| *** nati_ueno has quit IRC | 00:04 | |
| *** nati_ueno has joined #openstack-neutron | 00:04 | |
| *** nati_ueno has quit IRC | 00:09 | |
| *** julim has quit IRC | 00:11 | |
| openstackgerrit | A change was merged to openstack/neutron: Remove deprecated fields in keystone auth middleware https://review.openstack.org/53985 | 00:12 |
|---|---|---|
| *** layer427expert has joined #openstack-neutron | 00:18 | |
| *** jsergent has quit IRC | 00:19 | |
| *** bvandenh has quit IRC | 00:33 | |
| *** otherwiseguy has quit IRC | 00:33 | |
| *** suresh12 has quit IRC | 00:33 | |
| *** suresh12 has joined #openstack-neutron | 00:34 | |
| *** networkstatic has joined #openstack-neutron | 00:35 | |
| *** suresh12 has quit IRC | 00:35 | |
| *** suresh12 has joined #openstack-neutron | 00:35 | |
| *** wes_ has quit IRC | 00:36 | |
| *** networkstatic is now known as networks_ | 00:37 | |
| *** networkstatic has joined #openstack-neutron | 00:38 | |
| *** matsuhashi has quit IRC | 00:47 | |
| *** matsuhashi has joined #openstack-neutron | 00:48 | |
| *** matsuhashi has quit IRC | 00:53 | |
| *** matsuhashi has joined #openstack-neutron | 00:58 | |
| *** beagles has quit IRC | 01:12 | |
| *** nati_ueno has joined #openstack-neutron | 01:12 | |
| *** LiJiansheng has joined #openstack-neutron | 01:12 | |
| *** dkliban has quit IRC | 01:14 | |
| *** suresh12 has quit IRC | 01:16 | |
| *** suresh12 has joined #openstack-neutron | 01:17 | |
| *** suresh12 has quit IRC | 01:19 | |
| *** suresh12 has joined #openstack-neutron | 01:20 | |
| *** gatuus has quit IRC | 01:26 | |
| *** carl_baldwin has joined #openstack-neutron | 01:31 | |
| *** ljjjustin has joined #openstack-neutron | 01:31 | |
| *** layer427expert has quit IRC | 01:34 | |
| *** dzyu has joined #openstack-neutron | 01:41 | |
| *** berlin has joined #openstack-neutron | 01:42 | |
| *** nati_ueno has quit IRC | 01:47 | |
| *** nati_ueno has joined #openstack-neutron | 01:47 | |
| *** berlin has quit IRC | 01:50 | |
| *** matsuhashi has quit IRC | 02:03 | |
| *** matsuhashi has joined #openstack-neutron | 02:03 | |
| *** layer427expert has joined #openstack-neutron | 02:05 | |
| *** layer427expert has quit IRC | 02:06 | |
| *** carl_baldwin has quit IRC | 02:06 | |
| *** dkliban has joined #openstack-neutron | 02:13 | |
| *** dkliban has quit IRC | 02:18 | |
| openstackgerrit | Armando Migliaccio proposed a change to openstack/neutron: Lower severity of log trace for DB integrity error https://review.openstack.org/56496 | 02:24 |
| *** dkliban has joined #openstack-neutron | 02:30 | |
| *** nati_uen_ has joined #openstack-neutron | 02:35 | |
| *** nati_ueno has quit IRC | 02:38 | |
| *** layer427expert has joined #openstack-neutron | 02:52 | |
| *** layer427expert has quit IRC | 03:06 | |
| *** dkehn has quit IRC | 03:11 | |
| *** dkehn has joined #openstack-neutron | 03:13 | |
| *** layer427expert has joined #openstack-neutron | 03:13 | |
| openstackgerrit | Armando Migliaccio proposed a change to openstack/neutron: Add log statements for policy check failures https://review.openstack.org/54457 | 03:21 |
| *** dims has quit IRC | 03:22 | |
| *** armax has quit IRC | 03:23 | |
| *** matsuhashi has quit IRC | 03:31 | |
| *** matsuhashi has joined #openstack-neutron | 03:31 | |
| *** matsuhashi has quit IRC | 03:41 | |
| *** matsuhashi has joined #openstack-neutron | 03:42 | |
| *** enikanorov_ has joined #openstack-neutron | 03:43 | |
| *** enikanorov___ has joined #openstack-neutron | 03:45 | |
| *** matsuhashi has quit IRC | 03:47 | |
| *** rdo_ has joined #openstack-neutron | 03:47 | |
| *** silvrax_ has joined #openstack-neutron | 03:52 | |
| *** LiJiansheng has quit IRC | 03:52 | |
| *** silvrax has quit IRC | 03:52 | |
| *** enikanorov__ has quit IRC | 03:52 | |
| *** enikanorov has quit IRC | 03:52 | |
| *** matrohon_ has quit IRC | 03:52 | |
| *** Alssi has quit IRC | 03:52 | |
| *** rdo has quit IRC | 03:52 | |
| *** fandikurnia01 has joined #openstack-neutron | 03:58 | |
| *** layer427expert has quit IRC | 04:00 | |
| *** matrohon_ has joined #openstack-neutron | 04:00 | |
| *** LiJiansheng has joined #openstack-neutron | 04:03 | |
| *** nati_uen_ has quit IRC | 04:04 | |
| *** nati_ueno has joined #openstack-neutron | 04:05 | |
| *** nati_ueno has quit IRC | 04:09 | |
| *** suresh12 has quit IRC | 04:15 | |
| *** suresh12 has joined #openstack-neutron | 04:15 | |
| *** jianingy is now known as jianingy_afk | 04:19 | |
| *** suresh12 has quit IRC | 04:20 | |
| *** layer427expert has joined #openstack-neutron | 04:21 | |
| *** yamahata__ has quit IRC | 04:25 | |
| *** yuikotakada has joined #openstack-neutron | 04:26 | |
| *** yamahata has joined #openstack-neutron | 04:27 | |
| *** nati_ueno has joined #openstack-neutron | 04:33 | |
| *** yamahata has quit IRC | 04:35 | |
| *** armax has joined #openstack-neutron | 04:36 | |
| *** armax has left #openstack-neutron | 04:37 | |
| *** chandankumar has joined #openstack-neutron | 04:39 | |
| *** nati_ueno has quit IRC | 04:43 | |
| openstackgerrit | Zhang Hua proposed a change to openstack/neutron: Raise max header size to accommodate large tokens https://review.openstack.org/56511 | 04:43 |
| *** coolsvap has joined #openstack-neutron | 04:46 | |
| *** nati_ueno has joined #openstack-neutron | 04:49 | |
| *** alexpilotti has joined #openstack-neutron | 04:55 | |
| *** matsuhashi has joined #openstack-neutron | 04:56 | |
| *** layer427expert has quit IRC | 04:56 | |
| *** chandankumar has quit IRC | 04:57 | |
| *** layer427expert has joined #openstack-neutron | 04:57 | |
| *** layer427expert has quit IRC | 04:58 | |
| *** LiJiansheng has quit IRC | 04:58 | |
| *** nati_ueno has quit IRC | 05:02 | |
| openstackgerrit | Dazhao Yu proposed a change to openstack/neutron: Calculate stateless IPv6 address https://review.openstack.org/56184 | 05:06 |
| *** nati_ueno has joined #openstack-neutron | 05:09 | |
| *** chandankumar has joined #openstack-neutron | 05:09 | |
| *** LiJiansheng has joined #openstack-neutron | 05:10 | |
| *** suresh12 has joined #openstack-neutron | 05:26 | |
| *** suresh12 has quit IRC | 05:30 | |
| *** marun has joined #openstack-neutron | 05:50 | |
| *** nati_ueno has quit IRC | 06:03 | |
| *** x86brandon has quit IRC | 06:09 | |
| *** layer427expert has joined #openstack-neutron | 06:15 | |
| *** coolsvap has quit IRC | 06:26 | |
| *** dkliban has quit IRC | 06:30 | |
| *** nati_ueno has joined #openstack-neutron | 06:35 | |
| *** LiJiansheng has quit IRC | 06:39 | |
| *** amritanshu_RnD has joined #openstack-neutron | 06:39 | |
| openstackgerrit | Jenkins proposed a change to openstack/neutron: Imported Translations from Transifex https://review.openstack.org/54680 | 06:40 |
| marun | nati_ueno: ping | 06:41 |
| nati_ueno | sorry in mtg | 06:41 |
| *** coolsvap has joined #openstack-neutron | 06:45 | |
| *** openstackgerrit has quit IRC | 06:46 | |
| *** openstackgerrit has joined #openstack-neutron | 06:46 | |
| *** suresh12 has joined #openstack-neutron | 06:48 | |
| *** mjbright_ has quit IRC | 06:56 | |
| openstackgerrit | Ann Kamyshnikova proposed a change to openstack/neutron: Fix downgrade in migration https://review.openstack.org/53064 | 06:56 |
| openstackgerrit | Ann Kamyshnikova proposed a change to openstack/neutron: Fix downgrade in migration https://review.openstack.org/53064 | 06:58 |
| *** rwsu is now known as rwsu-pto | 07:04 | |
| *** jprovazn has joined #openstack-neutron | 07:08 | |
| marun | nati_ueno: what tz are you in? | 07:17 |
| *** gdubreui has quit IRC | 07:18 | |
| nati_ueno | marun: JST (japan) | 07:19 |
| marun | nati_ueno: ah, that explains it. You're usually in PST right? | 07:19 |
| marun | nati_ueno: How long are you going to be in JST? | 07:19 |
| nati_ueno | marun: yeah, i'll back to PST 11/23 | 07:20 |
| marun | nati_ueno: cool. I'm in CST until 11/30. | 07:20 |
| marun | nati_ueno: when you have a chance, would like to chat about the request id stuff. i'm not clear on some of the specifics. | 07:22 |
| nati_ueno | marun: sure. could you mail me your question? | 07:22 |
| marun | er, ok. | 07:23 |
| marun | i was hoping for a more interactive query, but whatever works for you. | 07:23 |
| nati_ueno | marun: yeah, I'll also ping you | 07:24 |
| nati_ueno | on IRC | 07:24 |
| marun | nati_ueno: ok | 07:25 |
| marun | btw, i have 3 emails for you. is there a particular one you prefer? | 07:25 |
| nati_ueno | nachi@ntti3.com is for business :) | 07:28 |
| nati_ueno | nati.ueno@gmail.com is for private | 07:28 |
| *** jroovers has quit IRC | 07:30 | |
| *** networkstatic has quit IRC | 07:35 | |
| *** suresh12 has quit IRC | 07:37 | |
| *** suresh12 has joined #openstack-neutron | 07:37 | |
| openstackgerrit | Oleg Bondarev proposed a change to openstack/neutron: ML2 plugin should not delete ports on subnet deletion https://review.openstack.org/54918 | 07:37 |
| nati_ueno | marun: https://github.com/openstack/neutron/blob/master/neutron/openstack/common/rpc/amqp.py#L535 | 07:37 |
| nati_ueno | marun: https://github.com/openstack/neutron/blob/master/neutron/openstack/common/rpc/amqp.py#L297 | 07:38 |
| marun | nati_ueno: I'm afraid my question wasn't regarding amqp semantics. | 07:38 |
| nati_ueno | marun: I assume you are interested in this code | 07:38 |
| marun | nati_ueno: I wasn't clear on whether all agent actions would be triggered by amqp or not. | 07:38 |
| nati_ueno | marun: some action isn't triggered by rpc | 07:39 |
| marun | nati_ueno: For example. doesn't the ovs agent react to ovs interface changes (in the daemon loop) and configure a port if it is to be managed by neutron? | 07:39 |
| nati_ueno | marun: yeah, it could be possible. | 07:40 |
| marun | nati_ueno: So my question is, is traceability on such actions possible? | 07:40 |
| nati_ueno | marun: I don't think so. | 07:41 |
| marun | Ideally all agent actions could be traceable. Currently if a failure in vm launch occurs we have to manually correlate. | 07:41 |
| marun | The request id in the log as proposed solves part of the problem, but I'm hoping we can do better. | 07:41 |
| nati_ueno | marun: We can do it by using context | 07:41 |
| marun | context from where though? | 07:41 |
| nati_ueno | marun: eaach context may have context id, so we can trace it | 07:41 |
| nati_ueno | marun: context object | 07:41 |
| nati_ueno | marun: yeah, I got your point now | 07:42 |
| marun | if communication is through db, we may need to associate the context data with db state | 07:42 |
| *** suresh12 has quit IRC | 07:42 | |
| nati_ueno | marun: context is not only for db | 07:42 |
| marun | nati_ueno: sorry to be asking difficult questions, i know it's a muddy area | 07:42 |
| nati_ueno | marun: ha ha, your question is tend to difficult | 07:42 |
