- Users - libvirt List Archives

[libvirt-users] does kvm support SATA driver?
by lisiyang 16 Aug '11

16 Aug '11

Hi, I encounter a problem. My virtual machine is windows 7 . I install it using bus ide. There is not any problem until now. but when I want to attach a disk to it(bus=scsi ) , the os can’t find the suitable SCSI driver. I search the problem on the internet ,it seems that the windows7 don’t support SCSI driver. So ,I want attach the disk to vm using SATA. So I want to know that whether the kvm support SATA driver and whether the kvm support USB 2.0? Hope for reply. Thank in advance. Thanks & Best Regards, ---------------------------------------------------- Li, Si Yang(李思扬) Beijing University of Posts and Telecommunications , Beijing Internet e-mail:lisiyang_best@sina.com Address: Road 10 Xitucheng, Haidian District, Beijing 100876 ----------------------------------------------------

2 1

[libvirt-users] The saga continues ....
by William A. Mahaffey III 15 Aug '11

15 Aug '11

.... I have a 64-bit CentOS 5.6 VM running on a 64-bit FC14 host. Actually, I *HAD* a 64-bit CentOS 5.6 VM :-) .... After a read of the virt-install man page, I concluded that I could lose the current VM (but *NOT* its HDD image) & recreate it fairly easily w/ the right combo of arguments to virt-install. The proceedings: [root@Q6600:/etc, Mon Aug 15, 04:30 PM] 1163 # virsh list Id Name State ---------------------------------- 1 CentOS-5.6-64bitVM running [root@Q6600:/etc, Mon Aug 15, 04:30 PM] 1164 # virsh destroy CentOS-5.6-64bitVM Domain CentOS-5.6-64bitVM destroyed [root@Q6600:/etc, Mon Aug 15, 04:30 PM] 1165 # virsh list Id Name State ---------------------------------- [root@Q6600:/etc, Mon Aug 15, 04:30 PM] 1166 # virsh undefine CentOS-5.6-64bitVM Domain CentOS-5.6-64bitVM has been undefined [root@Q6600:/etc, Mon Aug 15, 04:30 PM] 1167 # virsh list Id Name State ---------------------------------- [root@Q6600:/etc, Mon Aug 15, 04:31 PM] 1169 # virt-install -n CentOS-5.6-64bitVM -r 2048 --arch=x86_64 --cpu core2duo,disable=vmx --vcpus=2,maxvcpus=2,sockets=1,cores=2,threads=1 --os-type=linux --os-variant=rhel5 --import --disk path=/home/VMs/CentOS.5.6-64bit-VM/disk.img,format=raw --autostart --prompt This will overwrite the existing path '/home/VMs/CentOS.5.6-64bit-VM/disk.img' Do you really want to use this disk (yes or no) no What would you like to use as the disk (file path)? quit How large would you like the disk (quit) to be (in gigabytes)? ^CInstallation aborted at user request [root@Q6600:/etc, Mon Aug 15, 04:31 PM] 1170 # i.e. I chickened out at the ominous-sounding message about wiping out my hard-won disk image .... I thought (from the man-page) that the '--import' command would skip any attempts at installing anything, define the VM from the arguments supplied, & boot it from the already-prepped disk image .... Am I right here ? If not, is there a recommended way to re-create my VM (mostly) from the extant disk image w/o having to redo whatever initialization I have done so far (setup a user, install compilers & dev.env., install Intel compiler, etc.) ? Also, is there a recommended way for the host to copy that disk image somewhere so I can try to reuse it if my future attempts at re-creation don't pan out ? TIA .... -- William A. Mahaffey III ---------------------------------------------------------------------- "The M1 Garand is without doubt the finest implement of war ever devised by man." -- Gen. George S. Patton Jr.

1 0

[libvirt-users] Non-persistent disk
by Xianqing Yu 15 Aug '11

15 Aug '11

Hi Libvirt team, I am trying to create a new domain from XML file by using libvirt. I want this new domain using non-persistent disk as its disk. Non-persistent disk means that when the domain is running, all the modification would not be written to disk file. So next time, when you boot VM from the same disk file, you would have a fresh disk. I check the website but I could not find anything about this. I wonder if libvirt support this function. FYI, if I use KVM, I can use command "qemu-kvm -snapshot" to boot a VM on non-persistent disk. If I use VMware, I can set disk type to Independ-disk, non-persistent. Thanks, Xianqing Yu ------ Graduate Research Assistant, Cyber Defense Lab Department of Computer Science North Carolina State University, Raleigh, NC E-mail: xyu6(a)ncsu.edu

2 1

[libvirt-users] How to send a break using "virsh console"??
by David Knierim 15 Aug '11

15 Aug '11

I have a KVM guest running on CentOS 5.6 that is in a bad state. I would like to dump the kernel information using magic-sysrq commands, but I cannot figure out how to send a break on the virtual serial port provided by "virsh console". The guest is a standard install that will respond if I can send a break on the virtual serial port. Is it possible to send a break on the virtual serial port? Thanks, David

