I'd suggest to have a look at the openstack Operations Guide which contains
information on scaling approaches
(http://docs.openstack.org/ops/)
In particular, the use of cells is a good way to build large scale
infrastructures (see
Michael,
It is desirable that a project, such as Trove, in incubation would be able to
install and run on Redhat derived systems as well as
Ubuntu/Debian.
With the combination of packstack/RDO, RedHat based deployment is becoming a
lot easier and I would hope that this is considered to
Doug,
Can you advise on what the plan/policy will be for Havana ?
- Will I be able to run Havana core components such as Nova with
Grizzly ceilometer ?
- Will I be able to run Havana ceilometer with Grizzly core components
(with reduced functionality compared to the
a key component of production, large scale
clouds.
Tim
From: Doug Hellmann [mailto:doug.hellm...@dreamhost.com]
Sent: 30 May 2013 22:35
To: Tim Bell
Cc: openstack@lists.launchpad.net
Subject: Re: [Openstack] [ceilometer] support for older versions of ceilometer
From what I understand
I have a hypervisor running KVM with two network interfaces to two independent
networks.
I could not find any documentation on how to configure a set up such as this so
the guests could also be configured with IPs on the
two networks.
Any pointers ?
Tim
smime.p7s
+2
I feel there is a difference between
- Is it possible ?
- Is it a good approach for use case X ?
Multi-site varies hugely depending on whether the aim is sharing the work or
recovery scenarios.
Defining a good model for multi-site, resilient deployments
I'd like to be able to
1. checkpoint a running virtual machine
2. run a test
3. rollback to the checkpoint from step 1
Has anyone had experience of doing this using OpenStack (such as with
snapshots) ?
Tim
smime.p7s
Description: S/MIME cryptographic signature
a driver from the list in the event
of failing to keep up with additional functionality (but this is a different
question to the documentation).
Tim
From: Blair Bethwaite [mailto:blair.bethwa...@gmail.com]
Sent: 16 April 2013 04:51
To: John Griffith; Tim Bell
Cc: openstack
The Boson project was looking at this sort of problem
(https://wiki.openstack.org/wiki/Boson).
There is a session at the summit to review this and other activities since
it appears quotas are appearing in many projects and there is a clear need
for multi-cell and quota delegation (i.e.
We have some additional checks on the names of instances such as
- Maximum length (to be compatible with Active Directory)
- Restrict characters which are usable
Is there a configurable set of parameters for these checks (or a user exit
to define a functional check) ?
Is there a method to do a basic service status check for OpenStack
components ?
I'd like to able to validate that the daemons are able to handle an
unauthenticated request from a monitoring script. This is somewhat
equivalent to a network level ping or nagios probe, i.e. validate that there
Is there a list of devices which are currently compatible with cinder and
their relative functionality ?
Looking through the source code, there are EMC, IBM, NetApp, Nexenta but it
is not clear which models are supported or if there are other products also.
A functionality matrix (like
There is also Robin Hood (http://en.wikipedia.org/wiki/Robin_hood)
steal
from the rich and give to the poor
Personally, Havana got my vote
.
Tim
From: openstack-bounces+tim.bell=cern...@lists.launchpad.net
[mailto:openstack-bounces+tim.bell=cern...@lists.launchpad.net] On
Sounds like activation key management would be a great feature for the
Windows cloud-init so this could be passed in as meta data.
Tim
From: openstack-bounces+tim.bell=cern...@lists.launchpad.net
[mailto:openstack-bounces+tim.bell=cern...@lists.launchpad.net] On Behalf Of
Balamurugan V
We run Active Directory with Keystone at CERN.
The configuration is documented by Jose in the Wiki at
http://wiki.openstack.org/HowtoIntegrateKeystonewithAD.
Not sure if all the patches made it into Folsom though.
Tim
From: openstack-bounces+tim.bell=cern...@lists.launchpad.net
A few of you have had trouble reading the document. The updated link is at
https://docs.google.com/document/d/1yD8TfqUik2dt5xo_jMVHMl7tw9oIJnEndLK8YqET
oyo/edit
If you can't read it, please send me a note and I'll share it.
Tim
From: Tim Bell
Sent: 09 January 2013 09:52
To: 'Sean
Updating the DB for the project would be one step but there are other
potential things to consider such as attached volumes.
Building up the set of things that need to be done in the doc would be
useful but eventually we need to get to an openstack command that can do
this (or do the equivalent
Sean,
Is the aim to provide co-ordination for the OpenStack user groups worldwide
?
