I like GRE for the tenant network, floating or provider network for specific
accounts.
Contact me off line if you need to talk.
Remo
On Jun 13, 2014, at 20:07, gustavo panizzo wrote:
> On 06/13/2014 07:28 PM, Michael Gale wrote:
>> Hello,
>>
>>Does anyone have any experience mixing a
On 06/13/2014 07:28 PM, Michael Gale wrote:
> Hello,
>
> Does anyone have any experience mixing and matching Provider network
> installations with Tenant network installations all using VLANs?
i have been doing it in prod since grizzly, it works, no problems there :)
i don't use many float i
I have but u would use gre.
Inviato da iPhone ()
> Il giorno Jun 13, 2014, alle ore 15:28, Michael Gale
> ha scritto:
>
> Hello,
>
> Does anyone have any experience mixing and matching Provider network
> installations with Tenant network installations all using VLANs?
>
> Assuming:
>
Hello,
Does anyone have any experience mixing and matching Provider network
installations with Tenant network installations all using VLANs?
Assuming:
- Icehouse release
- VLAN instead of GRE
For a large internal cloud we planed to use a Provider network VLAN for
almost all VM's. This is
Resolved: Turned out to be an iptables issue. I stopped iptables and
the install went through.
On Fri Jun 13 16:59:00 2014, Eric Berg wrote:
I'm installing a new RDO compute host and seeing this message in the
logs. Nothing helpful has turned up on google.
nova.openstack.common.db.sqlal
On 06/13/2014 03:45 PM, Craig Jellick wrote:
> We use AD as the identity backend and MySQL as the assignment backend, but I
> don't see how the backends would affect what I would want to do.
i do identity and assignment in ldap. that way i can do things like you
asked
the link of Adam's blog yo
I'm installing a new RDO compute host and seeing this message in the
logs. Nothing helpful has turned up on google.
nova.openstack.common.db.sqlalchemy.session [-] SQL connection
failed. infinite attempts left
Any ideas as to what's up here? It's a packstack install from the
controller h
Answering my own question. Looks like this will walk me through exactly
what I want:
http://adam.younglogic.com/2013/11/policy-enforcement-openstack/
/Craig J
On 6/13/14 11:45 AM, "Craig Jellick" wrote:
>We use AD as the identity backend and MySQL as the assignment backend,
>but I don't see
We use AD as the identity backend and MySQL as the assignment backend, but I
don't see how the backends would affect what I would want to do.
Sent from my iPhone
> On Jun 13, 2014, at 11:22 AM, "gustavo panizzo "
> wrote:
>
>> On 06/13/2014 02:57 PM, Craig Jellick wrote:
>> Has anyone setup a
On 06/13/2014 02:57 PM, Craig Jellick wrote:
> Has anyone setup a "project admin" rule for keystone?
> Let me explain what I mean by that rule to be clear:
> it should allow a user to add and remove other users to projects to
> */which he belongs/*. Meaning, as a project admin for project foo, I
>
Has anyone setup a "project admin" rule for keystone?
Let me explain what I mean by that rule to be clear:
it should allow a user to add and remove other users to projects to which he
belongs. Meaning, as a project admin for project foo, I should be able to
add/remove users to project foo, but no
On 06/13/2014 04:13 AM, Ricardo Carrillo Cruz wrote:
> Hi there
>
> I installed a one node Icehouse cloud at home, with Neutron and GRE.
> Everything is ok, except I can't ssh to ubuntu instances.
> The problem is that cloud-init is unable to reach the metadata service, thus
> the
> keypair is ne
Hi, is there any nova/neutron option that i can enable to create instances
without subnets?
i created the network without subnet and no problem there, but when i try
to create/boot i get the error
Error: Network 2a5a6826-f9df-4767-8885-21d92541fddc requires a subnet in
order to boot instances on.
Hi,
can you please elaborate?
Ty
Cumprimentos,
Tiago Sousa
On Fri, Jun 13, 2014 at 5:53 PM, Remo Mattei wrote:
> You have to use the provider network
>
> Ciao
> On Jun 13, 2014, at 9:46, Tiago Sousa wrote:
>
> > Hi, is there any nova/neutron option that i can enable to create
> instances wi
Hi Kumar,
I have also faced this during Havana times, but did not see the issue from
Icehouse.
If possible try with a later release.
