Network is unreachable on instances

[Sergey Murashov]

ISO: {"build_id": "2014-08-01_15-12-05", "ostf_sha": "061b999c014db816c2da12546dcff1ae3a5a6ec0", "build_number": "547", "auth_required": true, "api": "1.0", "nailgun_sha": "3a67c786f5a07b5c2597c0649247c6a9db6c0185", "production": "docker", "fuelmain_sha": "7990f5bfa7fea5b74ebf0402b1918109b9bc505b", "astute_sha": "5a93fa8f9abbc087ee1c9cca894d781a03167094", "feature_groups": ["mirantis"], "release": "5.1", "fuellib_sha": "004ab7ed0e6e4268f3657d47836d3629f339bd9a"}

Steps To Reproduce:
1. Install OS(Simple mode, CentOS, Neutron GRE, Murano).
2. Create Private Network 'test'.
3. Add this private network to default router.
4. Crate VM and assign this VM with new private network 'test'.

Actual Result:
If we go on instance, we can see that network is unreachable on instance.

In fact we can see that VM has no IP address (probably it is DHCP agent?)

[Alexander Ignatov (aignatov)]

As per comment https://bugs.launchpad.net/mos/+bug/1348649/comments/19

Issue is reproduced again on next cluster configuration:
Simple, Ubuntu, Neutron GRE, 1 controller, 2 compute

Instance can't get ip - http://paste.openstack.org/show/88165/

Logs are attached

[Anastasia Kuznetsova (akuznetsova)]

This bug reproduced on ISO 405:
{"build_id": "2014-08-06_02-01-17", "ostf_sha": "be71965998364bf8e6415bd38b75c84b63aab867", "build_number": "405", "auth_required": true, "api": "1.0", "nailgun_sha": "f64b06c788e2b92fcb8e678ea6d0c9b86e8d0ab7", "production": "docker", "fuelmain_sha": "124ea87f1ac1c06e27613fe3b31fd5fc6b39e82d", "astute_sha": "99a790ad1b7526cbbd5bf8add0cb2b4e503fccd4", "feature_groups": ["mirantis"], "release": "5.1", "fuellib_sha": "513ec5cdcdef74c7419d5bae967b9edc7da8dbd7"}

Ubuntu, Neutron GRE, 1 controller, 1 compute

[Eugene Nikanorov (enikanorov)]

As Ilya found out, the issue was with custom network created via Horizon.

It appears that neutron creates the network with first available segmentation type, which is 'local' for the deployment, if it's not specified explicitly in the command.

tenant_network_types option in /etc/neutron/plugin.ini is what affecting this behavior.

We need types local, flat to be moved at the end of the list, so main network type (gre or vlan or vxlan) is used first.

[Alexander Ignatov (aignatov)]

fix proposed for master https://review.openstack.org/#/c/112646/

[Dennis Dmitriev (ddmitriev)]

Reproduced on ISO #408, system test 'deploy_neutron_gre' , 1 controller, 2 computes, neutron+gre.

master:/var/log/docker-logs/ostf.log :
...
fuel_health.tests.smoke.test_nova_create_instance_with_connectivity: INFO: is address is 10.108.35.139
fuel_health.tests.smoke.test_nova_create_instance_with_connectivity: DEBUG: 10.108.35.139
...
fuel_health.nmanager: DEBUG: Traceback (most recent call last):
  File "/usr/lib/python2.6/site-packages/fuel_health/nmanager.py", line 299, in retry_command
    result = method(*args, **kwargs)
  File "/usr/lib/python2.6/site-packages/fuel_health/common/ssh.py", line 168, in exec_command
    strerror=''.join(err_data).join(out_data))
SSHExecCommandFailed: Command 'ping -q -c1 -w10 10.108.35.139', exit status: 1, Error:
PING 10.108.35.139 (10.108.35.139) 56(84) bytes of data.

--- 10.108.35.139 ping statistics ---
4 packets transmitted, 0 received, +3 errors, 100% packet loss, time 2999ms
pipe 4

fuel_health.nmanager: DEBUG: Command 'ping -q -c1 -w10 10.108.35.139', exit status: 1, Error:
PING 10.108.35.139 (10.108.35.139) 56(84) bytes of data.

--- 10.108.35.139 ping statistics ---
4 packets transmitted, 0 received, +3 errors, 100% packet loss, time 2999ms
pipe 4
. Another effort needed.
...