Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2013-12-19

| Channels | #puppet-openstack index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
23:08 xarses dmsimard: that works now?
23:10 dmsimard xarses: Wouldn't have had commited otherwise :p
23:13 xarses waiting for the other bot
00:27 prad joined #puppet-openstack
00:52 openstackgerrit A change was merged to stackforge/puppet-cinder: Using contain_class instead of include_class  https://review.openstack.org/62799
00:55 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-ceilometer: Updated Modulefile for Stackforge/Icehouse  https://review.openstack.org/63032
00:55 openstackgerrit Sabari Murugesan proposed a change to stackforge/puppet-neutron: Add support for neutron nvp plugin  https://review.openstack.org/63033
00:56 otherwiseguy joined #puppet-openstack
01:01 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-ceilometer: Updated Modulefile for release candidate 2  https://review.openstack.org/63036
01:19 rwsu joined #puppet-openstack
01:49 slowe joined #puppet-openstack
01:56 markvoelker1 joined #puppet-openstack
02:12 rongze joined #puppet-openstack
03:44 badiane_ka joined #puppet-openstack
05:18 rongze joined #puppet-openstack
05:42 openstackgerrit Chip Baeseman proposed a change to stackforge/puppet-openstack: Allow running of proxy and storage on the same node.  https://review.openstack.org/62746
05:54 openstackgerrit Chip Baeseman proposed a change to stackforge/puppet-openstack: Allow running of proxy and storage on the same node.  https://review.openstack.org/62746
05:59 openstackgerrit Sabari Murugesan proposed a change to stackforge/puppet-neutron: Add support for neutron nvp plugin  https://review.openstack.org/63033
06:12 tnoor joined #puppet-openstack
06:20 rongze joined #puppet-openstack
06:27 rongze joined #puppet-openstack
06:37 rongze joined #puppet-openstack
06:37 tnoor joined #puppet-openstack
06:43 JonnyNomad joined #puppet-openstack
07:11 xarses joined #puppet-openstack
07:12 JonnyNomad joined #puppet-openstack
07:49 rongze joined #puppet-openstack
08:01 JonnyNomad joined #puppet-openstack
08:05 EmilienM good morning puppies
08:06 EmilienM michchap: if you have some time, I would like you to have a look at https://review.openstack.org/#/c/62535/ and https://review.openstack.org/#/c/55578/ (I made deep review yesterday)
08:11 slowe joined #puppet-openstack
08:16 mattymo puppies :)
08:38 openstackgerrit Benedikt Trefzer proposed a change to stackforge/puppet-neutron: include class neutron to ml2 plugin class Without this inclusion the check for the correct core_plugin (line 147ff) fails since $::neutron::core_plugin is not available in the context.  https://review.openstack.org/63068
08:39 openstackgerrit Benedikt Trefzer proposed a change to stackforge/puppet-neutron: Include class neutron to ml2 plugin class  https://review.openstack.org/63068
08:43 slowe Using 3.0.0-rc1 of puppet-keystone and it seems that when using allowed_hosts with keystone::db:mysql, it ignores setting permissions for localhost. Is this is a known issue? Perhaps I'm doing something wrong.
08:49 michchap yay puppies
08:53 EmilienM \o/
08:56 mattymo michchap, did you ever move closer to using ruby-openstack in puppet?
08:59 michchap ruby-openstack?
09:00 michchap mattymo: I don't think I was ever working on it.
09:00 mattymo oh my mistake
09:00 mattymo someone else from cisco was looking at it
09:52 mattymo michchap, EmilienM any of you try to deploy swift with Ubuntu in Havana yet?
09:52 EmilienM mattymo: yes, even icehouse
09:52 mattymo we're running into this: https://bugs.launchpad.net/swift/+bug/1262166
09:53 EmilienM damn
09:54 mattymo if you look at the original description for the bug, you can see some analysis of common functions called
09:54 EmilienM mattymo: i'm checking with our swift devs
09:55 mattymo it looks like for an empty swift server when balancing a single node server, it calls sorted() 3.1 million times
09:56 mattymo it's supposed to reduce the # of rebalance tries, but our py devs in Fuel think swift balancer is trying to reevaluate far too many times in the least optimal way
09:57 mattymo and apparently it doesn't break in Python 2.6 for CentOS, but in Python 2.7 the sort function acts differently
09:57 EmilienM mattymo:  they don't think it's a bug - it's an improvement; because rings are more balanced.
