Ask Your Question

affo's profile - activity

2017-06-10 12:33:55 -0500 received badge  Famous Question (source)
2015-11-21 12:45:47 -0500 received badge  Famous Question (source)
2015-11-21 12:45:47 -0500 received badge  Notable Question (source)
2015-07-12 06:21:05 -0500 received badge  Famous Question (source)
2015-05-20 04:17:00 -0500 commented question DevStack Setup with 3 nodes (Controller - Network - Compute)

answer by GLaupre is very valid. If you want a sample of local.conf for controller and compute nodes look at my repo: https://github.com/affear/fakestack. Note that I am using nova-network...

2015-02-27 10:14:48 -0500 commented question FakeDriver live_migration issue, POWERSTATE=NOSTATE
2015-02-27 08:53:50 -0500 received badge  Nice Question (source)
2015-02-18 03:10:08 -0500 received badge  Popular Question (source)
2015-02-17 03:30:33 -0500 asked a question FakeDriver live_migration issue, POWERSTATE=NOSTATE

Hi everybody, is there any known problem with live migration in FakeDriver?
I am migrating an instance, but after migration, the instance goes to Power State NOSTATE.
Is it normal?

Thank you in advance

2015-02-02 04:12:01 -0500 commented answer How to develop for Openstack

thank you so much

2015-02-01 07:27:52 -0500 received badge  Famous Question (source)
2015-01-30 07:07:32 -0500 received badge  Notable Question (source)
2015-01-30 03:43:40 -0500 commented answer How to develop for Openstack

thank you, I edited my question. I hope I made my issue clearer

2015-01-30 03:20:48 -0500 received badge  Popular Question (source)
2015-01-29 11:13:33 -0500 asked a question How to develop for Openstack

Hi everybody! I am starting developing a feature for OpenStack Nova.
I have already opened a blueprint on launchpad.

I read developer's guide but I think I missed an important point mainly because I don't know Gerrit, I think.

I am in the case of a __long-lived topic branch__.
My issue is that I have to work on my branch in team with other people.
How can I push changes and make them visible to my team if I do not use a fork of nova (by my github organization, in my case) and I use a clone of the official nova from OpenStack?

Moreover, I have a testing environment built with DevStack. How can I make DevStack use the changes I made if they are not pushed?

Do I have to __fork__ nova?
In these case, how does the normal workflow of commit, reviewing, rebasing work?

Thank you in advance

2015-01-04 07:38:19 -0500 received badge  Famous Question (source)
2015-01-03 00:51:28 -0500 received badge  Nice Question (source)
2014-12-10 22:00:31 -0500 received badge  Notable Question (source)
2014-12-10 07:36:07 -0500 received badge  Popular Question (source)
2014-12-10 06:29:51 -0500 commented answer Installing with DevStack in Docker container

thank you, see the edit please

2014-12-10 04:56:08 -0500 asked a question Installing with DevStack in Docker container

Hi everybody,
I am trying to install OpenStack into a Docker container.

The problem happens at almost the end of the installation, at line 1352: service_check.
Devstack checks services and the problem is that nova-api doesn't start!
Why? Because the command

sudo nova-rootwrap /etc/nova/rootwrap.conf iptables-save -c

gives

iptables-save v1.4.21: Cannot initialize: Permission denied (you must be root)

I know that there is a known problem with Docker and iptables... But I couldn't find a solution except from running commands with --privileged=true.
Unfortunately, I do want my image to be built automatically from Dockerfile, and it is __not possible__ to RUN commands as privileged from file.
Actually, running with --privileged=true, gives this error on mysql service start:

/usr/sbin/mysqld: error while loading shared libraries: libaio.so.1: cannot open shared object file: Permission denied

And mysql has to be up during OpenStack installation (obviously).

Can you please help me?
Do you have any idea on how should I make iptables-save -c run?

