Hi,<div><br></div><div>First, thank you very much for your time and your efforts. I totally agree that using the vz commands would be a better approach, there is a special driver format (XML) that pass to the driver the template in XML format. We can use that in a ruby/python script to create the containers.....</div>
<div><br></div><div>BTW, we've just included OpenVZ as a project proposal for GSoC, I'll keep you updated if a student apply for this project.</div><div><br></div><div>Cheers</div><div><br></div><div>Ruben </div>
<div>
<br><br><div class="gmail_quote">On Thu, Mar 10, 2011 at 6:33 PM, <span dir="ltr"><<a href="mailto:knawnd@gmail.com">knawnd@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
Hi!<br>
<br>
Spent some time in attempts to implement OpenVZ support in OpenNebula via libvirt I realized that for the moment probably it was not the most efficient way to do that since libvirt provides pretty poor OpenVZ support. In particular:<br>
1) some of openvz specific parameters/options are missing from libvirt docs (e.g. [1]) despite of they are mentioned in "Example guest domain XML configuration" at libvirt openvz driver page [2] (I submitted a bug: [3] );<br>
2) as far as I understand there is no way to specify assigned disk space and disk inodes by libvirt means (feature request: [4]);<br>
3) it looks like network device veth is wrongly assigned inside openvz container (at least I couldn't find the proper way to do that: [5]) and there is no way to use another network device - venet (<a href="http://wiki.openvz.org/Venet" target="_blank">http://wiki.openvz.org/Venet</a>).<br>
4) total (max) and used (current) memory parameters for certain VM always have a zero value ([6]).<br>
Libvirt-users mail-list doesn't help too much to clarify details;<br>
Most of issues mentioned above as well as e.g. [7] had been discussed before in libvirt mail-list (some of them more than a year ago) but I couldn't find any submitted bugs and that and those issues still exist.<br>
<br>
For the moment I have done not so much progress in implementing openvz support in ONE via libvirt:<br>
$ git status<br>
# On branch one-openvz<br>
# Changes to be committed:<br>
# (use "git reset HEAD <file>..." to unstage)<br>
#<br>
# new file: src/im_mad/remotes/ovz.d/architecture.sh<br>
# new file: src/im_mad/remotes/ovz.d/cpu.sh<br>
# new file: src/im_mad/remotes/ovz.d/name.sh<br>
# new file: src/im_mad/remotes/ovz.d/ovz.rb<br>
# new file: src/vmm/LibVirtDriverOVZ.cc<br>
# new file: src/vmm_mad/remotes/ovz/cancel<br>
# new file: src/vmm_mad/remotes/ovz/deploy<br>
# new file: src/vmm_mad/remotes/ovz/migrate<br>
# new file: src/vmm_mad/remotes/ovz/ovzrc<br>
# new file: src/vmm_mad/remotes/ovz/restore<br>
# new file: src/vmm_mad/remotes/ovz/save<br>
# new file: src/vmm_mad/remotes/ovz/shutdown<br>
# new file: src/vmm_mad/remotes/poll_vmm.rb<br>
# deleted: src/vmm_mad/remotes/poll_xen_kvm.rb<br>
# new file: src/vmm_mad/ssh/vmm_ssh_ovz.conf<br>
#<br>
# Changed but not updated:<br>
# (use "git add <file>..." to update what will be committed)<br>
#<br>
# modified: include/LibVirtDriver.h<br>
# modified: install.sh<br>
# modified: share/etc/oned.conf<br>
# modified: src/vmm/LibVirtDriverOVZ.cc<br>
# modified: src/vmm/SConstruct<br>
# modified: src/vmm/VirtualMachineManager.cc<br>
# modified: src/vmm_mad/ssh/vmm_ssh_ovz.conf<br>
<br>
Taking all mentioned above into account I incline to an opinion it would be better to implement openvz support in OpenNebula by native openvz means like vzctl/vzlist/vzcalc/vzmigrate/vzdump commands just invoking those commands on remote host.<br>
<br>
I would like to know other opinions on that.<br>
<br>
Best regards,<br>
Nikolay.<br>
<br>
[1] <a href="http://libvirt.org/formatdomain.html" target="_blank">http://libvirt.org/formatdomain.html</a><br>
[2] <a href="http://libvirt.org/drvopenvz.html" target="_blank">http://libvirt.org/drvopenvz.html</a><br>
[3] <a href="https://bugzilla.redhat.com/show_bug.cgi?id=683355" target="_blank">https://bugzilla.redhat.com/show_bug.cgi?id=683355</a><br>
[4] <a href="https://bugzilla.redhat.com/show_bug.cgi?id=683512" target="_blank">https://bugzilla.redhat.com/show_bug.cgi?id=683512</a><br>
[5] <a href="https://bugzilla.redhat.com/show_bug.cgi?id=683834" target="_blank">https://bugzilla.redhat.com/show_bug.cgi?id=683834</a><br>
[6] <a href="https://bugzilla.redhat.com/show_bug.cgi?id=683851" target="_blank">https://bugzilla.redhat.com/show_bug.cgi?id=683851</a><br>
[7] <a href="https://bugzilla.redhat.com/show_bug.cgi?id=683834" target="_blank">https://bugzilla.redhat.com/show_bug.cgi?id=683834</a><br>
_______________________________________________<br>
Users mailing list<br>
<a href="mailto:Users@lists.opennebula.org" target="_blank">Users@lists.opennebula.org</a><br>
<a href="http://lists.opennebula.org/listinfo.cgi/users-opennebula.org" target="_blank">http://lists.opennebula.org/listinfo.cgi/users-opennebula.org</a><br>
</blockquote></div><br><br clear="all"><br>-- <br>Dr. Ruben Santiago Montero<br>Associate Professor (Profesor Titular), Complutense University of Madrid<br><br>URL: <a href="http://dsa-research.org/doku.php?id=people:ruben" target="_blank">http://dsa-research.org/doku.php?id=people:ruben</a><br>
Weblog: <a href="http://blog.dsa-research.org/?author=7" target="_blank">http://blog.dsa-research.org/?author=7</a><br>
</div>