Ask Your Question

Guillaume's profile - activity

2016-06-28 13:08:24 -0500 received badge  Great Answer (source)
2016-06-28 13:08:24 -0500 received badge  Guru (source)
2015-11-19 09:29:43 -0500 received badge  Enlightened (source)
2015-11-19 09:29:43 -0500 received badge  Good Answer (source)
2015-11-17 13:01:59 -0500 received badge  Nice Answer (source)
2015-09-16 09:54:18 -0500 received badge  Necromancer (source)
2015-08-05 11:22:44 -0500 received badge  Teacher (source)
2015-06-03 03:38:34 -0500 received badge  Famous Question (source)
2015-05-30 03:25:32 -0500 received badge  Notable Question (source)
2015-05-29 13:49:54 -0500 commented answer cinder-api and bind_host parameter

Thanks for pointing me to the cinder-volume A/A problem.

For cinder-api, even if i set my_ip to 192.168.100.205, the bind occurs on *:8776 and not on 192.168.100.205:8776. To work around the problem, i set the cinder endpoint and the VIP to port 8777.

2015-05-29 12:47:32 -0500 received badge  Popular Question (source)
2015-05-28 20:29:46 -0500 asked a question cinder-api and bind_host parameter

Hi,

My 3 controllers run all OpenStack APIs and HAProxy service. I must tell to cinder-api to run on a specific IP address to avoid conflict with HAProxy (port 8776) but i can't find how. For Glance, i used the bind_host variable to do that but it doesn't works with Cinder.

I looked in this documentation but i found nothing.

Thanks for your help.

2015-05-20 11:25:15 -0500 answered a question nova image-list returns HTTP 500

I had the same issue with Kilo and HAProxy setup. glance image-list worked fine but nova image-list gave me the same error as yours.

I solved the problem by setting registry_host value to point to my virtual IP in glance-api.conf :

/etc/glance/glance-api.conf

[DEFAULT]

registry_host = vip-int

Thanks to this post : https://ask.openstack.org/en/question...

2015-03-27 19:10:49 -0500 received badge  Supporter (source)
2015-03-22 10:56:49 -0500 received badge  Notable Question (source)
2015-03-22 10:56:49 -0500 received badge  Famous Question (source)
2015-03-22 10:27:55 -0500 answered a question swift juno permission denied error

Hi,

Define the right SELinux security context to /srv/node solved the problem for me : restorecon -R /srv/node

https://access.redhat.com/documentati...

2014-08-11 00:24:34 -0500 received badge  Popular Question (source)
2014-08-10 23:43:21 -0500 answered a question devstack: [Neutron] instances unreachables

Hi,

I faced this problem when trying to get Icehouse working on VirtualBox. I could ping the public host interface (192.168.50.50) from my instance (10.0.0.4) but i was unable to ping my physical router (192.168.50.1) from my instance or my internal router (interfaces on 10.0.0.0/24 and 192.168.50.0/24).

I solve my problem by enabling promiscuity mode in my VirtualBox network interface settings.

2014-08-10 14:05:51 -0500 answered a question Instance forever spawning with RDO quickstart example "running an instance"

Hi,

I already had this problem. Are you installing OpenStack on a virtual machine? If yes, you have to change the hypervisor type in /etc/nova/nova.conf. Search for virt_type and change it from 'kvm' to 'qemu'.

2014-08-03 14:43:15 -0500 received badge  Enthusiast
2014-07-27 20:50:35 -0500 received badge  Editor (source)
2014-07-27 20:50:35 -0500 edited question Terminating LXC instance fails

LXC instance fails to terminate with this message in nova-compute.log :

libvirtError: Failed to kill process 1618: Permission denied

1618 ? 00:00:00 dhclient

If i kill this process before trying to terminate the instance, it works.

I have this problem with the official Ubuntu 14.04.1 cloud image and also with my own LXC images created with lxc-create.

Thanks for your help.
Host system: Ubuntu 14.04.1, Icehouse and LXC 1.0.5

UPDATE 1
I replaced dhclient with dhcpcd in my Ubuntu image and the problem is gone. The instance get his IP from Nova and it can be terminated without error. For dhclient, maybe something in dhclient.conf could solve the problem?