Ask Your Question

Elangovan Anganann's profile - activity

2018-10-13 12:24:02 -0500 received badge  Famous Question (source)
2018-06-06 07:33:38 -0500 received badge  Famous Question (source)
2018-02-13 11:55:32 -0500 answered a question opendaylight , openstack mitaka 3nodes integration on centos7

You can create a separate VM for ODL. You need to clean all the neutron resources from the openstack nodes and integrate neutron-server with ODL.

2018-02-13 11:49:28 -0500 commented question How to bind nova VM to a port created in OpenDayLight Controller?

Abdul Malik, What versions of OpenStack and ODL are you using? Are you using OpenStack Pike with ODL Nitrogen?

2017-10-25 13:39:37 -0500 received badge  Famous Question (source)
2017-09-26 03:38:28 -0500 received badge  Famous Question (source)
2017-09-20 21:50:53 -0500 received badge  Notable Question (source)
2017-09-20 21:50:53 -0500 received badge  Famous Question (source)
2017-09-20 11:26:49 -0500 received badge  Student (source)
2017-09-20 09:24:15 -0500 received badge  Notable Question (source)
2017-08-14 16:57:43 -0500 received badge  Notable Question (source)
2017-08-14 07:23:47 -0500 received badge  Popular Question (source)
2017-08-10 23:28:30 -0500 received badge  Popular Question (source)
2017-08-10 11:38:38 -0500 asked a question Ocata - Discovering versions from the identity service failed

Below error comes when I try to run "openstack token issue" command. If I restart httpd service, "openstack token issue" command runs successfully for some time. Then, again the same problem occurs.

[root@controller1 ~ (admin)]# openstack token issue
Discovering versions from the identity service failed when creating the password plugin. Attempting to determine version from URL.
Service Unavailable (HTTP 503)

Even curl command is not responding.

[root@controller1 ~ (admin)]# curl http://10.0.0.12:35357/v3

When I restart httpd, everything works for some time. After some time, same issue as described above happens.

[root@controller1 ~ (admin)]# systemctl restart httpd

[root@controller1 ~ (admin)]# curl http://10.0.0.12:35357/v3
{"version": {"status": "stable", "updated": "2017-02-22T00:00:00Z", "media-types": [{"base": "application/json", "type": "application/vnd.openstack.identity-v3+json"}], "id": "v3.8", "links": [{"href": "http://10.0.0.12:35357/v3/", "rel": "self"}]}}

[root@controller1 ~ (admin)]# openstack token issue
+------------+---------------------------------------------------------------------------------------------------------------------------+
| Field      | Value                                                                                                                     |
+------------+---------------------------------------------------------------------------------------------------------------------------+
| expires    | 2017-08-10T17:29:06+0000                                                                                                  |
| id         | gAAAAABZjInSp0NlbMOcMg7JSr7QQgp13eNrESOuEiki1ANFBYwlYjwBIj7rfLR01PvYFz968R7oGxEv3L4YPSMNcsLXFebVagJeBYW2                  |
|            | -qb5APOYhdGPPVdUmIYL0QegG3ogaJTAY-cWAaPolwe421IFFxvwVIDRGe-fiykah0b9zkPLEEY70IY                                           |
| project_id | 96c9b397c67047269b9a59470cf35622                                                                                          |
| user_id    | 35e74dfcd02946d1ae95467f14c3bfb8                                                                                          |
+------------+---------------------------------------------------------------------------------------------------------------------------+

I don't see any errors logged in /var/log/httpd/keystone.log (or) /var/log/keystone/keystone.log NOTE: I enabled debug=true in /etc/keystone/keystone.conf

Any idea why this happens?

2017-08-06 20:49:46 -0500 received badge  Famous Question (source)
2017-07-26 02:23:54 -0500 received badge  Notable Question (source)
2017-07-25 09:00:49 -0500 answered a question IPv6 Support in Ocata

IPv6 doesn't seem to work with Neutron Linux Bridge Agent (provider configuration) in OpenStack Ocata. It seem to work with Neutron OpenvSwitch Agent (provider configuration) in OpenStack Ocata.

