_net"
>> TYPE = FIXED
>> BRIDGE = virbr0
>> LEASES = [IP=192.168.122.23]
>>
>
> I guess you have tried including the ip in the NIC section of the VM
> template? And had a look to http://opennebula.org/documentation:rel2.0:vgg
>
> NIC = [
> network = "name_of_the_virtual_network",   #Optional, XEN, KVM ip  =
> "ip_address",#Optional, XEN, KVM
> bridge  = "name_of_bridge_to_bind_if", #Optional, XEN, KVM target  =
> "device_name_to_map_if", #Optional, KVM
> mac = "HW_address",#Optional, XEN, KVM script  =
> "path_to_script_to_bring_up_if"] #Optional, KVM
>
>
> You might need to use the contextualization feature as well:
> http://opennebula.org/documentation:rel2.0:cong
>
>
>
>> Nomather what i try, the virtual machine will always get a random IP
>> located in the subnet of the virbr0 bridge. Although using onevm show,
>> the IP i have requested is reported as assigned, inside the virtual
>> machine things are different.
>>
>> Can there be issues regarding the host/guest OS? I am running CentOS as
>>  host and currently experimenting with Ubuntu JeOs and Desktop editions
>> as guest systems.The network configuration inside the hosts states that
>> the IP is assigned by DHCP.
>>
> Another way may be to configure your guests directly with the fixed IP
> address you want, but this is not so elegant as letting opennebula do it
> for you, of course.
>>
>> P.S. And just of curiosity(i am new to bridging/networking) how
>> difficult it is to perform some bridging or connection in order to be
>> able to ping the virtual machine from any opennebula node or client, not
>> just from the current host of the virtual machine?
> Not too difficult but it can get tricky the first time. You may need to
> look on http://wiki.libvirt.org/page/Networking
> if you're using kvm, to check that libvirt is correctly setup for that.
>
> Good luck,
>
>
> Hector
>
>>
>> Thank you,
>> Daniel
>>
>>
>> ___
>> Users mailing list
>> Users@lists.opennebula.org
>> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>>
>>
>
>
> --
> Hector
> ___
> Users mailing list
> Users@lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>
>



___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

Re: [one-users] Source RPM package

[Daniel . MOLDOVAN]

Extracted from Opennebula express installer( the CentOs install file)

http://dev.opennebula.org/attachments/download/222/opennebula-2.0-1.x86_64.rpm
http://dev.opennebula.org/attachments/download/223/opennebula-2.0-1.i386.rpm

În Joi, Noiembrie 18, 2010 9:31 pm, Pham Kim Long a scris:
> Hi Jaime,
>
>
> It's for CentOS 5.5.
>
>
> Thanks
> Long
>
>
> From: j.me...@gmail.com [mailto:j.me...@gmail.com] On Behalf Of Jaime
> Melis
> Sent: Friday, November 19, 2010 12:38 AM
> To: Phạm Kim Long
> Cc: users@lists.opennebula.org
> Subject: Re: [one-users] Source RPM package
>
>
> Hello,
>
>
> for what distribution do you want the source package?
>
> regards, Jaime
> On Thu, Nov 18, 2010 at 1:03 PM, Pham Kim Long
> mailto:lon...@fpt.net>> wrote:
> Hi
>
>
> Is there a source RPM package available for ONE 2.0? I looked at
> http://dev.opennebula.org/projects/packages/files
> and found only 2.0 beta source.
>
> Thanks OpenNebula team for the great work.
> Long
>
>
>
> ___
> Users mailing list
> Users@lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>
> ___
> Users mailing list
> Users@lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>


___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

[one-users] Virtual Machine Static IP Address

[Daniel . MOLDOVAN]

Hy,

Is there any method of assigning a fixed/static IP to a virtual machine? I
have tried anything, even creating a virtual network with only one IP as
lease like below and assign this network to the virtual machine in the vm
template.

