Ask Your Question

arthur.lutz's profile - activity

2018-10-18 02:17:53 -0600 received badge  Famous Question (source)
2018-05-23 02:42:46 -0600 received badge  Notable Question (source)
2018-03-29 05:09:18 -0600 received badge  Popular Question (source)
2018-03-20 06:06:42 -0600 marked best answer [neutron] "ext-net" appears in "Available networks" but is unusable


I have an (almost) working setup of icehouse using neutron with ml2 and openvswitch (vlan config). The working workflow to create an VM with network access is the following : use a tenant-net that has a router that is both switched to tenant-net and ext-net, and then assign it a floating IP (which comes from ext-net).

The workflow that doesn't work is : create a VM directly on ext-net, which is available in the "Launch instance" > "Networking" > "Available Networks". This VM then gets a floating IP which is pingable but no access is possible.

I've read that it is not recommended to launch VMs directly on ext-net, which seems ok, but why does this option appear in the dashboard (that my users will be using) ? Is this configurable ? Should I file a bug/feature request ?

2018-03-20 05:56:54 -0600 asked a question [ocata] [magnum] [neutron] failing to create a cluster on "provider" network

On our ocata openstack install, we've chosen to configure the neutron "option1" "Provider networks" (so not the option with vxlan

This is quite nice and simple for VMs, they get the flat internal IP and we get less network complexity.

Now, we're trying to install and use magnum (which by the way doesn't seem to mention that it requires heat - which we installed for it). The "test the install" part says that creating a subnet is optional so we skipped that.

The template looks like this (trying to replace the external-network from public to provider)

magnum cluster-template-create --name swarm-cluster-template     \
                 --image fedora-atomic-ocata                \
                 --keypair arthur \
                 --external-network provider  \
                 --dns-nameserver   \
                --master-flavor m1.small   \
                --flavor m1.small  \
                 --coe swarm

For the creation of the cluster we get an error :

  magnum cluster-create --name swarm-cluster \
                        --cluster-template swarm-cluster-template

Here is the status_reason :

 ERROR: Failed to validate: : : HEAT-E99001 Service neutron is not available for resource type Magnum::Optional::Neutron::RouterInterface, reason: Required extension router in neutron service is not available.

So my question is this : is it possible to get magnum to work with the provider networks / flat approach or if we want to use it do we have to switch to the vxlan neutron approach ?

2017-10-16 13:55:57 -0600 received badge  Popular Question (source)
2017-10-16 13:55:57 -0600 received badge  Famous Question (source)
2017-10-16 13:55:57 -0600 received badge  Notable Question (source)
2017-08-17 09:49:36 -0600 asked a question [ocata] Can't add a resource_provider in placement-api

On a new install of ocata, I can't get new instances to schedule

I'm getting a bunch 404 and 409 on the placement-api :

"GET /resource_providers/3c476ac4-770f-4e12-b2cb-d3aca5eed4c3" status: 404 len: 125 microversion: 1.0

"POST /resource_providers" status: 409 len: 644 microversion: 1.0

"GET /resource_providers/3c476ac4-770f-4e12-b2cb-d3aca5eed4c3" status: 404 len: 125 microversion: 1.0

"GET /resource_providers/3c476ac4-770f-4e12-b2cb-d3aca5eed4c3/allocations" status: 404 len: 193 microversion: 1.0

On the compute side I'm getting :

Another thread already created a resource provider with the UUID 3c476ac4-770f-4e12-b2cb-d3aca5eed4c3. Grabbing that record from the placement API.

Unable to refresh my resource provider record

Final resource view: phys_ram=32159MB used_ram=512MB phys_disk=426GB used_disk=0GB total_vcpus=16 used_vcpus=0 pci_stats=[]

Any tips on how to debug this ?

Update : I solved this one, the resource had a different UUID in the nova_api resource_providers table (reinstall of the compute node), so I did :

mysql nova_api update resource_providers set uuid="3c476ac4-770f-4e12-b2cb-d3aca5eed4c3" WHERE name=''

2017-08-17 09:16:49 -0600 received badge  Critic (source)
2015-06-02 22:59:02 -0600 received badge  Nice Question (source)
2015-02-07 21:39:41 -0600 received badge  Famous Question (source)
2014-12-18 02:00:15 -0600 commented question serve SPICE html5 console using https

nope, no fix found. Have to find the time to adress this issue.