Thank you in advance

2014-12-04 07:06:02 -0500 received badge  Nice Question (source)
2014-11-25 15:55:10 -0500 received badge  Notable Question (source)
2014-11-25 15:55:10 -0500 received badge  Famous Question (source)
2014-11-15 08:54:40 -0500 received badge  Popular Question (source)
2014-11-13 04:25:54 -0500 commented question nova scheduler scheduling instance on invalid host

re-imaged?

2014-11-13 04:25:54 -0500 received badge  Commentator
2014-11-13 02:52:24 -0500 asked a question How to create tons of compute nodes with FakeDrivers

Hi everybody,
I want to stress test nova for purpose of scheduling.
I want to reach a number of compute nodes between 50 and 100.

I know that I can use nova.virt.fake to achieve this, but the problem is: should I have to create a VM for each fake compute node?

In this case, what configuration do you suggest to me in order to have the smallest VMs possible?
Spawning a hundred of VMs is a tough job!

I tried to spawn the VMs using Vagrant, but, on my laptop (I have also the possibility to use a more powerfu machines (Dell PowerEdge T320)), every machine takes at least 1 minute to be up.
I paste here a draft of Vagrantfile:

Vagrant.configure(VAGRANTFILE_API_VERSION) do |config|
# defining controller
config.vm.define "controller" do |controller|
    controller.vm.box = DEFAULT_BOX
    controller.vm.hostname = "controller"
end

# defining tons of computes
NO_COMPUTE.times do |index|
    config.vm.define "compute#{index}" do |compute_node|
        compute_node.vm.box = DEFAULT_BOX
        compute_node.vm.hostname = "compute#{index}"
    end
end

end

I have also another question: can you suggest me a good place where to retrieve information on how to deploy your openstack code to a local testing environment?
I also read something about ironic, but I didn't understand a lot about it... Is it something that could help me in your opinion?
I read nova developer's guide, but there is almost nothing about how can I deploy my code to a local DevStack deploy on VMs.

Thank you in advance

2014-10-30 03:05:37 -0500 received badge  Notable Question (source)
2014-10-27 02:42:03 -0500 commented question nova scheduler scheduling instance on invalid host

I can tell you that the scheduler reads metrics from db and chooses the host. If the bad host has ever written something, it could be chosen! The scheduler should check the timestamp of the metrics understanding that something is going bad. Can you explain exactly how nova-compute has crashed?

2014-10-25 15:14:06 -0500 received badge  Notable Question (source)
2014-10-25 09:13:23 -0500 received badge  Popular Question (source)
2014-10-24 07:22:05 -0500 received badge  Popular Question (source)
2014-10-24 07:10:01 -0500 received badge  Famous Question (source)
2014-10-24 07:02:06 -0500 received badge  Supporter (source)
2014-10-24 06:59:28 -0500 commented answer How are VM-related actions handled by Nova?

Thank you very much for your answer!

I edited the post to better explain point 2

2014-10-24 06:59:15 -0500 edited question How are VM-related actions handled by Nova?

Hi everybody,
we are developing for academic interest a service for Nova which tries to achieve resource optimization.
We already saw that somebody noticed that Nova hasn't nothing which does this.

Given this, a first question: do you know if the state of the art is still unchanged? Or there is some work related which is not included in the link above?

Going further, to develop the service, we have to know how actions on VMs (such as create, delete, resize and so on) are handled by Nova.
We have already read (and understood) a lot of code from github.
For instance, we understood that (correct me if I am wrong, please):

  • the creation and the resize of a VM passes through the scheduler (only to determine the destination host)
  • the delete of a VM doesn't cause any other service to take autonomic decisions such as consolidating VMs. For instance, if host A has only one VM, some service could live migrate the instance to host B and put host A in standby to optimize resources.

The real problem is that Nova is really really big and we are scared of missing something crucial.

So, the second question is: do you know of some documentation in OpenStack which well describes this processes? Do we have to continue on this direction and continue looking at the code?

Thank you in advance