fails to connect to VM that is created on more than 1 NUMA node

[Sergii]

Detailed bug description:
When I try to launch VM on more than 1 NUMA node using hw:numa_nodes=2 on flavor, with DPDK enabled - there are no connection via SSH or Ping.

Environment: MOS 9.0 ISO 459, hardware lab
1 controller + 1 compute with DPDK (In our case node-1.domain.tld) + 1 compute.

Steps to reproduce:
Run this script on freshly deployed environment. Just replace value for compute_node_with_feature.
http://paste.openstack.org/show/510446/

Expected result:
Instance is available via SSH, ping.

Actual result:
Instance is not available via SSH, ping.

Additional information:
On compute with feature

ovs-vsctl get Open_vSwitch . other_config:pmd-cpu-mask=9c2709827 it is ([0, 1, 2, 5, 11, 12, 15, 20, 21, 22, 25, 30, 31, 32, 35]) cpu's

vcpu_pin_set from nova.conf - 6,26,7,27,8,28,9,29,16,36,17,37,18,38,19,39,3,23,4,24,13,33

node 0 cpus: 0 1 2 3 4 20 21 22 23 24
node 1 cpus: 5 6 7 8 9 25 26 27 28 29
node 2 cpus: 10 11 12 13 14 30 31 32 33 34
node 3 cpus: 15 16 17 18 19 35 36 37 38 39

Diagnostic snapshot - http://mos-scale-share.mirantis.com/fuel-snapshot-2016-06-10_08-23-56.tar.gz

[Sergey Matov (smatov)]

Confirmed on 9.0 release.
This case is quite narrow. Additional information should be added to release notes.
Main option is to step on 16.04 packages.
So it won't be fixed in 9.0 since DPDK consider to be experimental.

[Dmitry Kalashnik (dkalashnik)]

Commit https://review.fuel-infra.org/#/c/22332

[Nastya Urlapova (aurlapova)]

Moved to "Fix released" regarding Dima's comment.