On Thu, 30 Oct 2014, Ruben S. Montero wrote:



On Thu, Oct 30, 2014 at 2:57 PM, Steven Timm <t...@fnal.gov> wrote:
      I think I see what went wrong by looking at the documentation.
      The value as shown in onehost show -x
      is now labeled FREE_MEM instead of FREEMEMORY.
      (The current docs also correctly show the variable as FREE_MEM).
      I will edit my template and see if that has an effect.

      Also it looks like RANK is deprecated in favor of SCHED_RANK
      and SCHED_DS_RANK is that correct?



Yes, 

 

      Any way short of a text editor and onetemplate update to
      fix all the legacy templates that have it wrong?

 

All of this should be automatically handled by the upgrade process... 

It wasn't, the last time we tried a test upgrade using the
upgrade scripts.  There were a lot of deprecated fields that stayed
in the machine templates and some required ones that didn't get in.

Steve




      Steve Timm


      On Thu, 30 Oct 2014, Ruben S. Montero wrote:

            Hi Steven,
            This should be working as long it is in the host
            information. Could you send
            the onehost show -x for one of the hosts? (You can
            send it to me offlist if
            there is any sensible information).

            The Rank expression is supposed to work with any
            variable under the
            following XPaths:

                "/HOST/TEMPLATE/",
                "/HOST/HOST_SHARE/",
                "/HOST/",
                "/HOST/CLUSTER_TEMPLATE/

            Cheers

            Ruben

            On Wed, Oct 29, 2014 at 10:34 PM, Steven Timm
            <t...@fnal.gov> wrote:

                  Under OpenNebula 3 we have been using
            RANK=FREEMEMORY
                  to launch the next virtual machine on the
            hypervisor which
                  has the most memory available at the time.  I
            am using the
                  same in OpenNebula 4.8 but it does not appear
            to be working,
                  even though FREEMEMORY is still a documented
            selection variable
                  in the requirements section.  In particular
                  all the RANKS of the machines show up as "0"
            in the sched.log
                  when you look at the allocation cycle and
            there is no
                  "FREEMEMORY"
                  variable reported in the output of
                  onehost show xxxx
                  under opennebula 4.8.

                  Are there other selection variables which
            could have
                  the same desired effect, i.e. a horizontal
            fill of the cloud?

                  Steve Timm


                 
            ------------------------------------------------------------------
                  Steven C. Timm, Ph.D  (630) 840-8525
                  t...@fnal.gov  http://home.fnal.gov/~timm/
                  Fermilab Scientific Computing Division,
            Scientific Computing
                  Services Quad.
                  Grid and Cloud Services Dept., Associate Dept.
            Head for Cloud
                  Computing
                 
            _______________________________________________
                  Users mailing list
                  Users@lists.opennebula.org
                 
            http://lists.opennebula.org/listinfo.cgi/users-opennebula.org




            --
            Ruben S. Montero, PhD
            Project co-Lead and Chief ArchitectOpenNebula -
            Flexible Enterprise Cloud
            Made Simple
            www.OpenNebula.org | rsmont...@opennebula.org |
            @OpenNebula



      ------------------------------------------------------------------
      Steven C. Timm, Ph.D  (630) 840-8525
      t...@fnal.gov  http://home.fnal.gov/~timm/
      Fermilab Scientific Computing Division, Scientific Computing
      Services Quad.
      Grid and Cloud Services Dept., Associate Dept. Head for Cloud
      Computing




--
Ruben S. Montero, PhD
Project co-Lead and Chief ArchitectOpenNebula - Flexible Enterprise Cloud
Made Simple
www.OpenNebula.org | rsmont...@opennebula.org | @OpenNebula



------------------------------------------------------------------
Steven C. Timm, Ph.D  (630) 840-8525
t...@fnal.gov  http://home.fnal.gov/~timm/
Fermilab Scientific Computing Division, Scientific Computing Services Quad.
Grid and Cloud Services Dept., Associate Dept. Head for Cloud Computing
_______________________________________________
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

Reply via email to