| marun | nati_ueno: 'may you live in interesting times (or be asked difficult questions)' ;) | 07:43 |
| nati_ueno | may be :) | 07:43 |
| marun | nati_ueno: I don't see why the patch shouldn't go in, but I'd like to see either that bug extended to investigate the possibility of maximizing traceability for all agent actions or maybe create a blueprint to see it done. | 07:44 |
| marun | nati_ueno: i.e. not let it fall by the wayside | 07:44 |
| nati_ueno | marun: We should limit the scope of the bug report | 07:44 |
| nati_ueno | marun: then let's write bp for this | 07:45 |
| nati_ueno | for more wider discussion | 07:45 |
| marun | nati_ueno: ok, fair enough. I'll ask Akihiro if he wants to be responsible or pass it to someone else. | 07:46 |
| nati_ueno | marun: gotcha | 07:46 |
| *** cyeoh has quit IRC | 07:50 | |
| *** cyeoh has joined #openstack-neutron | 07:51 | |
| *** nati_ueno has quit IRC | 07:52 | |
| *** layer427expert has quit IRC | 07:56 | |
| *** matsuhashi has quit IRC | 08:01 | |
| *** matsuhashi has joined #openstack-neutron | 08:02 | |
| *** fandikurnia01 has quit IRC | 08:02 | |
| *** fandikurnia01 has joined #openstack-neutron | 08:03 | |
| *** networkstatic has joined #openstack-neutron | 08:09 | |
| *** suresh12 has joined #openstack-neutron | 08:11 | |
| *** suresh12 has quit IRC | 08:12 | |
| *** suresh12 has joined #openstack-neutron | 08:13 | |
| *** suresh12 has quit IRC | 08:17 | |
| *** changbl has quit IRC | 08:24 | |
| *** salv-orlando has joined #openstack-neutron | 08:26 | |
| *** jlibosva has joined #openstack-neutron | 08:37 | |
| *** rossella_s has joined #openstack-neutron | 08:41 | |
| *** Alssi has joined #openstack-neutron | 08:44 | |
| *** changbl has joined #openstack-neutron | 08:57 | |
| *** jpich has joined #openstack-neutron | 08:59 | |
| *** LiJiansheng has joined #openstack-neutron | 09:05 | |
| *** changbl has quit IRC | 09:05 | |
| *** changbl has joined #openstack-neutron | 09:15 | |
| *** ygbo has joined #openstack-neutron | 09:15 | |
| *** bvandenh has joined #openstack-neutron | 09:15 | |
| *** xianghui has joined #openstack-neutron | 09:20 | |
| *** networkstatic has quit IRC | 09:22 | |
| *** matsuhashi has quit IRC | 09:25 | |
| *** matsuhashi has joined #openstack-neutron | 09:25 | |
| *** safchain has joined #openstack-neutron | 09:26 | |
| *** xianghui has quit IRC | 09:26 | |
| *** nati_ueno has joined #openstack-neutron | 09:27 | |
| *** changbl has quit IRC | 09:28 | |
| *** matsuhashi has quit IRC | 09:30 | |
| openstackgerrit | Ann Kamyshnikova proposed a change to openstack/neutron: Update lockutils and fixture in openstack.common https://review.openstack.org/47557 | 09:30 |
| *** dzyu has quit IRC | 09:34 | |
| *** nati_ueno has quit IRC | 09:42 | |
| openstackgerrit | Ann Kamyshnikova proposed a change to openstack/neutron: Add testing of migrations from oslo https://review.openstack.org/46927 | 09:43 |
| openstackgerrit | Ann Kamyshnikova proposed a change to openstack/neutron: Implement testing of migrations https://review.openstack.org/48594 | 09:43 |
| *** ljjjustin has quit IRC | 09:44 | |
| *** LiJiansheng has quit IRC | 09:46 | |
| *** amritanshu_RnD has quit IRC | 09:50 | |
| *** LiJiansheng has joined #openstack-neutron | 10:04 | |
| *** ppenjoy has joined #openstack-neutron | 10:08 | |
| *** ppenjoy has left #openstack-neutron | 10:08 | |
| *** markvoelker has quit IRC | 10:12 | |
| openstackgerrit | Dazhao Yu proposed a change to openstack/neutron: Calculate stateless IPv6 address https://review.openstack.org/56184 | 10:21 |
| *** yuikotakada has quit IRC | 10:24 | |
| openstackgerrit | A change was merged to openstack/neutron: Cleanup HACKING.rst https://review.openstack.org/55885 | 10:24 |
| *** markvoelker has joined #openstack-neutron | 10:25 | |
| anteaya | nati_ueno marun thank you for conducting your discussion in channel, it is really helpful to supporting my goal of getting more Neutron devs talking to each other, thank you | 10:46 |
| anteaya | newcomers can see the conversation and learn from it | 10:47 |
| anteaya | when I am asked by other projects what is happening in Neutron I am better informed | 10:47 |
| anteaya | and the channel is logged so people can read the back logs if they missed the original conversation | 10:47 |
| anteaya | http://eavesdrop.openstack.org/irclogs/ | 10:47 |
| anteaya | all very helpful for addressing the current Neutron status | 10:48 |
| anteaya | thank you both | 10:48 |
| anteaya | if you haven't read the Neutron email yet: http://lists.openstack.org/pipermail/openstack-dev/2013-November/019251.html | 10:49 |
| anteaya | please read this | 10:49 |
| anteaya | and thank you | 10:49 |
| anteaya | let me know what you need to find out more about them or understand them better | 10:51 |
| marun | anteaya: likely that all of the bugs are race conditions that markmcclain and salv-orlando are working on already. they could confirm/deny. | 10:57 |
| *** marun has quit IRC | 10:59 | |
| salv-orlando | 1249065 and 1251448 appear to be issues in nova/neutron integration | 11:00 |
| salv-orlando | we are working on that too, but that's a separate issue from the race bw neutron db ops and RPC calls from agents | 11:00 |
| *** marun has joined #openstack-neutron | 11:06 | |
| anteaya | marun: are salv-orlando and markmcclain the only neutron devs able to address bugs | 11:08 |
| anteaya | there were soooo many people in all the neutron design summit sessions | 11:08 |
| anteaya | if salv-orlando and markmcclain are the only ones doing the work, we could have held the summit sessions in a cafe and saved on space | 11:09 |
| anteaya | salv-orlando: let me know how I can help | 11:09 |
| anteaya | salv-orlando: do you have any folks from nova to talk to from their side? | 11:09 |
| anteaya | if not, I can round up some | 11:09 |
| salv-orlando | everyone on the core team can address this bugs + at least 5/6 non-core devs which have great experience | 11:09 |
| marun | anteaya: Bugs? No. The race conditions? Yes. I talked with Mark at the summit about how to best to assist in fixing the race conditions, and he made it clear that they had it in hand and that brining new people on board wouldn't be time efficient. | 11:09 |
| salv-orlando | nati_ueno and marun have fixed tons of bugs too | 11:10 |
| anteaya | marun: let me talk to markmcclain about this | 11:10 |
| anteaya | marun: if you are willing to help, I want you to be empowered to help | 11:10 |
| anteaya | the race conditions seem to be growing faster than they can be contained | 11:10 |
| anteaya | and thank you | 11:10 |
| marun | anteaya: Uh, I have a good relationship with other core members. | 11:10 |
| anteaya | great | 11:10 |
| anteaya | so 4 | 11:10 |
| anteaya | good to hear that | 11:11 |
| marun | anteaya: If I ask Mark directly whether he can help fix, and he says 'no', I trust him at his word. | 11:11 |
| anteaya | that is fine | 11:11 |
| anteaya | I need to talk to him myself | 11:11 |
| marun | Have we met? | 11:11 |
| anteaya | my goal it to be able to get neutron development to scale | 11:11 |
| anteaya | I think talking to markmcclain about this will be a part of that | 11:11 |
| anteaya | I don't know | 11:12 |
| anteaya | my name is Anita Kuno | 11:12 |
| anteaya | I am from infra | 11:12 |
| *** fandikurnia01 has quit IRC | 11:12 | |
| anteaya | given the current state of the gate, after the tuesday summit session on neutron tempest I moved in here | 11:12 |
| anteaya | since there was a strong lack of willing people coming forward in that session | 11:12 |
| *** pcm_ has joined #openstack-neutron | 11:13 | |
| anteaya | which is very bad for the rest of openstack | 11:13 |
| *** pcm_ has quit IRC | 11:14 | |
| marun | strong lack? | 11:14 |
| *** pcm_ has joined #openstack-neutron | 11:14 | |
| marun | I seem to remember a whole bunch of people approaching you when you announced the tempest sprint on Friday. | 11:15 |
| anteaya | yes | 11:19 |
| anteaya | the Friday announcement was a direct outcome of my disappointment from the Tuesday meeting | 11:20 |
| anteaya | had the Tuesday meeting gone well | 11:20 |
| anteaya | I would be creating puppet manifest in infra, working with salt, and addressing storyboard as I had been planning to do prior to that meeting | 11:20 |
| anteaya | and the announcement for the code sprint did go well | 11:21 |
| *** LiJiansheng has quit IRC | 11:26 | |
| openstackgerrit | Salvatore Orlando proposed a change to openstack/neutron: NVP plugin:fix delete sec group when backend is out of sync https://review.openstack.org/56591 | 11:27 |
| anteaya | salv-orlando: okay, I don't see you in #openstack-nova | 11:39 |
| salv-orlando | I can log in there too, I usually ping nova-devs in openstack-dev, but one more channel won't hurt | 11:40 |
| anteaya | I will convey your thoughts that 1249065 and 1251448 seem to be integration related and see if some nova people are able to help | 11:40 |
| salv-orlando | Perhaps the nova devs would like a more thorough analysis of the root cause beyond just "integration related". I can triage them in my afternoon/evening and provide more info | 11:41 |
| anteaya | okay, I will look for conversations in -dev as well | 11:42 |
| anteaya | if that works, no need to change, I just would like to know where you are so that I can follow and support | 11:42 |
| anteaya | salv-orlando: that would be awesome thank you | 11:42 |
| anteaya | salv-orlando: and I think that both you and markmcclain need to delegate more | 11:44 |
| anteaya | there are people willing to learn how to help | 11:44 |
| anteaya | but they need some guidance and direction | 11:44 |
| anteaya | and I don't have the knowledge to convey that | 11:44 |
| anteaya | you and markmcclain don't scale | 11:45 |
| anteaya | and this might be one of the bottlenecks that is interfering with the project to meet current demands | 11:45 |
| salv-orlando | Personally I would not entirely agree on the delegation issue. Disclaimer: I am not looking for argument or fights. | 11:46 |
| anteaya | how to get the work done then? | 11:46 |
| anteaya | the amount of people who came forward on friday indicate there is interest to help | 11:47 |
| salv-orlando | I have only a partial perspective here, but rather than a problem of development which does not scale, I see a need for resifting priorities. | 11:47 |
| anteaya | if the ones who hold the vision don't give away responsibility to do tasks and oversee the work | 11:47 |
| anteaya | then the ability to grow is finite | 11:47 |
| * anteaya listens | 11:48 | |
| salv-orlando | Tasks have always been given away and distributed across the various developers. If you look at the number on this statistics website, stackalatycs. | 11:48 |
