Maybe it is just me, but most reviews seem to take hours to days to
complete. I'm been sitting here waiting for a one line change to get
a second "+2 / approved" so I can redeploy our test cluster for the
last 2 hours.
When lots of people are active, reviews can be approved in quickly -
especiall
On Mon, Mar 12, 2012 at 07:54:38PM -0700, Justin Santa Barbara wrote:
> I just today was able to diagnose a libvirt hang. It appears to be
> (similar to) a known bug in libvirt, likely fixed in the latest
> Fedora, but it does not appear to be fixed in Ubuntu Oneirc; I think
> the fix is in Precis
On 03/13/2012 02:32 PM, Renuka Apte wrote:
> You need to install the xenapi plugins:
>
> If you are using devstack (https://github.com/openstack-dev/devstack), the
> code below from build_domU.sh does this for you.
I'm part of the debian Openstack packaging team, and I worked with Mike
on packagi
Dear all,
I had my first sleepless night last night after a conversation I had in
work regarding an OpenStack installation. We are a planning an OpenStack
install but I've failed to demonstrate OpenStack running in our environment
and yet packages are now being tagged up with RC. As big a support
2012/3/12 Josh Kearney :
> Is this is really a problem that needs solving? I'd like to believe
> that no member of Nova Core would approve something that they aren't
> familiar with.
That's not the point. The review process isn't just about finding enough
people who agree with your change, it's ju
Obviously, a better review algorithm is required,
To avoid hungry condition of some branches.
Now the whole system feels more like a client/server model.
Maybe a peer to peer like model behaves better.
在 2012-3-13，15:13，"Jesse Andrews" 写道：
> Maybe it is just me, but most reviews seem to take
Hi Joe,
On 12/03/12 21:59, Joe Gordon wrote:
> I have noticed that some Gerrit branches get approved very quickly,
> sometimes in a matter of minutes. While most of the time these
> branches are vetted properly, the window for reviewing can be so small
> that a non-trivial branch lands but witho
On 03/13/2012 09:28 AM, Kevin Jackson wrote:
> Dear all,
> I had my first sleepless night last night after a conversation I had in work
> regarding an OpenStack installation.
> Are my problems OpenStack's or Ubuntu's packaging? I would love to speak to
> someone who has this running and direct
2012/3/13 Jesse Andrews :
> Maybe it is just me, but most reviews seem to take hours to days to
> complete. I'm been sitting here waiting for a one line change to get
> a second "+2 / approved" so I can redeploy our test cluster for the
> last 2 hours.
Can we keep this discussion separate, please
One extra concern. Since I work in the UK, most things happen while I am
sleeping.
Not sure I know of a good solution to that problem. A 12 hour window seems
stupidly long, and I don't fancy becoming nocturnal.
Having said that, the core reviewers look to be doing a fine job at the moment,
so
Cheers Padraig - I'll grab a Fedora install and compare notes.
I guess if Fedora has an installation candidate, the problem is probably
Ubuntu packaging - at least I can direct my issues at Ubuntu rather than
OpenStack as a whole...
Kev
2012/3/13 Pádraig Brady
> On 03/13/2012 09:28 AM, Kevin J
Your stacktrace:
['XENAPI_PLUGIN_FAILURE', 'non-zero exit', '', '']
2012-03-13 05:41:32 ERROR nova.virt.xenapi.host [-] The call to
host_data returned an error: ['XENAPI_PLUGIN_FAILURE', 'non-zero
exit', '', ''].
2012-03-13 05:41:32 CRITICAL nova [-] 'str' object does not support
item assignment
(
> On Behalf Of Thomas Goirand
> which installs the above in the correct folders (which are *NOT* the same as
> for CentOS). I'd love to have a bit of help from Citrix to test all this by
> the way.
I hear you, and want to help out :-)
I hope we can setup some Jenkins tests on various XenServer/X
2012/3/13 John Garbutt :
> One extra concern. Since I work in the UK, most things happen while I
> am sleeping.
>
> Not sure I know of a good solution to that problem. A 12 hour window
> seems stupidly long,
No. No, it doesn't.
--
Soren Hansen | http://linux2go.dk/
Senior Software En
> I'm not using CentOS, but Debian (project Kronos). And in there, you just
> need to do:
> apt-get install nova-xcp-plugins nova-xcp-network
> which installs the above in the correct folders (which are *NOT* the same as
> for CentOS). I'd love to have a bit of help from Citrix to test all this by
Hi Kevin, sorry for the hard time you're having.
However, most of the methods you described, are NOT meant for
production deployments (not saying all, because I haven't tried
them all).
