Hi I may have found the issue with image name in libvirt. I don't know if it breaks anything else in libvirt (i.e. update image). Once I made following change, I'm able to capture new image AND reload newly captured image to a computer with no issues. Here is diff file: (also attached).
# diff /tmp/libvirt.pm-RC3 Module/Provisioning/libvirt.pm 3c3 < # $Id$ --- > # $Id: libvirt.pm 1341477 2012-05-22 13:47:34Z fapeeler $ 325,326c325,327 < my $new_image_name = $self->get_new_image_name(); < $self->data->set_image_name($new_image_name); --- > my $new_image_name = $image_name; > #my $new_image_name = $self->get_new_image_name(); > #$self->data->set_image_name($new_image_name); I still cannot make reservation for KVM-based image. Getting 'Available Times' dialog box every time when selecting KVM-based image. I think it has something todo with imagetype. It's 29, which is bare-metal, using partimage, i.e. xCAT. -- Thank you, Dmitri Chebotarov Virtual Computing Lab Systems Engineer, TSD - Ent Servers & Messaging 223 Aquia Building, Ffx, MSN: 1B5 Phone: (703) 993-6175 Fax: (703) 993-3404 On Friday, June 8, 2012 at 12:08 , Dmitri Chebotarov wrote: > Hi > > > It looks like image name changes during image capture process. The capture > starts with one image name and finishes with a different image name: > > > start: > > VCL IMAGE Creation Started: centos5image-LinuxCentOS62KVMv399-v0 > > finish: > > VCL IMAGE Creation Completed: centos5-LinuxCentOS62KVMv3-99-v0 > > Imagename in 'vcl.image' table and 'vcl.imagerevision' table is different. > Once I update the name in 'vcl.imagerevision' table to match the name in > 'vcl.image', I can reload newly captured image fine. > > > I.e.: > > SELECT imagename FROM imagerevision where imageid=99 and production=1 > centos5-LinuxCentOS62KVMv3-99-v0 SELECT name FROM image where id=99 > centos5image-LinuxCentOS62KVMv399-v0 > > I'm attaching image capture log file. > > > I've tried to trace where the image name changes, but cannot find it in the > code (libvirt.pm). > > > > > -- > Thank you, > > Dmitri Chebotarov > Virtual Computing Lab Systems Engineer, TSD - Ent Servers & Messaging > 223 Aquia Building, Ffx, MSN: 1B5 > Phone: (703) 993-6175 > Fax: (703) 993-3404 > > > On Thursday, June 7, 2012 at 17:15 , Dmitri Chebotarov wrote: > > > Hi > > > > Have anyone else tried Libvirt support in VCL 2.3? > > > > I'm getting this error, not sure if it's specific to my configuration (# > > svn up At revision 1347799) > > > > During base image capture VCL creates two files per VM in "Virtual Disk > > Path" folder (I'm using qcow2 disk format): > > > > qemu-img convert "/var/lib/libvirt/images/Linux_CentOS_62.img" -O qcow2 > > "/var/lib/libvirt/images/baseImages/centos5image-linuxkvm898-v0.qcow2" > > two files created: > > > > centos5image-linuxkvm898-v0.qcow2 > > centos5-linuxkvm8-98-v0.xml > > > > The content of centos5-linuxkvm8-98-v0.xml is: > > > > # cat centos5-linuxkvm8-98-v0.xml HASH(0x154d9a50) > > > > When I try to reload a computer with new image, VCL is trying to use .xml > > file as disk image and reload fails. > > > > Once I delete .xml and rename image file, it works fine: > > > > rm -f centos5-linuxkvm8-98-v0.xml > > mv centos5image-linuxkvm898-v0.qcow2 centos5-linuxkvm8-98-v0.qcow2 > > > > > > Llog file: > > > > |27378|484:476|reload|utils.pm:insertloadlog(3701)|inserted computer=98, > > loadimageblade, sandbox-vcl-20 must be reloaded with > > centos5-linuxkvm8-98-v0 2012-06-07 > > 16:42:46|27378|484:476|reload|new.pm:reload_image(616)|calling > > VCL::Module::Provisioning::libvirt->does_image_exist() 2012-06-07 > > 16:42:46|27378|484:476|reload|OS.pm:find_files(2801)|attempting to find > > files on BC1-12, base directory path: > > '/var/lib/libvirt/images/baseImages/', pattern: centos5-linuxkvm8-98-v0.*, > > command: /usr/bin/find "/var/lib/libvirt/images/baseImages/" -type f -iname > > "centos5-linuxkvm8-98-v0.*" 2012-06-07 > > 16:42:46|27378|484:476|reload|OS.pm:find_files(2829)|files found: 1, base > > directory: '/var/lib/libvirt/images/baseImages/', pattern: > > 'centos5-linuxkvm8-98-v0.