NAME = "VM2_net"
TYPE = FIXED
BRIDGE = virbr0
LEASES = [IP=192.168.122.23]

Nomather what i try, the virtual machine will always get a random IP
located in the subnet of the virbr0 bridge. Although using onevm show, the
IP i have requested is reported as assigned, inside the virtual machine
things are different.

Can there be issues regarding the host/guest OS? I am running CentOS as
host and currently experimenting with Ubuntu JeOs and Desktop editions as
guest systems.The network configuration inside the hosts states that the
IP is assigned by DHCP.

P.S. And just of curiosity(i am new to bridging/networking) how difficult
it is to perform some bridging or connection in order to be able to ping 
the virtual machine from any opennebula node or client, not just from the
current host of the virtual machine?

Thank you,
Daniel

___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

Re: [one-users] VM Templates

[Daniel . MOLDOVAN]

Hy,

I have encountered the same problem. Do not specify Kernel because this is
mainly used in conjunction with the XEN hypervisor and it indicates the
physical path to the HOST kernel used to boot the virtual machine, not of
the kernel inside the VM. So if the Host is using Ubuntu, only the Ubuntu
distributions can be loaded by using the Kernel attribute.

And about the no bootable device found, i have avoided it only by:
- defining also Image disk
 example: DISK = [type=cdrom ,
source=/home/utcn/Downloads/jeos-8.04.3-jeos-i386.iso, target=hdb]
- stating that ROOT=hdb and booting from the cdrom. From it i could
install the linux distribution on the "disk" device and then use it as a
normal boot device.


I think that because the virtual disk was not created by opennebula( i
used AQUEMU GUI), it did not recognized it properly in order to boot from
it.

P.S. If anywone has a more elegant solution to this "no bootable device
problem" problem fell free to share.

Thanks,
Daniel


> I've some problems to create the exact vm template for any distribution
> different from Ubuntu. I want to try OpenSuse for my vm's images but or i
> get status failed [1] o if I delete the  KERNEL variable in the template
> I
> get "No bootable device". the image isnt corrupted, infact i can use it
> using kvm from console (kvm -hda /public/images/osuse.qcow2 -m 1024). Some
> suggestions?
>
> PS: It's possible to add in the website some tested and working
> templates?
>
>
> [1] ~~~ vm.log ~
>
>
> Wed Nov 17 17:29:15 2010 [VMM][I]: Generating deployment file:
> /var/lib/one/26/deployment.0
> Wed Nov 17 17:29:47 2010 [VMM][I]: Command execution fail:
> /tmp/one/vmm/kvm/deploy /var/lib/one//26/images/deployment.0
> Wed Nov 17 17:29:47 2010 [VMM][I]: STDERR follows.
> Wed Nov 17 17:29:47 2010 [VMM][I]: error: Failed to create domain from
> /var/lib/one//26/images/deployment.0
> Wed Nov 17 17:29:47 2010 [VMM][I]: error: monitor socket did not show up.:
>  Connection refused
> Wed Nov 17 17:29:47 2010 [VMM][I]: ExitCode: 255
> Wed Nov 17 17:29:47 2010 [VMM][E]: Error deploying virtual machine
> Wed Nov 17 17:29:48 2010 [DiM][I]: New VM state is FAILED
>
>
> ~~~ vm template ~
>
>
> NAME   = osuse test
> MEMORY =1024
> CPU=1
>
>
> OS = [
> BOOT="hd",
> KERNEL="/vmlinuz", # i made a link on / pointing at
> /boot/"kernel-file"
> INITRD="/initrd", #i made a link on / pointing at
> /boot/"initrd-image-file"
> ROOT="sda2" #mount execution on the vm states  "/dev/sda2 on / type
> ext4" ]
>
>
> DISK = [
> SOURCE = "/public/images/osuse.qcow2",
> TARGET = "sda",
> CLONE = "no",
> SAVE = "no",
> DRIVER = "qcow2"
> ]
>
>
>
> INPUT = [ type = "tablet", bus = "usb" ]
> GRAPHICS = [TYPE = "vnc", LISTEN = "localhost"]
>
>
> FEATURES=[ acpi="no" ]
>
>
>
> ~~~ generated deployment ~
>
>
>
> 
> one-26
> 524288
> 
> hvm
> /vmlinuz
> /initrd
> root=/dev/sda2 
> 
> 
> 
> /usr/bin/kvm
> 
> 
> 
> 
> 
> 
> 
> 
> ___
> Users mailing list
> Users@lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>