You might want to look at projects which aim to automate production deployments.
I can point you to the one I
Hi everyone,
The US moved to DST over the last week-end, which inevitably creates a
disturbance in the Force.
Please remember that the *meetings time are set in UTC time*. That means
the US people are having their OpenStack meetings one hour later than usual.
Cheers,
--
Thierry Carrez (ttx)
Re
Hello everyone,
Our weekly project & release status meeting will take place at *21:00
UTC* this Tuesday in #openstack-meeting on IRC. Old Essex and
newly-elected Folsom PTLs are welcome to join -- if you can't make it,
please name a substitute on [2].
We'll review the RC1 bug lists and evaluate E
Hi Andi,
Sure - the methods aren't meant for automated production installs, but to
get to a world where I can automate using Orchestra or variations on PXE
booting, I first need to get to a world where a manual install works.
I was getting desperate in having a demonstration available to those that
2012/3/12 Vishvananda Ishaya :
> Diagnostics are a bunch of key/value pairs, but unfortunately there doesn't
> seem to be an example of what they are. If someone running xen could post
> the result of get_diagnostics so that it could be added to a docstring and
> implemented for libvirt, that woul
Justin Santa Barbara wrote:
> Which operating system(s) are we aiming to support for Essex? Is the
> plan to backport the latest libvirt to Oneric, or are we going to wait
> for Precise?
The question is the other way around: which operating systems aim to
support Essex ? We try to set the depende
Venkat and Monsyne, feel free to coordinate with Jason Koelker and Mark
McLoughlin (cc'd) on the openstack-common stuff (which I very much support!)
Best,
-jay
On 03/13/2012 01:19 AM, Swaminathan Venkataraman wrote:
Hi,
I'd be happy to take a stab at that. I guess, the process would be to
move
Joe Gordon wrote:
> I have noticed that some Gerrit branches get approved very quickly,
> sometimes in a matter of minutes. While most of the time these
> branches are vetted properly, the window for reviewing can be so small
> that a non-trivial branch lands but without enough vetting. If someo
On 03/12/2012 08:58 PM, Adam Young wrote:
Also, these global objects force us to do a bunch of hacks in unit
tests. We need to do tricks to ensure the object is initialized as
we want. We also need to save and restore its state between runs.
I don't agree with the statement that
Hi Kevin,
As others have said, I can only speak to what I'm familiar with ( the
Managed I.T. packages).
They should be fairly up-to-date, there has been very few bugfixes to the
stable branch that I'm aware of since I last updated the packages.
Re EC2 compatibility, they are as compatible as the
2012/3/9 Hengqing Hu :
> Last few weeks I spend some time to speed up unittest in Nova.
This is great stuff! Thanks!
--
Soren Hansen | http://linux2go.dk/
Senior Software Engineer | http://www.cisco.com/
Ubuntu Developer | http://www.ubuntu.com/
OpenStack Developer | htt
On Tue, Mar 13, 2012 at 7:43 AM, Kevin Jackson
wrote:
> Hi Andi,
> Sure - the methods aren't meant for automated production installs, but to
> get to a world where I can automate using Orchestra or variations on PXE
> booting,
Well.. setting up a distributed system where all the pieces connect,
i
Thanks a lot for the points Andi (and the links), but not to stray away
from my original post: I don't want complexity at this stage - I want a
single server that works. I haven't even got that.
I had a multi node set up working (to some degree of butchery) of an early
Essex-1 release using Orches
On Mar 12, 2012, at 2:45 AM, Thierry Carrez wrote:
> Boris Renski Jr. wrote:
>> While I like the simplicity and elegance of the newly proposed
>> structure, I don’t see how it does away with the evils of the
>> pay-to-play model…. Which is what you purport we are striving to
>> achieve. What you,
Actually we can create such architecture without writing code. "Ring of
ring" is a classic Ring configured in specific way.
On Mon, Mar 12, 2012 at 9:49 PM, John Dickinson wrote:
> On Mar 11, 2012, at 12:16 AM, Dmitry Ukov wrote:
>
> > Hi all,
> > I want to introduce some ideas about Swift.
> >
On 03/13/2012 09:00 AM, Kevin Jackson wrote:
Thanks a lot for the points Andi (and the links), but not to stray away
from my original post: I don't want complexity at this stage - I want a
single server that works. I haven't even got that.
