Re: [one-users] Little help with first building cloud
Hello Matheus On 20.12.2011 14:41, matheus tor4 wrote: At the moment I got this error message: Tue Dec 20 10:20:09 2011 [InM][I]: Monitoring host PacsOnCloud-FrontEnd (0) Tue Dec 20 10:20:10 2011 [InM][I]: Command execution fail: 'if [ -x /var/tmp/one/im/run_probes ]; then /var/tmp/one/im/run_probes kvm 0 PacsOnCloud-FrontEnd; else exit 42; fi' Tue Dec 20 10:20:10 2011 [InM][I]: ExitCode: 42 Tue Dec 20 10:20:10 2011 [InM][E]: Error monitoring host 0 : MONITOR FAILURE 0 - Is your sunstone server running on the same system as the front-end? With which user is sunstone running, as root or as oneadmin? Is root from the front-end also able to password less login to the cluster nodes? bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Virtual Machine Lock Manager
Hello Upendra On 15.12.2011 13:54, Upendra Moturi wrote: Hello Fabian Can you please explain me the work flow locking. i.e how one is throwing error when an image is registered with it. You need to create a persistent image (define in the image template), then it can only be used from one VM at the same time. bye Fabian ___ 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?
Hello Biro On 15.12.2011 09:54, biro lehel wrote: Hello Fabian. Thanks again for your reply. I really appreciate you for taking the time. You're welcome. 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: I do not see any bullets, this is probably only available when viewed in HTML. I read (and also write) e-mails as text only, so the part below looks quite confusing to me and those it is very hard to answer, but I try. As I understand so far, OpenNebula has two types of users: the administrator, who basically has control over everything, Only everything regarding the management of the VMs, but depending on who did the installation of the OS (Operating System) inside the VM, he may not have access to it. But as he can control the virtual hardware (the VM), he could eventually circumvent security measures done inside the OS of the VM. 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 This is not the duty of OpenNebula, this is something which needs to be done by the administrator of the OS inside the VM. This depends a lot of the used OS inside this VMs, but tools should be available. 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 As above, this service provisioning and user management of the web service depends on the person who creates and runs this web service. This is independent of OpenNebula, as OpenNebula only provides the VMs to run any OS in it. As I already wrote, OpenNebula is just an abstraction layer between physical computer hardware and the OS you run inside the VM. Without the OpenNebula cloud platform you would just install physical computers with the OS of your choice and the services and applications you would like to run. There you also need to create the necessary system / application to manage end users visiting your web service. 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 The private cloud does just provide you with virtual computers to run your OS and application of choice on it. This helps to better use the physical computer with more virtual machines on it. It gives you more flexibility with the available hardware resources to run more then one OS installation a the same time. installing OpenNebula the performance needs? Is there any OpenNebula does reduce the performance of your hardware a little bit, as the additional layer also needs some capacity of the physical hardware, but I guess this can be ignored. Your hardware can be used more flexible with OpenNebula (or any other cloud abstraction layer), as you can use more then one OS (in a VM) in parallel on the same hardware. 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 OpenNebula VM's), are new VM's created automatically (it the physical resources allow this) to support the performace needs? Or the only way of creating No, OpenNebula does not out of the box start new VMs when the currently running VMs are at a capacity limit. You need to build your own monitoring system, which does monitor your web service and act on the needs of more performance. This monitoring can then use OpenNebula to start additional VMs with your service / application. But additional VMs can only be started when there is enough physical hardware (eg. cluster nodes) available to support more VMs. It can not give you more raw hardware power as when your service / application would run directly on several physical servers instead. But it gives you more flexibility. VM's is the manual one?Can OpenNebula be installed on any type of physical network, or does it have some special needs? The front end can be any i386 or amd64 (preferred) compatible computer which supports a current Linux distribution. But for the cluster nodes it would probably help if you use a CPU with VT support. Check the needs through the Virtualization Subsystem 3.0 [1] with the details of the type of virtualization you would like to
Re: [one-users] Concurrent disk access
Hello Richard On 12.12.2011 00:55, richard -rw- weinberger wrote: On Mon, Dec 12, 2011 at 12:50 AM, Steven Timmt...@fnal.gov wrote: What kind of transfer method are you using? shared, ssh, lvm? You can load the .raw file into the image repository, make it persistent, and that will take care of it. I'm using shared. But if I use the image repository, it will *copy* my image at least once, right? My images are very big.500GiB. Yes, it will be copied, unless you use NFS or any other shared storage for the images repository. Using a registered image is the only chance for OpenNebula to know that a certain image is being used by a VM. This will be updated in the images table in the database. OpenNebula does not care about a more then once used path to a disk in the VM template. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Virtual Machine Lock Manager
Hello Upendra On 12.12.2011 08:04, Upendra Moturi wrote: Is there any locking mechanism to lock vms ,so that there are no two vms using same hard disk Register the hard disk image in the image repository (with 'oneimage register ...') and then in the VM template use the registered image. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] error status on disk image
Hello Wojciech On 12.12.2011 15:18, Wojciech Giel wrote: TIMESTAMP Mon Dec 12 14:07:32 2011 MESSAGE Error copying image in the repository: Not allowed to copy image file /var/lib/one/templates/linux_generic.img oneadmin is the owner of /var/lib/one/templates/ directory. what might be a problem ? Check the permissions on the folder of the image repository, as far as I know, at least oneadmin:cloud should be able to write there. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] [help] How to connet to a virtual machine?
Hello Cat On 11.12.2011 03:12, cat fa wrote: You meat I should set up DHCP server on my host? You did write, that your server got his IP 1.185.2.21 trough DHCP, so on then I guess on the LAN with probably 1.185.2.0/24 is already a DHCP server running, so this one should also provide the IP addresses to your VM, as they are through the bridge in the same LAN. Login in through VNC and check if they not already got an IP with 'ifconfig' or 'ifconfig -a' to see all interfaces including the assigned MAC address (which you then could register in the DHCP to assign the corresponding IP address). bye Fabian ___ 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?
Hello Lehel On 11.12.2011 09:42, biro lehel wrote: Hello Fabian. First of all, thanks for your answer. You're welcome. So, are you telling me, that there is no way for an application to exploit the advantages of OpenNebula? What is Not directly, but through the setup with pre-installed VMs which could be started on an as needed basis. But in this VMs your (cloud) application needs to be installed and able to run. The misunderstanding probably comes from the term Cloud which is used for at least two different types of clouds. One type of Cloud is to offer a platform to distribute VMs (virtual machines) of pre-installed systems on a cluster of hardware servers which are able to run concurrent VMs. This is was OpenNebula can provide. And the other type of Cloud is an application cloud, like eg. Google Apps [1], which does offer applications (eg. mail, calendar, docs) for a certain user group, but on a shared bigger platform. I do not know if there are any frameworks around to create such application clouds. As far as I know, they are all custom built depending on the needed services. [1] http://www.google.com/apps/ Unfortunately the term Cloud is such a hype nowadays, that it is used for a lot of stuff and those helping to confuse many. For example Apples iCloud [2] actually is just a centralized storage with some added features to be able to sync apps or pictures to your other iDevices. But this is probably mostly done by the client device, which needs to also connect to the iCloud and waits for new content in your account. [2] http://www.apple.com/icloud/ the use of it then? :) Basically, all I want to do is the following: when I will have OpenNebula set up and running (on a small scale), I will try to experiment and exemplify its benefits, by the means of an application that uses the private cloud. Tests, performance benefits, see how the nodes communicate, etc. But I don't understand exactly how a web application (for instance) is made and written such that it can use OpenNebula, it can exploit the benefits of the running VM's, so that it can be more performant. How is this whole integration done? How can an application make use of OpenNebula? Couldn't it be published somehow, such that its final users (clients) could use it as a service (through OpenNebula), in a way that is totally transparent to them? I think what I'm referring to is exactly this communication with the outside world that you were writing about. Does this application which you would like to offer to clients already exist, or is this something you are developing? As far as I understand it, you would like to create something like Google Apps and then offer it to potential customers, right? bye Fabian ___ 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?