| salv-orlando | sorry I interrupted the sentence in the middle | 11:49 |
| * anteaya continues to listen | 11:49 | |
| salv-orlando | You will notice that there are a large number of recurring contributors, where recurring means merging at least two patches per milestone. Some of these contributors are "peripheral", | 11:49 |
| * anteaya hopes there is more | 11:51 | |
| salv-orlando | meaning they only contribute code to a specific plugin or driver. However about 3/4 of the developers also perform "core" contributions, which is either the API layer, the authZ framework, one of the FOSSs plugins, one of the common services, or one of the components of the frameworks of agents which makes part of the default neutrons service offering. | 11:51 |
| salv-orlando | The problem in my opinion, is that the focus of the development team has drifted towards non-core feature (and I have not said 'vendors' be careful) - such as L4/L7 services. | 11:52 |
| * anteaya admires salv-orlando's choice in words | 11:52 | |
| salv-orlando | Icehouse needs to be about stabilisation of the L2/L3 framework and providing those use cases which are not yet met wrt nova-network | 11:53 |
| salv-orlando | So - as far I see it, people will have to rebalance their contribution from the non-core items to the core one. | 11:53 |
| anteaya | how do we refocus on L2/L3? | 11:54 |
| salv-orlando | People like marun are completely inline with this, he's been working in the past 3 months on optimisation of the L2 agent | 11:54 |
| anteaya | awesome | 11:54 |
| salv-orlando | markmcclain, which should be online in a short while, is preparing the roadmap draft, I think. | 11:54 |
| anteaya | fantastic | 11:55 |
| anteaya | I look forward to hearing his thoughts | 11:55 |
| anteaya | and learning how I can support the people who come forward in the focus on L2/L3 | 11:55 |
| salv-orlando | it would be important to understand the ins and outs the three agents: l2, dhcp, and l3 | 11:56 |
| anteaya | great | 11:56 |
| salv-orlando | and how they interoperate with the API server | 11:56 |
| anteaya | can you point me to some documentation? | 11:57 |
| anteaya | something that would be good for someone with no networking experience? | 11:57 |
| anteaya | sounds like a good place for me to start | 11:57 |
| anteaya | thank you | 11:57 |
| salv-orlando | I need to find the links I had. I'm not sure we have official technical documentation regarding the architecture, but I am sure there are plenty of blog posts from other members of the community. | 11:58 |
| *** coolsvap has quit IRC | 11:58 | |
| anteaya | anything that would give me a place to start | 11:59 |
| anteaya | that is not so complex that I need to study it for 3 days to track down all the acroymns | 11:59 |
| anteaya | it doesn't have to be the official documentation | 12:02 |
| anteaya | I actually find the official documentation too dry for me to follow a lot of the time | 12:02 |
| anteaya | I prefer the unique personalities of blog posts | 12:02 |
| openstackgerrit | Ofer Barkai proposed a change to openstack/neutron: Enhance floating IP router lookup https://review.openstack.org/55987 | 12:05 |
| *** jorisroovers has joined #openstack-neutron | 12:06 | |
| anteaya | hey jorisroovers | 12:10 |
| jorisroovers | hi anteaya | 12:10 |
| jorisroovers | :-) | 12:10 |
| anteaya | how are you doing with tox? | 12:10 |
| jorisroovers | I was hoping you would be around :-) | 12:10 |
| anteaya | yes | 12:10 |
| jorisroovers | I didn't get to it yesterday | 12:10 |
| anteaya | I had to take a quiet day yesterday | 12:10 |
| anteaya | no problem | 12:10 |
| jorisroovers | that makes 2 of us :-) | 12:10 |
| anteaya | how does today look for you? | 12:10 |
| anteaya | :D | 12:10 |
| jorisroovers | so I'm currently in the process of getting them to run again | 12:11 |
| jorisroovers | just installing the dependencies | 12:11 |
| jorisroovers | I did a devstack reinstall, as you suggested | 12:11 |
| anteaya | go you! | 12:11 |
| anteaya | great | 12:11 |
| anteaya | yeah, don't get attached to your devstacks | 12:11 |
| jorisroovers | I did run into the same issue as before during the devstack install | 12:11 |
| anteaya | you will have to bring them downn and create a new one often | 12:11 |
| anteaya | can you paste the output please? | 12:12 |
| jorisroovers | sorry, but don't have it | 12:12 |
| jorisroovers | I have the related bug though | 12:12 |
| jorisroovers | sec | 12:12 |
| jorisroovers | s -l | 12:12 |
| anteaya | k | 12:12 |
| jorisroovers | https://bugs.launchpad.net/devstack/+bug/1248923 | 12:14 |
| jorisroovers | so I fixed it quick and dirty by just chowning /etc/nova | 12:15 |
| anteaya | rarrrrrrr | 12:17 |
| anteaya | okay | 12:17 |
| jorisroovers | ok, just ran tox again | 12:17 |
| jorisroovers | http://paste.openstack.org/show/53093/ | 12:17 |
| jorisroovers | new issues | 12:17 |
| anteaya | looks like I am firing up a devstack and digging into this | 12:17 |
| anteaya | I don't expect dtroyer to be awake for a few hours | 12:17 |
| anteaya | let me get a walk in and some breakfast and then I will look into this | 12:17 |
| jorisroovers | he's 'the devstack guy' | 12:17 |
| anteaya | is that fair? | 12:17 |
| anteaya | he is the devstack guy | 12:18 |
| jorisroovers | :-) | 12:18 |
| anteaya | introduce yourself | 12:18 |
| jorisroovers | ok, will do | 12:18 |
| anteaya | this could be your first patch or review to openstack | 12:18 |
| jorisroovers | enjoy your breakfast and walk | 12:18 |
| jorisroovers | :-) | 12:18 |
| anteaya | thanks | 12:19 |
| anteaya | jorisroovers: this looks to be a result of some Ubuntu packaging | 12:23 |
| anteaya | jorisroovers: can you fire up a vm with a fedora image and do a straight devstack install? (the dependencies may be different) | 12:23 |
| jorisroovers | ok | 12:23 |
| jorisroovers | on it :-) | 12:23 |
| anteaya | don't do any sudos other than to install dependencies and save all outputs and logs | 12:23 |
| anteaya | if this is ubuntu packaging we need to include chuck in this for the solution | 12:24 |
| anteaya | it will take you awhile, and thanks | 12:24 |
| jorisroovers | no worries, thanks for all the patience. I feel we are not a whole lot further than on Monday, but I know that this is just how it works :-) | 12:26 |
| *** alexpilotti has quit IRC | 12:49 | |
| *** dims has joined #openstack-neutron | 13:18 | |
| *** julim has joined #openstack-neutron | 13:46 | |
| openstackgerrit | A change was merged to openstack/neutron: Fix import log_handler error with publish_errors set https://review.openstack.org/55346 | 13:52 |
| *** HenryG has joined #openstack-neutron | 13:52 | |
| *** openstackgerrit has quit IRC | 13:53 | |
| *** openstackgerrit has joined #openstack-neutron | 13:53 | |
| *** changbl has joined #openstack-neutron | 13:55 | |
| anteaya | jorisroovers: you are absolutely correct | 13:57 |
| anteaya | this is why at least having someone to talk to make such a huge difference when getting started with something new | 13:58 |
| jorisroovers | :-) | 13:58 |
| openstackgerrit | Yves-Gwenael Bourhis proposed a change to openstack/neutron: Make dnsmaq aware of all names https://review.openstack.org/52930 | 13:58 |
| jorisroovers | just started the devstack install here | 13:58 |
| anteaya | go you! | 13:58 |
| jorisroovers | had some issues getting Vbox Guest Additions to work on Fedora | 13:58 |
| anteaya | writing letters of invitation to canadian embassies here | 13:58 |
| anteaya | my first time having to do this | 13:58 |
| *** rkukura has joined #openstack-neutron | 13:59 | |
| anteaya | any thought of using virtualbox and vagrant? | 13:59 |
| anteaya | hi rkukura | 13:59 |
| jorisroovers | is that to arange visas? | 13:59 |
| anteaya | are you bob that I met the last day of the summit? | 13:59 |
| anteaya | yeah | 13:59 |
| rkukura | anteaya: good morning | 13:59 |
| anteaya | morning | 13:59 |
| rkukura | anteaya: yes | 13:59 |
| anteaya | so far only 1 I have to write for sure, and one I am waiting to hear back from | 13:59 |
| anteaya | great | 13:59 |
| jorisroovers | mm, just saw that devstack is only tested on fedora 18, I'm running 19 now. Fingers crossed | 14:00 |
| anteaya | glad to put a nick to a face | 14:00 |
| anteaya | how was your journey home? | 14:00 |
| anteaya | jorisroovers: you might shake out some new bugs :D | 14:01 |
| rkukura | anteaya: uneventful - slept 8 hours on the plane - how about yours? | 14:01 |
| openstackgerrit | Yves-Gwenael Bourhis proposed a change to openstack/neutron: Make dnsmaq aware of all names https://review.openstack.org/52930 | 14:01 |
| * jorisroovers wishes he could sleep on the plane for more than an hour | 14:02 | |
| anteaya | rkukura: nice | 14:07 |
| anteaya | mine was better than the flight that had to turn back to HK after reaching Tokyo airspace | 14:07 |
| rkukura | anteaya: made up for getting no sleep at all my last night in HK | 14:07 |
| rkukura | anteaya: we had a couple people on that one I think | 14:08 |
| anteaya | but next time I hope to sit next to someone who understands the concept of personal space and also is familiar with how to use kleenex | 14:08 |
| anteaya | wasn't a lot of napkins and kleneex in HK, guess they don't use them much | 14:08 |
| anteaya | rkukura: ouch, that's no fun | 14:08 |
| anteaya | rkukura: yeah fungi was on that one and someone named ctracey | 14:09 |
| anteaya | whom I don't know | 14:09 |
| anteaya | flying to perth I had the option of going through HK | 14:10 |
| anteaya | I choose to take US to Sydney instead | 14:10 |
| anteaya | I will see if that was a good choice | 14:10 |
| *** alexpilotti has joined #openstack-neutron | 14:10 | |
| *** jecarey has quit IRC | 14:12 | |
| * anteaya looks at the snow outside | 14:12 | |
| *** russellb is now known as rustlebee | 14:13 | |
| anteaya | rustlebee: oh yeah, Friday | 14:13 |
| *** anteaya is now known as anttteaya | 14:14 | |
| *** x86brandon has joined #openstack-neutron | 14:22 | |
| jorisroovers | mm, anttteaya, having issues as well | 14:24 |
| jorisroovers | http://paste.openstack.org/show/53121/ | 14:24 |
| anttteaya | jorisroovers: I don't recall dealing with certs in any of my installtions with devstack | 14:25 |
| jorisroovers | mm, yeah | 14:25 |
| anttteaya | what instructions are you following that you are dealing with certs? | 14:25 |
| jorisroovers | I think it is a SELinux thing | 14:25 |
| anttteaya | oh | 14:25 |
| anttteaya | disable selinux | 14:25 |
| anttteaya | sorry I have only installed devstack on fedora once or twice | 14:26 |
| anttteaya | I think it is posted on the fedora/devstack instructions somewhere | 14:26 |