2017-07-22 04:54:25 -0500 commented answer IPv6 Support in Ocata

These chapters doesn't talk about provider network interface configuration. As per link text, we cannot have IPv6 address to NIC that is connected to provider network. Openstack disables IPv6 for linux bridges.

2017-07-22 04:25:11 -0500 received badge  Commentator
2017-07-22 04:25:11 -0500 commented answer IPv6 Support in Ocata

Suppose eth0 is the provide network interface; does eth0 have both IPv4 and IPv6 addresses? Could you please provide ifconfig output of a controller (network) node? Also provide the output of "rpm -qa | grep openstack-neutron" from a controller (network) node. Thanks.

2017-07-22 03:37:28 -0500 commented question DHCP not providing IPv6 address to VM

Iuowei, That reference article using Openvswitch (NOT Linux Bridge).

2017-07-21 11:05:17 -0500 asked a question DHCP not providing IPv6 address to VM

OpenStack Environment Details: I have an Ocata OpenStack HA environment (3 controllers, 2 compute). There are two NICs in each node: one private network and one public network. Private network is used for openstack inter-node communication. Public network is used for API access. All the NICs have IPv4 addresses (no IPv6 addresses configured as OpenStack disables IPv6 on Linux Bridges in the OS). Both private and public networks are FLAT networks (No VLANs).

Issue: I created a provider network (linux bridge driver) on the public network with two subnets: IPv4 subnet and IPv6 subnet (IPv6 Address mode: dhcpv6-stateful, IPv6 RA mode: dhcpv6 stateful). When I created a VM, OpenStack assigned IPv4 and IPv6 addresses. These addresses were visible in "openstack server list" command output.

But inside the VM, I could see only IPv4 address; I don't see IPv6 address. When I did tcpdump to track the DHCP communication between VM and controllers, I saw only IPv4 addresses getting offered to the VM. IPv6 addresses were not offered. I want the VM to get both IPv4 and IPv6 addresses.

How to achieve this?

Update: I delete IPv4 subnet and allowed the network with only IPv6 subnet. When I created a VM on this network, tcpdump show DHCP discover is broadcasted and these discover packets reached controller nodes. But DHCP didn't respond. Why this behavior?

[root@controller1 ~]# rpm -qa | grep dnsmasq
dnsmasq-2.66-21.el7.x86_64
dnsmasq-utils-2.66-21.el7.x86_64
2017-07-19 12:31:16 -0500 received badge  Famous Question (source)
2017-07-17 13:07:03 -0500 received badge  Popular Question (source)
2017-07-16 09:12:54 -0500 received badge  Notable Question (source)
2017-07-15 12:45:02 -0500 received badge  Popular Question (source)
2017-07-15 04:23:17 -0500 asked a question IPv6 Support in Ocata

I have an openstack cloud (ocata) in HA configuration (3 controllers, 2 compute nodes). At the moment, it supports only IPv4 (TypeDrivers: vxlan, flat) (Mechanism Drivers: linuxbridge,l2population). Now, I want to enabled IPv6 support for my tenants. What is the procedure to accomplish IPv6 Support?

Suppose eth0 is used for public internet connectivity and as interface to provider network, do we need to provide IPv4 and IPv6 address to eth0. I am using linux bridge agent.

[root@controller1 ~ ]# rpm -qa | grep openstack-neutron
openstack-neutron-ml2-10.0.1-1.el7.noarch
openstack-neutron-linuxbridge-10.0.1-1.el7.noarch
openstack-neutron-common-10.0.1-1.el7.noarch
openstack-neutron-10.0.1-1.el7.noarch
2017-07-13 23:40:28 -0500 answered a question Can't force delete an instance

You can delete the VM from compute node using virsh and update Nova tables in mysql.