Regards,
Sudhakar
On Fri, Jun 13, 2014 at 2:02 PM, wrote:
> Hi All,
>
>
>
> Any solution found so far ?
>
>
>
> Thanks
>
> Kumar
>
>
>
> *From:* Gnan Kumar, Ya
Remo,
I have a network node (running l3 agent) plus 4 compute nodes. The compute
nodes have physical interfaces on the storage net.
Are you saying I'd have to add a physical interface to the network node for
the storage network and run this storage net though the network node?
Thanks,
Marc
On
then you need to do that in the l3 config file. You can use the * and use
anyname you want to map it.
Bye,
Remo
On Jun 13, 2014, at 9:20, Marc Soda wrote:
> ML2.
>
>
> On Fri, Jun 13, 2014 at 12:16 PM, Remo Mattei wrote:
> are you using ML2 or not?
> On Jun 13, 2014, at 8:52, Marc Soda w
are you using ML2 or not?
On Jun 13, 2014, at 8:52, Marc Soda wrote:
> My compute nodes all have a separate NIC for a storage network. How can I
> map Neutron interfaces to this physical interface?
>
> !DSPAM:1,539b2310109581051517306!
> ___
> Maili
Actually the requirement is for an NFS share to be mounted to several VMs
over this network. I would like to setup block storage here eventually,
though.
On Fri, Jun 13, 2014 at 12:16 PM, Martinx - ジェームズ wrote:
> Hey Marc,
>
> When you say, storage network, do you mean your iSCSI traffic
> (c
ML2.
On Fri, Jun 13, 2014 at 12:16 PM, Remo Mattei wrote:
> are you using ML2 or not?
> On Jun 13, 2014, at 8:52, Marc Soda wrote:
>
> > My compute nodes all have a separate NIC for a storage network. How can
> I map Neutron interfaces to this physical interface?
> >
> > !DSPAM:1,539b23101095
Hey Marc,
When you say, storage network, do you mean your iSCSI traffic
(cinder-volumes) of block storage services?
Cheers!
Thiago
On 13 June 2014 12:52, Marc Soda wrote:
> My compute nodes all have a separate NIC for a storage network. How can I
> map Neutron interfaces to this physical in
That's what I needed to know. So, when configuring the compute nodes, I can go
"native" rather than using qemu on them, which I understand is preferable from
a performance standpoint, correct?
From: Remo Mattei [mailto:r...@italy1.com]
Sent: Friday, June 13, 2014 9:40 AM
To: O'Reilly, Dan
Cc: o
Each compute node will have nova.conf where you specify that info. So if you
use kvm, or qemu etc.. so you can have two compute with different language if
you think in that way. One speaking italian, one english.
Bye,
Remo
On Jun 13, 2014, at 8:38, O'Reilly, Dan wrote:
> But does the contro
On 06/13/2014 11:22 AM, Michael Hearn wrote:
Horizon gurus
Release: icehouse
Token Type : PKI
Identity Backend: LDAP
Monitoring the authentication traffic generated by Horizon to LDAP, I
was surprised to see that after the initial logon, and under the
'Project' tab, I was still seeing
My compute nodes all have a separate NIC for a storage network. How can I
map Neutron interfaces to this physical interface?
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsub
But does the controller speak the same to any compute node, regardless of the
underlying hypervisor, or is there something special for KVM vs Xen, for
example? What I'm trying to do is not configure qemu if I don't need to.
From: Remo Mattei [mailto:r...@italy1.com]
Sent: Friday, June 13, 2014
Right. So, to do that, do I have to use something like qemu, or does the
controller care? Which is another way of asking, does the controller need to
know that 1 node is Xen, one is KVM and one is VMware? And if so, how is that
configured?
From: Remo Mattei [mailto:r...@italy1.com]
Sent: Fri
Horizon gurus
Release: icehouse
Token Type : PKI
Identity Backend: LDAP
Monitoring the authentication traffic generated by Horizon to LDAP, I was
surprised to see that after the initial logon, and under the 'Project' tab,
I was still seeing calls out to LDAP each time I entered a link rel
Nope the controller does not care where it runs. AZ allows you to select the
dest and its’ pretty nice since you can have a sever with SSD or VMware etc…
On Jun 13, 2014, at 8:12, O'Reilly, Dan wrote:
> Right. So, to do that, do I have to use something like qemu, or does the
> controller car
Ah, yes. But does that still affect how Openstack (specifically, the
controller and compute nodes) must be configured to use Xen/KVM/VMware?
