[one-users] overcommit of cpu and memory under kvm
Mark Wagner
mwagner at intelius.com
Wed Aug 8 17:15:37 PDT 2012
After tracing the execution of the IM host update I found the problem.
Alas, it was my fault.
/var/lib/one/remotes/im/run_probes runs all the scripts in
/var/lib/one/remotes/im/kvm.d/ (paths translated to /var/tmp/one/... on
the hosts).
We use cfengine to manage files so there was a kvm.rb.cfsaved in
/var/lib/one/remotes/im/kvm.d/. Its output would supersede the output of
kvm.rb.
Thanks,
Mark
On 08/03/2012 04:26 PM, Ruben S. Montero wrote:
> Hi,
>
> If you are seeing FOO=bar in the template that shows that the
> informations flow is right, and one.db updated. Could it be that the
> values generated by the modified kvm.rb change from execution to
> execution?
>
> Cheers
>
> Ruben
> On Fri, Aug 3, 2012 at 12:56 AM, Mark Wagner <mwagner at intelius.com
> <mailto:mwagner at intelius.com>> wrote:
>
>
> On 08/02/2012 01:14 PM, Ruben S. Montero wrote:
>
>
> When you modified the scripts you must execute
>
> onehost sync
>
> to propagate the changes to the hosts. The scripts are copied
> next time the host is monitored. Maybe you are missing that part.
>
>
> I've run "onehost sync". There wasn't a change. Just to be sure
> the script was being executed I added "FOO=bar" to the output.
>
> Here is "onehost show"
>
> oneadmin at onefe1:~$ onehost show 11
> HOST 11 INFORMATION
> ID : 11
> NAME : hpblade1-1
> CLUSTER : -
> STATE : MONITORED
> IM_MAD : im_kvm
> VM_MAD : vmm_kvm
> VN_MAD : dummy
> LAST MONITORING TIME : 08/02 15:52:11
>
> HOST SHARES
> TOTAL MEM : 424.4G
> USED MEM (REAL) : 6.1G
> USED MEM (ALLOCATED) : 12G
> TOTAL CPU : 21600
> USED CPU (REAL) : 2160
> USED CPU (ALLOCATED) : 1200
> RUNNING VMS : 3
>
> MONITORING INFORMATION
> ARCH="x86_64"
> CPUSPEED="2933"
> FOO="bar"
> FREECPU="19440.0"
> FREEMEMORY="387163512"
> HOSTNAME="hpblade1-1"
> HYPERVISOR="kvm"
> MODELNAME="Intel(R) Xeon(R) CPU X5670 @ 2.93GHz"
> NETRX="2048157672079"
> NETTX="2337238135328"
> TOTALCPU="21600"
> TOTALMEMORY="445056984"
> USEDCPU="2160.0"
> USEDMEMORY="6432608"
>
> Here is the result of running kvm.rb on the host:
>
> oneadmin at hpblade1-1:~$ /var/tmp/one/im/kvm.d/kvm.rb
> HYPERVISOR=kvm
> TOTALCPU=2400
> CPUSPEED=2933
> TOTALMEMORY=49450776
> USEDMEMORY=6428340
> FREEMEMORY=43022436
> FREECPU=2174.4
> USEDCPU=225.6
> NETRX=2052917131737
> NETTX=2341806785454
> FOO=bar
>
> Could there be something preventing the one.db from being updated?
> Could I flush it somehow?
>
> On Thu, Aug 2, 2012 at 8:57 PM, Mark Wagner
> <mwagner at intelius.com <mailto:mwagner at intelius.com>
> <mailto:mwagner at intelius.com <mailto:mwagner at intelius.com>>>
> wrote:
>
> I've modified /var/lib/one/remotes/im/kvm.d/kvm.rb to allow
> overcommiting CPU and memory but I'm encountering some
> inconsistencies.
>
> When I run the script on the host I get:
>
> TOTALMEMORY=247253880
>
> However, onehost show gives
>
> TOTALMEMORY="445056984"
>
> How could the onehost be getting that value?
>
> --
> Mark Wagner | mwagner at intelius.com <mailto:mwagner at intelius.com>
> System Administrator | Intelius Inc.
>
> _______________________________________________
> Users mailing list
> Users at lists.opennebula.org <mailto:Users at lists.opennebula.org>
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>
>
>
> --
> Ruben S. Montero, PhD
> Project co-Lead and Chief Architect
> OpenNebula - The Open Source Solution for Data Center Virtualization
> www.OpenNebula.org <http://www.OpenNebula.org> |
> rsmontero at opennebula.org <mailto:rsmontero at opennebula.org> | @OpenNebula
--
Mark Wagner | mwagner at intelius.com
System Administrator | Intelius Inc.
More information about the Users
mailing list