Hello Lehel On 11.12.2011 14:33, biro lehel wrote: what I've been referring to. I will have OpenNebula set up, and (as common sense would tell) I will have my application installed on the created VM's. My question only referred to: how can I install an application on these VM's (should I only just copy it, or is it more complex than this), or stuff like: Look at the VM like at any other physical computer. It is just a container (eg. a virtual computer) where you can install the OS of your choice. The installation of your application inside the OS of your VM needs to be done the same as you would do it on a physical computer. But the installation of the OS in the VM needs to be done first. See my recent posting Re: Creating virtual machines from scratch [1] to this mailing list. [1] http://lists.opennebula.org/pipermail/users-opennebula.org/2011-December/007156.html Look at an OpenNeubla cluster / cloud like on an additional abstraction layer between a physical computer and your OS installation. An example: If you have 3 computers, you can install on each one the OS of your choice and run it, but then you have only 3 concurrent running OS installation available. With OpenNebula you need to install Linux on all 3 computers (1x front-end and 2x cluster nodes). The cluster nodes also need to support some kind of hypervisor (eg. KVM or XEN). Then you install OpenNebula on the front-end and then adjust the configuration for the shared file systems to be used by the cluster nodes. Then you can create VMs (virtual machines / virtual computers) and deploy them through the front-end (with Sunstone you also have a web GUI). Now you can create as many VMs as the two cluster nodes can support (depending on CPU power an available memory). You even can stop or terminate VMs and reuse them (with persistent image) at a later time. can the different tiers of the application (interface, business logic, and data repository) be on different VM's, but Sure, they can. most importantly: how can an end-user (not the administrator, but a potencial client) use the application? Or there is no such thing as the end-user / client concept in OpenNebula, since the only user is the administrator who has control over the infrastructure? If OpenNebula provides IaaS support, I In OpenNebula the administrator has full control over the running VMs, eg. he can stop (pause), resume or even shutdown / destroy them. OpenNebula also knows users, which eg. could create their own VMs (with their choice of OS installation) or can use pre-created shared system image to boot a VM. But as far as I know, out of the box OpenNebula is not able to provide virtualization on application level. But it has a very open and flexible design and you should be able to customize it to your needs, eg. with contextualization. suppose this means that he does not have control over the application only as a service, but rather he, as the admin, has control over the whole physical application? What do you understand as physical application? OpenNebula controls the distribution and monitoring of the VMs. It will place a newly created VM on a cluster node which has the requested requirements and resources available. It also manages all the system images (persistent and public / shared) and network interfaces (done through bridges) which the VMs need to run. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] [help]Could not create domain
Hello Cat On 10.12.2011 02:06, cat fa wrote: I modified the user and group in /etc/libvirt/qemu.conf , is that correct? No, qemu should run as root, to be able to use the kernel KVM stuff. You need to adjust it libvirtd.conf. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] [help] How to connet to a virtual machine?
Hello Cat On 10.12.2011 19:28, cat fa wrote: I create a virtual network in SunStone with leases like 1.185.2.22 1.185.2.23 1.185.2.24 This IP addresses are only used with contextualization. If your VM does not support contextualization, then you need to assign the IP address inside the VM or assign through DHCP from the bridged Network. My host abtains its own ip (1.185.2.21) through DHCP. Then your VM should also get an IP from there. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] [help] How to connet to a virtual machine?
Hello On 10.12.2011 21:35, Fabian Wenk wrote: This IP addresses are only used with contextualization. If your VM does not support contextualization, then you need to assign the IP address inside the VM or assign through DHCP from the bridged Network. I just missed the fact, that this IP addresses also create corresponding MAC addresses which are uses by your VMs. So you could create the VM template which includes one of the existing MAC addresses, so it will stay the same for this VM. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Creating virtual machines from scratch
Hello Richard On 08.12.2011 17:45, richard -rw- weinberger wrote: I'm a bit confused how to create a vm from scratch. Assume I want a vm running with CentOS6 and a new virtual hard disk of 500GiB. How can I create a new disk using OpenNebula (especially with Sunstone)? I do not know how to do this steps in Sunstone, I did it with the command line tools. Create the image manually (outside of OpenNebula) with this steps (for KVM): On a system which has KVM available: qemu-img create -f raw servername.img 10G qemu-system-x86_64 -hda servername.img -cdrom /path/to/install.iso -boot d -m 512 Connect through VNC for installation, the above command will report you the used port (default 5900), see blow as qemu-system-x86_64 listen only on localhost for VNC qemu-system-x86_64 servername.img -m 512 # to test after install Connect through VNC login and run 'poweroff' as root or with sudo Now on the front-end: Create an image template (servername-image.one) oneimage register servername-image.one Create an VM template for the host (servername.one) onevm create servername.one Connect to VNC on the cluster node: I do not know about your workstation, but from my Mac client I use Chicken [1], which supports connection through ssh. I guess there is a VNC client for the OS of your workstation available which also can do this. Else you could run it with manual ssh forwarding like this: ssh -L localhost:5900:localhost:5907 server-with-KVM replace 5907 with the port which qemu-system-x86_64 as reported and then use the local VNC client to connect to localhost port 5900. [1] http://sourceforge.net/projects/chicken/ In my setup each vm will have it's own disk image, thus no disk image needs to be copied. Is there a way to enforce this? The best is to register each image in the Image Repository with 'oneimage register ...' A final question, is it possible to change the boot order of a vm? Do I really have to delete and recreate it? You need to shutdown and recreate the VM. Best done with the command line tool 'onevm create template' and the template you can modify. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Access network by name (custom attributes)
Hello Tomáš On 09.12.2011 10:42, Tomáš Nechutný wrote: possible to define network by name (instead of id)? I tried NETWORK=local, NETWORK=\local\ and same with NAME instead of NETWORK_ID=1, but it doesn't work. As far as I know with OpenNebula 3.0 only IDs can be uses. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Little help with first building cloud