I had a multi node set up working (to some degree of but
I have some ideas about Swift multi-cluster organisation. Please give
feedback on
https://lists.launchpad.net/openstack/msg08518.html
On Sat, Mar 10, 2012 at 5:58 PM, John Leach wrote:
> Hi,
>
> I'm looking at deploying swift across two data-centres - I'd like to
> write 3 replicas: two in one
- Original Message -
> From: "Joe Gordon"
> To: openstack@lists.launchpad.net
> Sent: Monday, March 12, 2012 5:59:02 PM
> Subject: [Openstack] Gerrit minimum review time frame
>
> Hi All,
>
> I have noticed that some Gerrit branches get approved very quickly,
> sometimes in a matter o
Kevin.. would like to help you get this working.. and I understand the
confusion as I recently got all or our chef recipes in shape for RC1 and had to
work through the Keystone Light migration.
1. Where are you in the setup process?
2. Do you think you have a working Kesytone?
3. What services
Hi Joe,
On Mon, 2012-03-12 at 17:42 -0700, Joseph Heck wrote:
> I personally don't have a huge preference one way or the other. I've
> used a global object for configuration in applications, and find it
> immensely convenient, and the methods that Mark (and others) have put
> into the openstack/co
On Mon, Mar 12, 2012 at 2:10 PM, Eduardo Nunes wrote:
> I read all the docs about it but i dont get it, what release shold i use, i
> run the xcp under the linux or i run linux under the xcp?
You actually have two options for XCP. You can install XCP as a distro
itself (CD/PXE install) by downloa
Hi Kevin,
I like suggest you to start with devstack first. It is the shortest
way if you like to deploy a working openstack on a single machine
(takes just a few minutes). Anyway if you like to see a different
deployment scenario, it requires a proper planning and more time. In
the last two days I
On Tue, Mar 13, 2012, Mark McLoughlin wrote:
> On Mon, 2012-03-12 at 17:57 -0500, Josh Kearney wrote:
> > Is this is really a problem that needs solving? I'd like to believe that no
> > member of Nova Core would approve something that they aren't familiar with.
>
> I think the issue is more a cas
Thanks Marton,
I'm just running through another devstack install now. This failed for me
when I followed the instructions yesterday. Yours are in more detail so
hopefully some quirk of my environment might've caused an issue (don't see
why, but not ruling anything out).
My goal is a working exam
On Tue, Mar 13, 2012 at 3:20 PM, Marton Kiss wrote:
>
> My experience that success depends on proper OS environment and
> openstack component configuration.
>
Yup - I've also found from #openstack that anytime anybody attempts method
A, then method B without a complete operating system re-instal
On Tue, Mar 13, 2012 at 5:07 AM, Soren Hansen wrote:
> 2012/3/13 Jesse Andrews :
> > Maybe it is just me, but most reviews seem to take hours to days to
> > complete. I'm been sitting here waiting for a one line change to get
> > a second "+2 / approved" so I can redeploy our test cluster for th
In the wiki, we had laid out a target of the 1/3's being 4-5 members, so that
puts the target in 12-15 range. I think it's a reasonable size that allows
broad representation. I agree with Sean 20+ begins to be less effective.
In the current proposal with the Affiliate Member class, I'd imagine t
On Tue, Mar 13, 2012 at 7:17 AM, Thierry Carrez wrote:
> So I'd rather work on formally splitting the nova-core group into
> smaller topic-oriented review groups, which would ensure that code
> affecting a given area is reviewed by the experts of that area (rather
> than by whoever happens to be o
If u feel like it.
https://github.com/yahoo/Openstack-DevstackPy/wiki
Try it out :-)
On 3/13/12 8:29 AM, "Kevin Jackson" wrote:
Thanks Marton,
I'm just running through another devstack install now. This failed for me when
I followed the instructions yesterday. Yours are in more detail so ho
Sent from my iPhone
On Mar 13, 2012, at 5:17 AM, Thierry Carrez wrote:
> Joe Gordon wrote:
>> I have noticed that some Gerrit branches get approved very quickly,
>> sometimes in a matter of minutes. While most of the time these
>> branches are vetted properly, the window for reviewing can be
On Tue, Mar 13, 2012 at 3:58 PM, Josh Kearney wrote:
> On Tue, Mar 13, 2012 at 7:17 AM, Thierry Carrez
> wrote:
>>
>> So I'd rather work on formally splitting the nova-core group into
>> smaller topic-oriented review groups, which would ensure that code
>> affecting a given area is reviewed by th
There are a few inconsistencies currently with
the sqlalchemy version specifications in nova and glance.
tl;dr SQLAlchemy>=0.6.3 should be fine for Essex?
For e.g. the name "useexisting" was deprecated in
sqlalchemy-0.7.0 and changed to "extend_existing".
That name change was also made recently i
Hi all,
Thanks for those trying to help.