1 0

[libvirt-users] Ongoing VM saga ....
by William A. Mahaffey III 15 Aug '11

15 Aug '11

.... I have a 64-bit CentOS 5.6 VM running on a 64-bit FC-14 host. I was originally using the stock NAT networking on the guest, which could access the host, but nothing else on the network. I eventually found some links on how to setup routed networking on the guest using a specifically setup bridge. I followed the example a bit too closely, bridging my eth0 interface, which is the one I nornally use to access the host over my LAN. So eth0 temporarily ceased functioning & I had to undo my handiwork through the console on the host, which I virtually never use. I restarted networking manually a couple of time on the host, but eth0 was still fried. I then rebooted the host (still through the console) & that went AOK, took about 5 min. & all was well, eth0/SSH LAN access was back & I logged back in & tried to log into the VM to plot my next course of action & .... The VM address had moved. I had set it up to use static addressing (modified the ifcfg-eth0 on the VM & restarted networking, 2 or 3 times yesterday, installed compilers, etc. from the host, & all was well until I messed around w/ trying the roll-your-own bridge this A.M.). Which brings me to the real question: How do I figure out the IP address the VM is using so I can login & continue the festivities ? I found it yesterday through the VM (virt-viewer) console after setting up the VM the 1st time, but that doesn't work today either (black screen inside the viewer, no response to regular or ctrl-keys, etc.) :-/ .... 'virsh .... console' hangs virsh, etc. :-) .... Soooooo .... how do I recover that $&#&^$&#* IP address for the VM :-) ? TIA .... -- William A. Mahaffey III ---------------------------------------------------------------------- "The M1 Garand is without doubt the finest implement of war ever devised by man." -- Gen. George S. Patton Jr.

1 0

[libvirt-users] How to setup routed networking in CLI ....
by William A. Mahaffey III 14 Aug '11

14 Aug '11

.... I have a 64-bit CentOS 5.6 VM running on a 64-bit FC14 host. I have the stock networking (NAT), but I need routed to be able to access other boxen on my LAN. I looked at the docs on the libvirtd website, but their example of how to setup routed networking is for the GUI, which doesn't work for me (text in buttons, etc. doesn't display, no gfx card on host). The man pages are pretty good, but lack pointers on the myriad options & their interaction for networking & are a bit short on examples. Any hints on doing that (setup routed networking under CLI) for me ? URL's, etc. ? TIA .... -- William A. Mahaffey III ---------------------------------------------------------------------- "The M1 Garand is without doubt the finest implement of war ever devised by man." -- Gen. George S. Patton Jr.

1 0

[libvirt-users] VNC ports
by Jeff 13 Aug '11

13 Aug '11

I am needing to explicitly assign VNC ports to my virtual guests so that I can connect to them externally via a VNC client, and not virt-viewer. Since I will have multiple KVM Hosts with access to the same shared storage (to facilitate migrations) I need to ensure I have unique VNC ports assigned to each guest so that I don't overlap somehow. I had planned to use a range of ports starting at 51000 for this purpose. While it works to assign port 51001, for example, to a guest and connect via my VNC client, when I run 'virsh vncdisplay guestname' it gives me back a different number that is specified in the XML file: #virsh vncdisplay win2k8guest :45101 #virsh dumpxml win2k8guest |grep vnc <graphics type='vnc' port='51001' autoport='no' /> Is there a logical reason for this that I'm not getting? I know when I have the port set to 5907, for example, it returns something expected: #virsh vncdisplay win2k8guest :7 But since I am likely going to need a good number of ports, I want to use a non-registered port range. Any insights?

3 3

[libvirt-users] How do I disable auto-logout ....
by William A. Mahaffey III 13 Aug '11

13 Aug '11

.... I have a 64-bit CentOS 5.6 VM running on a 64-bit FC14 host. I notice that it has an automatic logout feature after a period of idleness, see below: [wam@centos-5, CFD, 5:52:45pm] 1183 % ll -tr LIST LIST.Update /usr/local/bin/P* /usr/local/bin/*WAM* ; ps -u wam -lf ; uname -a ; date ls: No match. F S UID PID PPID C PRI NI P SZ:RSS WCHAN STIME TTY TIME CMD 5 S wam 2270 2266 0 15 0 * 22540:1788 ? 15:48 ? 00:00:04 sshd: wam@pts/0 0 S wam 2271 2270 0 15 0 * 3144:2720 rt_sig 15:48 pts/0 00:00:00 -tcsh 0 R wam 18249 2271 0 15 0 0 2621:896 - 17:52 pts/0 00:00:00 ps -u wam -lf Linux centos-5.6-vm 2.6.18-238.el5 #1 SMP Thu Jan 13 15:51:15 EST 2011 x86_64 x86_64 x86_64 GNU/Linux Fri Aug 12 17:52:50 CDT 2011 [wam@centos-5, CFD, 5:52:50pm] 1184 % auto-logout Connection to 192.168.122.213 closed. [wam@Q6600, ~, 6:52:51pm] 1014 % [wam@Q6600, ~, 9:40:49pm] 1014 % [wam@Q6600, ~, 9:40:49pm] 1014 % ssh2 -l wam 192.168.122.213 Last login: Fri Aug 12 15:48:11 2011 from 192.168.122.1 [wam@centos-5, ~, 9:40:54pm] 1135 % I checked man pages on both host & VM, but came up blank. I have another box running CentOS 5.6 native & it doesn't do this, so I suspect it is something w/ the VM environment. How do I disable that :-) ? Thanks in advance. -- William A. Mahaffey III ---------------------------------------------------------------------- "The M1 Garand is without doubt the finest implement of war ever devised by man." -- Gen. George S. Patton Jr.