| anttteaya | but disable selinux | 14:26 |
| anttteaya | see what that does | 14:26 |
| jorisroovers | ok, try #27 or something like that :-) | 14:28 |
| haleyb | anteaya: regarding your delegation/lack of interest, there are a number of people looking at the agents from various angles, from races, to scaling, to stability. I know those of us at hpcloud have been pushing our end-of-H-release fixes upstream, and we're only a small part of the total - it does take a village to build Neutron :) | 14:30 |
| haleyb | And the race conditions mark and others are chasing are sometimes the hardest to | 14:30 |
| haleyb | track down, I'm surprised we haven't tripped over them here yet | 14:30 |
| *** carl_baldwin has joined #openstack-neutron | 14:33 | |
| jorisroovers | anttteaya, mm, same issue :( | 14:38 |
| *** richardboswell has joined #openstack-neutron | 14:39 | |
| jorisroovers | anttteaya, downloading fedora 18 | 14:40 |
| jpich | jorisroovers: Does it stop for a long time before showing the timeout? I've seen this a few times on VMs where I've run devstack multiple times. "Ctrl-C" helps things move on | 14:41 |
| jpich | if you time it well :-) I think it's having trouble overwriting existing certs | 14:41 |
| anttteaya | hello haleyb thanks for the introduction | 14:41 |
| anttteaya | glad to know what you are working on | 14:41 |
| *** steven-weston has joined #openstack-neutron | 14:41 | |
| jpich | jorisroovers: I'm running devstack on Fedora 19 just fine, FWIW | 14:41 |
| anttteaya | and yes I agree on both points it takes a village to build a project, for me the project is openstack | 14:41 |
| jorisroovers | well, I read on the intertubes that hitting enter a couple of times could help, so that is what I did | 14:42 |
| jorisroovers | jpich, ok let me try that | 14:42 |
| anttteaya | and race conditions are very hard to track down, which is why I am keen to hear markmcclain's perspective on them | 14:42 |
| * jorisroovers spins VM back up | 14:42 | |
| haleyb | anttteaya: we met in HK too, there were three of us from HP there (4 if counting yourself) | 14:42 |
| anttteaya | so we can get many eyes on them | 14:42 |
| anttteaya | welcome jpich | 14:42 |
| anttteaya | jpich: are you feeling better yet? | 14:42 |
| jpich | jorisroovers: Maybe deleting that directory would work just fine as well :-) | 14:42 |
| jpich | anttteaya: Hello! Yep, once I was home, all good :) | 14:43 |
| anttteaya | haleyb: I do, yes I remember, Carl indicated interest in coming to Montreal | 14:43 |
| anttteaya | jpich: so glad to hear that | 14:43 |
| *** bswrchrd has quit IRC | 14:43 | |
| jorisroovers | jpich, you mean deleting '/home/cisco/devstack/accrc/' ? | 14:43 |
| anttteaya | ill while traveling is the worst kind of ill | 14:43 |
| jpich | anttteaya: Thanks :) | 14:43 |
| jpich | jorisroovers: Yes... though I've usually gone the Ctrl-C route so I can't guarantee anything - I'll try that myself next time I see it | 14:44 |
| anttteaya | jorisroovers: jpich works with fedora way more than I | 14:44 |
| anttteaya | jorisroovers: she is sure to know stuff about how to get devstack working | 14:44 |
| jpich | jorisroovers: IMO devstack should be recreating the files if they're not there (...seeing how it's trying to recreate them if they're already there and all) | 14:44 |
| jpich | :) | 14:45 |
| anttteaya | :D | 14:45 |
| anttteaya | glad to have you here | 14:45 |
| jorisroovers | ok, gonna try the CTRL+C route (more of an ubuntu guy myself) | 14:45 |
| *** dkliban has joined #openstack-neutron | 14:45 | |
| * anttteaya goes back to writing letters of invitation | 14:46 | |
| anttteaya | obondarev: you around? | 14:49 |
| jorisroovers | anttteaya, jpich, ok, so now all of the sudden devstack finished without me having to do anything :-) | 14:49 |
| * jorisroovers runs tox | 14:50 | |
| anttteaya | jorisroovers: isn't that great news | 14:50 |
| anttteaya | can you paste the last 20 or so lines of devstack output | 14:50 |
| anttteaya | when it finished? | 14:50 |
| anttteaya | or you could run tox | 14:50 |
| anttteaya | :D | 14:50 |
| jorisroovers | anttteaya: http://paste.openstack.org/show/53128/ | 14:51 |
| anttteaya | awesome thanks | 14:51 |
| anttteaya | fantastic output | 14:52 |
| anttteaya | is tox running for you? | 14:52 |
| jorisroovers | nope, trying to install the dependencies now | 14:52 |
| anttteaya | okay | 14:52 |
| anttteaya | what dependencies for tox? | 14:52 |
| jorisroovers | uhu | 14:53 |
| *** nati_ueno has joined #openstack-neutron | 14:54 | |
| jorisroovers | anttteaya, jpich, ok that didn't really help | 14:55 |
| jorisroovers | http://paste.openstack.org/show/53129/ | 14:55 |
| jpich | jorisroovers: There's an issue with pypi at the moment, you probably want to use our openstack mirror for now | 14:56 |
| *** bswrchrd has joined #openstack-neutron | 14:57 | |
| anttteaya | make sure the development packages of libxml2 and libxslt are installed | 14:57 |
| anttteaya | did you install these dependencies jorisroovers? | 14:58 |
| *** jecarey has joined #openstack-neutron | 14:58 | |
| jorisroovers | yep | 14:58 |
| anttteaya | okay | 14:58 |
| jpich | jorisroovers: Personally I set up a ~/.pip/pip.conf as described in http://jacobian.org/writing/when-pypi-goes-down/ with this URL -> http://pypi.openstack.org/openstack | 14:58 |
| *** richardboswell has quit IRC | 14:59 | |
| anttteaya | what jpich said | 14:59 |
| jpich | https://pypi.python.org/pypi -> 502, cf http://status.python.org/ | 14:59 |
| anttteaya | that points you to the openstack mirror | 14:59 |
| anttteaya | so after pypi is back up again, you might need to rename that file to download pip packages that are not on our mirror | 15:00 |
| anttteaya | we just have the packages needed for openstack | 15:00 |
| jorisroovers | i | 15:01 |
| jorisroovers | sorry, that was supposed to be in vim ;-) | 15:02 |
| anttteaya | :D | 15:02 |
| jorisroovers | ok, jpich, thanks, I just did that | 15:02 |
| jorisroovers | seems to work :-) | 15:02 |
| anttteaya | try pip installing tox again | 15:02 |
| anttteaya | and then running tox after rm -rf /opt/stack/nova/.tox | 15:02 |
| jpich | jorisroovers: Great, yw | 15:03 |
| openstackgerrit | Oleg Bondarev proposed a change to openstack/neutron: Add state reporting to the metadata agent https://review.openstack.org/56628 | 15:03 |
| * jorisroovers is running tox again after following anttteaya's instructions | 15:04 | |
| *** beagles has joined #openstack-neutron | 15:05 | |
| anttteaya | welcome beagles | 15:06 |
| jorisroovers | anttteaya, mm, it seems like there is still some dependency missing, I think the libxml one | 15:08 |
| jorisroovers | http://paste.openstack.org/show/53140/ | 15:08 |
| *** nplanel has quit IRC | 15:10 | |
| * anttteaya clicks | 15:12 | |
| anttteaya | what command did you run that this was the output? | 15:13 |
| *** nplanel has joined #openstack-neutron | 15:13 | |
| jorisroovers | tox | 15:14 |
| anttteaya | really? | 15:14 |
| jorisroovers | jep | 15:15 |
| anttteaya | jpich to you usually get this kind of output when running tox in devstack on fedora? | 15:15 |
| anttteaya | *do you | 15:15 |
| anttteaya | http://paste.openstack.org/show/53140/ | 15:15 |
| jpich | This looks like dependencies are still missing | 15:16 |
| anttteaya | fatal error: libxml/xmlversion.h: No such file or directory | 15:16 |
| anttteaya | so yeah, jorisroovers I agree with you | 15:17 |
| jorisroovers | uhu | 15:17 |
| *** armax has joined #openstack-neutron | 15:17 | |
| anttteaya | what kind of output do you get when you install libxml? | 15:17 |
| jorisroovers | I just installed libxml | 15:17 |
| anttteaya | hello armax | 15:17 |
| jorisroovers | I only installed libxml2 before | 15:17 |
| armax | hello | 15:17 |
| jorisroovers | trying again | 15:17 |
| anttteaya | is there a libxml-dev package? | 15:17 |
| anttteaya | armax: how goes your day so far? | 15:18 |
| * anttteaya thinks in terms of ubuntu packaing | 15:18 | |
| armax | I have just stepped in the office, I am not yet properly caffeinated so…bad :) | 15:18 |
| anttteaya | does fedora do -dev packages, jpich | 15:18 |
| jorisroovers | aha, there is :-) | 15:18 |
| jorisroovers | just installed it | 15:18 |
| jpich | Of course :-) | 15:18 |
| anttteaya | armax: ha ha ha, happy caffeine to you | 15:18 |
| anttteaya | okay | 15:18 |
| armax | brb | 15:18 |
| anttteaya | k | 15:19 |
| jpich | jorisroovers: Glad you found it | 15:19 |
| jorisroovers | mm, same issue :( | 15:21 |
| *** nati_ueno has quit IRC | 15:21 | |
| anttteaya | jorisroovers: paste please | 15:22 |
| *** jlibosva has quit IRC | 15:23 | |
| jorisroovers | ok, I did not install devel for libxslt | 15:23 |
| jorisroovers | I did now | 15:23 |
| jorisroovers | retrying | 15:23 |
| dims | anttteaya, salv-orlando - https://review.openstack.org/#/c/56629/ asking infra to switch on large-ops as voting job | 15:23 |
| jorisroovers | anttteaya, as a heads up, I will be travelling next week. So probably won't be around often. Will be back the week after. | 15:24 |
| jorisroovers | just to let you know that I'm not abandoning it after the first week ;-) | 15:24 |
| *** nati_ueno has joined #openstack-neutron | 15:26 | |
| *** nati_ueno has quit IRC | 15:28 | |
| anttteaya | dims you rock!! | 15:32 |
| anttteaya | thank you soo much | 15:32 |
| anttteaya | dims: do we know anything about this failure? http://bit.ly/1cwcXWK | 15:32 |
| anttteaya | it seems to have occured after 56075 was merged | 15:33 |
| anttteaya | jorisroovers: fair enough, I will be traveling the week of Nov. 25 and the week after that | 15:33 |
| anttteaya | I have to drive with my parents on their annual migration south | 15:33 |
| *** steven-weston_ has joined #openstack-neutron | 15:33 | |
| anttteaya | didn't plan on it, decided it was time, they didn't resist | 15:34 |
| anttteaya | so good choice, bad timing | 15:34 |
| anttteaya | sigh | 15:34 |
| jorisroovers | mm, sql dependency error now | 15:34 |
| anttteaya | so I will be checking -neutron logs during the evenings but will have little time for daily stuff | 15:34 |
| anttteaya | but you know others and jpich is great with getting started with devstack | 15:35 |
| jorisroovers | :-) | 15:35 |
| anttteaya | so be sure to ask for help | 15:35 |
| jorisroovers | will do, thanks | 15:35 |
| anttteaya | jorisroovers: can I see the paste of it | 15:35 |
| anttteaya | great | 15:35 |
| anttteaya | don't want your progress to be slowed down by me | 15:35 |
| anttteaya | too many good people here to call on | 15:35 |
| armax | anttteaya: back | 15:35 |
| *** carl_baldwin has quit IRC | 15:35 | |
| anttteaya | armax: better now? | 15:35 |