2017-07-13 23:37:48 -0500 answered a question vcpu number does not reflect cpu_allocation_ratio option

Deafult CPU allocation ratio 0.0 means 16.0. You have set it to 4.0. If you have 100 cpu threads, you allocate 400 vCPUs to the VMs. Create VMs and see the overcommit in action. Once vCPU usage reaches 400, you won't be able to create VMs. nova host-describe shows actual CPU threads available and vCPUs used.

2017-07-11 10:22:14 -0500 asked a question Allowed address Pairs

Linux Bridge Agent is used in this deployment (Controller, compute1, compute2). I have created 5 VMs (vm1, vm2, vm3, vm4, vm5) on a single self-service network. Three VMs (vm1, vm2, vm3) are hosted on compute1 node and Two VMs (vm4, vm5) are hosted on compute2 node. I added a virtual ip (10.10.10.10) to vm4. I am able to ping the virtual ip (10.10.10.10) from (vm4, vm5) but NOT from (vm1, vm2, vm3).

Why VIP is not pingable from VMs hosted on other compute node?

2017-07-11 02:27:10 -0500 received badge  Notable Question (source)
2017-07-10 00:58:13 -0500 received badge  Famous Question (source)
2017-07-08 21:23:21 -0500 received badge  Notable Question (source)
2017-07-08 21:00:42 -0500 received badge  Popular Question (source)
2017-07-08 02:53:03 -0500 received badge  Popular Question (source)
2017-07-07 22:16:54 -0500 marked best answer ceilometer empty meter-list output

Ceilometer meter-list always provides empty list.

[root@controller1 ~ (admin)]# IMAGE_ID=$(glance image-list | grep 'cirros' | awk '{ print $2 }')
[root@controller1 ~ (admin)]# glance image-download $IMAGE_ID > /tmp/cirros.img
You have new mail in /var/spool/mail/root
[root@controller1 ~ (admin)]# ceilometer meter-list
+------+------+------+-------------+---------+------------+
| Name | Type | Unit | Resource ID | User ID | Project ID |
+------+------+------+-------------+---------+------------+
+------+------+------+-------------+---------+------------+

I am using mongodb as the backend in OpenStack (Ocata). Mongodb don't have any data.

> show dbs
admin       0.000GB
ceilometer  0.000GB
local       0.000GB
> use ceilometer
switched to db ceilometer
> db
ceilometer
> show collections
meter
resource
> show dbs
admin       0.000GB
ceilometer  0.000GB
local       0.000GB
> show profile
db.system.profile is empty
Use db.setProfilingLevel(2) will enable profiling
Use db.system.profile.find() to show raw profile entries
> show collections
meter
resource
> db.meter.find()
> show collections
meter
resource
> db.resource.find()
>

I have provided /etc/ceilometer/ceilometer.conf below

[root@controller1 ~ (admin)]# cat /etc/ceilometer/ceilometer.conf
[DEFAULT]
auth_strategy = keystone
transport_url = rabbit://openstack:RABBIT_PASS@10.0.0.11
meter_dispatcher=database


[database]
connection = mongodb://ceilometer:CEILOMETER_DBPASS@10.0.0.33:27017/ceilometer

[keystone_authtoken]
auth_uri = http://10.0.0.11:5000
auth_url = http://10.0.0.11:35357
memcached_servers = 10.0.0.12:11211,10.0.0.13:11211,10.0.0.14:11211
auth_type = password
project_domain_name = default
user_domain_name = default
project_name = service
username = ceilometer
password = CEILOMETER_PASS
service_token_roles_required = true

[service_credentials]
auth_type = password
auth_url = http://10.0.0.11:5000/v3
project_domain_name = default
user_domain_name = default
project_name = service
username = ceilometer
password = CEILOMETER_PASS
interface = internalURL
region_name = RegionOne

What could be the issue?

2017-07-07 22:16:54 -0500 edited answer ceilometer empty meter-list output

I found the issue. Looks like in Ocata release, default value of meter_dispatchers and event_dispatchers is not "database" in ceilometer.conf. Once I added the following configurations, Ceilometer started working.

meter_dispatchers=database
event_dispatchers=database