Hi,
The current handling of the 'SQL server has gone away' is different across
OpenStack projects (eg. Nova uses MySQLPingListener, whereas Glance retries the
db operation to recover connection).
I am curious to know if the fix implemented in
https://review.openstack.org/5552 will be used acros
Hi All,
I had recently submitted changes to the Python novaclient repo for adding the
support to specify user_data, availability_zone and security_groups in the
'nova boot' command. The OpenStack API v1.1 spec for the create server API does
not mention these parameters. Shouldn't these paramete
Trinath,
You probably have a wrong config setting. Please check the value of
'network_api_class' in your nova.conf and confirm that the module with the
specified class exists.
The default value of this parameter is "nova.network.api.API", so you
should expect class 'API' in nova/network/api.py of
Greg,
IMO, this qualifies as a bug.
To debug what's causing this scenario is your case, I would suggest
checking the database connection setting in nova.conf on your Compute &
Controller host. AFAIK, it either means Compute's status is not being
updated (Compute service down) or its being updated
Swapnil,
>From the n-api logs available on the bug, it looks like the EC2 port
(default 8773) is being used by some other service. You can check this
using 'netstat -pan |grep 8773'.
You can either configure the EC2 port (by specifying 'ec2_listen_port' in
nova.conf) or configure the other service
Kazim,
In any case, the command should have returned either a successful or failed
response. It might be helpful if you paste the exact output of this command.
To move forward, I would suggest having a look at the Glance API logs,
which would indicate what is going wrong in your case. You might a
Javier,
The stack trace you provided indicates that the allocate network call from
Compute (to nova-network) timed out. I would suggest checking the
nova-network logs, that might give some insight.
Also, IMO, the create instance request is being served by Compute means the
RabbitMQ is working. You
Hi Paras,
AFAIK, this should be possible in grizzly.
Here is the related bug: https://bugs.launchpad.net/nova/+bug/1074054
--
Thanks & Regards,
Unmesh G.
On Wed, Apr 17, 2013 at 10:04 PM, Paras pradhan wrote:
> Hi,
>
> If we boot a instance from cinder volume. Is it possible to setup live
> m
Kamaill,
I think that is caused because of the wrong HTTP header in the second curl
command. It should be 'X-Auth-Token' instead of 'XX-Auth-Token'.
Give it a try and get back if you have any issue.
-Unmesh.
On Tue, Apr 23, 2013 at 1:17 PM, Study Kamaill wrote:
> THis is the output :: --
Arindam,
You can do it using 'nova-manage service disable --host
--service ' command. Similarly, you can enable the network service
in controller node.
-Unmesh.
On Thu, Apr 25, 2013 at 9:16 PM, Arindam Choudhury wrote:
> Hi,
>
> I mistakenly installed nova-network in a compute node and its
Hi,
AFAIK, there are no additional changes required (other than restarting the
service so that it loads the new conf settings). Assuming you have done
that, can you check what is the 'volume_driver' from the service start up
logs?
~Unmesh.
On Mon, Jul 15, 2013 at 10:44 AM, zan tosh wrote:
> R
Hi Justin,
Looking at the endpoints in the auth response (v2.0/tokens), probably the
URL should be
http://openstack.cloud.fandingo.org:8776/v1/4457698fa26d4ce59f1811e91521f6c4/volumes/detail
instead
of
http://openstack.cloud.kerrygroup.net:8776/v1/4457698fa26d4ce59f1811e91521f6c4/volumes/detail.
12 matches
Mail list logo