*' |27378|484:476|reload| command: '/usr/bin/find > > "/var/lib/libvirt/images/baseImages/" -type f -iname > > "centos5-linuxkvm8-98-v0.*"', output: |27378|484:476|reload| > > /var/lib/libvirt/images/baseImages/centos5-linuxkvm8-98-v0.xml 2012-06-07 > > 16:42:46 |27378|484:476|reload|libvirt.pm:get_master_image_file_path(1011)|found master image file on BC1-12: /var/lib/libvirt/images/baseImages/centos5-linuxkvm8-98-v0.xml 2012-06-07 16:42:46|27378|484:476|reload|Module.pm:new(203)|VCL::Module::Semaphore object created, address: 2179cfe0 2012-06-07 16:42:46|27378|484:476|reload|Module.pm:code_loop_timeout(879)|attempting to open lockfile, maximum of 7200 seconds 2012-06-07 16:42:46|27378|484:476|reload|Semaphore.pm:open_lockfile(175)|opened and obtained an exclusive lock on file: /tmp/master-centos5-linuxkvm8-98-v0.semaphore 2012-06-07 16:42:46|27378|484:476|reload|Semaphore.pm:open_lockfile(182)|wrote to file: /tmp/master-centos5-linuxkvm8-98-v0.semaphore, contents: |27378|484:476|reload| '27378 vcld 484:476 reload sandbox-vcl-20>BC1-12 centos5-linuxkvm8-98-v0 vclreload' 2012-06-07 16:42:46|27378|484:476|reload|Module.pm:code_loop_timeout(899)|attempting to open lockfile, code returned true 2012-06-07 16:42:46|27378|484:476|reload|Module.pm :get_semaphore(977)|created 'master:centos5-linuxkvm8-98-v0' Semaphore object, memory address: 2179cfe0 2012-06-07 16:42:46|27378|484:476|reload|utils.pm:run_ssh_command(4928)|executing SSH command on BC1-12: |27378|484:476|reload| /usr/bin/ssh -i /etc/vcl/vcl.key -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o ConnectionAttempts=1 -o ConnectTimeout=3 -l root -p 22 -x BC1-12 'stat \/var\/lib\/libvirt\/images\/baseImages\/centos5\-linuxkvm8\-98\-v0\.xml' 2>&1 2012-06-07 16:42:46|27378|484:476|reload|utils.pm:run_ssh_command(5046)|run_ssh_command output: |27378|484:476|reload| File: `/var/lib/libvirt/images/baseImages/centos5-linuxkvm8-98-v0.xml' |27378|484:476|reload| Size: 16 Blocks: 0 IO Block: 65536 regular file |27378|484:476|reload| Device: 1bh/27d Inode: 24549393 Links: 1 |27378|484:476|reload| Access: (0644/-rw-r--r--) Uid: ( 0/ root) Gid: ( 0/ root) |27378|484:476|reload| Access: 2012-06-07 16:45:02.507151000 -0400 |27378|484:476|reload| Modify: 2012-06-07 16:45 :02.507153000 -0400 |27378|484:476|reload| Change: 2012-06-07 16:45:02.507153000 -0400 2012-06-07 16:42:46|27378|484:476|reload|utils.pm:run_ssh_command(5060)|SSH command executed on BC1-12, returning (0, "File: `/var/lib/libvirt/images...") 2012-06-07 16:42:46|27378|484:476|reload|Linux.pm:file_exists(1335)|'/var/lib/libvirt/images/baseImages/centos5-linuxkvm8-98-v0.xml' exists on BC1-12, files: 1, directories: 0, links: 0 2012-06-07 16:42:46|27378|484:476|reload|libvirt.pm:does_image_exist(478)|centos5-linuxkvm8-98-v0 image exists on BC1-12: /var/lib/libvirt/images/baseImages/centos5-linuxkvm8-98-v0.xml > > > > -- > > Thank you, > > > > Dmitri Chebotarov > > Virtual Computing Lab Systems Engineer, TSD - Ent Servers & Messaging > > 223 Aquia Building, Ffx, MSN: 1B5 > > Phone: (703) 993-6175 > > Fax: (703) 993-3404 > > > > > > On Thursday, June 7, 2012 at 14:18 , Dmitri Chebotarov wrote: > > > > > > > > Hi > > > > > > > > > > > > Has anyone tried KVM support in 2.3? > > > > > > I got it working to a point where I can capture a Linux image and able to > > > reload a VM computer with new KVM-based image via 'Manage Computers'. > > > The VM is available and I confirmed that VM is 'running' on KVM host. > > > > > > However when I make reservation of the KVM image, VCL tries to load the > > > image using xCAT, not libvirt. > > > > > > The image is listed with type 46, which is Bare Metal RedHat. > > > Do I need to add new OSid / OSinstalltype for libvirt? > > > > > > -- > > > Thank you, > > > > > > Dmitri Chebotarov > > > Virtual Computing Lab Systems Engineer, TSD - Ent Servers & Messaging > > > 223 Aquia Building, Ffx, MSN: 1B5 > > > Phone: (703) 993-6175 > > > Fax: (703) 993-3404 > > > > > > > > Attachments: > - capture1.txt >
