- Users - libvirt List Archives

[libvirt-users] serial console locking for a domain (concurrent access = garbled output)
by Dusty Mabe 01 May '12

01 May '12

Hi, I am seeing an issue in libvirt-0.9.4-23.el6_2.7.x86_64.rpm where my virsh console output for a domain gets garbled because I have virt-manager running and it is also trying to access the serial device for the guest domain. Basically there is concurrent serial access and half the text ends up on one end while the rest ends up on the other end. Closing virt-manager does not help. Disconnecting and reconnecting using virsh console does not help. The only way to recover is to destroy the domain and then start the guest again and never start virt-manager. Note that the domain does not have VNC graphics so virt-manager defaults to try to access the serial port when started for a particular domain. Should there be locking within libvirt to only allow for a single client to access a specific serial port of a domain? Thanks, Dusty Mabe

2 2

[libvirt-users] Console to RHEL6.1 container
by Sukadev Bhattiprolu 01 May '12

01 May '12

We are trying to create RHEL6.1 container and having trouble getting a proper console to the container. I have used lxc containers in the past and lxc has a script (lxc-fedora) to setup a basic Fedora container. Is there something similar or docs with virtmgr ? We started out making a copy of the root fs and using that copy as the root for the container. I made a couple of tweaks to the container rootfs: 1. ensured that /dev/ptmx is a symlink to pts/ptmx 2. Added 'newinstance' option to the devpts mount in fstab The container xml file is attached. If devpts entry in container's fstab has the 'newinstance' option, I get a brief "domain test01 started" message on the stdout. When I run 'virsh consle test01', I only see a few messages --- Setting hostname node03: [ OK ] Setting up Logical Volume Management: No volume groups found [ OK ] Checking filesystems [ OK ] mount: can't find / in /etc/fstab or /etc/mtab Mounting local filesystems: [ OK ] Enabling local filesystem quotas: [ OK ] Enabling /etc/fstab swaps: [ OK ] --- and then a hang wihtout a shell. At ths point sshd or any of the other services don't appear to be running. If I remove the 'newinstance' entry in fstab and run 'virsh start test01' the container starts up ok. I get all "console" messages on whichever is my current /dev/pts/0 ! At this point sshd and many other services are running and I can ssh in. Even in this case, though 'virsh console test01' shows the same _few_ messages as above and no login shell. Back to the newinstance case, by putting some debug messages in the container's /etc/rc.sysinit, I see that devpts is already mounted with newinstance. With lxc, in setup_pts(), we unmount devpts in the container and then mount with newinstance. I wonder what that sequence is with virtmgr. I double checked that /dev/ptmx in container is a symlink to pts/ptmx and that CONFIG_DEVPTS_MULTIPLE_INSTANCES=y. Appreciate any tips on how I can get access to the console of the RHEL6.1 container. Suka Suka

1 1

[libvirt-users] question of iSCSI pool on the same iSCSI target
by david_zhu 28 Apr '12

28 Apr '12

3 2

[libvirt-users] Is there a way for telling libvirt to auto define an xml under /etc/libvirt/qemu?
by Raoul Scarazzini 26 Apr '12

26 Apr '12

Hi everybody, like in the subject: is there a way for telling libvirt to auto define an xml under /etc/libvirt/qemu? I mean, sometimes it happens to have the /etc/libvirt/qemu shared between nodes, say with a NFS share, since I can set a new xml on a node and I want o see the virtual machine on all the others, is there a way to automate this or am I obligated to launch virsh define <xml> on EACH node? Thank you so much, -- Raoul Scarazzini Solution Architect MMUL: Niente è impossibile da realizzare, se lo pensi bene! +39 3281776712 rasca(a)mmul.it http://www.mmul.it

4 3

[libvirt-users] Installing the python libvirt bindings in a virtualenv
by Daniel Gonzalez 26 Apr '12