Ryan, JC and I have been putting together the structure for the OpenStack
user committee (as described on the foundation list at
Lei,
At CERN, we run our RHEL-based OpenStack instance with the puppetlabs
modules. They work very well and simplify the multi-node deployments.
We've got around 200 hypervisors deployed with this method.
Tim
-Original Message-
From:
We were also interested in this function but could not find an easy way to do
it.
The operation becomes more complex when there are attached volumes and
potentially different permissions between the two projects.
Tim
From: openstack-bounces+tim.bell=cern...@lists.launchpad.net
There is a use case for base metal hardware metering in the private cloud
where the user allocated the machine does not have root access to kill the
metering.
Being able to create a single metering infrastructure for the entire private
cloud, virtual or bare-metal allocation, is a need
Have you had a look at cloud-init ?
Tim
- Reply message -
From: Ray Sun qsun01...@cienet.com.cn
To: openstack openstack@lists.launchpad.net
Subject: [Openstack] Is there any method to Activate Windows during Launch a
new Instance?
Date: Sat, Nov 3, 2012 03:04
I create a windows 7
Puppet is great for this sort of thing. There are various ways of querying
parameters and making choices about them. A typical example would be where
you want to adjust a configuration parameter due to memory configuration or
network.
Writing the puppet configuration is not difficult...
+1 for Boson ... this is a key area for CERN too. When you have a fixed budget
and no credit cards, quota management is a strong requirement!
Tim
-Original Message-
From: openstack-bounces+tim.bell=cern...@lists.launchpad.net
If we purchase an OUI, is there a mechanism within Quantum to only allocate
Mac addresses with that prefix ?
Tim
From: openstack-bounces+tim.bell=cern...@lists.launchpad.net
[mailto:openstack-bounces+tim.bell=cern...@lists.launchpad.net] On Behalf Of
Salvatore Orlando
Sent: 20 October
Cloud-init is also available on RHEL systems too. We're using it extensively
to contextualise VMs on Scientific Linux.
We've even tried a windows version but this is not as functional.
Tim
From: openstack-bounces+tim.bell=cern...@lists.launchpad.net
scripts to run during initialization to retrieve and install
ssh keys and to let the user run various scripts.
Tim
-Original Message-
From: Blair Bethwaite [mailto:blair.bethwa...@gmail.com]
Sent: 03 October 2012 14:39
To: Kiall Mac Innes
Cc: Tim Bell; openstack@lists.launchpad.net
Were the developer and operator lists included in your analysis ?
- http://lists.openstack.org/pipermail/openstack-dev/
- http://lists.openstack.org/pipermail/openstack-operators/
Full set is at http://wiki.openstack.org/MailingLists
Tim
-Original Message-
From:
I am not sure if I understand your question but Microsoft are making
substantial contributions towards OpenStack such as the work on Hyper-V. The
meeting logs are http://www.gossamer-threads.com/lists/openstack/dev/17175.
Peter Pouliot could give more details.
Tim
From:
There has been some load balancing discussion and more is due at the summit.
The various current activities are summarised in
http://wiki.openstack.org/Quantum/LBaaS
Can you explain what you mean by storage devices with respect to Quantum ?
The storage activities are underway as part of
Can we integrate the registration process for the CLA with that of the
membership ?
There is likely to be some additional questions such as affiliation and
interest in OpenStack, so it cannot be automatic but I would hope that most
contributors to the code base would also be interested in other
I think a new release should contains details of how to do the upgrade
(rather than discovering as we try it)
I should aim that the deliverables for each of the projects in a new version
includes in the release notes:
A. dependencies (i.e. does glance folsom need to talk to horizon folsom or
deployment experience.
I'm happy to help to consolidate the current pain points. We should gather the
data for both Diablo and Essex based deployments and it may well be that many
of them are addressed in Folsom (but we should check to be sure).
Tim Bell
CERN
On 27 Jul 2012, at 19:38
I thought that the v3 API supports domains as a group of tenants which would
make the question rather different.
Thus, I guess the question is
A. Should there be users in multiple tenants in a single domain ?
B. Should there be users in multiple domains ?
There are clear
Joyce wrote:
I could see service users and security / operations teams having a need to
span many domains.
-Matt
On Tue, Jul 17, 2012 at 11:24 PM, Tim Bell tim.b...@cern.ch wrote:
I thought that the v3 API supports domains as a group of tenants which would
make the question rather different
.
My question is if the user id name space is per domain or per IaaS instance
? I suspect this has significant implications in areas such as Horizon login
and API compatibility.