09:57 EmilienM mattymo: sorry to be that guy, I'm not sure it's the right place here
09:58 mattymo it shouldn't take 8 minutes
09:58 EmilienM not sure people here could help you on swift
09:58 mattymo well, I'm just wondering if you encountered it
09:58 mattymo but if your swift devs at your company want to see the bug, we welcome it)
09:58 EmilienM mattymo: not yet :(
09:58 EmilienM mattymo: one of them is looking at it.
09:59 EmilienM they also use to discuss on #openstack-dev or #openstack-swift
09:59 mattymo yeah I'm there
09:59 mattymo it's just still early for most devs. I don't expect a lot of help until the US wakes up
10:04 michchap mattymo: I haven't worked on swift personally.
10:05 EmilienM michchap: i answered to the patch I set -2, I don't want to abandon it. IMO there is a bug somewhere
10:08 mattymo michchap, we want to obsolete swift.. it's too primitive and slow
10:09 mattymo but it's still a core project
10:09 michchap EmilienM: Maybe he thought -2 meant to abandon
10:09 EmilienM mattymo: to replace by what ?
10:09 EmilienM mattymo: Ceph ?
10:09 michchap mattymo: ceph is our default, we don't deploy swift unless someone asks for it
10:10 EmilienM michchap: that's why i answered. I would like to see at least a bug in LP, here we are digging the issue with neutron devs, I'm quite sure he's not wrong at all. The patch was just not ready for merge.
10:10 michchap mattymo: Swift is useful but for specific tasks
10:10 mattymo yeah we use ceph too
10:10 mattymo but we only offer Fuel with Ceph with Inktank support
10:11 michchap yep
10:11 michchap I think that is the way things are going
10:11 mattymo but if you are unpaid you can use ceph all you want
10:11 mattymo we're also battling a clone instance bug with ceph
10:11 michchap bug in ceph or bug in the driver?
10:12 mattymo https://bugs.launchpad.net/fuel/+bug/1260911
10:12 mattymo CoW copying bug in glance
10:14 derekh joined #puppet-openstack
10:23 rongze joined #puppet-openstack
10:40 mjeanson joined #puppet-openstack
10:44 sanman joined #puppet-openstack
12:04 rongze joined #puppet-openstack
12:24 rongze_ joined #puppet-openstack
12:54 prad joined #puppet-openstack
13:32 djschaap joined #puppet-openstack
13:41 djschaap EmilienM: good afternoon (morning for me)
13:41 EmilienM djschaap: o/
13:41 EmilienM Doug
13:41 EmilienM good to see you here :)
13:42 djschaap very new here. reading yesterday's chat logs.
13:42 EmilienM djschaap: cool, welcome here
13:42 djschaap I'll file a bug report on the neutron firewall_driver issue.
13:42 EmilienM djschaap: that's awesome
13:42 EmilienM djschaap: i did not want to be "mad" by -2 the patch, it was just to avoid accidental merge
13:43 djschaap I'll restore the change, too.  Still learning the gerritt workflow.
13:43 EmilienM djschaap: but in the other hand, I don't want to abandon this stuff
13:43 EmilienM djschaap: I have one question though
13:43 djschaap I understand!
13:43 EmilienM djschaap: could you confirm that you made SG working with ML2 in Havana by setting firewall_driver in ml2 config file ?
13:45 djschaap Yes. Only after I set firewall_driver in ml2's config (symlinked from /etc/neutron/plugin.ini) did the neutron securitygroup API respond.
13:46 djschaap I don't know if it actually works end-to-end - but without that, the API was missing.
13:47 EmilienM djschaap: my question is, if you set the flag in ovs.ini, is it working ?
13:50 dprince joined #puppet-openstack
13:50 djschaap If firewall_driver is set ONLY in ovs.ini, the securitygroup API returns 404 Not Found.
13:54 EmilienM dsockwell: that's weird
14:04 djschaap bug 1262678 is now filed for firewall_driver
14:17 EmilienM djschaap: awesome
14:18 EmilienM djschaap: i'm working on investigation now.
14:22 piliman974 joined #puppet-openstack
14:32 EmilienM djschaap: one quesiton
14:32 EmilienM if you move
14:32 EmilienM [securitygroup]
14:32 EmilienM firewall_driver=neutron.agent.linux.iptables_firewall.OVSHybridIptablesFirewallDriver
14:32 EmilienM to ovs.ini config file ?
14:32 EmilienM does it work ?
14:32 mattymo joined #puppet-openstack
14:35 djschaap You mean /etc/neutron/plugins/openvswitch/ovs_neutron_plugin.ini right?
14:36 djschaap I've always had it set there.