___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

Re: [one-users] VMs stuck in pending state

[Daniel . MOLDOVAN]

Hy,

Maybe there is a problem with the host. Onehost top/show does not indicate
an error on the host monitoring side? I am indicating this because
deployement.0 or disk.0 are created when the virtual machine is deployed
on a host(in prolog state). And have you tried to do a force deploy using
"onevm deploy vm_id host_id"? Because maybe you are using onevm create
(not "onevm submit") and the scheduler does not deploy the virtual machine
by itself.

I hope i haven't indicated things that you already have tried, but nothing
else comes to mind. A VM should work event if no network is specified for
it, so i think that the network bridge does not influence this.

Also, because the oned.log usually grows up fast you may try to delete the
one.db( and all other folders  ) from /var/lib/one and the logs from
/var/log/one and try a clean test. Add a host, see if it is monitored ok
and then do a onevm submit and onevm deploy

best
Daniel


> Hi,
>
>
>> 1. Where is the OS section of the virtual machine template?
>>
> Indeed, there isn't one! I was blindly following the tutorial in
> http://opennebula.org/documentation:rel2.0:vmg and the example given had
> no OS section... anyways I added one now.
>
>> 2. What hypervisor are you using? With XEN the virtual machine might
>> not boot if the xen kernel was not loaded .
> I'm using KVM.
>
>
>> I am particularly using KVM so something like this is in my virtual
>> machine template:
>>
>> NAME   = vm-example
>> CPU= 0.5
>> MEMORY = 512
>> OS = [
>> boot = hd, ROOT = hda
>> ]
>> DISK   = [
>> type = disk, clone=no, size = 1024, format = ext3, source=
>> "/home/oneadmin/Desktop/vm/disk.0",
>> target   = hda, readonly = "no" ]
>>
>>
>
> Thank you Daniel. I tried using your template and just changing the
> source path, but I'm still facing the same problem.
>
> One thing I noticed though is that in the node /var/lib/one/ has no
> sub-directories with VM ids. There is no images/deployment.0 or disk.0
> created? I wonder why? The oned.log still doesn't signal anything
> unusual.
>
>> As for the network, i am using the "virbr0" bridge created by the
>> opennebula express install, not eth0.
>
> How did you manage to use virbr0? did you make any changes to what was
> done by default by opennebula? I can see the bridge virbr0 added to both
> the frontend and node, but both have this same IP 192.168.122.1...
>
> Thank you.
>
>
> --
> Zaina
>
>
> ___
> Users mailing list
> Users@lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>
>


___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

Re: [one-users] VMs stuck in pending state

[Daniel . MOLDOVAN]

Hi,

1. Where is the OS section of the virtual machine template?
2. What hypervisor are you using? With XEN the virtual machine might not
boot if the xen kernel was not loaded .

I am particularly using KVM so something like this is in my virtual
machine template:

NAME   = vm-example
CPU= 0.5
MEMORY = 512
OS = [
   boot = hd,
   ROOT = hda
   ]
DISK   = [
   type = disk,
   clone=no,
   size = 1024,
   format = ext3,
   source= "/home/oneadmin/Desktop/vm/disk.0",
   target   = hda,
   readonly = "no"
   ]

As for the network, i am using the "virbr0" bridge created by the
opennebula express install, not eth0.

Also, use http://opennebula.org/documentation:rel2.0:template for properly
defining a virtual machine template with respect to one hypervisor or the
other.

