kfox1111's profile - activity

2015-04-12 01:42:30 -0600 received badge  Famous Question (source)
2015-04-12 01:41:31 -0600 received badge  Taxonomist
2014-12-21 05:56:19 -0600 received badge  Famous Question (source)
2014-06-22 12:54:53 -0600 received badge  Famous Question (source)
2014-05-07 12:32:17 -0600 received badge  Notable Question (source)
2014-04-22 20:10:19 -0600 received badge  Notable Question (source)
2014-03-10 12:23:22 -0600 received badge  Popular Question (source)
2014-03-10 11:02:32 -0600 answered a question RPM is the latest version, but openstack is old version.

First you need to decide what you are trying to accomplish:

  • If you want the absolute newest code, you need to use git.
  • If you want stability, the latest stable packages are usually best.

Icehouse is not released yet, so RDO 'Icehouse' packages are for testing but are not always as up to date as git.

2014-03-08 07:38:18 -0600 received badge  Student (source)
2014-03-07 14:34:35 -0600 asked a question GroupAntiAffinityFilter Heat

I'm working on creating a set of MongoDB heat templates. I would like to ensure that as Mongo replica set's come and go, they are placed on seperate hosts.

The some older documentation makes it sound like you can specify a group name like "MongoDB-ReplicaSets" and have the scheduler figure out what instances are tagged that way. Some newer documentation makes it sound like you must specify the set of all instances you want it on different hosts from. Which is it? The group name only is much better then needing to know all the instances, which I'm not sure I have access to.

It is also unclear, how I would specify that I want the GroupAntiAffinityFilter and not the GroupAffinity filter in the template. What is the correct nova hint for this?

Maybe there is a better way to do this too. Any suggestions?

Thanks, Kevin

2014-02-24 14:46:29 -0600 answered a question metadata agent throwing AttributeError: 'HTTPClient' object has no attribute 'auth_tenant_id' with latest release
2014-02-24 13:59:23 -0600 answered a question metadata agent throwing AttributeError: 'HTTPClient' object has no attribute 'auth_tenant_id' with latest release

I'm seeing the same thing. Bug seems to be:


Need an RDO fix.

yum downgrade <the neutron packages you had installed> and restarting the metadata agent seems to work so far for us as a workaround until RDO upgrades their packages.

2014-01-31 18:21:41 -0600 answered a question Can you execute a program on the guest os via openstack API?

You probably want: http://docs.openstack.org/user-guide/content/user-data.html

But, if you want to do it at any time, then the official API, if you could call it that, is to ssh execute the command on the host.

2014-01-06 11:48:20 -0600 commented answer policy management

I'm hoping to let it be able to see running instances so that it can clean out images that have not been used in a while. I'm generating updated images automatically that are up to date so they launch quickly, but don't want old images piling up. It shouldn't be able to delete instances though.

2014-01-06 11:44:14 -0600 received badge  Popular Question (source)
2014-01-03 17:45:13 -0600 asked a question policy management

I have Jenkins building images for me. I would like it to be able to upload images to glance for one of my tenants and also to be able to look at running instances to see images used so it can delete images from glance that haven't been used in a while. I really don't need or want this user to be able to do anything else.

It looks like I could edit the policy.json file for every OpenStack project (nova, cinder, glance, heat...) but that looks like a ton of work, and hard to maintain when updated policy comes out.

Is there a better way to do this? To disable all but a few api calls for a user on a tenant? Maybe a keystone policy api or trusts?

2014-01-03 16:33:21 -0600 received badge  Supporter (source)
2013-12-21 23:26:32 -0600 received badge  Teacher (source)
2013-12-20 19:47:42 -0600 answered a question How to Upgrade an OpenStack Swift?
2013-12-20 18:35:15 -0600 answered a question Why floating IP?

One use case it enables is rolling updates. You start with a floating ip on vm A. You create a new vm B, check it out and make sure it is doing what you want. You then pause vm A and move the floating ip and any other resources onto vm B. If everything works, you delete vm A. If it doesn't, you move stuff back to vm A and delete vm B.

As far as the user is concerned, they only ever have to know about the floating ip, which doesn't change.

2013-11-21 15:33:32 -0600 received badge  Notable Question (source)
2013-10-18 02:01:34 -0600 received badge  Popular Question (source)
2013-10-16 17:14:24 -0600 answered a question heat security group id

OS::Neutron::Firewall seems intended for something else. I'm not sure though.

Setting VpcId on the security group to the neutron NetworkId seemed to make this work. Seems alittle sketchy though. Can someone garantee this behaviour is safe?

Thanks, Kevin

2013-10-16 14:39:07 -0600 asked a question heat security group id

How do you get a security group ID out of a created AWS::EC2::SecurityGroup?

OS::Neutron::Port only seems to take ID's, and Ref on the security group only gives you the name. I also want to pass it as an output, and again, it only gives you the name, not the ID.

Thanks, Kevin