Tim
From: Joseph Heck [mailto:he...@me.com]
Sent: 18 July 2012 20:17
To: Tim Bell
Cc: Adam Young; openstack
+1 The corporate LDAP should be read-only for a source of user, roles and
attributes. Updating the corporate LDAP is not an option in many
environments which can significantly benefit from the structured directory
information available.
Thus, at minimum, allow a r/o LDAP and local DB store for
Vish,
How would the Nova migration from Essex to Folsom take place ? I'm wondering
how we can validate Folsom without risking an existing Essex installation
via some sort of clone/migrate operation.
What is your assessment of the risk that Cinder is less stable than Nova
volume ?
Option 1
Would it be possible a bit earlier ? This would be 10pm in Europe so it
would limit the participation.
Tim
-Original Message-
From: openstack-bounces+tim.bell=cern...@lists.launchpad.net
[mailto:openstack-bounces+tim.bell=cern...@lists.launchpad.net] On Behalf
Of Narayan Desai
HPC is often used as a general term but it is actually many different facets
depending on the computing model.
CERN is at the centre of a server grid of 100,000s of servers called WLCG
(http://wlcg.web.cern.ch) for analyzing the data from the Large Hadron
Collider. The servers are located at
The following may also be worth scanning:
- http://forums.openstack.org/
- Mailing lists on http://wiki.openstack.org/MailingLists (although quite a
few of them are quiet so would not affect the numbers much)
Tim
-Original Message-
From:
BTW, isn't the lack of option for a dashboard user to change their own
password considered as a missing functionality ?
Personally, if it was there, I'd like to turn it off (since we aim to use
Active Directory/LDAP) but for the standalone case, this would seem a
reasonable request.
Tim
]
Sent: 04 June 2012 21:50
To: Tim Bell
Cc: Guillermo Alvarado; openstack@lists.launchpad.net
Subject: Re: [Openstack] Adding panel to a dashboard - Horizon
Keystone doesn't currently support querying for capabilities (i.e. what is
the Keystone auth backend) so Horizon defines a dict
Sounds like the folsom solution is the right way to go.. If the backend can
do it, offer it..
Tim
From: Tres Henry [mailto:t...@treshenry.net]
Sent: 04 June 2012 22:27
To: Tim Bell
Cc: Guillermo Alvarado; openstack@lists.launchpad.net
Subject: Re: [Openstack] Adding panel
If youre using puppet for configuration management, we would strongly
recommend the puppetlabs openstack modules. They handle all the distribution
differences too...
They're working great for us with Scientific Linux so CentOS should work
well too...
Tim
CERN
-Original Message-
I would favour a field for the type of migration which allows future
expansion. we've already got migrate, live migrate and block migrate but
hypervisors may have further different flavours too in the future and the
API should support the full set of options while encouraging convergence
when
-tenant per user. Correct accounting is critical.
This does require extra code but it is relevant given the use cases.
Tim Bell
CERN
From: openstack-bounces+tim.bell=cern...@lists.launchpad.net
[mailto:openstack-bounces+tim.bell=cern...@lists.launchpad.net] On Behalf Of
Caitlin Bestler
CERN is more of a high throughput computing environment rather than a high
performance site. Although we do multi-core, our programs are not the large
scale floating point programs of some other sciences, so GPUs etc. are not a
significant gain for us.
CERN is currently targeting a
Fully agreed. Academic and Research sites have extensive X.509
infrastructure that we would not wish to duplicate.
Are you only looking at user certificates or are host certificates in the
scope too ?
Tim
From: openstack-bounces+tim.bell=cern...@lists.launchpad.net
I would also vote for an application prefix and error number for easy
searching. It is also great when you have to write the problem determination
guides.
Tim
-Original Message-
From: openstack-bounces+tim.bell=cern...@lists.launchpad.net
Andy,
Thanks for this work. OCCI is an interesting and open option as the world
moves towards federated clouds. I hope that this will continue the momentum
towards an open API which can address multiple backend IaaS solutions
transparently.
Tim Bell
CERN
From: openstack-bounces+tim.bell
To: Tim Bell
Cc: David Kranz; Andrew Clay Shafer;
openstack-operat...@lists.openstack.org; Duncan McGreggor; openstack
Subject: Re: [Openstack] [Ops] OpenStack and Operations: Input from the Wild
Hi Tim,
Could you elaborate more on 'performance metrics'? Like what kind of
metrics are considered
Splitting monitoring into
1. Gathering of metrics (availability, performance) and reporting in a
standard fashion should be part of OpenStack.