| jorisroovers | http://paste.openstack.org/show/53142/ | 15:36 |
| jpich | jorisroovers, anttteaya: Right, I am happy to help. #openstack-101 is also a good place for new contributors to ask questions, if you get stuck | 15:36 |
| armax | anttteaya: caffeine hasn't kicked yet, so in a few minutes probably yes :) | 15:36 |
| anttteaya | dims: that failure was a check failure, not a gate failure - I'm still curious though | 15:36 |
| *** steven-weston has quit IRC | 15:37 | |
| *** steven-weston_ is now known as steven-weston | 15:37 | |
| jorisroovers | \me gotta run in a few | 15:37 |
| * anttteaya sets a timer for taking to armax | 15:37 | |
| anttteaya | :D | 15:37 |
| *** networkstatic has joined #openstack-neutron | 15:37 | |
| anttteaya | jorisroovers: fair enough | 15:37 |
| anttteaya | and as jpich suggests join -101 | 15:37 |
| sc68cal | anttteaya: is there an IRC room for the devstack folk? | 15:37 |
| sc68cal | sorry to bother you | 15:37 |
| armax | anttteaya: what can I do for you? | 15:37 |
| anttteaya | sc68cal: never a bother | 15:38 |
| anttteaya | always happy to hear from you | 15:38 |
| anttteaya | I'd suggest -qa | 15:38 |
| anttteaya | or -dev | 15:38 |
| jpich | jorisroovers: Probably another devel package missing | 15:38 |
| anttteaya | nothing, just saying hi | 15:38 |
| anttteaya | :D | 15:38 |
| anttteaya | now that the caffeine has kicked in | 15:38 |
| jorisroovers | jpich, agreed, now determining which one | 15:38 |
| anttteaya | sc68cal: dtroyer writes most of devstack | 15:39 |
| anttteaya | sc68cal: sdague and mtreinish would be up on what is happening with it too | 15:39 |
| anttteaya | sc68cal: here is the list of devstack core folks: https://review.openstack.org/#/admin/groups/50,members | 15:39 |
| sc68cal | yeah i've just got a wierd permission denied error when devstack runs, nova.conf is apparently getting locked down 700 nova:nova | 15:40 |
| sc68cal | * /etc/nova | 15:40 |
| anttteaya | sc68cal jorisroovers has the same issue | 15:40 |
| sc68cal | fun fun fun | 15:41 |
| anttteaya | sc68cal: https://bugs.launchpad.net/devstack/+bug/1248923 | 15:41 |
| anttteaya | yup | 15:41 |
| anttteaya | especially for a beginner | 15:41 |
| anttteaya | I am trying to find out if it is ubuntu based or affects both ubuntu and fedora | 15:41 |
| anttteaya | jorisroovers is trying a fedora install now, but his first run at it | 15:42 |
| anttteaya | for extra fun pypi was down | 15:42 |
| anttteaya | though is now back up | 15:42 |
| jorisroovers | murphy's law :-) | 15:42 |
| anttteaya | it's the only constant I know of | 15:42 |
| sc68cal | I remember when PBR was all screwed up | 15:42 |
| sc68cal | like mid-july or so | 15:42 |
| anttteaya | I don't remember at time when it wasn't | 15:43 |
| * anttteaya trolls mordred | 15:43 | |
| sc68cal | well I'll just set devstack back to a known good commit hash for the time being and wait it out | 15:43 |
| openstackgerrit | Yves-Gwenael Bourhis proposed a change to openstack/neutron: Make dnsmaq aware of all names https://review.openstack.org/52930 | 15:44 |
| * mordred throws a goat at anttteaya | 15:44 | |
| * anttteaya throws a duck at mordred's goat | 15:45 | |
| anttteaya | sc68cal: that is an option | 15:45 |
| *** gatuus has joined #openstack-neutron | 15:46 | |
| anttteaya | asking in -qa | 15:47 |
| anttteaya | no takers so far, will let you know if I find anything | 15:48 |
| sc68cal | I should just hop in there and idle instead of always pestering you | 15:48 |
| openstackgerrit | Yves-Gwenael Bourhis proposed a change to openstack/neutron: Make dnsmaq aware of all names https://review.openstack.org/52930 | 15:49 |
| *** rustlebee is now known as drumkilla | 15:50 | |
| *** drumkilla is now known as rustlebee | 15:50 | |
| *** beagles is now known as b3nt_pin | 15:51 | |
| *** rustlebee is now known as drumkilla | 15:51 | |
| jorisroovers | *happy dance* | 15:52 |
| jorisroovers | the tests are running :D | 15:52 |
| jorisroovers | anttteaya, gotta run | 15:52 |
| anttteaya | w00t! | 15:55 |
| anttteaya | sc68cal: never a pester, and always a good idea to just jump in | 15:55 |
| anttteaya | sc68cal: looks like this issue might be just ubuntu based | 15:55 |
| anttteaya | sc68cal: can I sell you on trying a fedora image? | 15:56 |
| sc68cal | possibly - though not quickly - using vagrant_devstack | 15:56 |
| anttteaya | okay | 15:56 |
| anttteaya | let me know what you decide | 15:56 |
| anttteaya | if I can get two positives for fedora, I'll go chat with chuck | 15:57 |
| * anttteaya goes to check the fire | 15:57 | |
| *** jorisroovers has quit IRC | 15:57 | |
| *** carl_baldwin has joined #openstack-neutron | 16:03 | |
| sc68cal | Hey, I've gotten stuck on an issue for one of my reviews - where tox blows up on a weird error when the full suite is run, but running the test that fails on its own does not error out | 16:22 |
| sc68cal | Details - http://paste.openstack.org/show/53159/ | 16:24 |
| *** layer427expert has joined #openstack-neutron | 16:25 | |
| *** x86brandon_ has joined #openstack-neutron | 16:26 | |
| *** x86brandon has quit IRC | 16:26 | |
| *** x86brandon_ is now known as x86brandon | 16:26 | |
| sc68cal | attempting to rebase, see if that fixes the issue..... | 16:28 |
| *** layer427expert has quit IRC | 16:30 | |
| anttteaya | go sc68cal | 16:30 |
| * anttteaya clicks | 16:30 | |
| anttteaya | sc68cal: did you see mestery's comment | 16:31 |
| sc68cal | yeah about ml2 | 16:31 |
| sc68cal | ? | 16:31 |
| *** markmcclain has joined #openstack-neutron | 16:31 | |
| sc68cal | my concern is that whatever is wrong will just follow me when I port it over to ml2 | 16:32 |
| anttteaya | let's see if mestery is available for a discussion about it | 16:32 |
| anttteaya | kyle? are you around? | 16:33 |
| sc68cal | mestery: ping | 16:33 |
| *** coolsvap has joined #openstack-neutron | 16:38 | |
| sc68cal | nope- rebase doesn't fix | 16:41 |
| * markmcclain reads catches up on scrollback | 16:41 | |
| markmcclain | sc68cal: still working on my coffee, but to make sure I'm current.. when you run a single test it passes as a suite failure.. correct? | 16:45 |
| dkehn | markmcclain: https://etherpad.openstack.org/p/icehouse-summit-neutron-pain-points, where we having a meeting sometime? | 16:45 |
| sc68cal | markmcclain: when I run the test that fails in jenkins on my local machine, it passes, but when I just run tox and all the tests get run, then it gives that wierd mock assignment error and the tests fail | 16:46 |
| *** nplanel has quit IRC | 16:46 | |
| openstackgerrit | Jenkins proposed a change to openstack/python-neutronclient: Updated from global requirements https://review.openstack.org/56654 | 16:51 |
| *** jsergent has joined #openstack-neutron | 16:51 | |
| *** jpich has quit IRC | 16:55 | |
| mestery | sc68cal: What comment of mine are you looking for? | 16:56 |
| sc68cal | mestery: just the QOS API, the OVS based implementation | 16:56 |
| sc68cal | trying to figure out that weird jenkins error | 16:56 |
| sc68cal | I'm worried that whatever is triggering the failure is just going to follow me over into the review for the ML2 based implementation | 16:57 |
| sc68cal | plus it's driving me bonkers | 16:57 |
| *** enikanorov has joined #openstack-neutron | 16:57 | |
| sc68cal | mestery: http://paste.openstack.org/show/53159/ | 16:57 |
| markmcclain | dkehn: sorry I missed a bit of presentation… what meeting? | 16:58 |
| *** enikanorov___ has quit IRC | 16:58 | |
| dkehn | markmcclain: about logging and the other items and assignments | 16:58 |
| mestery | sc68cal: I've not see that error, and this only happens with your QoS patches and not a stock run with OVS? | 16:59 |
| sc68cal | mestery: With my QoS patches, and only when you run the full test suite | 16:59 |
| markmcclain | dkehn: I know marun has been traveling post summit | 16:59 |
| sc68cal | mestery: running the tests by themselves, they pass | 16:59 |
| dkehn | markmcclain: np, just didn't want to miss something | 16:59 |
| markmcclain | I don't know if there is a dedicated meeting post summit | 16:59 |
| markmcclain | but my plan is to incorporate the output of that session into our plan for the cycle | 17:00 |
| dkehn | markmcclain: please communicate whatever is decided against plan, just don't wnat to drop the ball on logging | 17:01 |
| markmcclain | dkehn: agreed | 17:02 |
| markmcclain | good logging is a must | 17:02 |
| *** jprovazn has quit IRC | 17:02 | |
| dkehn | markmcclain: sorry sent the wrong link this is the one I was referring to | 17:03 |
| dkehn | https://etherpad.openstack.org/p/icehouse-summit-qa-neutron | 17:03 |
| dkehn | markmcclain: see Tenant Isolation & Parallel Testing | 17:04 |
| markmcclain | ah.. now things are clearer | 17:04 |
| dkehn | markmcclain: sorry about that, jet lag has a way fof creeping in | 17:04 |
| markmcclain | yeah I know the feeling | 17:05 |
| marun | dkehn, markmcclain: not sure we came to any conclusion. I guess following up on the mailing list is the next step? | 17:07 |
| *** decede has joined #openstack-neutron | 17:07 | |
| *** safchain has quit IRC | 17:08 | |
| markmcclain | yeah.. maybe start with a strawman proposal | 17:09 |
| marun | markmcclain: okely | 17:10 |
| marun | markmcclain: I'm assuming you're talking about moving to an sdn controller-based solution for the default open source plugin. | 17:11 |
| markmcclain | marun: for how to improve the logging | 17:11 |
| markmcclain | for SDN side of it.. we need a mechdriver or plugin for one of the opensource ones first | 17:11 |
| marun | markmcclain: ah, was reading the wrong link | 17:12 |
| markmcclain | no worries.. it's after midnight there for you right? | 17:12 |
| marun | markmcclain, dkehn: we already have a patch in play: https://review.openstack.org/#/c/52128/ | 17:12 |
| markmcclain | right.. that's step one for request logging | 17:13 |
| marun | And there should be a follow-on blueprint for ensuring traceability at the agent level as well | 17:13 |
| * marun goes to actually read the link | 17:13 | |
| marun | ah, right | 17:13 |
| markmcclain | we still have areas in the code which log too much of the wrong info and generate noise | 17:13 |
| markmcclain | or don't log at the correct level | 17:13 |
| marun | Does this mean we can go ahead and make polling minimization the default? :) | 17:14 |
| markmcclain | so we should identify the areas that need level changes and messages that are vague/misleading | 17:14 |
| markmcclain | +1 | 17:14 |
| marun | awesome | 17:14 |
| markmcclain | I'd be happy to switch | 17:14 |
| markmcclain | that way it will have the whole cycle to bake | 17:14 |