2014-12-17 16:07:14 -0600 received badge  Notable Question (source)
2014-12-17 16:07:14 -0600 received badge  Popular Question (source)
2014-12-15 23:19:13 -0600 received badge  Famous Question (source)
2014-12-08 07:20:37 -0600 commented answer Spice Client

This means that one can only use spice via a web browser, not through a client such as spicec which might have better performance.

2014-12-05 13:22:54 -0600 received badge  Notable Question (source)
2014-12-05 13:22:54 -0600 received badge  Famous Question (source)
2014-12-04 07:38:55 -0600 received badge  Popular Question (source)
2014-12-04 04:25:38 -0600 asked a question FlavorDiskTooSmall : is there a way to define a minimal flavor on in the image in glance

I've started using glance image properties to have better metadata about my images in glance

So I was wondering if there was a way to define the minimal flavor an image can run on in those properties. Updating them using glance image-update.

ps. while I'm at it, is there a way to use these properties to define the default username one has to use to login via ssh on the VM ?

2014-11-12 06:51:14 -0600 commented question [neutron] "ext-net" appears in "Available networks" but is unusable

up on this question?

2014-10-30 09:02:28 -0600 asked a question serve SPICE html5 console using https

We run our horizon UI on https and so when a browser disables mixed content (which it should) the spice console doesn't appear in the console tab. We're doing this on icehouse but the documentation doesn't seem to be different in juno.

I can't seem to find any configuration for this in or do I have to "just" replace http by https ?

I went to read the code /usr/share/pyshared/nova/cmd/ and there is mention of ssl_only, cert and key options. Where do I have to put these ?

Thanks in advance for this. Am willing to contribute some documentation.

2014-10-30 06:14:37 -0600 received badge  Notable Question (source)
2014-10-02 14:29:23 -0600 received badge  Popular Question (source)
2014-09-11 07:53:44 -0600 received badge  Enthusiast
2014-08-26 02:29:48 -0600 received badge  Citizen Patrol (source)
2014-06-16 15:14:43 -0600 received badge  Famous Question (source)
2013-11-04 01:44:26 -0600 received badge  Famous Question (source)
2013-11-01 09:57:11 -0600 received badge  Notable Question (source)
2013-11-01 08:13:25 -0600 received badge  Popular Question (source)
2013-10-31 10:19:21 -0600 asked a question Migration from essex to havana


I know that this question about migration from folsom to grizzly remains unanswered but I'll give it a go anyway.

We have an essex infrastructure in place. Any chance of migrating bits of this to a newer version of OpenStack ? If not, are there any recommended guidelines on how to make two infrastructures co-exist and the steps to take to "manually" migrate some machines from one to the other ?

2013-08-28 22:48:08 -0600 received badge  Notable Question (source)
2013-08-28 22:48:08 -0600 received badge  Popular Question (source)
2013-06-20 10:04:02 -0600 received badge  Famous Question (source)
2013-06-11 10:24:08 -0600 received badge  Notable Question (source)
2013-05-22 11:47:03 -0600 asked a question activating ""

I am trying to get the LDAP DNS backend working on Essex. In our infrastructure we already have a powerdns coupled with ldap to serve our domain names.

Reading /usr/share/pyshared/nova/network/ makes me think that the code could fit our needs if I were to active it and point to the right sub-branch of our LDAP.

How do I enable it ? in nova.conf ? Is there any documentation for this ?

2013-05-22 10:56:38 -0600 received badge  Popular Question (source)
2013-05-21 11:18:12 -0600 received badge  Student (source)
2013-05-21 11:10:36 -0600 received badge  Scholar (source)
2013-05-21 09:03:54 -0600 asked a question How does dnsmasq configuration get reloaded when dhcp_domain is set

On a running openstack deployement I am trying to modify dhcp_domain in /etc/nova/nova.conf but this isn't taken into account when I start an instance.

What service do I have to restart or which command do I have to issue to get it to run ?

I am looking at the --domain= in dnsmasq instances (I have over 30 on each compute node) but no new dnsmasq with my new domain name seems to pop up.

Additional question : is there a way to apply this new domain name to all running instances ?

2013-05-21 08:59:53 -0600 received badge  Supporter (source)