2. Best practice sensors should sample the metrics and provide alarms
for issues which could cause service impacts. Posting of these
sites will become factors in keeping the
early adopters enthusiastic. These are the user stories that will drive the
next wave of OpenStack growth as much as expanding the feature set.
Tim Bell
CERN
From: openstack-bounces+tim.bell=cern...@lists.launchpad.net
[mailto:openstack-bounces
It would be useful if there was a glossary of terms related to Openstack. It
is easy to get confused as many words are overloaded or slightly different
between different parts of Openstack.
There is also the page on identity at
This does bring up a more generic problem of sharing the
availability/performance code for all of the OpenStack components.
At the design summit, this was proposed as one of the example use cases of
the OpenStack community forge (I forget the exact name) but it was intended
as a place for
Fully agree with the prefix for the cell... there should be storage-cells
and compute-cells with different goals in terms of data locality and
availability, zone has become too overloaded...
Tim
-Original Message-
From: openstack-bounces+tim.bell=cern...@lists.launchpad.net
-Microsoft hypervisor ?
- Are there other sites who are affected by this proposal who would be
willing to invest effort to maintain the Hyper-V support ?
Tim Bell
CERN
-Original Message-
From: openstack-bounces+tim.bell=cern...@lists.launchpad.net
[mailto:openstack-bounces+tim.bell=cern
Looks interesting.
I wonder if it can be made more generic so that it would satisfy a similar
set of requirements for Chef (or whatever else comes along). I would
suspect that the general requirements for tables etc. are not puppet
specific, although the implementations may vary.
I would hope
A Dropbox like 'sync' function would be very interesting.. does anyone know
one which is compatible with OpenStack Swift ?
Tim
From: openstack-bounces+tim.bell=cern...@lists.launchpad.net
[mailto:openstack-bounces+tim.bell=cern...@lists.launchpad.net] On Behalf Of
Kuo Hugo
Sent: 19 December
Thierry,
I'm not clear on who will be maintaining the stable/diablo branch. The people
such as EPEL for RedHat systems need to have something with the appropriate bug
fixes back ported.
There are an increasing number of sites looking to deploy in production and
cannot follow the latest
We need more than 'just' packaging it is using the testing, documentation
and above all care to produce *and* maintain a stable release that production
sites can rely on for 6-12 months and know that others are relying on it too.
Who is going to make the judgement that a bug fix to the
The stable team with Duncan's additions would fully address my concerns.
Tim
___
Mailing list: https://launchpad.net/~openstack
Post to : openstack@lists.launchpad.net
Unsubscribe : https://launchpad.net/~openstack
More help :
We would be very interested to know where we can get a recent, stable Diablo++
release for RHEL 6. We appreciate the speed of development for Essex but there
is also a need for a pre-tested base on Diablo+fixes stable functionality for
the potential production users to integrate to.
Tim Bell
Let's not raise the bar too high for Satellite The points mentioned (CI,
QA, etc.) should be clearly documented with a y/n for the projects to aid
selection but I would much rather try out a non-CI tool in the directory than
have to write it all myself from scratch.
Ultimately, the aim
.
- Reply message -
From: George Reese george.re...@enstratus.com
To: Tim Bell tim.b...@cern.ch
Cc: openstack@lists.launchpad.net openstack@lists.launchpad.net
Subject: [Openstack] API Spec
Date: Sat, Aug 27, 2011 20:47
A cloud platform simply isn't functional without an API. It is a core
useful for these disaster recovery scenarios.
Tim Bell
CERN
___
Mailing list: https://launchpad.net/~openstack
Post to : openstack@lists.launchpad.net
Unsubscribe : https://launchpad.net/~openstack
More help : https://help.launchpad.net/ListHelp
The terms for snapshot and backup seem completely reasonable.
From a server consolidation/disaster recovery point of view, the dream would
be to be able to do backup and snapshot mirroring, almost like an Oracle
dataguard, to a remote site.
From my understanding of the upcoming Swift
to the other interfaces
as required.
Tim Bell
CERN
___
Mailing list: https://launchpad.net/~openstack
Post to : openstack@lists.launchpad.net
Unsubscribe : https://launchpad.net/~openstack
More help : https://help.launchpad.net/ListHelp
this situation, whereas using Erlang
may create some more difficulty further down the line. Anything that makes
porting/rebuilding more difficult needs to be carefully thought through.
Tim Bell
CERN
___
Mailing list: https://launchpad.net/~openstack
Post
70 matches
Mail list logo