Indeed, it was a stupid Unix right issue. The python scripts of nova for
XCP aren't chmod +x in the Git (well, in fact, some are, some aren't
chmod +x). Now it does seem to work.
I'm testing more now, the debian packaging already chmod +x the python
files
On 03/13/2012 06:57 PM, John Garbutt wrote:
> A few possible issues:
> * you might not have the execute permissions set on the xenhost plugin file?
> On XenServer that means it fails to register the plugin.
Yup, it was that. Thanks for the hint. IMO, the error message should
have been a little bi
Hi , folks
I'm confusing about the usage of memcached on Swift-proxy serve. There's
the description in Swift dev-doc
http://swift.openstack.org/deployment_guide.html?highlight=token#memcached-considerations
Memcached Considerations
Several of the Services rely on Memcached for caching certain ty
Hi JE,
There are two places in the current master branch that use tpool:
NWFilterFirewall and XenAPISession. Are they safe?
I think if it's a pure C-based API call, then monkey patch should not
mess with it and it shouldn't try to reschedule among co-routines,
right?
After examining, the code, I
On Tue, Mar 13, 2012 at 5:49 PM, Kuo Hugo wrote:
> Also , this consideration keeps in my mind . But it seems not like what I'm
> thinking while authenticate by Keystone. Swift-proxy does not cache
> validated tokens in memcached . It always try to validate client's token by
> query keystone for
On Tue, Mar 13, 2012, Yun Mao wrote:
> There are two places in the current master branch that use tpool:
> NWFilterFirewall and XenAPISession. Are they safe?
I've looked at XenAPISession and it appears to be safe. It doesn't use
logging nor any other locks.
It does use other Python modules, but
I noticed that in the configuration file for yagi that there is no option
for storing messages for a time period but rather a limit of max_messages.
Does anyone know if there is a way to configure yagi to keep messages for a
set amount of time?
###yagi/etc/yagi.conf###
[consumer:notifications.wa
On 03/13/2012 02:28 AM, Kevin Jackson wrote:
I have tried the following package/scripted installations:
1) Ubuntu 11.10 with Diablo (Nova PPA) - some degree of success -
failures around Keystone and EC2 API
2) Ubuntu 11.10 with Managedit PPA - better degree of success, but no
EC2 compatibili
+1
I actually thought this same thought just yesterday. There have been 2-3 times
(in nova) when I was about to respond to an inline comment on one of my own
branches, saved as a draft, and then bam, it was approved by a couple people
and going in.
I'd love for there to be an unconference at
I'm just going to beat Thierry to the punch here - we need some assistance
squishing some bugs in Keystone. Keystone's behind in getting things closed out
for the RC1 cut (https://launchpad.net/keystone/+milestone/essex-rc1).
We've had a few new bugs crop up over the past day or two, and we need
I certainly understand your position Thierry. However, I think it is
important that we target one 'golden' platform, and that we take
responsibility for any issues with that platform. Otherwise we simply
end up pointing fingers and being blocked on backports, and the end
result is a system that j
On Mar 13, 2012, at 11:54 AM, Josh Kearney wrote:
> On Tue, Mar 13, 2012 at 5:07 AM, Soren Hansen wrote:
> 2012/3/13 Jesse Andrews :
> > Maybe it is just me, but most reviews seem to take hours to days to
> > complete. I'm been sitting here waiting for a one line change to get
> > a second "+2
On a different note,
Who all out there is providing public packages for OpenStack?
background:
Anne Gentle brought up wanting to reference package based installs for Keystone
(and other components) in the keystone meeting this morning (the second one,
when I got my timezones right). I wanted
On Tue, Mar 13, 2012, Lorin Hochstein wrote:
> As long as we're talking about review issues, I'd really like to have a
> set of reviewing guidelines. Even better would be to have them in
> checklist form to make sure we don't miss things as we do reviews
> (checklists are great: http://gawande.com
On 03/13/2012 03:08 PM, Joseph Heck wrote:
* https://bugs.launchpad.net/keystone/+bug/954089 (service-list returns empty
set for TemplatedCatalog backend)
On it.
* https://bugs.launchpad.net/keystone/+bug/954087 (endpoint-list with
TemplatedCatalog backend raises AttributeError)
Adam G's
On 03/13/2012 03:25 PM, Joseph Heck wrote:
> On a different note,
>
> Who all out there is providing public packages for OpenStack?
The Fedora Project is providing packages for both Fedora and EPEL (Extra
Packages for Enterprise Linux).
