Ask Your Question
0

Instance Error: Launcing instance gives error

asked 2015-02-15 00:37:05 -0600

anil987 gravatar image

updated 2015-02-16 10:28:54 -0600

I am launching instance on openstack but instance always ends in error state

Here are logs of nova-scheduler

2015-02-15 11:56:10.137 879 INFO nova.scheduler.filter_scheduler [req-e645a7ed-6272-4caa-b3b9-6c9fe866b382 dedc4b8f9f6c445e8c958002398a3758 7ab5428b376c4de683a65c6487b98e06] Attempting to build 1 instance(s) uuids: [u'e27a1874-2e44-4ac7-b65a-f99419cf1121']

2015-02-15 11:56:10.172 879 WARNING nova.scheduler.filters.compute_filter [req-e645a7ed-6272-4caa-b3b9-6c9fe866b382 dedc4b8f9f6c445e8c958002398a3758 7ab5428b376c4de683a65c6487b98e06] ram:1489 disk:7168 io_ops:0 instances:0 has not been heard from in a while

2015-02-15 11:56:10.173 879 INFO nova.filters [req-e645a7ed-6272-4caa-b3b9-6c9fe866b382 dedc4b8f9f6c445e8c958002398a3758 7ab5428b376c4de683a65c6487b98e06] Filter ComputeFilter returned 0 hosts

2015-02-15 11:56:10.177 879 WARNING nova.scheduler.driver [req-e645a7ed-6272-4caa-b3b9-6c9fe866b382 dedc4b8f9f6c445e8c958002398a3758 7ab5428b376c4de683a65c6487b98e06] [instance: e27a1874-2e44-4ac7-b65a-f99419cf1121] Setting instance to ERROR state.

Update

Here is log of nova-compute

2015-02-16 21:44:12.884 1475 INFO oslo.messaging._drivers.impl_rabbit [-] Reconnecting to AMQP server on controller:5672

2015-02-16 21:44:12.885 1475 INFO oslo.messaging._drivers.impl_rabbit [-] Delaying reconnect for 1.0 seconds...

2015-02-16 21:44:16.905 1475 ERROR oslo.messaging._drivers.impl_rabbit [-] AMQP server on controller:5672 is unreachable: [Errno 113] EHOSTUNREACH. Trying again in 30 seconds.

2015-02-16 21:44:46.931 1475 INFO oslo.messaging._drivers.impl_rabbit [-] Reconnecting to AMQP server on controller:5672

2015-02-16 21:44:46.931 1475 INFO oslo.messaging._drivers.impl_rabbit [-] Delaying reconnect for 1.0 seconds...

2015-02-16 21:44:50.944 1475 ERROR oslo.messaging._drivers.impl_rabbit [-] AMQP server on controller:5672 is unreachable: [Errno 113] EHOSTUNREACH. Trying again in 30 seconds.

edit retag flag offensive close merge delete

Comments

Restart all the services and then check the status of your instances.

Varsha gravatar imageVarsha ( 2015-02-15 00:57:35 -0600 )edit

I have restarted all the services, but still I am getting the same error

anil987 gravatar imageanil987 ( 2015-02-15 03:03:15 -0600 )edit

Your compute hosts have enough resources? Rabbitmq/qpid is able to transmit data between controller and compute node? All the services are up when launching nova service-list?

Eduardo Gonzalez gravatar imageEduardo Gonzalez ( 2015-02-16 03:40:35 -0600 )edit

How can I check "Rabbitmq/qpid is able to transmit data between controller and compute node"?

Thanks

anil987 gravatar imageanil987 ( 2015-02-16 10:56:24 -0600 )edit

7 answers

Sort by ยป oldest newest most voted
0

answered 2015-02-17 03:50:03 -0600

Hi anil, once again check the username and password of rabbitmq-sever in /etc/nova/nova.conf. we have to set the Replace RABBIT_PASS with a suitable password. like given below

rabbitmqctl change_password guest RABBIT_PASS , then restart the all nova services and rabbitmq services could you please send log files once again?

edit flag offensive delete link more
0

answered 2015-02-17 02:02:15 -0600

RabbitMQ is unable to establish connection with controller. Maybe you got a firewall blocking this connection, open the ports or disable firewall, also check rabbitmq configuration in both hosts and restart the service.

edit flag offensive delete link more
0

answered 2015-02-16 11:03:18 -0600

Hey! Check for the correct username and password in /etc/nova/nova.conf. Copy the rabbitMQ setting from [rabbitmq] section to [default] section in nova.conf. Restart the rabbitmq-server and all other services.

edit flag offensive delete link more

Comments

In my nova.conf there is no [rabbitmq] section. RabbitMq credentials are already in [default] section. I have restarted all the services but it doesn't work. I am planning to reinstall openstack from the beginning.

Thanks

anil987 gravatar imageanil987 ( 2015-02-17 02:37:39 -0600 )edit
0

answered 2015-02-16 03:30:16 -0600

Hi Anil,

  1. I would like to see what nova-compute.log file talking about.
  2. How about trying with other images? Is that yielding you the expected result?