| marun | There's one more patch we have to merge first - the respawn interval is not configurable (stupid omission) so that if the monitor dies (ovs restarting, accidental process death, etc) the agent reverts to polling. | 17:15 |
| marun | but that's easy to fix. i'll make sure to get it submitted asap | 17:15 |
| markmcclain | marun: awesome | 17:17 |
| *** devlaps has joined #openstack-neutron | 17:21 | |
| gatuus | VPNaaS w/OpenSwan anyone? | 17:22 |
| *** ygbo has quit IRC | 17:24 | |
| mestery | markmcclain marun: I have a good start on the ODL MechanismDriver for ML2, FYI. Was reading your conversation. | 17:24 |
| markmcclain | mestery: awesome | 17:24 |
| mestery | Working with the ODL guys to ensure they have the appropriate APIs on their side for this to work, and that things are running smoothly there. | 17:24 |
| marun | mestery: awesome | 17:25 |
| mestery | Yes, it's only 160 LOC, which is nice. :) | 17:25 |
| markmcclain | gatuus: there are feature differences between openswan and strongswan | 17:25 |
| markmcclain | for consistency we had to pick 1 | 17:25 |
| marun | i bet devstack deployment of odl is going to be the harder part :) | 17:25 |
| mestery | marun: Already have that working too. | 17:26 |
| mestery | I have a modified devstack which can either deploy ODL for you, or use an existing one. | 17:26 |
| gatuus | I'm having problems getting ipsce pluto write into a neutron dir | 17:26 |
| gatuus | *ipsec | 17:27 |
| mestery | I just need the ODL guys to tie into their NeutronAPIService with a SB thing like OVSDB, and then I need to handle port binding. | 17:27 |
| markmcclain | pcm_: have you seen gatuus issue? | 17:28 |
| mestery | marun: https://github.com/CiscoSystems/neutron/blob/odl_ml2/neutron/plugins/ml2/drivers/mechanism_odl.py | 17:28 |
| *** networkstatic has quit IRC | 17:30 | |
| marun | mestery: nice | 17:31 |
| marun | mestery: those assert statements are just for dev? | 17:31 |
| *** networkstatic has joined #openstack-neutron | 17:32 | |
| * mestery nods. | 17:32 | |
| mestery | Although, the NCS driver from which this was modeled has them in there too, so perhaps they shoudl be removed there as well. | 17:32 |
| marun | mestery: what functionality does the driver actually enable? it looks like it's just syncing data between the two, and i'm not clear on how that enables advanced functionality | 17:32 |
| marun | mestery: yeah, I'd -1 any change that included assert in running code | 17:32 |
| marun | mestery: throwing explicit exceptions is much preferable | 17:33 |
| mestery | marun: That's all it does, and then ODL itself makes decisions on how to setup the virtual tenant networks. | 17:33 |
| * mestery agrees. | 17:33 | |
| *** jgrimm has joined #openstack-neutron | 17:33 | |
| marun | mestery: hmmm. i guess i'm looking at it from a 'parity with ovs/lb' perspective. is the intention to allow setup through neutron apis or to push the setup to odl? | 17:34 |
| marun | mestery: and (apologies in advance), what does the nvp plugin support? | 17:34 |
| mestery | marun: Push the setup to ODL. For now, we're not dealing with L3 for example, we'll use namespaces for that. But ODL could do that in the future. | 17:34 |
| marun | mestery: Hmmm | 17:35 |
| marun | mestery: I guess, just like parity with nova, the key is going to ensure we have a representative suite of tests that can target both ovs/lb in tempest. | 17:35 |
| *** jgrimm is now known as reaper | 17:35 | |
| marun | mestery: and then ensure that odl can pass them | 17:36 |
| *** reaper is now known as jgrimm | 17:36 | |
| *** gatuus has quit IRC | 17:36 | |
| marun | odl -> odl integration with neutron | 17:36 |
| mestery | marun: That's exactly correct. | 17:36 |
| *** gatuus has joined #openstack-neutron | 17:36 | |
| marun | mestery: awesome | 17:36 |
| *** suresh12 has joined #openstack-neutron | 17:36 | |
| *** suresh12 has quit IRC | 17:36 | |
| *** suresh12 has joined #openstack-neutron | 17:37 | |
| mestery | marun: I've been working on getting Tempest running locally on my devstack setups, I've hit a bunch of issues with the network tests where the DHCP agent is querying a network which was already deleted. | 17:38 |
| mestery | Otherwise, they seem to run ok. | 17:38 |
| marun | mestery: I'm afraid we have a ways to go to good coverage in tempest, though. Part of the work will be writing new tests. | 17:40 |
| *** layer427expert has joined #openstack-neutron | 17:41 | |
| markmcclain | marun: that's a lot of what we'll be trying to close in the Montreal sprint | 17:41 |
| marun | markmcclain: before, I hope | 17:41 |
| *** suresh12 has quit IRC | 17:41 | |
| markmcclain | we'll defnitely be doing things leading up to it | 17:43 |
| *** gatuus has quit IRC | 17:43 | |
| *** suresh12 has joined #openstack-neutron | 17:45 | |
| *** CaptainTacoSauce has joined #openstack-neutron | 17:51 | |
| *** layer427expert has quit IRC | 17:54 | |
| sc68cal | Comcast is hoping that we can start adding tempest tests for v6 functionality | 18:13 |
| mestery | sc68cal: Awesome! | 18:15 |
| sc68cal | since we've been hitting some issues and we're not sure if it's something we did | 18:17 |
| sc68cal | or if it's legit bugs | 18:17 |
| *** jog0 is now known as flashgordon | 18:20 | |
| salv-orlando | mestery: could you have hit the bugs for which armax has pushed patches under review? | 18:21 |
| *** jecarey has quit IRC | 18:21 | |
| *** suresh12 has quit IRC | 18:29 | |
| *** suresh12 has joined #openstack-neutron | 18:30 | |
| armax | salv-orlando: can you fill me in? | 18:31 |
| armax | oh I see: mestery what issues are you experiencing? | 18:32 |
| salv-orlando | ok | 18:32 |
| *** marun has quit IRC | 18:34 | |
| *** jianingy_afk has quit IRC | 18:34 | |
| *** suresh12 has quit IRC | 18:35 | |
| mestery | armax salv-orlando: Apologies, had stepped out for a few minutes. | 18:36 |
| armax | mestery:no worries | 18:36 |
| mestery | armax: I am hitting an issue with my tempest runs where the DHCP agent is querying for a network which has already been deleted. | 18:36 |
| armax | right, but does that lead to any tempest failures? | 18:36 |
| mestery | Well, for me they seem to fail at this point. Perhaps I've mistaken the cause, let me go and grab a screen shot. | 18:37 |
| *** jianingy_afk has joined #openstack-neutron | 18:38 | |
| armax | unfortunately the interaction between the server and the agents is fundamentally broken | 18:38 |
| armax | for instance, a network may be about to be deleted, but the dhcp agent may still go the the server and check whether a dhcp port exists for it | 18:39 |
| mestery | Yes, markmcclain has mentioned that to me. | 18:39 |
| * mestery nods and has seen that. | 18:40 | |
| armax | so it may go and try to create the port even though the network is about to be deleted | 18:40 |
| armax | so something funky like this may happen | 18:40 |
| armax | unless we introduce some sort of intermediate state | 18:40 |
| mestery | armax: Thanks for the explanation. | 18:40 |
| armax | that said, if you see persistent failures in your tempest runs | 18:41 |
| armax | that's a different matter | 18:41 |
| armax | and I'd like to understand more | 18:41 |
| mestery | armax: I'm trying to familiarize myself with the tempest stuff as well as part of this exercise. | 18:41 |
| mestery | armax: OK, let me see what else I can do now, resetting things up now. | 18:41 |
| armax | cool | 18:41 |
| mestery | armax: So to just run the network tempest tests, I do this: testr run --parallel tempest.api.network | 18:41 |
| mestery | That's correct? | 18:42 |
| armax | I use tox rather than tester directly | 18:42 |
| armax | but that sounds about right | 18:43 |
| mestery | OK, cool. Any reason? Should I be using tox? The README indicates testr, just curious. | 18:43 |
| armax | no particular reason…that runs the whole lot | 18:44 |
| mestery | thanks | 18:45 |
| mestery | So on this run, I had 2 LoadBalancer tests fail and one VPN test fail. | 18:46 |
| mestery | Also, I see networks, subnets and ports from the tempest run still dangling around. Is that expected? | 18:47 |
| *** suresh12 has joined #openstack-neutron | 18:51 | |
| *** rossella_s has quit IRC | 18:53 | |
| *** alexpilotti has quit IRC | 19:00 | |
| *** carl_baldwin has quit IRC | 19:00 | |
| *** alexpilotti has joined #openstack-neutron | 19:04 | |
| *** dims has quit IRC | 19:14 | |
| *** dims has joined #openstack-neutron | 19:14 | |
| *** jecarey has joined #openstack-neutron | 19:27 | |
| *** layer427expert has joined #openstack-neutron | 19:32 | |
| sc68cal | markmcclain: Should I file a bug for the mock assignment error issue? Does it sound like a bug or is it something I've not done correctly in my code | 19:32 |
| markmcclain | sc68cal: if the issue occurs on the master branch then it is a bug | 19:33 |
| markmcclain | if the issue only occurs in your working branch with code changes then it is something about the change | 19:33 |
| markmcclain | when you run all of the tests they aren't always executed in the same order | 19:34 |
| markmcclain | so it could be a test ordering issue that has revealed this bug | 19:34 |
| sc68cal | is there an easy way with tox to set a breakpoint on that portion of the code? I'm a litle new to tox, compared to nosetests | 19:35 |
| *** jianingy_afk has quit IRC | 19:35 | |
| *** chandankumar has quit IRC | 19:35 | |
| *** chandankumar has joined #openstack-neutron | 19:37 | |
| *** alexpilotti has quit IRC | 19:45 | |
| *** steven-weston_ has joined #openstack-neutron | 19:46 | |
| *** steven-weston has quit IRC | 19:49 | |
| *** steven-weston_ is now known as steven-weston | 19:49 | |
| dkehn | sc68cal: I use ./tox/py27/bin/activate; python -m unittest neutron.tests.unit<yourtestclass>; note set the pdb.set_trace() before hand | 19:52 |
| *** nati_ueno has joined #openstack-neutron | 19:59 | |
| *** markmcclain has quit IRC | 20:05 | |
| salv-orlando | sc68cal: are you using one of those fake extension managers in your tests? | 20:14 |
| openstackgerrit | Carl Baldwin proposed a change to openstack/neutron: Optionally delete namespaces when they are no longer needed https://review.openstack.org/56114 | 20:16 |
| openstackgerrit | Carl Baldwin proposed a change to openstack/neutron: Optionally delete namespaces when they are no longer needed https://review.openstack.org/56114 | 20:18 |
| *** carl_baldwin has joined #openstack-neutron | 20:19 | |
| *** nati_ueno has quit IRC | 20:25 | |
| *** nati_ueno has joined #openstack-neutron | 20:25 | |
| anttteaya | it appears that the conversation follows markmcclain | 20:29 |
| mestery | anttteaya: What does that mean? | 20:30 |
| anttteaya | markmcclain showed up everyone started talking | 20:31 |
| anttteaya | the channel exploded with life | 20:31 |