26 Apr '12

Hi, I have a virtualenv which I am using for production (I need python 2.7 and certain libraries which can not be easily installed on the host environment). Now I am trying to install also libvirt, but I have not succeeded yet. The problem that I have now looks quite difficult to solve. This is the script I am using to install libvirt (bash script): install_libvirt_ ( ) { local myvirtualenv="$1" local libvirt_tag=v0.9.9 cd $TMP_DIR git clone git://libvirt.org/libvirt.git cd libvirt git checkout $libvirt_tag mkdir -p $VIRTUALENVS_DIR/$myvirtualenv/usr sudo apt-get install autopoint # configure: error: You must install the GnuTLS library in order to compile and run libvirt sudo apt-get install -y gnutls-bin gnutls-dev # configure: error: You must install device-mapper-devel/libdevmapper >= 1.0.0 to compile libvirt sudo apt-get install -y libdevmapper-dev libdevmapper # configure: error: You must install python-devel to build Python bindings sudo apt-get install -y python-all-dev ./autogen.sh --prefix=$VIRTUALENVS_DIR/$myvirtualenv/usr --enable-compile-warnings=error make make install } But this is failing with error message: checking for python script directory... ${prefix}/lib/python2.7/site-packages checking for python extension module directory... ${exec_prefix}/lib/python2.7/site-packages configure: error: You must install python-devel to build Python bindings python-devel does not exist. I have used python-all-dev, but it has not solved the problem. Has somebody succeeded in installing libvirt-python inside a virtualenv? Thanks, Daniel Gonzalez

2 2

[libvirt-users] Snapshot system: really confusing.
by NoxDaFox 25 Apr '12

25 Apr '12

Greetings, I am developing a platform for test and debug use. The typical scenario I want to realize is the following: I run a .qcow2 image via QEMU-KVM, the running OS is not relevant. I want to take several snapshots of both Disk and RAM status; same file, separate files doesn't matter. I just need a way to have consistent information about them (better with XML description) and the data as well. I want to store these snapshots on my storage devices to use them whenever and wherever I want. Is it possible to store a single snapshot providing both the memory and disks state in a file (maybe a .qcow2 file)? Is there any way to get a unique interface which handles my snapshots? I was used to use the virDomainSnapshotCreateXML() defining the destination file in the XML with <disk> fields. After updating libvirt it was not working anymore, I thought was a bug but then I realized it was intentional. The function complains about the fact that the <disk> parameter is not accepted anymore. So I started guessing how to solve reading the API documentation and I fall in a completely nebulous world. For what I got: - virDomainSnapshotCreateXML(): According to flags can take system checkpoints (really useful) and disks snapshots. System checkpoints: What I need but I didn't find any way to retrieve the storage file; I'm only able to get the snapshot pointer, quite useless as from its pointer I can only print the XML description. Disk snapshot: here through the XML description I can specify a target file where to write the information. But I still need the memory status. - virDomainSaveFlags(): No matter how I set the flags, it halts the guest; I need it still running afterward and I think that reverting the domain and restarting for my scenarios is unaffordable. - virDomainCoreDump(): Does what I need for the memory, but doesn't give any XML description and doesn't supply any useful interface to handle it, I just get back an int that represent the exit status of the routine. (other functions really similar) The question is: why all this confusion? I absolutely understand the problematic that realizing a multiplatform snapshots management raises; but I think that for an API purpose what is implemented here is completely confusing the developer. Regards, NoxDaFox.

2 1

[libvirt-users] Non-Root not able to use openReadOnly()
by vipul borikar 24 Apr '12

24 Apr '12

Hi, I have a code : import libvirt con = libvirt.openReadOnly(None) print con.listDomainsID() When i run the above code using root i get the proper output... but when i use non-root user it givers be a empty list [ ] If i change the 2nd line to con = libvirt.open(qemu:///system) It works. I am using RHEL 6.1 with default kvm and libvirt. unix_sock_group = "virt" in libvirtd.conf where virt is a group of non-root user group = "virt" in qemu.conf is set. Any Ideas???? -- Thanks Vipul Borikar "Our task must be to free ourselves...by widening our circle of compassion to embrace all living creatures and the whole of nature and its beauty."