Hope this si somewhat usefull,
Daniel

> Hi,
>
>
> I'm following this tutorial:
> http://opennebula.org/documentation:rel2.0:vmg
>
>
> The VMs are hanging in the pend state,
> $ onevm list
> ID USER NAME STAT CPU MEMHOSTNAMETIME
> 11 oneadmin ttylinux pend   0  0K 00 00:05:23
>
>
> I read in previous emails in this list that it could be related to
> insufficient memory/CPU in the node, but this is the output of
>
> $ onehost list:
> ID NAME CLUSTER  RVM   TCPU   FCPU   ACPUTMEMFMEM STAT
>  2 cluster02 default0200200200  2G1.7G
> on
>
> and I created a VM using the template below:
>
> NAME   = ttylinux
> CPU= 0.1
> MEMORY = 64
>
>
> DISK   = [
> source   = "/home/x/ttylinux/ttylinux.img", target   = "hda", files=
> "/bin/bash /home/x/ttylinux/id_rda.pub",
> readonly = "no" ]
>
> NIC= [ NETWORK = "test" ]
>
>
> FEATURES=[ acpi="no" ]
>
>
> So there are apparently enough resources to run the VM right? What could
> I
> doing wrong?
>
> The oned.log says nothing special:
> Tue Nov 16 18:34:38 2010 [ReM][D]: HostPoolInfo method invoked
> Tue Nov 16 18:34:43 2010 [ReM][D]: VirtualMachinePoolInfo method invoked
> Tue Nov 16 18:36:25 2010 [InM][I]: Monitoring host cluster02 (2)
> Tue Nov 16 18:36:29 2010 [InM][D]: Host 2 successfully monitored.
> Tue Nov 16 18:38:56 2010 [ReM][D]: VirtualMachinePoolInfo method invoked
> Tue Nov 16 18:39:02 2010 [ReM][D]: HostPoolInfo method invoked
> Tue Nov 16 18:39:02 2010 [ReM][D]: HostInfo method invoked
> Tue Nov 16 18:44:10 2010 [VMM][I]: --Mark--
> Tue Nov 16 18:44:10 2010 [InM][I]: --Mark--
>
>
> The user oneadmin can ssh without password to cluster02.
> This is what my small_network.net looks like:
> NAME = "test"
> TYPE = FIXED
>
>
> BRIDGE = eth0
> LEASES = [ IP="x.x.x.x"]
>
>
> is it okay to have eth0 as a bridge? and should I add more lease IPs?
>
> I'm crossing my fingers that my explanation provides enough background. I
>  would appreciate any help. Thank you.
>
> --
> Zaina
>
>
> ___
> Users mailing list
> Users@lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>
>


___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

Re: [one-users] Differences between Xen, KVM and Vmware

[Daniel . MOLDOVAN]

Hi,

There are differences at least in terms of virtual machine template
definition. There are fields supported/required by one hypervisor and not
the other. See http://opennebula.org/documentation:rel2.0:template

And i have also noticed on my ScientificLinux instalation( i think also on
CentOS, both being similar ) that when using the XEN hypervisor, onehost
info returnes "used memory" as "free memory". So if i am using 300MB i get
a Free Memory of 300MB.

Regards,
Daniel

> Hi,
>
>
> I'm still evaluating Open Nebula and the different drivers that it
> support.
>
> My question is if there is any difference between Xen, KVM and VMware
> from the Open Nebula perspective?
>
> I know that, for example, Xen and Vmware are different softwares and have
> different virtualization processes. But from Open Nebula perspective, are
> there features which are available only in one driver? If yes, is there a
> comparison table?
>
> Thanks,
>
>
> Fernando.
> ___
> Users mailing list
> Users@lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>
>


___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

Re: [one-users] Express Installation Script.

[Daniel . MOLDOVAN]

I am sorry. I haven't said that 755 are the wrong permissions. Just that
being new using opennebula i have encountered some "permission denied"
errors and by making the folder world-writable i have avoided them. I have
just mentioned as a quick fix to give everybody access to /var/lib/one. I
assume my problem originated from the oneadmin user being created without
necessary access rights. Because oneadmin also needs write permissions and
i think(excuse me if i am wrong) 755 means read-only for anyone other than
root. So my oneadmin user created from the install script does not have
write access needed for every opennebula operation as the one.db and
virtual machines files are there(images and deployment files).