14:36 EmilienM djschaap: yeah that place
14:37 EmilienM ok that what I would like know
14:37 djschaap I think that's necessary for the actual filtering to happen - but I haven't got that far yet.
14:47 EmilienM djschaap: i'm not sure your bug is relevant now.
14:47 EmilienM djschaap: maybe a packaging issue though
14:48 EmilienM djschaap: you use red hat
14:48 EmilienM djschaap: and it seems that you have a symlink for config file
14:56 djschaap that is possible. I'd like to find the neutron code that determines which APIs to load/enable.
14:58 EmilienM djschaap: neutron/plugins/ml2/plugin.py line 95
15:00 djschaap thx
15:03 prad joined #puppet-openstack
15:03 dmsimard joined #puppet-openstack
15:04 EmilienM djschaap: https://github.com/openstack-dev/devstack/blob/master/lib/neutron_plugins/ml2#L92-103
15:10 djschaap EmilienM: that confirms my guess - ...not.a.real.FirewallDriver is enough
15:12 EmilienM djschaap: i'm digging with neutron devs now, i'll let you know
15:16 otherwiseguy joined #puppet-openstack
15:24 openstackgerrit A change was merged to stackforge/puppet-nova: Fix wrong vif driver by default  https://review.openstack.org/61708
15:25 EmilienM a good one got merged ^ thx fc__ . This is an update for icehouse though, no need to be backported.
15:25 openstackgerrit A change was merged to stackforge/puppet-nova: Using contain_class instead of include_class  https://review.openstack.org/62802
15:28 EmilienM djschaap: can you confirm that when you put a dummy value in the firewall_driver param, it works ?
15:31 djschaap EmilienM: just tried. yes, a dummy value in ml2.ini is sufficient to get the security-group extension loaded.
15:34 EmilienM djschaap: far enough, i'll produce a patch i think to fix it asap.
15:34 EmilienM and backport
15:34 EmilienM djschaap: can I keep in you in the loop to help me with testing ?
15:36 djschaap EmilienM: please do! My initial change may have the right code/tests - but the comments are definitely not correct now.
15:38 EmilienM djschaap: cool thank you, I appreciate
15:42 rongze joined #puppet-openstack
15:44 openstackgerrit François Charlier proposed a change to stackforge/puppet-horizon: Introduce keystone_url parameter  https://review.openstack.org/61177
15:48 openstackgerrit François Charlier proposed a change to stackforge/puppet-horizon: Introduce keystone_url parameter  https://review.openstack.org/61177
15:48 openstackgerrit Gonéri Le Bouder proposed a change to stackforge/puppet-horizon: horizon replaced nodejs by a python alternative  https://review.openstack.org/63154
16:12 cbaesema joined #puppet-openstack
16:16 rongze joined #puppet-openstack
16:21 xarses joined #puppet-openstack
16:22 rongze joined #puppet-openstack
16:33 EmilienM djschaap: if it's good for you, I take the bug https://bugs.launchpad.net/puppet-neutron/+bug/1262678
16:33 EmilienM actually I already took it, but if you don't mind, I work on a patch
16:43 djschaap EmilienM: you're welcome to it
16:55 tnoor joined #puppet-openstack
16:55 mattymo joined #puppet-openstack
17:00 badiane_ka joined #puppet-openstack
17:02 angdraug joined #puppet-openstack
17:04 tnoor joined #puppet-openstack
17:04 mattymo joined #puppet-openstack
17:15 tnoor joined #puppet-openstack
17:15 mattymo joined #puppet-openstack
17:30 marun joined #puppet-openstack
17:32 openstackgerrit Sabari Murugesan proposed a change to stackforge/puppet-nova: Add support for VMware driver  https://review.openstack.org/63181
17:40 rongze joined #puppet-openstack
17:44 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-ceilometer: Updated Modulefile for release candidate 2  https://review.openstack.org/63036
17:46 hogepodge joined #puppet-openstack
17:49 slowe joined #puppet-openstack
18:01 brdude joined #puppet-openstack
18:02 hogepodge dprince Hey, what's up with Smokestack? I haven't seen jobs pass through it in quite a while.
18:02 hogepodge More curious than anything.
18:04 dprince hogepodge: hi, So... the short version is I've been focussed on other things and just haven't had time to keep up. It went down a week or so after Hong Kong.
18:05 dprince hogepodge: I probably should have sent out a status note or something. It isn't like I don't care, or I changed my mind, etc. I've just been focussed on other things.
18:05 hogepodge Ah. I see. If there's anything I can do to help, let me know. We have new hardware at PL, and I could look into devoting some of it to running Smokestack here.