Hello Matheus On 07.12.2011 01:44, matheus tor4 wrote: Make the changes on files like oned.conf using my standart user will brings for me troubles in future, or not? Configuration files in /etc/ usually are changed with the root user, normal users should not be able to write or depending on the content even be able to read it. What you recommend? - Put the rights on the oneadmin user, or - Use root user to make changes (painlessly) What kind of changes? On my system the /etc/one/oned.conf belongs to root (rw) with only read permissions for the cloud group. The startup script /etc/init.d/opennebula takes care to start the OpenNebula daemons with the user oneadmin (eg. dropping privileges). But this is something which could depend on the used Linux distribution if you used a distribution provided package. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] create image vm
Hello Dian On 07.12.2011 20:02, Dian Djaelani wrote: ow sory i not have sever with GUI ok thanks for advise i want tray install ubuntu with gui for manage my VM Ah, I probably missed to tell, that qemu-system-x86_64 listen only on localhost for VNC. But you do not need GUI on the server. I do not know about your workstation, but from my Mac client I use Chicken [1], which supports connection through ssh. I guess there is a VNC client for the OS of your workstation available which also can do this. Else you could run it with manual ssh forwarding like this: ssh -L localhost:5900:localhost:5907 server-with-KVM replace 5907 with the port which qemu-system-x86_64 as reported and then use the local VNC client to connect to localhost port 5900. [1] http://sourceforge.net/projects/chicken/ bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] modifying a virtual machine resources
Hello Davood On 29.11.2011 23:23, davood ghatreh wrote: Does anyone know how to managea virtual machine resources? consider that I create a vps with one CPU, and after a wile, i decide to increase its CPU's to two. Is it possible in OpenNebula? i dont want to re-deploy ma machine, and dont want to loose any file of existing virtual machine in this process. As far as I know, it is not possible. You need to shutdown the VM, modify the template an recreate the VM again. It is like with real hardware, if you need to change something, you need to shutdown, do the modification and boot up again. If you need to be able to modify a VM without loosing anything, you should use persistent images out of the Image Repository. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] A question about the utilization of VMs
Hello Adnan On 04.12.2011 18:18, Adnan Pasic wrote: I can't really tell you what lookbusy does, as it's not a programme coded by me, but found on the internet to fulfill its duties ;) Or it does not properly as you would need it. Also, the website where I downloaded it from doesn't say anything on how the programme is really working, and if it's possible to somehow upgrade it with for the purpose of, e.g. filling zeros. Google pointed me to [1], is this the one you are using? If yes, you could ask the developer, if he could adjust it for your needs (eg. filling the memory with random data, or even changing it during runtime), or ask some other developer which knows about C, if he could do it for you. [1] http://www.devin.com/lookbusy/ So, do you maybe know a programme, or could tell me what to do in my case? This is becoming frustrating, as I'm almost finished with my thesis and need only a couple more measurements!!! Sorry, I do not know any such program. But I hope one of the possibilities I mention above are an option for you. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Little help with first building cloud
Hello Matheus On 06.12.2011 20:26, matheus tor4 wrote: I dont create any group named 'cloud'. Is it created automatically? oneadmin@PacsOnCloud-FrontEnd:/etc/one$ id oneadmin uid=107(oneadmin) gid=118(oneadmin) grupos=118(oneadmin) What changes I have to do ? It does also work with the group oneadmin instead of cloud, just adjust your libvirtd.conf according. The OpenNebula documentation usualy talks about oneadmin:cloud, so you just have to use oneadmin:oneadmin instead. About the permissions, is the images folder and the VM_DIR on a NFS storage? If yes, you need to adjust this exports in /etc/exports according to allow root to read/write files there, as with KVM the VMs will be run with root privileges. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] modifying a virtual machine resources
Hello Davood On 06.12.2011 21:32, davood ghatreh wrote: Do you know how this save as feature works? I can use it. when I save as a VM, it goes to my images, and after a while of being LOCKED, its status changes to FAILED. how can i save a machine as an image? If you want to use the 'onevm saveas vm_id disk_id image_name' on a running VM, then you need to be able properly shutdown this running VM in OpenNebula and wait until the image is written to the Image Repository. The cluster node needs to be able to write to the Image folder. The error you see, could be because this writing is not possible. Check your log files. Alternatively you could do the following steps to create an image into the Image Repository: Create the image manually (outside of OpenNebula) with this steps (for KVM): On a system which has KVM available: qemu-img create -f raw servername.img 10G qemu-system-x86_64 -hda servername.img -cdrom /path/to/install.iso -boot d -m 512 Connect through VNC for installation, the above command will report you the used port (default 5900) qemu-system-x86_64 servername.img -m 512 # to test after install Connect through VNC login and run 'poweroff' as root or with sudo Now on the front-end: Create an image template (servername-image.one) oneimage register servername-image.one Create an VM template for the host (servername.one) onevm create servername.one Hope this helps. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Little help with first building cloud
Hello Matheus On 23.11.2011 12:39, matheus tor4 wrote: I have only two servers (With VT) and two Core2Duo (Without VT). I want to build a little private cloud. I have a system with Core2Duo CPU which does support VT, so I am not sure if this is not available for all Core2Duo CPUs or not, but check out [1], this could probably help you to enable VT on this systems (if available) too. [1] http://en.wikipedia.org/wiki/X86_virtualization#Intel_virtualization_.28VT-x.29 Else as pointed out in a other recent post to this mailing list, you could use XEN on this two systems without VT, but then you are limited to paravirtualization. As far as I know, the guest OS in the VM needs to support this. My doubt is the following: - Can I use a server as Front-End + Image Repository + Cluster at the same time? You can, I have running the front end and the cluster node on a single system. When adding the local cluster node, I used 'onehost create localhost im_kvm vmm_kvm tm_nfs'. Also adding the other systems as cluster nodes will be possible. - Or, It's more advantageous use a Core2Duo as a Front-End and release all resources of the server to be use by cloud? This depends on the usage of the OpenNebula cloud as a whole. On a rather small installation, it is probably not a problem to use the front end also as a cluster node. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Problem with DataBlock iamges
Hello Salma On 24.11.2011 15:13, salma rebai wrote: Now, when I added DISK = [ IMAGE_ID = 72, TARGET=hdb ] to my VM template. I succeed to create the VM. Its state is « runn ». But when i have access to the VM and i check partition disk , I don't find any mounting to the DATABLOCK Disk --- ubuntu@ubuntu-KVM:~$ df -h Sys. de fichiersTaille Uti. Disp. Uti% Monté sur /dev/sda1 8,7G 2,7G 5,6G 32% / none 492M 584K 491M 1% /dev none 499M 164K 499M 1% /dev/shm none 499M 96K 499M 1% /var/run none 499M 0 499M 0% /var/lock OpenNebula and the hypervisor just gives the VM the hardware you have configured. The usage depends on the OS configuration in the VM. So you need to mount this second disk in the OS inside the VM, as you would do it on a physical system when you have added a second hard disk for the first time. Eventually you also need to create the file system on this second disk (also from within the VM). bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] A question about the utilization of VMs