But i insist, this might just be due to an incorrect installation process
performed by me and not due to incorrect access rights.

Regards,
Daniel


> Hello,
>
>
> thank you both for reporting these issues. There is indeed a problem with
>  /var/lock/one and /var/run/one directories getting removed on system
> restart. The problem is not OpenNebula Express but the binary packages
> themselves. We have opened an issue to provide init scripts which will
> recreate these directories.
>
> This will be fixed in the upcoming 2.0.1 maintanence release.
>
>
> There is one other thing that is still unclear to us: Daniel, you said
> that /var/lib/one wasn't created with the correct permissions, could you
> please elaborate? We think 755 are the correct permissions for that
> directory and not 777. Why do you need it to be world-writable?
>
> Regards,
> Jaime
>
>
> On Sat, Nov 6, 2010 at 8:49 AM,  wrote:
>
>
>> I think that the error is related to "export ONE_AUTH=/$HOME/.one-auth"
>>  and it means that there is no one_auth file. This file should contain
>> username:pasword of the opennebula user. but it might also be from some
>>  broken/missing ruby dependencies. You can also try install ruby-full
>> from a package manager.
>>
>> Now about installing nebula. Using the express install is the easiest
>> way of getting things working(I am still talking of installing on a
>> Ubuntu
>> distribution). 1. You run the install on the client machine(using sudo
>> install.sh or something). The install script creates a "oneadmin" user
>> and generates a "rsa" key for this user. When it generates the
>> "node-install.sh" it copies
>> that key to the node-install script.
>>
>> 2. If you add a node using "tm_ssh"( i have only used ssh, never nfs
>> because i am new at this too) then OpenNebula will get the host info by
>> using a "scp" to copy the "remotes" folder in the node /tmp/one and
>> then will connect using "ssh" to the node and call those ruby scripts.
>>
>> NOTE1: one start must be done from the oneadmin user( so log in as
>> oneadmin , because this is why the instalation script creates it ) and
>> DO
>> NOT USE sudo. If you use SUDO for "sudo start one", nebula will try  to
>> ssh as root. NOTE2: to log in as oneadmin (i don't know the default
>> oneadmin password) i do a sudo passwd oneadmin and input another
>> password.
>>
>> 3. The commands:
>>
 1. export ONE_AUTH=/$HOME/.one-auth
 2. export ONE_XMLRPC=http://localhost:2633/RPC2
 3. sudo mkdir /var/run/one
 4. sudo mkdir /var/lock/one
 5. sudo chmod 0777 /var/run/one
 6. sudo chmod 0777 /var/lock/one
 7. one start

>> Need to be performed on the client machine from the "oneadmin" user
>> logged in. NOTE1: The install script should had created the
>> /$HOME/.one-auth
>> containing oneadmin:oneadmin inside. NOTE2: The password in the one_auth
>> file does NOT NEED TO match the password of the oneadmin user. They are
>> two separate things. The one_auth file is used for opennebula requests
>> for client validation.
>>
>> 4. To install opennebula-node just run the node-install.sh on each
>> node. The node-install script also creates a oneadmin user. And more
>> important, it creates a  $HOME/.ssh (hidden folder, use Ctrl+H to see it
>> in a file manager). In this folder it creates(if not already existing) a
>> file called authorized_keys. Here the "rsa" key generated on the client
>> is placed. This file contains all the "rsa" keys used by anywone which
>> wants to be able to connect remotely to this node trough ssh. If the key
>> is not present a password is requested when issuing a ssh.
>>
>> NOTE1: after running node-install, generate a password for oneadmin
>> user and log in as oneadmin. If you remain logged as other user the
>> nebula client will not be able to connect to the node to get info.
>>
>> NOTE2: this steps only enable onehost add and onevm submit methods to
>> work. Migrate and onevm stop will fail because when migrating the nebula
>>  nodes communicate directly. And when issuing a stop the node will try
>> to save the state of the virtual machine and copy back the machine to
>> the nebula client. This two methods will fail because the nodes do not
>> have the "rsa" key of the other nodes in their
>> $HOME/.ssh/a

Re: [one-users] Installation of OpenNebula 2.0 from Source

[Daniel . MOLDOVAN]

You will encounter the same problem with the folders being deleted. Create
a script which creates them and run it at system reboot.