| anttteaya | which is awesome | 20:31 |
| anttteaya | I am just observing what I see | 20:31 |
| mestery | Sometimes silence is golden. | 20:31 |
| anttteaya | sometimes | 20:31 |
| anttteaya | when the gate is stable | 20:32 |
| *** layer427expert has quit IRC | 20:32 | |
| openstackgerrit | A change was merged to openstack/neutron: Refactor configuring of floating ips on a router https://review.openstack.org/42412 | 20:40 |
| *** nati_ueno has quit IRC | 20:42 | |
| sc68cal | dkehn: thanks!! | 20:44 |
| anttteaya | mestery: you had asked earlier about tox and testr | 20:44 |
| mestery | anttteaya: Yes. armax game me some help there. | 20:45 |
| anttteaya | tox creates a virtual environment: http://tox.readthedocs.org/en/latest/index.html# | 20:45 |
| sc68cal | salv-orlando: not sure - possibly | 20:45 |
| anttteaya | testr allows you to run tests in parallel: https://testrepository.readthedocs.org/en/latest/MANUAL.html#parallel-testing | 20:45 |
| anttteaya | most testr commands are available to be invoked via tox | 20:46 |
| anttteaya | but I have never tried invoking --parallel via tox, so I don't know if that one can | 20:46 |
| anttteaya | great | 20:46 |
| mestery | thanks! | 20:46 |
| anttteaya | your welcome | 20:46 |
| anttteaya | *you're | 20:47 |
| *** nplanel has joined #openstack-neutron | 20:49 | |
| sc68cal | I might have missed this | 20:50 |
| sc68cal | self.addCleanup(mock.patch.stopall) | 20:50 |
| *** markmcclain has joined #openstack-neutron | 20:51 | |
| sc68cal | over in my setUp method for OpenvswitchQoSTestCase - since I create a mock rpcapi class | 20:51 |
| salv-orlando | if you see errors like 'object mock.Mock(…)' does not have attribute xyz' that's the cause, probably | 20:51 |
| salv-orlando | perhaps I should just go an check your logs | 20:51 |
| salv-orlando | wow I actually even starred your review - I'm getting so old I might a personal organiser soon | 20:52 |
| *** steven-weston has quit IRC | 20:53 | |
| sc68cal | salv-orlando: yeah that's it | 20:53 |
| sc68cal | or at least sounds like it | 20:53 |
| sc68cal | running full tox now | 20:53 |
| *** mlavalle has joined #openstack-neutron | 20:53 | |
| *** steven-weston has joined #openstack-neutron | 20:55 | |
| salv-orlando | sc68cal: yes it looks like tath's the problem | 20:55 |
| sc68cal | argh I'm such a dumbass | 20:55 |
| salv-orlando | It's also me being a lazy sloth. Very slow in reviews this week. | 20:57 |
| *** steven-weston has quit IRC | 21:03 | |
| sc68cal | yep that was it. tests pass | 21:14 |
| sc68cal | one less mystery in the universe | 21:14 |
| anttteaya | congratulations for passing tests | 21:16 |
| sc68cal | thanks - I'll remember to add that line when I create a unit test for the ML2 plugin | 21:18 |
| mestery | We'll also remember to look for it in the review as well. :) | 21:18 |
| sc68cal | nah, I'd put my money on me doing something new and stupid | 21:19 |
| *** markmcclain has quit IRC | 21:19 | |
| *** alexpilotti has joined #openstack-neutron | 21:25 | |
| *** gatuus has joined #openstack-neutron | 21:27 | |
| *** networkstatic has quit IRC | 21:28 | |
| *** catintheroof has joined #openstack-neutron | 21:36 | |
| *** dims has quit IRC | 21:37 | |
| catintheroof | Hi guys, im having problems accesing metadata from instances using this neutron OVS setup http://ow.ly/qScF6 can anyone help me understand something ? | 21:39 |
| *** suresh12 has quit IRC | 21:40 | |
| *** suresh12 has joined #openstack-neutron | 21:41 | |
| *** suresh12 has quit IRC | 21:46 | |
| *** layer427expert has joined #openstack-neutron | 21:46 | |
| bswrchrd | yea so I figured out what my issues were salv-orlando and mestery | 21:46 |
| *** bswrchrd has left #openstack-neutron | 21:47 | |
| salv-orlando | so? | 21:47 |
| mestery | What was it? | 21:47 |
| *** richardboswell has joined #openstack-neutron | 21:48 | |
| *** markmcclain has joined #openstack-neutron | 21:48 | |
| catintheroof | Hi guys, im having problems accesing metadata from instances using this neutron OVS setup http://ow.ly/qScF6 can anyone help me understand something ? | 22:01 |
| *** layer427expert has quit IRC | 22:01 | |
| anttteaya | hi catintheroof | 22:04 |
| anttteaya | based on what you shared, I don't have enough information to help you | 22:04 |
| *** x86brandon has quit IRC | 22:04 | |
| anttteaya | can you explain what you are trying to accomplish? | 22:05 |
| catintheroof | hi anttteaya can i give you more information ? | 22:05 |
| anttteaya | please do | 22:05 |
| anttteaya | if you need to share code please use paste.openstack.org | 22:05 |
| *** layer427expert has joined #openstack-neutron | 22:05 | |
| catintheroof | anttteaya, ok nice!! heres the deal ... i have vms created with neutron agent and OVS with this exact same schema that i provided , its VLAN MODE | 22:06 |
| *** suresh12 has joined #openstack-neutron | 22:06 | |
| catintheroof | anttteaya, and a default GW for that VM VLAN casted on a switch | 22:07 |
| catintheroof | anttteaya, so if i try to access any other network packets go through all the taps, and bridges inside de compute | 22:07 |
| catintheroof | anttteaya, and get to the default gw where it gets routed ... till here am i explaining myself ? | 22:08 |
| anttteaya | well I am not sure what the question is yet | 22:08 |
| anttteaya | is there a question? | 22:08 |
| catintheroof | anttteaya, of course... the question is this ... is regarding metadata, the instances cant access the metadata, so i issue the regular DNAT iptables rule to be able to acces it ... and then ... | 22:09 |
| anttteaya | and then what? | 22:09 |
| catintheroof | anttteaya, i see the original package exit the TAP ... the DNATED package exiting the qbrXXX but the package never hits the qvbXXXX interface | 22:10 |
| catintheroof | anttteaya, and we dont have an idea why ... since its not an iptables issue | 22:10 |
| anttteaya | can you paste an output of the command you run and the resultant code please? | 22:11 |
| anttteaya | are you trying to fix code or is this a support issue? | 22:11 |
| *** richardboswell has quit IRC | 22:12 | |
| catintheroof | anttteaya, its simple, im just tcpdumping the tap interface, the qbrXXX interface and the qvbXXX interface ... when i issue an ... for example SSH i see the packets go through all interfaces ... but when i curl metadata ip, it gets DNATED by iptables but the packages stays inside qbrXXX | 22:12 |
| *** richardboswell has joined #openstack-neutron | 22:12 | |
| anttteaya | not simple for me | 22:13 |
| anttteaya | I know nothing about the code base | 22:13 |
| anttteaya | I am facilitating communication between this project and other openstack projects | 22:13 |
| anttteaya | I am trying to determine if I can help you | 22:13 |
| anttteaya | so far I don't know how to help you | 22:13 |
| anttteaya | are you working with code? | 22:14 |
| anttteaya | if so can you show me which file? | 22:14 |
| catintheroof | anttteaya, im not touching any code base ... so, its i think ... a bridges / interfaces question | 22:14 |
| anttteaya | then I can find out who wrote it and see if they are avavilable | 22:14 |
| anttteaya | well so far noone has jumped into the conversation to help | 22:14 |
| anttteaya | and I don't know who in the channel has the knowledge you are looking for | 22:15 |
| anttteaya | and since I don't have a file to look at, I can't find someone that way | 22:15 |
| anttteaya | you are welcome to wait to see if someone answers your question | 22:15 |
| anttteaya | best I can offer | 22:15 |
| catintheroof | anttteaya, well ... if i cant find any networking help regarding neutron in the neutron list ... i dont know where to look then | 22:16 |
| * anttteaya nodes | 22:16 | |
| * anttteaya nods | 22:16 | |
| *** markmcclain has quit IRC | 22:16 | |
| anttteaya | all I can do is commiserate | 22:16 |
| *** nati_ueno has joined #openstack-neutron | 22:17 | |
| catintheroof | anttteaya, thanks anyway ... for trying | 22:17 |
| anttteaya | thanks catintheroof | 22:18 |
| *** drumkilla has quit IRC | 22:19 | |
| flashgordon | http://logstash.openstack.org/#eyJzZWFyY2giOiJcIkNvbm5lY3Rpb25GYWlsZWQ6IENvbm5lY3Rpb24gdG8gbmV1dHJvbiBmYWlsZWQ6IE1heGltdW0gYXR0ZW1wdHMgcmVhY2hlZFwiIiwiZmllbGRzIjpbXSwib2Zmc2V0IjowLCJ0aW1lZnJhbWUiOiI4NjQwMCIsImdyYXBobW9kZSI6ImNvdW50IiwidGltZSI6eyJ1c2VyX2ludGVydmFsIjowfSwic3RhbXAiOjEzODQ1NTM5NzA2NjF9 | 22:20 |
| flashgordon | that doesn't sound good | 22:20 |
| *** russellb has joined #openstack-neutron | 22:20 | |
| *** nati_ueno has quit IRC | 22:21 | |
| *** nplanel has quit IRC | 22:21 | |
| *** coolsvap has quit IRC | 22:28 | |
| salv-orlando | catintheroof: folsom, grizzly or havana? | 22:28 |
| catintheroof | salv-orlando, havana | 22:29 |
| catintheroof | salv-orlando, can you help me ? | 22:29 |
| salv-orlando | catintheroof: when an instance sends a packet to 169.254.169.254, the metadata proxy hijacks that packet and sends it to the metadata agent via unix socket. the metadata agent then is supposed to be in the same network as nova-metatada and will be able to reach it directly | 22:30 |
| salv-orlando | so instead of natting, there is the metadata proxy process performing the bridge between the tenant and management networks | 22:31 |
| openstackgerrit | Duarte Nunes proposed a change to openstack/neutron: MidoNet: Added support for the admin_state_up flag https://review.openstack.org/54275 | 22:32 |
| salv-orlando | what about your config. Which services are you running? the metadata-agent is needed | 22:32 |
| catintheroof | salv-orlando, but thats using a network node, but we are using vlan mode, so how do i do to use metadata service, since my GW is the switch where the vlan is crated ? | 22:32 |
| salv-orlando | ok - so you do not have a l3 agent running, but you are a pure-L2 use case? | 22:33 |
| catintheroof | EXACTLY | 22:33 |
| salv-orlando | you said vlan mode, just to be sure, are you running neutron or nova-network (because both have a vlan mode) | 22:34 |
| salv-orlando | seems neutron, but just want to confirm | 22:34 |
| catintheroof | salv-orlando, neutron | 22:34 |
| catintheroof | salv-orlando, with OVS hybrid | 22:34 |
| *** markmcclain has joined #openstack-neutron | 22:35 | |
| anttteaya | flashgordon: 1158 hits? | 22:35 |
| anttteaya | 1158 hits today? | 22:35 |
| salv-orlando | anttteaya: perhaps you wanted to post that in openstack-qa | 22:36 |
| catintheroof | salv-orlando, any ideas ? | 22:36 |
| salv-orlando | catintheroof: let me see if I got it right - you have a subnet whose gateway_ip is configured on your switch, and is then NATted to the address of the metadata ip, changing port from 80 to 8775? | 22:37 |
| catintheroof | salv-orlando, yes | 22:37 |