Thanks!
From: Remo Mattei [mailto:r...@italy1.com]
Sent: Friday, June 13, 2014 9:08 AM
To: O'Reilly, Dan
Cc: openstack@lists.openstack.org
Subject: Re: [O
Availability Zone.
On Jun 13, 2014, at 8:03, O'Reilly, Dan wrote:
> AZ?
>
> From: Remo Mattei [mailto:r...@italy1.com]
> Sent: Friday, June 13, 2014 9:00 AM
> To: O'Reilly, Dan
> Cc: openstack@lists.openstack.org
> Subject: Re: [Openstack] Compute servers configuration - best practices
>
>
No,
you need to configure your nova.conf file on each node so then you are set.
On Jun 13, 2014, at 8:10, O'Reilly, Dan wrote:
> Ah, yes. But does that still affect how Openstack (specifically, the
> controller and compute nodes) must be configured to use Xen/KVM/VMware?
>
> Thanks!
>
>
AZ?
From: Remo Mattei [mailto:r...@italy1.com]
Sent: Friday, June 13, 2014 9:00 AM
To: O'Reilly, Dan
Cc: openstack@lists.openstack.org
Subject: Re: [Openstack] Compute servers configuration - best practices
Why don't you use AZ for this. So you can select where you want to run a
specific VM. The
Why don’t you use AZ for this. So you can select where you want to run a
specific VM. Then you have VM on KVM, XEN etc.. based on the AZ.
Remo
On Jun 13, 2014, at 7:41, O'Reilly, Dan wrote:
> I realize there have been some discussions on this in the recent past, but it
> has become somewhat
I realize there have been some discussions on this in the recent past, but it
has become somewhat difficult to follow. So, I would like to take a fresh
approach with a fresh question.
I have the following configuration:
- Controller with identity, dashboard, orchestration, image, te
George and Anne,
Thank you. I'll dig into the security guide and look forward to the
architecture guide next month.
//Daniel
On Thu, Jun 12, 2014 at 4:07 PM, Anne Gentle wrote:
>
>
>
> On Thu, Jun 12, 2014 at 8:51 AM, George Mihaiescu > wrote:
>
>> Hi Daniel,
>>
>>
>>
>> It’s recommended t
Hello,
I don' t know is this bug or not, so I wrote it to list.
Compute nodes in horizon shows with short hostname in "Hypervisors"
section. In other sections like "Host Aggregates" and "Instances" it
shows with full hostname. When I perform migration it shows short
hostname also.
It will be
I am in IceHouse and as an ADMIN I am trying to list the instances on a
specific tenant.
I have the following tenants:
# keystone tenant-list
+--+-+-+
|id| name | enabled |
+--+---
On 13/06/14 13:22, Mark Kirkwood wrote:
On 06/06/14 13:32, Mark Kirkwood wrote:
On 06/06/14 11:22, Mark Kirkwood wrote:
On 06/06/14 07:58, Pete Zaitcev wrote:
On Thu, 05 Jun 2014 17:14:43 +1200
Mark Kirkwood wrote:
The swift logs on the storage don't show anything odd (200
responses to
the
Hi ,
Can nova boot several servers from just one cinder volume?
Or Can I boot several servers from a bootable volume without making
serveral volumes of same image?
I don't know whether the copy-on-write(cow) block image is supported by
openstack now,
which I means a base image volume can be crea
Hi All,
Any solution found so far ?
Thanks
Kumar
From: Gnan Kumar, Yalla
Sent: Wednesday, June 11, 2014 1:07 PM
To: openstack@lists.openstack.org
Subject: High CPU usage
Hi All,
I have a three node Havana setup of openstack. Whenever I launch an instance
and terminate it, the load on the CPU
On Jun 12, 2014, at 11:16 PM, Pete Zaitcev wrote:
> On Thu, 12 Jun 2014 09:52:59 +0100
> Diogo Vieira wrote:
>
>> Ok, I guess that might work, but I have one problem with that approach.
>> For a service I'm developing I have to know the public URL for an object
>> in the store. For that I use
Hi there
I installed a one node Icehouse cloud at home, with Neutron and GRE.
Everything is ok, except I can't ssh to ubuntu instances.
The problem is that cloud-init is unable to reach the metadata service,
thus the keypair is never injected into the instance.
This is the error:
cloud-init sta
43 matches
Mail list logo
