nova hypervisor show give overcommit data

asked 2015-05-29 02:31:32 -0600

lafada gravatar image

I am using Openstack JUNO, one of the ESX Cluster in my node is giving overcommit data in nova hypervisor-show command.

root@localhost:~# nova hypervisor-show 2
+---------------------------+-----------------------------------------------+
| Property                  | Value                                         |
+---------------------------+-----------------------------------------------+
| cpu_info_model            | ["Intel(R) Xeon(R) CPU E5-2650 v2 @ 2.60GHz"] |
| cpu_info_topology_cores   | 16                                            |
| cpu_info_topology_threads | 32                                            |
| cpu_info_vendor           | ["XX"]                                        |
| current_workload          | 0                                             |
| disk_available_least      | -                                             |
| free_disk_gb              | -801                                          |
| free_ram_mb               | 120016                                        |
| host_ip                   | X.X.X.X                                       |
| hypervisor_hostname       | My-Cluster                                    |
| hypervisor_type           | VMware vCenter Server                         |
| hypervisor_version        | 5005000                                       |
| id                        | 23                                            |
| local_gb                  | 499                                           |
| local_gb_used             | 1300                                          |
| memory_mb                 | 253136                                        |
| memory_mb_used            | 133120                                        |
| running_vms               | 35                                            |
| service_disabled_reason   | -                                             |
| service_host              | test-vcenter                                  |
| service_id                | 56                                            |
| state                     | up                                            |
| status                    | enabled                                       |
| vcpus                     | 32                                            |
| vcpus_used                | 65                                            |
+---------------------------+-----------------------------------------------+

I can see, vcpu is 32 and vcpu_used is 65, how this overcommit data store in nova?

I gave cpu ration as 10.0, but as per my understanding, this ratio used by nova-scheduler to spanwn vm on particular host.

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted
1

answered 2015-05-29 10:26:56 -0600

jdexter gravatar image

The overcommit is not stored in the scheduler, if you look in the nova database for the hypervisor the compute nodes are reporting true numbers not the inflated numbers.

The overcommit number is used by the scheduler when it goes to determine the hypervisor host.

edit flag offensive delete link more

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2015-05-29 02:31:32 -0600

Seen: 853 times

Last updated: May 29 '15