18:06 hogepodge Oh, I figured you were slammed with work. I'm embarrassed I still don't have a stable Havana release out. Not for want of doing it, it just takes time and I'm being pulled in a million directions here. So I totally understand.
18:08 dprince hogepodge: physical (cloud) resource aren't my problem. I've got enough of those for the moment. It is the time to constantly keep packages, etc in sync.
18:08 slowe Quick Q: in version 3.0.0-rc1 of puppet-nova/puppet-cinder, is the expectation that they will both use the same RabbitMQ configuration?
18:09 dprince hogepodge: I expect to spend more time on packaging sort of things in the new year so perhaps then I'll be able to revisit SmokeStack.
18:10 dtalton joined #puppet-openstack
18:12 openstackgerrit Sabari Murugesan proposed a change to stackforge/puppet-nova: Add support for VMware driver  https://review.openstack.org/63181
18:23 xarses joined #puppet-openstack
18:40 rongze joined #puppet-openstack
18:42 hogepodge slowe: Only one should set up RabbitMQ
18:42 hogepodge Either can, and the rest of the modules just need to know the connection informaiton.
18:42 hogepodge s/informaiton/information
18:43 hogepodge I'm not sure why both can configure, but that's how it is right now.
18:43 slowe hogepodge: So I should use *either* nova::rabbitmq or cinder::rabbitmq, but not both.
18:43 hogepodge Correct
18:43 hogepodge Trying to use both would lead to a catalog compilation error.
18:44 slowe hogepodge: It looks like one possible reason why you can use both is that they both attempt to declare RabbitMQ::Server. I'm also trying to determine if OpenStack requires that Cinder and Nova share the same message queue configuration.
18:44 slowe hogepodge: Sorry...one reason you can't use both...
18:45 hogepodge Yeah, that's why. They'll both use the same configuration. All of the projects use different channels in the same message queue
18:46 slowe hogepodge: Excellent, thanks for the clarification. I have it working by just using nova::rabbitmq, but wanted to double-check.
18:46 hogepodge I do the same. I think that cinder also has it to support storage-only environments.
18:47 hogepodge It's something that should probably be moved out (along with a bunch of other things) into a standard OpenStack library.
18:48 slowe hogepodge: Agreed, makes sense.
18:52 cbaesema joined #puppet-openstack
19:21 xarses dachary: any thoughts on https://review.openstack.org/#/c/63008/
20:32 otherwiseguy joined #puppet-openstack
20:33 rmoe joined #puppet-openstack
21:04 michchap_ joined #puppet-openstack
21:16 dachary xarses: thanks for pinging me ;-)
21:17 openstackgerrit A change was merged to stackforge/puppet-ceph: spec:system - Change VM IP to fix virtualbox/vagrant error  https://review.openstack.org/63008
21:22 EmilienM mgagne: good news, I figured out how to fix the problem with ML2 / security group. Bad news, when you will see the patch you will think "WTF"
21:22 mgagne EmilienM: I read the email
21:22 mgagne EmilienM: this could be simplified to: firewall_driver=True
21:23 EmilienM mgagne: yep
21:23 EmilienM mgagne: and if true, add a dummy driver
21:23 mgagne EmilienM: what?
21:23 EmilienM mgagne: it's temporary as far i know, it will change (again) in icehouse
21:24 EmilienM mgagne: in mL2 conf, we should configure firewall_driver=something to enable the security group
21:24 mgagne EmilienM: that's what I said about
21:24 EmilienM mgagne: and you want to add True, why n ot
21:25 mgagne EmilienM: pourquoi inventer un nom débile alors qu'on a déjà des valeurs plus jolies à utiliser
21:25 EmilienM \o/
21:26 mgagne EmilienM: il faut creuser dans le code, le check est seulement: Si PAS neutron...None, active le firewall
21:26 EmilienM mgagne: I do agree True is more simple to use, thank you for your feedback
21:55 cbaesema feedback, thoughts and or reviews on these would be appreciated
21:55 cbaesema https://review.openstack.org/#/c/62746/
21:55 cbaesema https://review.openstack.org/#/c/62744/
21:56 cbaesema the purpose is to allow storage and swift proxy to coexist on the same node
22:02 hogepodge LGTM.
22:10 cbaesema @hogepodge: thank you for reviewing
22:25 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-neutron: Configure security group when using ML2 plugin  https://review.openstack.org/63228

| Channels | #puppet-openstack index | Today | | Search | Google Search | Plain-Text | summary