Hello Adnan On 24.11.2011 21:16, Adnan Pasic wrote: Is there a way to circumvent that? For my diploma thesis I need to utilize deployed VMs for a period of up to 12 hours. However, if KSM is active, the physical host doesn't hold this utilization, but adjusts the used memory pages. I do not know, but I guess your 'lookbusy' tool is just allocating the memory, but not really using it (eg. filling it up with random data). I guess using the allocated memory should help you even with KSM turned on. I am not sure about freeing the used memory after the process has stopped. Usually this memory is just marked as free from the kernel, but will only be purged over time or when an other process is requesting memory. Probably your 'lookbusy' process needs to clear (fill with zeros?) the memory before ending. Then I guess KSM on the cluster node will do its work and reduces the memory usage of the VM. Since I want to measure the power consumption of the physical host under full VM utilization, I need the real memory to stay utilized as well, yet I Does your 'lookbusy' process also use CPU cycles (eg. creating full load on the VM) or only memory? The most power consumption difference you will see on different CPU load and not memory allocation. Also the amount disk I/O does affect the power consumption. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] FW: Disk copy error when creating VM
Hello Sergio On 25.11.2011 10:25, Sergio Garcia Villalonga wrote: ERROR=[ MESSAGE=Error excuting image transfer script: Error copying mehmet:/var/lib/one/images/31dd9191dd9e5ac0cc86954f86f75dda to localhost:/var/lib/one//6/images/disk.0, TIMESTAMP=Fri Nov 25 09:38:08 2011 ] Did you try manually to copy this file on the cluster node with the oneadmin user? Could be, that there is a read error when copying this files, eg. from a hardware or filesystem problems. Or is the disk full on the cluster node? bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Little help with first building cloud
Hello Matheus On 27.11.2011 17:49, matheus tor4 wrote: At the time I'm using the second strategy: Using a Core2Duo as a Front-End and release all resources of the server to be use by cloud. I choose this option because it is more close to the proposal of official OpenNebula Documentation. I hope it works now! So I use one of my servers (With VT) as a Image Repository too. This is up to you, but I would keep the image repository on the front end or eventually on the second system without VT. So you could use the two systems with VT as two identical cluster nodes. And now I have a little more question: With a NFS server configured. I have the Image Repository sharing its directory /var/lib/one. My doubt is: - Have I to insert the shared directory in the file /etc/fstab on all my nodes? So the node already have this directory avaiable when the machine starts. Yes, and also on the front end. - Or Haven't I to do this? Does the directory only be mount when it is necessary? (sudo mount ...) It needs to be already mounted when OpenNebula will deploy a VM. And, if I edit the /etc/fstab, Do I have to do this in the node ImageRepository+Cluster at the same time? I hope not, because the directory is on the machine I am not sure about OpenNebula 3.0, but with 2.2 the image repository and also the VM_DIR folder need to be mounted in the same path as they are on the front end. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] error create VM KVM
Hello Dian On 15.11.2011 20:54, Dian Djaelani wrote: anyone can help me ?? i`m try to build opennebula but error when create virtual machine with KVM hypervisor Wed Nov 16 02:31:07 2011 [VMM][I]: error: unknown OS type hvm From where does this hvm come? Could be a type somewhere in your VM template. Is the image you have really able to run on KVM? Eg. was it created with the qemu/kvm tools? bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Looking at moving to OpenNebula with a few questions
Hello Donny On 27.10.2011 18:48, Donny Brooks wrote: Currently all machines have their network cards bonded and vlans passed over the trunked interface as we have approximately 20 vlans we use. This should be fairly simple to do with OpenNebula correct? Yes, you need to create bridge interfaces for each VLAN you need. And then in OpenNebula create the networks (with 'onevnet create template'), more details are in [1]. [1] http://www.opennebula.org/documentation:rel3.0:vgg I have a mix of local and network storage. Should OpenNebula be able to handle both local and SAN storage? This is (currently) not really possible. But let me discuss something else. As far as I guess you probably would like to have persistent VMs, right? The main idea of OpenNebula is to use it for on demand cloud computing, but persistent VMs are also possible. To accomplish persistent VMs, the following broad steps are needed: 1. register the OS image for the VM in the image repository 'oneimage register image_template' 2. create start the VM with using the registered image in the vm_templat. 'onevm create vm_template' Please also read the threads Storage subsystem: which one? [2] and Stopping persistent image corruption [3] in the mailing list archive. They have some more details about the whole working of OpenNebula with persistent VMs and storage, which should help you with your decisions. [2] http://lists.opennebula.org/pipermail/users-opennebula.org/2011-October/thread.html#6616 [3] http://lists.opennebula.org/pipermail/users-opennebula.org/2011-October/thread.html#6617 All raid is hardware based. With the current setup what is the best way to set it up for best fault tolerance/speed/space? What is the best OS to start with? We currently use Centos 5.5 on all 3 nodes but would prefer Fedora or similar. Debian would be doable also. About the RAID and storage solutions the thread linked above should also help with your decisions. Regarding the Linux distribution, I guess you should use the one which fits the available know-how in your team. Would I be able to import the existing virtual machines that are running into OpenNebula? Yes, you need to register a persistent image ('oneimage register image_template') with the current image as start and then create the VM, as described above. We are a small state government agency with little to no IT budget so I have to work with what I have. Please keep that in mind before suggesting why not buy such and such Thanks in advance for the input. The OpenNebula documentation describes the front end (where the OpenNebula deamons run) and cluster nodes (where the VMs run), but it is possible to combine it on a single system, see [4]. So for example you could then use the system xen-test as your test system with front end and cluster node on the same server. And for production use the system xen1 as front end and cluster node, and xen2 only as cluster node. [4] http://lists.opennebula.org/pipermail/users-opennebula.org/2011-October/006748.html I hope this will give you some more insight into OpenNebula and points you to some possible ways to migrate your current setup. Feel free to discuss further things on the mailing list. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Storage subsystem: which one?