> Hi,
>
>
> I have been struggling to get Opennebula 2.0 Installation with Express
> script, there seems to lot of issues with the script(/var/lock/one &
> /var/run/one get deleted on every reboot ) . Hence I started looking into
>  installing it from source. I see that theres no straight forward
> documentation for the same. Also i didn't find any node specific
> installation information for installing from source. Any help or good
> documentation for the same will be highly appreciated
>
> Correct me if I am wrong as I am new to Opennebula
>
>
> Regards,
> kinglabs ___
> Users mailing list
> Users@lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>


___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

Re: [one-users] Express Installation Script.

[Daniel . MOLDOVAN]

I think that the error is related to "export ONE_AUTH=/$HOME/.one-auth"
and it means that there is no one_auth file. This file should contain
username:pasword of the opennebula user. but it might also be from some
broken/missing ruby dependencies. You can also try install ruby-full from
a package manager.

Now about installing nebula. Using the express install is the easiest way
of getting things working(I am still talking of installing on a Ubuntu
distribution).
1. You run the install on the client machine(using sudo install.sh or
something). The install script creates a "oneadmin" user and generates a
"rsa" key for this user. When it generates the "node-install.sh" it copies
that key to the node-install script.

2. If you add a node using "tm_ssh"( i have only used ssh, never nfs
because i am new at this too) then OpenNebula will get the host info by
using a "scp" to copy the "remotes" folder in the node /tmp/one and then
will connect using "ssh" to the node and call those ruby scripts.

NOTE1: one start must be done from the oneadmin user( so log in as
oneadmin , because this is why the instalation script creates it ) and DO
NOT USE sudo. If you use SUDO for "sudo start one", nebula will try  to
ssh as root.
NOTE2: to log in as oneadmin (i don't know the default oneadmin password)
i do a sudo passwd oneadmin and input another password.

3. The commands:
>> 1. export ONE_AUTH=/$HOME/.one-auth
>> 2. export ONE_XMLRPC=http://localhost:2633/RPC2
>> 3. sudo mkdir /var/run/one
>> 4. sudo mkdir /var/lock/one
>> 5. sudo chmod 0777 /var/run/one
>> 6. sudo chmod 0777 /var/lock/one
>> 7. one start
Need to be performed on the client machine from the "oneadmin" user
logged in.
NOTE1: The install script should had created the /$HOME/.one-auth
containing oneadmin:oneadmin inside.
NOTE2: The password in the one_auth file does NOT NEED TO match the
password of the oneadmin user. They are two separate things. The one_auth
file is used for opennebula requests for client validation.

4. To install opennebula-node just run the node-install.sh on each node.
The node-install script also creates a oneadmin user. And more important,
it creates a  $HOME/.ssh (hidden folder, use Ctrl+H to see it in a file
manager). In this folder it creates(if not already existing) a file called
authorized_keys. Here the "rsa" key generated on the client is placed.
This file contains all the "rsa" keys used by anywone which wants to be
able to connect remotely to this node trough ssh. If the key is not
present a password is requested when issuing a ssh.

NOTE1: after running node-install, generate a password for oneadmin user
and log in as oneadmin. If you remain logged as other user the nebula
client will not be able to connect to the node to get info.