> background:
>
> Anne Gentle brought up wanting to refer
On 03/13/2012 03:08 PM, Joseph Heck wrote:
> Additionally, I'd like to ask anyone who's running under latest RHEL/Fedora
> and/or Debian/Ubuntu systems to try the packages and see if there's any
> explicit notes that we should be making related to installing Keystone. I
> haven't tried the Preci
Hi Jophep,
I can speak from Debian side. We provide OpenStack packages, but with
our target being Essex (maybe Folsom) + Debian Wheezy. In the case of
keystone, we support multiple database options (sqlite, mysql and
postgresql) that can be selected at install time (both interactively and
also
On 03/13/2012 10:55 PM, Todd Deshane wrote:
> On Mon, Mar 12, 2012 at 2:10 PM, Eduardo Nunes
> wrote:
>> I read all the docs about it but i dont get it, what release shold i use, i
>> run the xcp under the linux or i run linux under the xcp?
>
> You actually have two options for XCP. You can ins
max_messages actually corresponds to how many messages we try to send to the
configured recipient before it moves onto the next queue in the list.
Basically, it's super rudimentary QoS.
As for expiration, that's controlled by redis itself. Under the persistence
configuration section, you can se
Adam (and others I've not instantly replied to),
Thanks for the reply. I understand that OpenStack (moving target) on
Ubuntu (moving target) is no mean feat, but I no doubt expect a lot of
people are in the same boat as me. Ubuntu (and other distros) have set a
certain level of expectation for in
I understand the sentiment that leads one to want "minimum" review times; there
have been occasions where it's frustrated me that something I objected to got
merged without my seeing it or while I was trying to comment.
However, I'd like to present several arguments against a minimum review time
Just my 2 cents.
Shouldn't openstack be dependent only on a known set of package versions and
not a distro?
It seems like it should be the job of openstack to define those versions and
let the distro's match that as they choice.
If 12.04 matches these versions, then it should work, if somethin
yes, in your yagi.conf, in the persistence section, set the entry_ttl value
like so:
For 120 second expiration:
[persistence]
entry_ttl = 120
The value is in seconds, so for 1 hr use 3600, etc.
On Mar 13, 2012, at 1:50 PM, Craig Vyvial wrote:
> I noticed that in the configuration file fo
Is there a reason there isn't a unique contraint on display_names per
account for volumes, instances, etc? I see this potentially causing
some operational problems and customer confusion.Also, currently
the python-nova-client will only list the first one when you use the
"show" command.
What
I took a hack and made all C-based libvirt calls using tpool. It
wasn't that bad..
https://github.com/maoy/nova/commit/757bfcb239c62c23c8455a507389efe9c1a2676e
>From my limited testing, it works fine. At least snapshot doesn't
block the compute node anymore.
Yun
On Tue, Mar 13, 2012 at 2:04 PM,
Oh thats awesome! Looks like i just overlooked the definition of some of
the configuration parameters.
I found this:
default('entry_ttl', 60*60*24*30)
Thanks for that explanation its exactly what i needed.
-Craig Vyvial
On Tue, Mar 13, 2012 at 3:22 PM, Matt Dietz wrote:
> max_messages ac
On Tue, Mar 13, 2012 at 4:18 PM, Thomas Goirand wrote:
> On 03/13/2012 10:55 PM, Todd Deshane wrote:
>> On Mon, Mar 12, 2012 at 2:10 PM, Eduardo Nunes
>> wrote:
>>> I read all the docs about it but i dont get it, what release shold i use, i
>>> run the xcp under the linux or i run linux under th
On 12/03/12 17:42, John Dickinson wrote:
>
> On Mar 10, 2012, at 5:58 AM, John Leach wrote:
>>
>> I think what I need here is hierarchical zones - I'd define one
>> parent zone per data-centre, and then multiple child zones within
>> each (representing racks or whatever).
>>
>> Swift would be co
On 03/13/2012 01:53 PM, Kevin Jackson wrote:
Whilst OpenStack is being developed, a lot of people's entry into
OpenStack is through deb packages (or management in here>) - therefore Ubuntu becomes unofficial (but vocal)
PR to OpenStack. If the Ubuntu debs don't install, it becomes Plan B
to in
> there have been occasions where it's frustrated me that something I
> objected to got merged without my seeing it or while I was trying to
> comment.
I've heard of this several times in the past as well - some merge prop is
mid-review, and then it merges. Is there an easy mechanism to flag a r
Sent this to kevin earlier, thought i would throw it out to the list.. here are
the steps i take to get a working keystone and glance on Ubuntu-12.04 using the
ubuntu packages.
http://paste.openstack.org/show/9101/
These steps produce a working keystone and glance.. not 100% sure they are the
79 matches
Mail list logo