Hello Humberto Sorry for the delay. On 18.10.2011 10:35, Humberto N. Castejon Martinez wrote: Thank you very much, Fabian and Carlos, for your help. Things are much more clear now, I think. You're welcome. *Sharing the image repository. If I understood right, the aim with sharing the image repository between the front-end and the workers is to increase performance by reducing (or eliminating) the time needed to transfer an image from the repository to the worker that will run an instance of such image. I have, however, a With the shared images folder you are able to distribute the transfer over time, as the VM does only read (eg. transfer over NFS and those over the network) the parts of the image on an as needed basis. But from the performance point of view, NFS is most often slower then access to the image on the local disk. This may be different with other storage solutions, eg. with a distributed FS over all the cluster nodes, or eg. an other backend storage solution with iSCSI and 10 GBit/s Ethernet to the cluster nodes. This stuff most often depends on the complete setup and network infrastructure you have available. The best would be, if you can do performance testing by yourself on your own site and infrastructure to find the best solution depending on your expectation and needs of the VM cluster. question/remark here. To really reduce or eliminate the transfer time, the image should already reside on the worker node or close to it. If the image resides on a central server (case of NFS, if I am not wrong) or on an external shared distributed storage space (case of MooseFS, GlusterFS, Lustre, and the like), there is still a need to transfer the image to the worker, right? In the case of a distributed storage solution like MooseFs, etc., the worker could itself be part of the distributed storage space. In that case, the image may already reside on the worker, although not necessarily, right? But using the worker as both a storage server and client may actually compromise performance, for what I have read. With a distributed file system, it depends how this is working with such stuff. An example (I do not have any experience with it, but this is how I would expect the work of such a distributed file system to be done): In the example we would have cluster node 1 to 10, all set up with eg. MooseFS. We also would have a permanent image, which is located on the MooseFS storage (which for redundancy is physically distributed over several cluster nodes, probably also in parts). For the example, we assume that the image is physically on node 3, 5 and 7. Now when you start a VM which will use this image, the VM will be started on node 1, in the beginning, it will read the image through MooseFS over the network from one or more of the nodes 3, 5 or 7 and it can be used immediately. Now I expect from MooseFS, that it will modify the distributed file system in such a way, that over time the image physically will be stored on node 1 and to do this in the background. After a while the whole image should be available from the local disk of node 1, and those have the same performance as from a normal local disk. If somebody has experience with such stuff, please tell if my idea is right or wrong. Am I totally wrong with my thoughts here? If not, do we really increase transfer performance by sharing the image repository using, e.g. NFS? Are there any performance numbers for the different cases that could be shared? * Sharing theVM_dir. Sharing theVM_dir between the front-end and the workers is not really needed, but it is more of a convenient solution, right? Sharing theVM_dir between the workers themselves might be needed for live migration. I say might because i have just seen that, for example, with KVM we may perform live migrations without a shared storage [2]. Has anyone experimented with this? I'm not sure, but I guess OpenNebula depends on a shared file system for live migration, independent of the used Hypervisor. Probably you could do live migration with KVM and local storage when you are using KVM without OpenNebula. Regarding the documentation, Carlos, it looks fine. I would only suggest the possibility of documenting the 3rd case where the image repository is not shared but theVM_dir is shared. I am not sure, but I think OpenNebula is currently not able to handle this two differently, as it is defined per cluster node with the 'onehost create ...' command where you define to use tm_nfs or tm_ssh. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Storage subsystem: which one?
Hello Carlos On 17.10.2011 11:34, Carlos Martín Sánchez wrote: Thank you for your great contributions to the list! You're welcome. I'd like to add that we tried to summarize the implications of the shared [1] and non-shared [2] approaches in the documentation, let us know if there are any big gaps we forgot about. Thank you for documenting it on the website. I think it is complete and mentions all important facts about the two storage possibilities. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Bacula and OpenNebula
Hello Richard On 17.10.2011 12:45, Richard Palmer wrote: Or perhaps the backup agent should run inside each virtual machine rather than backing up the vm image ?. Any advantages/disadvantages ? Personally I would run the bacula-fd inside the VM and back it up as I do it with a physical systems. The advantage is, that the OS in the VM is running without any interruption. With the stopped/suspended VM and backing up the image file, you have the advantage of doing a snapshot backup of the file system of the VM. But then it would probably also help to backup the memory of the running VM (does the OpenNebula or Hypervisor write this to disk when doing 'onevm suspend vm_id'?). But the disadvantage is the paused OS during the backup, any running services on it (eg. web sites) are not available then. And probably also currently open network connections can break. This depends of the duration of the backup. I guess the decision depends also of the usage of the VM. Both scenarios have advantages and disadvantages. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Stopping persistent image corruption
Hello Richard On 14.10.2011 12:19, Richard Palmer wrote: Ah right, makes sense. At the moment I'm still just using the VM templates to manage disc images and haven't got my image repository configured; had planned to do that when I move to 3.0 but perhaps I should get on with it now... Ok, If you just assign an image located anywhere on disk to a VM, then OpenNebula does not know, that this is a persistent image. But it is also strange, because when you start an VM with an image located eg. in /scratch/, then this will be copied to VM_DIR and the VM will be started with the copy there. For the images repository you do not need much, just a folder which is also on the shared storage (like VM_DIR) and available also on the cluster nodes. Create a template and then use the 'oneimage register /path/to/template'. Then in the VM template modify the DISK entry like this: DISK = [ IMAGE = name_of_image ] bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Error in VM deploy
Hello Rubens On 10.10.2011 21:27, Rubens Pinheiro wrote: Hello, I've configured two machines, one with opennebula and another configured to be a host. I've created the host in opennebula and it's all ok (status: MONITORED) But when I try to create a vm, there is a problem in the deployment. Here is the vm.log: Mon Oct 10 15:15:30 2011 [VMM][I]: Command execution fail: 'if [ -x /var/tmp/one/vmm/kvm/deploy ]; then /var/tmp/one/vmm/kvm/deploy /srv/cloud/one/var//5/images/deployment.0; else exit 42; fi' Mon Oct 10 15:15:30 2011 [VMM][I]: STDERR follows. Mon Oct 10 15:15:30 2011 [VMM][I]: error: Failed to create domain from /srv/cloud/one/var//5/images/deployment.0 Mon Oct 10 15:15:30 2011 [VMM][I]: error: unable to set user and group to '118:131' on '/srv/cloud/one/var//5/images/disk.0': No such file or directory Mon Oct 10 15:15:30 2011 [VMM][I]: ExitCode: 255 Mon Oct 10 15:15:30 2011 [VMM][E]: Error deploying virtual machine: error: Failed to create domain from /srv/cloud/one/var//5/images/deployment.0 I think the error is there: _Command execution fail: 'if [ -x /var/tmp/one/vmm/kvm/deploy ]; then /var/tmp/one/vmm/kvm/deploy /srv/cloud/one/var//5/images/deployment.0; else_ It's a sintax error? More details are in the line error: unable to set user and group to '118:131' on '/srv/cloud/one/var//5/images/disk.0': No such file or directory Is tries to set permissions to some user:group which fails. To which user:group do this UID:GID match on the cluster node? bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Storage subsystem: which one?