| anttteaya | I'm wrong | 22:37 |
| anttteaya | 1558 hits | 22:37 |
| salv-orlando | with the changes we added in grizzly for allowing metadata to work for overlapping IPs, you now need the metadata agent to add some headers which will allow nova to find the right metadata for your instance. | 22:38 |
| salv-orlando | And the metadata agent needs the metadata proxy... | 22:38 |
| salv-orlando | are you running at least the dhcp agent? | 22:38 |
| catintheroof | salv-orlando, flat injected | 22:38 |
| salv-orlando | catintheroof: thinking | 22:39 |
| catintheroof | salv-orlando, we are injecting ip and default gw into the instance | 22:39 |
| salv-orlando | yup; your use case does not fall either within the 'isolated metadata' or 'routed metadata' use cases; I'm trying to see if we can suit your use case somehow or we need code changes | 22:40 |
| catintheroof | salv-orlando, the weird thing is that iptables rules matches i see the original package exit the TAP ... the DNATED package exiting the qbrXXX but the package never hits the qvbXXXX interface | 22:40 |
| catintheroof | <catintheroof> anttteaya, and we dont have an idea why ... since its not an iptables issue | 22:40 |
| flashgordon | anttteaya: thaose aren't all bugs and multiple hits | 22:41 |
| salv-orlando | catintheroof: are you saying the packet appears to get lost in the bridge? | 22:41 |
| catintheroof | yeahp | 22:42 |
| anttteaya | flashgordon: what is the source? | 22:42 |
| catintheroof | salv-orlando, yeahp | 22:42 |
| anttteaya | flashgordon: if I am reading logstash correctly they are all large-ops failures | 22:42 |
| salv-orlando | I guess you can't see it either on br-int (I am not sure how well tcpdump works on veths)? | 22:42 |
| anttteaya | flashgordon: am I wrong? | 22:42 |
| flashgordon | anttteaya: not all | 22:42 |
| anttteaya | flashgordon: okay good | 22:43 |
| flashgordon | anttteaya: the issue is everything stacktraces all the time | 22:43 |
| flashgordon | so stacktraces aren't a strong sign something is wrong :/ | 22:43 |
| anttteaya | so the test was simply executed 1558 times today? | 22:44 |
| flashgordon | I am trying to figure out why http://logs.openstack.org/96/56496/1/gate/gate-tempest-devstack-vm-neutron-isolated/8df6c6c/logs/ failed | 22:44 |
| catintheroof | salv-orlando, can i write you and email ? to explain you ... | 22:44 |
| salv-orlando | Yep sure | 22:44 |
| salv-orlando | It's also pretty late my side of the world, so I might answer when I wake up tomorrow morning | 22:45 |
| catintheroof | salv-orlando, can you give me an email address ? | 22:45 |
| catintheroof | salv-orlando, so you can read it when you have time ? | 22:45 |
| salv-orlando | write to openstack-dev with [neutron] in the subject; this way you'll get help also from people which know a lot more on the topic | 22:45 |
| anttteaya | flashgordon: any leads so far? | 22:46 |
| flashgordon | anttteaya: :( | 22:46 |
| catintheroof | salv-orlando, i JUST DID !!! [openstack][neutron] problems accesing metadata on OVS VLAN mode (havana) and to the operators and general list also | 22:46 |
| catintheroof | salv-orlando, oh ... im not suscrived ... i will doit | 22:46 |
| catintheroof | salv-orlando, ill suscribe and then shoot the email ! | 22:47 |
| salv-orlando | no you're right about writing this to openstack | 22:47 |
| catintheroof | salv-orlando, thnaks ! | 22:47 |
| salv-orlando | it's actually better writing to openstack rather than openstack-dev | 22:47 |
| salv-orlando | I missed it because I check the openstack mailing list in the morning only | 22:47 |
| catintheroof | salv-orlando, writing to both, and operators also! just got rejected by the neutron list ! so you can find it in the general | 22:48 |
| catintheroof | salv-orlando, under the subject : [openstack][neutron] problems accesing metadata on OVS VLAN mode (havana) | 22:48 |
| salv-orlando | catintheroot: there should not be anymore a neutron list | 22:48 |
| anttteaya | catintheroof: if you write a support request to -dev you will be told not to | 22:48 |
| anttteaya | flashgordon might even tell you not too | 22:49 |
| anttteaya | *to | 22:49 |
| salv-orlando | cool I found your email, I will post a reply before I go to bed | 22:49 |
| flashgordon | catintheroof: yeah that -dev isn't for support its for future planning for developme nt | 22:52 |
| flashgordon | salv-orlando: don't promote bad dev ML behavior | 22:53 |
| flashgordon | please | 22:53 |
| salv-orlando | I think I said is actually better to write to openstack than openstack-dev after I found out it was not a dev request... | 22:54 |
| *** CaptainTacoSauce has quit IRC | 22:54 | |
| flashgordon | cool | 22:54 |
| flashgordon | anttteaya: found a fingerprint | 22:56 |
| flashgordon | http://logstash.openstack.org/#eyJzZWFyY2giOiJcIkNvbm5lY3Rpb24gdG8gbmV1dHJvbiBmYWlsZWQ6IE1heGltdW0gYXR0ZW1wdHMgcmVhY2hlZFwiICBBTkQgZmlsZW5hbWU6XCJsb2dzL3NjcmVlbi1uLXNjaC50eHRcIiIsImZpZWxkcyI6W10sIm9mZnNldCI6MCwidGltZWZyYW1lIjoiNjA0ODAwIiwiZ3JhcGhtb2RlIjoiY291bnQiLCJ0aW1lIjp7InVzZXJfaW50ZXJ2YWwiOjB9LCJzdGFtcCI6MTM4NDU1NjE2NTY2OCwibW9kZSI6IiIsImFuYWx5emVfZmllbGQiOiIifQ== | 22:56 |
| flashgordon | and its a new bug | 22:56 |
| flashgordon | staretd today :/ | 22:56 |
| flashgordon | at least in this manifestation | 22:56 |
| salv-orlando | catintheroof: I replies, but to add to bad ML behaviour, I forgot to remove openstack-dev from the recipients. So if you reply again, please remove it. | 22:59 |
| salv-orlando | hey bug 1291115 is back! | 23:00 |
| salv-orlando | I meant https://bugs.launchpad.net/neutron/+bug/1211915 | 23:00 |
| flashgordon | salv-orlando: its slightly different | 23:01 |
| flashgordon | opened a new one | 23:01 |
| flashgordon | https://bugs.launchpad.net/nova/+bug/1251784 | 23:01 |
| salv-orlando | I can think of this patch only: https://review.openstack.org/#/c/56075/ | 23:02 |
| *** layer427expert has quit IRC | 23:04 | |
| anttteaya | flashgordon: can I get a bitly? | 23:04 |
| flashgordon | perhaps but not sure how | 23:04 |
| flashgordon | anttteaya: bitly? | 23:04 |
| anttteaya | teh weechat doesn't like multiline urls | 23:04 |
| anttteaya | for the logstash url | 23:04 |
| flashgordon | https://bugs.launchpad.net/nova/+bug/1251784 | 23:05 |
| anttteaya | only the first line is linked, the link doesn't linewrap | 23:05 |
| anttteaya | yeah I got that one, for the logstash url | 23:05 |
| anttteaya | you said you found a fingerprint | 23:05 |
| *** gatuus has quit IRC | 23:05 | |
| *** jecarey has quit IRC | 23:06 | |
| anttteaya | or I can click the link from the bug report, that works too | 23:06 |
| anttteaya | that was kind of a sudden onset for this one | 23:07 |
| *** pcm_ has quit IRC | 23:07 | |
| anttteaya | salv-orlando: well that patch was merged on the 14th | 23:08 |
| salv-orlando | that was the patch for re-enabling large-ops. It's the only change in the nova/neutron interface in the last few days afaik | 23:09 |
| anttteaya | can we do a revert and check it? | 23:09 |
| *** layer427expert has joined #openstack-neutron | 23:09 | |
| salv-orlando | We can try, but I guess I won't be around to see the results as it's bedtime for me. | 23:09 |
| flashgordon | salv-orlando: I would like to be more sure about the why first | 23:10 |
| anttteaya | yeah | 23:10 |
| anttteaya | I can't believe how late you work, salv-orlando | 23:10 |
| flashgordon | also http://logs.openstack.org/92/55992/1/gate/gate-tempest-devstack-vm-neutron/8c49214/logs/screen-n-cpu.txt.gz?level=TRACE | 23:10 |
| *** dkliban has quit IRC | 23:10 | |
| flashgordon | what is that stacktrace of recurssion hell | 23:11 |
| flashgordon | that==bad | 23:11 |
| salv-orlando | flashgordon: I pushed a revert anyway… too late sorry :( | 23:11 |
| *** jgrimm has quit IRC | 23:11 | |
| *** layer427expert has quit IRC | 23:11 | |
| *** layer427expert has joined #openstack-neutron | 23:12 | |
| anttteaya | well the reversion hasn't merged | 23:12 |
| anttteaya | I don't think I have much more left in me tonight either | 23:12 |
| salv-orlando | flashgordon: it seems the recursion starts in _get_instance_nw_info, do you agree on that? | 23:14 |
| flashgordon | salv-orlando: https://review.openstack.org/#/c/56714/ | 23:15 |
| *** suresh12 has quit IRC | 23:17 | |
| *** suresh12 has joined #openstack-neutron | 23:18 | |
| *** suresh12 has quit IRC | 23:19 | |
| *** suresh12 has joined #openstack-neutron | 23:19 | |
| *** mlavalle has quit IRC | 23:21 | |
| salv-orlando | this seems more like it. | 23:22 |
| armax | I saw this one in the context of bug #1235435 | 23:23 |
| armax | see this one: https://review.openstack.org/#/c/56504/ | 23:23 |
| *** armax has left #openstack-neutron | 23:25 | |
| openstackgerrit | A change was merged to openstack/neutron: Imported Translations from Transifex https://review.openstack.org/54680 | 23:26 |
| anttteaya | I'm off, thanks for a great day everybody | 23:26 |
| anttteaya | night | 23:26 |
| markmcclain | night | 23:26 |
| salv-orlando | yeah I agree with armax, we are definitely seeing the infinite loop starting in get_instance_nw_info | 23:28 |
| *** gatuus has joined #openstack-neutron | 23:32 | |
| *** julim has quit IRC | 23:33 | |
| *** gatuus has quit IRC | 23:39 | |
| flashgordon | salv-orlando: so for https://review.openstack.org/#/c/56504/1 | 23:42 |
| flashgordon | do you think its the right fix? | 23:43 |
| flashgordon | because it looks prommising | 23:43 |
| *** carl_baldwin has quit IRC | 23:43 | |
| salv-orlando | I think so, but I don't know a lot about how nova caches network info. What armax says is true, there's a chance of infinite recursion, which might have been triggered by the removal of refresh_cache decorator | 23:44 |
| salv-orlando | idk if with this change we'll make performance worse by basically not caching anymore, but still better than not working at all | 23:45 |
| flashgordon | salv-orlando: it did pass the large-ops test :) | 23:47 |
| salv-orlando | flashgordon: it seems 56504 is still running the largeops test: https://jenkins01.openstack.org/job/gate-tempest-devstack-vm-neutron-large-ops/6347/ | 23:49 |
| flashgordon | salv-orlando: I ran a recheck | 23:49 |
| flashgordon | so yes it should be | 23:49 |
| salv-orlando | ok, let's wait for another run and then if all the jobs pass we can merge it | 23:50 |
| flashgordon | yup | 23:50 |
| flashgordon | I'll +2 it if it works and get another from somewhere | 23:51 |
| *** gatuus has joined #openstack-neutron | 23:54 | |
| salv-orlando | looking at devstack-gate and https://review.openstack.org/#/c/56465/ it seems we've just enabled parallel testing for neutron too. So this will be fun. | 23:55 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!