NOTE2: this steps only enable onehost add and onevm submit methods to
work. Migrate and onevm stop will fail because when migrating the nebula
nodes communicate directly. And when issuing a stop the node will try to
save the state of the virtual machine and copy back the machine to the
nebula client. This two methods will fail because the nodes do not have
the "rsa" key of the other nodes in their $HOME/.ssh/authorized_keys file.
And also the nebula client does not have the keys of the nebula nodes. So,
on each node, do a "ssh-keygen -t rsa". It will generate a id_rsa.pub.
Copy the key from the .pub file to the authorized_keys file on the nebula
client and of the other nebula nodes. Do this for each node. If the
authorized_keys file does not exist create it but see in the node-install
sh how is that created. VERRY IMPORTANT it must have certain access rights
and owner. A chmod 0600 and chown -R oneadmin $HOME/.ssh  is necesary. But
search in the node-install script. There are the correct values.

În Vin, Noiembrie 5, 2010 6:42 pm, KING LABS a scris:
> Hi Daniel,
>
>
> What you said is right , I am still struggling to get things right  I
>  dont find opennebula docs to be straigt forward for a newbei , can you
> ask you for help .
>
> I am hoping if you can brief me the steps to install opennebula from
> source or using express script in brief for me. I would really appreciate
> it.
>
> also can you help me to understand this error
>
> *onehost list*
> */usr/lib/one/ruby/OpenNebula.rb:77:in `initialize': ONE_AUTH file not
> present (RuntimeError)* * **from
> /usr/lib/one/ruby/client_utilities.rb:239:in `new'*
> * **from /usr/lib/one/ruby/client_utilities.rb:239:in `get_one_client'*
> * **from /usr/bin/onehost:343*
>
>
>
> Regards,
> Kanthi
>
>
> On Thu, Nov 4, 2010 at 7:06 PM,  wrote:
>
>
>> There are several peculiar issues regarding the opennebula node.
>> Firstly,
>> it copies some files in /tmp/ and they get deleted on system restart.
>> Second, i found that it did not create /var/lib/one with coresponding
>> access rights.
>>
>> So, in order to make the opennebula node work you need to take some
>> steps each time the node is restarted(i am using Ubuntu server): 1. Copy
>> the content

Re: [one-users] Express Installation Script.

[Daniel . MOLDOVAN]

There are several peculiar issues regarding the opennebula node. Firstly,
it copies some files in /tmp/ and they get deleted on system restart.
Second, i found that it did not create /var/lib/one with coresponding
access rights.

So, in order to make the opennebula node work you need to take some steps
each time the node is restarted(i am using Ubuntu server):
1. Copy the content of /usr/lib/one/remotes from the machine where nebula
client is installed to /tmp/one ( create /tmp/one) on the nebula node
2. sudo "mkdir /var/lib/one" (if folder does not exist)
3. sudo "chmod 0777 -R /var/lib/one" (just to be certain give anyone
rights to access one)

Also, i have found that on the nebula client machine i need to perform
some similar tasks after system restart because folders keept
disappearing:
1. export ONE_AUTH=/$HOME/.one-auth
2. export ONE_XMLRPC=http://localhost:2633/RPC2
3. sudo mkdir /var/run/one
4. sudo mkdir /var/lock/one
5. sudo chmod 0777 /var/run/one
6. sudo chmod 0777 /var/lock/one
7. one start

Hope this helps, because even with the express install configuring nebula
takes time. For example, the express install does not have any means of
adding the rsa key to the authorized_keys of the nebula client so even if
the deploy will work, the "stop" of a virtual machine fails. In this case
the authorization keys have to be passed manually.


> Hi All,
>
>
> I think that the express installation script has some issues, I have
> been struggling for successful installation of OpenNebula 2.0 for past few
>  days.
>
> The variable *ONE_LOCATION=/srv/cloud/one *is used in the script but I do
>  not see any such folder created by the script later. When I try to
> deploy vms , it get created in  /var/lib/one/ and the error logs shows that
> theres some permission issues.
>
> Please let me know if anyone has also face the same issue and got it
> resolved.
>
> Regards.
> KINGLABS
> ___
> Users mailing list
> Users@lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>


___
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org