Hello Humberto On 13.10.2011 11:03, Humberto N. Castejon Martinez wrote: Reading the Opennebula documentation, I believe there are two things I have to deal with: 1) The image repository, and whether it is shared or not between the front-end and the workers I have some persistent Images which are used for persistent VMs. I have the images folder shared with NFS and I use tm_nfs. When starting a VM with a persistent images, it creates only a soft link in VM_DIR which points to the image in the images folder. If you want to have copied the persistent image into the VM_DIR on the cluster node, you would need tm_ssh. But then on startup the whole image will be copied to the cluster node into VM_DIR and on shutdown it will be copied back to the images folder. 2) TheVM_DIR, that contains deployment files, etc for the VMs running on a worker. This directory may or not be shared between the front-end and the workers, but it should always be shared between the workers if we want live migration, right? If you use tm_ssh you do not need to share this, if you use tm_nfs, you need to have it shared. The same with the images folder. For live migration you need a shared images folder and VM_DIR. Some of the questions I have are these (sorry if some of them seem stupid :-)): They are not stupid, It took me some time to try out and see through how this things works, and so I had to change my setup a few times until OpenNebula an I were happy. - What are the implications of sharing or not the image repository between the front-end and the workers (apart from the need to transfer images to the worker nodes in the latter case)? See above. - What are the implications of sharing or not theVM_DIR between the front-end and the workers? Also above. - Can I use ZFS or MooseFs and still be able to live migrate VMs? MooseFS [1] is a fault tolerant, network distributed file system (eg. over several servers). I do not know if ZFS can do this. MooseFS is similar like NFS, only that the data are distributed over several servers (including your local server). I guess live migration should work. [1] http://www.moosefs.org/ - Will theVM_DIR always hold a (copy of a) VM image for every VM running on a worker? Must theVM-DIR be in the local hard drive of a worker or may it reside on an external shared storage? See above, if used with tm_nfs, it is on a external shared storage (NFS server). When used with tm_ssh, it is on the local disk and all images (public or persistent) will be copied in full through ssh. I guess two factors i should also consider when choosing a solution are the following, right? - The speed of transferring VM images - The speed of cloning VM images Yes. Access to a persistent image through NFS only transfers the data needed, cloning always creates a full copy. When you use tm_ssh, the image will always be copied in full through ssh (which also gives some CPU overhead on the front end and cluster node). I hope this helps and my information are correct, if not, could somebody from OpenNebula please correct me. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Stopping persistent image corruption
Hello Richard On 13.10.2011 12:06, Richard Palmer wrote: After stupidly launching a VM using a persistent disc image twice, I'm very grateful to e2fsck for cleaning up the ensuing filesystem corruption, but wondered if there is anything I could put in the template file to tell opennebula not to allow this to happen?. Some sort of unique instance flag ?. Strange, this should be done with this settings in the image template (with OpenNebula 2.2.1, do not know about 3.0): PUBLIC = NO PERSISTENT = YES Check the output of 'oneimage show image_id' when one VM is running with this image, it should show this (which should lock the image for any other VM): # oneimage show 6 IMAGE INFORMATION --- ID : 6 NAME : image-name TYPE : OS REGISTER TIME : 10/05 15:38:33 PUBLIC : No PERSISTENT : Yes SOURCE : /path/to/image-uniq_id STATE : used RUNNING_VMS: 1 bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Bugreport / Patch for MySQL with InnoDB (instead of MyISAM)
Hello Carlos On 29.09.2011 15:43, Carlos Martín Sánchez wrote: It's a bit too late to apply and test this patch for the final 3.0 release, Ok, no problem for me. If now somebody is running into this problem, it is at least documented with a few possible solutions or workarounds. but I've opened a ticket [1] to include it in the next release. Thank you. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
[one-users] Bugreport / Patch for MySQL with InnoDB (instead of MyISAM)
Hello According to the posting Re: [one-users] Opennebula 2.2.1 Failed to create database tables [1] from Max Hennig, I prepared the attached patches (for 2.2.x and 2.9.90), which solve to problem with the first start of oned when the database will be initialized. [1] http://lists.opennebula.org/pipermail/users-opennebula.org/2011-August/006260.html As far as I had the problem with the first start of oned, it could not create the tables, when in my.cnf the setting default_storage_engine = InnoDB is present. After removing it (and restarting MySQL), it was working, as MySQL then is using the default MyISAM storage engine. But there are reasons for using the InnoDB storage engine as default in MySQL. So it would be helpful to OpenNebula if this is also working. The attached patches only change all the VARCHAR(256) to VARCHAR(255). I did test the patch with OpenNebula 2.2.1 (MySQL with InnoDB) and it is working fine so far. But I guess this should also work with 2.9.90. It would probably help if somebody could test this with 2.9.90 and then do this changes in the source repository before the next RC or final build for 3.0. I do not know, if it is a good idea or not to have the upgrade script also do this modifications on an already running MySQL database. To do this, the three 'alter table ... VARCHAR(255);' commands from below would be needed (for an existing 2.2.1 database). But reducing the field length could cause some problems if it is filled to the limit. I even do not know, if oned or the one* commands do check the field length before entering data into the database. If yes, then this should also be adjusted there in the source code. To convert an already running MySQL opennebula database from MyISAM to InnoDB, I did the following steps (with OpenNebula 2.2.1). It is probably a good idea to stop OpenNebula during this modifications. Then first create a backup with: mysqldump -u root -p opennebula opennebula.mysql And then convert the tables with the mysql client: mysql -u root -p mysql use opennebula mysql alter table host_pool modify host_name VARCHAR(255); mysql alter table network_pool modify name VARCHAR(255); mysql alter table user_pool modify user_name VARCHAR(255); mysql alter table cluster_pool ENGINE=InnoDB; mysql alter table history ENGINE=InnoDB; mysql alter table host_pool ENGINE=InnoDB; mysql alter table host_shares ENGINE=InnoDB; mysql alter table image_pool ENGINE=InnoDB; mysql alter table leases ENGINE=InnoDB; mysql alter table network_pool ENGINE=InnoDB; mysql alter table user_pool ENGINE=InnoDB; mysql alter table vm_pool ENGINE=InnoDB; To check the current properties of a table the following MySQL command can be used: mysql show create table table_name; bye Fabian --- a/src/host/Host.cc 2011-06-08 15:15:46.0 +0200 +++ b/src/host/Host.cc 2011-09-28 15:42:11.0 +0200 @@ -56,7 +56,7 @@ tm_mad,last_mon_time, cluster, template; const char * Host::db_bootstrap = CREATE TABLE IF NOT EXISTS host_pool ( -oid INTEGER PRIMARY KEY,host_name VARCHAR(256), state INTEGER, +oid INTEGER PRIMARY KEY,host_name VARCHAR(255), state INTEGER, im_mad VARCHAR(128),vm_mad VARCHAR(128),tm_mad VARCHAR(128), last_mon_time INTEGER, cluster VARCHAR(128), template TEXT, UNIQUE(host_name)); --- a/src/um/User.cc2011-06-08 15:15:46.0 +0200 +++ b/src/um/User.cc2011-09-28 15:42:25.0 +0200 @@ -53,7 +53,7 @@ const char * User::db_names = oid,user_name,password,enabled; const char * User::db_bootstrap = CREATE TABLE IF NOT EXISTS user_pool ( -oid INTEGER PRIMARY KEY, user_name VARCHAR(256), password TEXT, +oid INTEGER PRIMARY KEY, user_name VARCHAR(255), password TEXT, enabled INTEGER, UNIQUE(user_name)); /* -- */ --- a/src/vnm/VirtualNetwork.cc 2011-06-08 15:15:46.0 +0200 +++ b/src/vnm/VirtualNetwork.cc 2011-09-28 15:42:39.0 +0200 @@ -78,7 +78,7 @@ const char * VirtualNetwork::db_bootstrap = CREATE TABLE IF NOT EXISTS network_pool ( - oid INTEGER PRIMARY KEY, uid INTEGER, name VARCHAR(256), type INTEGER, + oid INTEGER PRIMARY KEY, uid INTEGER, name VARCHAR(255), type INTEGER, bridge TEXT, public INTEGER, template TEXT, UNIQUE(name)); /* -- */ --- a/src/group/Group.cc2011-09-23 16:56:55.0 +0200 +++ b/src/group/Group.cc2011-09-28 17:27:03.0 +0200 @@ -27,7 +27,7 @@ const char * Group::db_names = oid, name, body; const char * Group::db_bootstrap = CREATE TABLE IF NOT EXISTS group_pool ( -oid INTEGER PRIMARY KEY, name VARCHAR(256), body TEXT, +oid INTEGER PRIMARY KEY, name VARCHAR(255), body TEXT, UNIQUE(name)); /* */ --- a/src/host/Host.cc 2011-09-23