2 1

[libvirt-users] Creating Domain and moving a kvm disk under libvirt control
by James Allsopp 20 Apr '12

20 Apr '12

Hi, I'm trying to start using libvirt on my debian machine, but am having some issues. Regardless of using command-line or gui interface with virt-manager, building the virtual machine stalls on 0 with no indication of why it has stopped working. How do I go about fixing this or diagnosing this problem. Also I've a .img file that work s with KVM, would there be anyway to import it for use with libvirt? Thanks James

2 2

[libvirt-users] Using Sys::Virt with Libvirt
by John Wayne 20 Apr '12

20 Apr '12

Anyone using Sys::Virt CPAN Perl Module with Libvirt? I am unable to install Sys::Virt using cpan, Sys::Virt cannot find Libvirt, even though I have it installed. I tried setting LD_LIBRARY_PATH and also PKG_CONFIG_PATH to point to Libvirt, and in both case Sys::Virt was not able to find Libvirt. During installation Sys::Virt times out with : Sys-Virt-0.9.4/INSTALL CPAN: File::Temp loaded ok (v0.22) CPAN.pm: Going to build D/DA/DANBERR/Sys-Virt-0.9.4.tar.gz libvirt >= 0.9.4 is required Warning: No success on command[/usr/bin/perl Makefile.PL INSTALLDIRS=site] Warning (usually harmless): 'YAML' not installed, will not store persistent state DANBERR/Sys-Virt-0.9.4.tar.gz /usr/bin/perl Makefile.PL INSTALLDIRS=site -- NOT OK Running make test Make had some problems, won't test Running make install Make had some problems, won't install Could not read '/root/.cpan/build/Sys-Virt-0.9.4-DiRtOZ/META.yml'. Falling back to other methods to determine prerequisites Failed during this command: DANBERR/Sys-Virt-0.9.4.tar.gz : writemakefile NO '/usr/bin/perl Makefile.PL INSTALLDIRS=site' returned status 256 I have tried with both Libvirt 0.9.11 and Sys::Virt 0.9.11 in addition to Libvirt 0.9.4 with Sys::Virt 0.9.4. Thanks

1 0

[libvirt-users] Error on running libvirt (dnsmasq)
by John Wayne 20 Apr '12

20 Apr '12

Hi, I am running libvirt on RHEL 6.2 and getting this below error when I start the daemon: /usr/sbin/libvirtd --verbose 2012-04-20 06:00:13.366+0000: 26900: info : libvirt version: 0.9.11 2012-04-20 06:00:13.366+0000: 26900: error : virExecWithHook:424 : Cannot find 'dnsmasq' in path: No such file or directory 2012-04-20 06:00:13.405+0000: 26900: warning : ebiptablesDriverInit:4066 : Could not find 'ebtables' executable 2012-04-20 06:00:13.419+0000: 26900: error : virCommandWait:2306 : internal error Child process (/bin/sh -c IPT=/sbin/ip6tables cmd='$IPT -n -L FORWARD' eval res=\$$"${cmd} 2>&1"$ if [ $? -ne 0 ]; then echo "Failure to execute command '${cmd}' : '${res}'."; exit 1;fi ) status unexpected: exit status 1 2012-04-20 06:00:13.419+0000: 26900: error : ebiptablesDriverInit:4101 : Testing of ip6tables command failed: Failure to execute command '$IPT -n -L FORWARD' : 'WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. ip6tables v1.4.7: can't initialize ip6tables table `filter': Address family not supported by protocol Perhaps ip6tables or your kernel needs to be upgraded.'. I compiled libvirt from the source, do I really need dnsmasq? Thanks

3 2