Try this link. Hope this helps you.

edit flag offensive delete link more

Comments

I have updated the nova-compute log in the question

anil987 gravatar imageanil987 ( 2015-02-16 10:31:03 -0600 )edit

Hey! your post of nova-compute.log doesn't say anything, my advise to you is, paste the complete nova-compute.log in paste.openstack.org . Look specifically at the areas that talks about your instance in the log file. Tell me, did you look at above mentioned link?

Yashpal Beppurana gravatar imageYashpal Beppurana ( 2015-02-16 12:06:42 -0600 )edit

I have earlier said there is no useful log in nova-compute. I have gone through the link, and looked in the log files mentioned in the link but I cannot find any useful information from them. Only thing that I can conclude from the link is that there are not enough resources in the system

anil987 gravatar imageanil987 ( 2015-02-16 13:04:46 -0600 )edit

But still I don't get why I am not able to ping controller from compute and vice versa, which was possible during installation.

anil987 gravatar imageanil987 ( 2015-02-16 13:16:06 -0600 )edit

Okay Anil. From my understanding by looking at all comments over here is, you have two-node setup i.e, one controller and one compute. Your problem is that you are unable to connect from one box(controller) to another(compute) and vice versa.

Yashpal Beppurana gravatar imageYashpal Beppurana ( 2015-02-17 04:32:42 -0600 )edit
0

answered 2015-02-15 14:37:10 -0600

Praveen N gravatar image

Check the status of services openstack-nova-scheduler and openstack-nova-conductor.Ensure they are active. Try to launch instances with different images and flavors combinations.For example Cirros with m1.small or m1.tiny and so on. Also check the same by uploading few cloud images and launching instance with these cloud images.

edit flag offensive delete link more

Comments

I have checked the status of services using command "sudo service servicename status" and all the services are running. I tried to create instance with different flavor with Cirros and Ubuntu 13.10 image but the unable to launch any instance. This time I am not able to terminate a instance

anil987 gravatar imageanil987 ( 2015-02-15 23:05:17 -0600 )edit
0

answered 2015-02-15 08:41:51 -0600

1-Cloud gravatar image

updated 2015-02-15 08:46:19 -0600

Is this all in one node or multinode? once verify all connections in nova.conf and set hostname in hosts file. like 192.168.122.10 controller <host_name>

edit flag offensive delete link more

Comments

It is two node installation. I have also set controller in /etc/hosts file. Can I install everything in one node? I am aware of devstack, but it is only designed for proof of concept purpose. Is there any other way to install everything in one node?

anil987 gravatar imageanil987 ( 2015-02-15 22:21:26 -0600 )edit

I have installed openstack through this guide http://docs.openstack.org/icehouse/in... on Ubuntu 14.04 system

anil987 gravatar imageanil987 ( 2015-02-15 23:11:22 -0600 )edit

Yes you can install everything in one node. then use packstack in centos for all in one setup.

1-Cloud gravatar image1-Cloud ( 2015-02-15 23:14:30 -0600 )edit

Ohh it is manual installation guide. your using 2 node setup so you need to set your 2nd compute in controller and also set ur controller ip in 2nd compute.

1-Cloud gravatar image1-Cloud ( 2015-02-15 23:18:37 -0600 )edit

I have entered the ip of controller and compute in /etc/hosts file of both node. During installation I was able to ping the compute from controller and vice-versa. But after installation, when I tried to launch the instance, I figured out that I am no more able to ping one from another.

anil987 gravatar imageanil987 ( 2015-02-15 23:45:08 -0600 )edit
0

answered 2015-02-15 02:15:47 -0600

1-Cloud gravatar image

Restart all nova services and check nova-compute is working or not. if it is working fine launch an instance then paste nova-compute logs here.

edit flag offensive delete link more

Comments

Thanks for reply

First I restarted nova-compute and tried to launch instance but was unable to to that. Also there was no useful log in nova-compute. Later I figure out that I am unable to ping compute from controller and vice-versa.

anil987 gravatar imageanil987 ( 2015-02-15 04:52:24 -0600 )edit

You are not able to ping because your instance is in error state.Once you are able to launch instance correctly, you can transmit and recieve packets.Try to launch instance with smaller image like cirros with suitable flavor.From error i feel the size of the image is more & its using more ram&cpu.

Varsha gravatar imageVarsha ( 2015-02-15 09:49:36 -0600 )edit

Please check rabbitmq configuration in both hosts and restart the service and check it is working or not.

1-Cloud gravatar image1-Cloud ( 2015-02-18 02:42:20 -0600 )edit

I have checked the rabbitmq configration in nova.conf in both host and restarted the service. This time it worked.

Thanks

anil987 gravatar imageanil987 ( 2015-02-18 06:09:31 -0600 )edit

Okay are you able to ping compute from controller?

1-Cloud gravatar image1-Cloud ( 2015-02-19 02:04:09 -0600 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2015-02-15 00:37:05 -0600

Seen: 2,235 times

Last updated: Feb 17 '15