Re: [one-users] Opennebula 3.0 RC1 and persistent images in KVM VMs
Hello Alberto On 25.09.2011 00:27, Alberto Picón Couselo wrote: We have some a problems using persistent KVM images in Opennebula 3.0 RC1. Our configuration is as follows: Opennebula Front-End Ubuntu LTS 10.04 KVM worker node Debian Queeze 6.0.2 NAS for NFS Shared storage Sat Sep 24 23:49:08 2011 [VMM][I]: Command execution fail: 'if [ -x /var/lib/one/remotes/vmm/kvm/deploy ]; then /var/lib/one/remotes/vmm/kvm/deploy /var/lib/one/212/images/deployment.0 tc-kvm-hv02 212 tc-kvm-hv02; else exit 42; fi' Sat Sep 24 23:49:08 2011 [VMM][I]: error: Failed to create domain from /var/lib/one/212/images/deployment.0 Sat Sep 24 23:49:08 2011 [VMM][I]: error: internal error process exited while connecting to monitor: qemu: could not open disk image /var/lib/one/212/images/disk.0: Permission denied Please, can you give us any clue regarding this issue?. Persistent mode for KVM VMs is essential for us... Is root allowed to read/write in the NFS mounted images folder? Check the options in /etc/exports on the NFS server. Eventually you also need to force the client (cluster node) to mount it using NFSv3 (instead of NFSv4). With persistent images, the images stays in the images folder and is only linked from the vm_id/images/ folder. KVM does run with root privileges. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] REG: onevnet list
Hello Sriviatsan On 22.09.2011 10:12, srivatsan jagannathan wrote: Having same problem with this configuration also, NAME = test150 TYPE = FIXED BRIDGE = eth1 LEASES = [IP=192.168.58.150] OUTPUT onevnet show 108 VIRTUAL NETWORK 108 INFORMATION ID: : 108 UID: : 0 PUBLIC: N VIRTUAL NETWORK TEMPLATE BRIDGE=eth1 LEASES=[ IP=192.168.58.150 ] NAME=test150 TYPE=FIXED LEASES INFORMATION LEASE=[ IP=192.168.58.100, MAC=02:00:c0:a8:3a:64, USED=0, VID=-1 ] LEASE=[ IP=192.168.58.150, MAC=02:00:c0:a8:3a:96, USED=0, VID=-1 ] I have just tried your setup on my system, and everything looks good: # cat test150.net NAME = test150 TYPE = FIXED BRIDGE = eth1 LEASES = [IP=192.168.58.150] # # onevnet create test150.net # # onevnet list ID USER NAME TYPE BRIDGE P #LEASES [...] 5 admintest150 Fixed eth1 N 0 # # onevnet show 5 VIRTUAL NETWORK 5 INFORMATION ID: : 5 UID: : 0 PUBLIC: N VIRTUAL NETWORK TEMPLATE BRIDGE=eth1 LEASES=[ IP=192.168.58.150 ] NAME=test150 TYPE=FIXED LEASES INFORMATION LEASE=[ IP=192.168.58.150, MAC=02:00:c0:a8:3a:96, USED=0, VID=-1 ] # Did the network named test150 already exists when you created it with this template? This could explain why you are seeing two LEASE lines (192.168.58.100 and 192.168.58.150). But it is rather strange, that your 'onevnet list' does not show any output at all. Could it be, that the database got hit by something which trashed it in some parts? What are you using as backend database, sqlite or MySQL? bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] REG: onevnet list
Hello Srivatsan On 20.09.11 12:06, srivatsan jagannathan wrote: Trying to add virtual network, type fixed see bottom. onevnet -v create X.net - return vnet-number (10) onevnet show vnet-number (works fine, list information) onevnet list --- displays nothing (only header information rest blank.) somehow strange, did you see any errors in the log files? #Now we'll use the cluster private network (physical) BRIDGE = virbr0 Does the interface virbr0 exist? Is it really a bridge interface? You can check with 'brctl show'. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Users Digest, Vol 43, Issue 32
Hello Bala On 19.09.11 18:11, bala suru wrote: But I'm facing the onemore problem that when I try to save the VM using the following command . onevm saveasvm-id disk-id imagename onevm shutdownvm-id Ok, this steps are correct. Does 'onevm list' still show your VM which you just did shutdown? Because: Here is the output of oneimage show 32 ID : 32 NAME : alliswell TYPE : OS REGISTER TIME : 09/13 16:12:13 PUBLIC : No PERSISTENT : No SOURCE : /srv/cloud/one/var/images/def71f75ceab7d2fa444927b8e1588633c547422 STATE : rdy RUNNING_VMS: 1 There is still is a VM running which does use this image. Also you need to change some other settings, like maybe PERSISTENT (see help of oneimage) to Yes. IMAGE TEMPLATE DEV_PREFIX=hd NAME=alliswell TYPE=OS Depending on your settings during the installation of this disk, you probably want to change DEV_PREFIX to sd (for SCSI disk) with the 'oneimage' command. This is a known bug of the saveas command in 2.2.1. Since I'm able to save/ clone the VM , I guess there is no problem with user rights access to the nfs shared folder . But your error says: cp: cannot stat `/srv/cloud/one/var/images/def71f75ceab7d2fa444927b8e1588633c547422': No such file or directory When the VM did not shutdown properly (from OpenNeubla's point of view), then the image is not copied/moved to the images folder. The entry with 'oneimage list' and 'oneimage show image_id' is there, but the file in you images folder is not yet. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] onevm save method error
Hello Bala On 20.09.11 13:44, bala suru wrote: I need to save modified VMs as new one - here can I do normal copy and save the VM which is running and modified ..? You could first register a persistent image, then create a persistent VM, run it and do your installation/updates. Then run 'onevm shutdown vm_id'. Now you could create an other VM using the same registered image (makes probably not much sense), or copy the image away from your image folder, create a new template and register the copied image with a new name. Then create a new persistent VM which will use the new image. Can I deploy the new VM using the save Image ..? Probably as described above. But to clone a running image, the 'onevm saveas vm_id disk_id image_name' command and then 'onevm shutdown vm_id' (as in your link described the VM and the OS running in the VM needs to support ACPI) are probably easier to use and you do not need to manually copy around the image. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] onevm saveas error
Hello Bala On 21.09.11 07:37, bala suru wrote: I have used a some simple way to copy the image running Vm image(insted of onevm saveas) , 1. I did some modification over the running image(creating some files) 2. issued onevm suspend 3. copied var/images/uuid file of the running image 4. Registered the above copied image as a new image 5 . launched the new VM using above image But I could not see any modified files on this VM ,,? This could have two reasons. First, is this a persistent image? Second, did the VM have enough time to write back the changes to the image? But I guess the second one is difficult to find out, as the OS in the VM and probably also the VM layer do some caching in memory and the write back to the image is delayed for a longer time. Are the steps which I followed are correct ..? I do not think, that this is really supported from ONE. You are taking an image away from a running VM (even in the state suspended), which could have an not cleanly saved state of the file system in the image. It is much better to use an OS and VM which does support ACPI and then use the 'onevm saveas vm_id disk_id image_name' and 'onevm shutdown vm_id' commands. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] changing virtual network online
