[one-users] No more than 144 vm instances?

Neumann, Steffen sneumann at ipb-halle.de
Thu Apr 28 21:37:38 PDT 2011


Hi,

144 is exactly the number of cores on your physical nodes.
The scheduler uses the available CPUs (cf. onehost list)
and whatever CPU resources are defined in your template.

You could reduce CPU to X<1 to allow overcomitting
of physical CPUs.

Yours,
Steffen
________________________________________
From: users-bounces at lists.opennebula.org [users-bounces at lists.opennebula.org] on behalf of Lars Kellogg-Stedman [lars at seas.harvard.edu]
Sent: 28 April 2011 23:24
To: users at lists.opennebula.org
Subject: [one-users] No more than 144 vm instances?

I have a cluster of three machines, each with 256GB of RAM and 48
cores.  I deployed 200 VM instances using occi-compute, and all 200
intances show up in "onevm list".  However, OpenNebula does not appear
to be willing to actual boot more than 144 of these instances at a
time.  After there were 144 running instances, no more were started;
when I killed one of the running instances, a new one started.  I am
creating these instances as myself, rather than as the "oneadmin"
user.

I thought that perhaps some sort of quota mechanism was kicking in,
but I can't see anywhere that such a limit is defined.
/etc/one/auth/auth.conf looks like this:

:database: sqlite://auth.db
:authentication: simple
:quota:
  :enabled: false
  :defaults:
    :cpu: 10.0
    :memory: 1048576
    :num_vms:

And I've never used the "oneauth quota ..." command.

--
Lars Kellogg-Stedman <lars at seas.harvard.edu>
Senior Technologist
Harvard University SEAS
Academic and Research Computing (ARC)
_______________________________________________
Users mailing list
Users at lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org



More information about the Users mailing list