2 2

[libvirt-users] Problems setting up a VM ....
by William A. Mahaffey III 12 Aug '11

12 Aug '11

.... New to this list & definitely new to virtualization :-/ .... I am trying to setup a 64-bit CentOS 5.6 VM on a 64-bit FC14 server host, all patched up as of a couple of weeks ago. uname -a on the host shows: [wam@Q6600, CFD, 9:24:23am] 540 % uname -a Linux Q6600 2.6.35.11-83.fc14.x86_64 #1 SMP Mon Feb 7 07:06:44 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux [wam@Q6600, CFD, 9:52:43am] 541 % I access the server by ssh on my LAN. I was working on this last Sunday & got stuck, see below. I posted this question to a couple of other lists & had no takers, so here I am here. The server runs in runlevel 3 & has no gfx card onboard, so I am working from the command line. I had a couple of abortive attempts at setup, which got as far as initializing my disk storage for the VM, then the following: [root@Q6600:/etc, Sun Aug 07, 05:08 PM] 1339 # virt-install -n CentOS-5.6-64bitVM -r 2048 --arch=x86_64 --cpu core2duo,disable=vmx --vcpus=1,maxvcpus=1,sockets=1,cores=2,threads=1 --os-type=linux --os-variant=rhel5 --disk path=/home/VMs/CentOS.5.6-64bit-VM/disk.img,size=50 --cdrom /home/ISOs/CentOS-5.6/64bit/CentOS-5.6-x86_64-bin-DVD-1of2.iso --autostart --prompt This will overwrite the existing path '/home/VMs/CentOS.5.6-64bit-VM/disk.img' Do you really want to use this disk (yes or no) yes Starting install... Creating domain... | 0 B 00:00 Gtk-Message: Failed to load module "pk-gtk-module": libpk-gtk-module.so: cannot open shared object file: No such file or directory Gtk-Message: Failed to load module "canberra-gtk-module": libcanberra-gtk-module.so: cannot open shared object file: No such file or directory (virt-viewer:20165): Pango-WARNING **: failed to choose a font, expect ugly output. engine-type='PangoRenderFc', script='latin' Domain installation still in progress. You can reconnect to the console to complete the installation process. [root@Q6600:/etc, Sun Aug 07, 05:14 PM] 1339 # virsh --connect qemu:///system Welcome to virsh, the virtualization interactive terminal. Type: 'help' for help with commands 'quit' to quit virsh # list Id Name State ---------------------------------- 11 CentOS-5.6-64bitVM running virsh # help The install began OK, right after the 'Pango-WARNING' message, in a separate window, quite readable, started booting & scrolling text past *FAST*, for about 15 sec., then hung, right after a line about 'serial port ....' (I couldn't click-&-drag to capture text from the VM window, going from memory). I let it sit for about 5 min. to be sure it was really hung, then clicked the VM window out, whereupon the message about 'installation still in progress' & the rest appeared. I poked around in virsh a bit as you can see, but couldn't figure anything out. I tried (re)connecting to the console of the VM & that hung, I had to kill the virsh command from another window. I attach syslog output from the proceedings. I am *sure* this is pilot error, but I am stuck for now. Any assistance appreciated :-) .... -- William A. Mahaffey III ---------------------------------------------------------------------- "The M1 Garand is without doubt the finest implement of war ever devised by man." -- Gen. George S. Patton Jr.

1 1

[libvirt-users] virStoragePoolRefresh() returns before or after all new vols on path discovered?
by Gareth S Bestor 12 Aug '11

12 Aug '11

Basically, does virStoragePoolRefresh() api 'block' until the pool has refreshed itself (eg added any/all new volumes found in the pool path) or can it return early before the refresh is fully 'done' ? - G Dr. Gareth S. Bestor IBM Senior Software Engineer Systems & Technology Group - Systems Management Standards 971-285-6375 (mobile) bestor(a)us.ibm.com

2 1