Hello Samuel On 21.09.11 14:45, samuel wrote: I've just wondering whether is it possible to change the virtual network that a virtual machine is attached once it has been working (deploy-run). With 'onevnet' you can change some settings of an network, eg. leases / mac addresses. But as I understand, you try to modify stuff which belongs to the VM itself. I've tried to modify the deployment.0 file but it did not affect the new restarted machine. Might be a problem with the underlying MySQL database that has to be also changed? As far as I know, it is currently not supported to change an already running VM. Use Case: *create a new virtual machine and just forgot to attach a virtual network *modify a virtual network (in case of VLAN change) *attach a new interface to a running machine. The steps which could work, and will give minimum downtime are probably the following: - update your VM template with the new / changed network / NIC - run 'onevm shutdown vm_id' (with a registered persistent disk image, all the modifications in the VM are preserved). - wait until shutdown has completed, check with 'onevm list' or better 'onevm top' - run 'onevm create template' and the same VM will boot up (now with a different VM ID, but with the same registered persistent disk image) bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Error monitoring host
Hello Humberto On 21.09.11 17:21, Humberto N. Castejon Martinez wrote: Wed Sep 21 17:03:44 2011 [InM][I]: Command execution fail: 'if [ -x /var/tmp/one/im/run_probes ]; then /var/tmp/one/im/run_probes kvm joker; else $ Wed Sep 21 17:03:44 2011 [InM][I]: STDERR follows. Wed Sep 21 17:03:44 2011 [InM][I]: Permission denied, please try again. Wed Sep 21 17:03:44 2011 [InM][I]: Permission denied, please try again. Wed Sep 21 17:03:44 2011 [InM][I]: Permission denied (publickey,password). Wed Sep 21 17:03:44 2011 [InM][I]: ExitCode: 255 Wed Sep 21 17:03:44 2011 [InM][E]: Error monitoring host 0 : MONITOR FAILURE 0 Could not monitor host joker. I guess the ssh login from the front end to the cluster node with the user oneadmin does not work. Try manually from the front end to use 'ssh -v oneadmin@joker' (the -v gives some verbose output, if you increase eg. with -vvv, it will give even more). It guess on the cluster node the file ~oneadmin/.ssh/id_dsa.pub is missing (or not readable for oneadmin). Check also the sshd log file on the cluster node. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] onevm saveas error
Hello On 15.09.2011 07:39, bharath pb wrote: I tried to save the running Vm image using onevm saveasvm_id dick_id imagename onevm shutdownvm_id Fri Sep 9 10:11:46 2011 [TM][E]: Error excuting image transfer script: cp: cannot stat `/srv/cloud/one/var//images/e0561a492c9aaac280479f2f0d85dcced9156fbf': No such file or directory Does the oneadmin user have write permissions in that path? Which path have you set in the VM_DIR option in oned.conf? What does 'oneimage list' and 'oneimage show id' say? Fri Sep 9 10:11:47 2011 [DiM][I]: New VM state is FAILED Fri Sep 9 10:11:47 2011 [TM][W]: Ignored: LOG - 44 tm_delete.sh: Deleting /srv/cloud/one/var//44/images I guess the registered images (which also happens when you run 'onevm saveas ...') should be somewhere else and not in the var folder, which is used by the running VMs. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Non clonable readonly shared OS image?
Hello Ismael On 07.09.2011 19:15, Ismael Farfán wrote: 2011/9/7 Roger Pau Monné: I don't think it's possible to launch an OS from a single image multiple times... Actually I can, the problem is that the contextualization needs to modify the hostname, interfaces, passwd, shadow... since many VMs do that it corrupts the FS (actually, only those files, I think). Usually every Unix like OS does write to the disk, mostly in /var/ eg. log, pid and lock files. Now when you have several VMs writing to the same disk based file system during the same time, it will corrupt the file system. There is no clean file locking available when a file system on a disk (or image) is mounted from more then one running OS. I guess your only chance is to create a Live CD based on your own installation (your distribution should provide tools to do this), then it should work, as an OS bootet from Live CD does create union mounts with a RAM disk, so writing to the file system is going into RAM and not written back to the disk (in this case the ISO image). Mainly I was wondering why with readonly=yes the VM doesn't boot at all, it fails before even calling kvm and I can't figure out why. I guess this is not supported from ONE if an OS image (not CD image) is set to read only, and it will abort at an early stage. Did you see anything in the log/one/VID.log file? I'll try modifying fstab to mount / as readonly and set readonly=no in the VM description file, maybe that'll work. I guess your OS will then not work properly and have some other strange problems, if the / is read only. 2011/9/7 Matthew Smith: Hi Have you tried doing this with a 'live CD' distribution image (which is by its nature normally a read only boot)? I haven't. The idea is that my OS image works as a live CD since I need to install some random stuff until it works as a virtual cluster. During the first phase (as long as you need to setup your system), just create a persistent disk image and run a single VM with it. When you have everything installed as needed, use the tools from your distribution to create a Live CD. Eventually you need to start your installation based on the Live CD provided from your distribution. When you have created your Live CD you have to register this ISO image in ONE. And then you can create several VMs which use the registered ISO image (as CD) as the boot file system. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Non clonable readonly shared OS image?
Hello Ismael On 08.09.2011 20:03, Ismael Farfán wrote: 2011/9/8 Fabian Wenk: Mainly I was wondering why with readonly=yes the VM doesn't boot at all, it fails before even calling kvm and I can't figure out why. I guess this is not supported from ONE if an OS image (not CD image) is set to read only, and it will abort at an early stage. Did you see anything in the log/one/VID.log file? I attached the log (almost the same as in the firs mail), whatever ONE does different while setting the image as readonly makes libvirt trow this error that I haven't been able to fix jet: failed to retrieve chardev info in qemu with 'info chardev' Google [42] does find many reports with this particular message, maybe you can find a hint there. [42] http://www.google.com/search?q=%22failed%20to%20retrieve%20chardev%20info%20in%20qemu%20with%20%27info%20chardev%27%22 According to your attached log file it fails when ONE tries to start kvm, probably because kvm does not support booting from a hd image which is set to read only. I guess all this parameters are delivered through libvirt to kvm. So kvm is able to honor such parameters as read only disk from the hardware point of view, so the OS in the VM does not have any possibility to write to the disk. I have seen similar errors with my first steps with ONE, which most often where of no real help. But sometimes I found a hint in the libvirt log file. I do not remember if I tried it with a hd image which was set to read only. bye Fabian ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org