| dims__ | jlvillal: do not look related in any way | 00:00 |
|---|---|---|
| mrda | hmmm | 00:00 |
| *** salv-orl_ has quit IRC | 00:00 | |
| jlvillal | mrda: Do you want to pursue this with lintan? | 00:00 |
| jlvillal | If he is working today. | 00:01 |
| jlvillal | mrda: I think he is. He worked yesterday (Monday his time zone) | 00:01 |
| *** dannywilson has quit IRC | 00:01 | |
| mrda | jlvillal: I can try, although I am a little predisposed with other urgent | 00:01 |
| dansmith | lemme try something | 00:01 |
| dansmith | dims__: you haven't reverted the release yet have you? | 00:02 |
| dims__ | dansmith: nope | 00:03 |
| dims__ | dansmith: jlvillal: recreated the problem in my travis environment (https://api.travis-ci.org/jobs/77079815/log.txt?deansi=true) had to do this (https://github.com/dims/ironic/commit/1239dada23179d68ebd77ff0ea36e50a76dda7a6) | 00:04 |
| *** rushil has quit IRC | 00:05 | |
| dansmith | dims__: ? | 00:05 |
| dansmith | dims__: that just means you get the new release, right? | 00:05 |
| mikal | What time does sdague knock off for the day? Is he ever still around at this sort of time? | 00:05 |
| dims__ | dansmith: yes, i was trying to figure out why i had not seen the failure in my travis env | 00:06 |
| *** emagana has quit IRC | 00:06 | |
| dansmith | okay | 00:06 |
| *** lucas-dinner has quit IRC | 00:06 | |
| jlvillal | dims__: nova seems to do the same thing with install_command. Or at least similar. | 00:06 |
| dims__ | jlvillal: yes, i had to remove the one in nova too | 00:06 |
| jlvillal | Oh. | 00:07 |
| dims__ | https://github.com/dims/nova/commit/b5483a71c96ab7e53e67a8aed76bb743b1ac11f9#diff-b91f3d5bd63fcd17221b267e851608e8 | 00:08 |
| *** ZZll-off has quit IRC | 00:08 | |
| *** tpeoples has quit IRC | 00:09 | |
| dims__ | when i use "-e git+https://github.com/openstack/oslo.versionedobjects.git#egg=oslo.versionedobjects" in requirements.txt that install_command screws things up | 00:09 |
| dims__ | so apologies jlvillal | 00:09 |
| jlvillal | dims__: Thank you. What do you think the plan should be. | 00:10 |
| dims__ | jlvillal: checking https://review.openstack.org/#/c/216470/ | 00:10 |
| jlvillal | dims__: I'm afraid I don't know anything about that code :( | 00:10 |
| dims__ | jlvillal: if that works then should see if there's any other jobs that fail | 00:10 |
| dansmith | dims__: jlvillal: yeah, this is going to be a lot of work | 00:10 |
| jlvillal | dims__: I have sent lintan an email asking him to look into it. | 00:10 |
| dansmith | you're still using old fields, which means you can't use the new registry until you do | 00:11 |
| dansmith | just to be clear, | 00:11 |
| *** garyk has joined #openstack-nova | 00:11 | |
| dansmith | this isn't something we broke in 0.8.0, but rather a bunch of screwups in the ironic objects code that are _just_now_ being sensitive to something we changed underneath | 00:11 |
| jlvillal | dansmith: Okay. Should the release be reverted until Ironic can fix it? Reverted like a change to globalrequirements? | 00:11 |
| dansmith | we just dipped into the registry in a new place, which is broken | 00:11 |
| dansmith | jlvillal: well, nova and cinder have been waiting for this release for a while now | 00:12 |
| dims__ | mikal: occasionally, sdague drops off around 5:30-6:00 i think | 00:12 |
| dansmith | jlvillal: both need it to accomplish things for liberty | 00:12 |
| jlvillal | Hmmm. Not sounding good for Ironic :( | 00:12 |
| dansmith | the thing that I suspect is causing this is a critical fix for proper backporting of things too, | 00:12 |
| dansmith | so it's not like removing it and re-releasing is a good idea either | 00:12 |
| dansmith | jlvillal: https://review.openstack.org/#/q/project:openstack/nova+topic:bp/use-oslo-objects,n,z | 00:13 |
| dansmith | that's the work I did to get us from nova's initial implementation to where we are now | 00:13 |
| jlvillal | dansmith: Thanks. Looking. | 00:14 |
| dansmith | jlvillal: honestly, the easier thing would probably be to revert the bit that | 00:14 |
| dansmith | caused ironic to start using the listbase object from o.vo | 00:14 |
| jlvillal | Okay. I'll try to chat with lintan later on and see if he can work on it. I'll look into it too. | 00:15 |
| jlvillal | dansmith: But since there isn't smoke in the air, I was thinking about leaving work ;) | 00:15 |
| dansmith | jlvillal: well, isn't this breaking ironic gate? | 00:16 |
| jlvillal | dansmith: Yes it is. And unfortunately not sure how many cores are around. | 00:16 |
| jlvillal | I'll plug away and see what I can figure out. | 00:16 |
| *** tjones has joined #openstack-nova | 00:16 | |
| jlvillal | I've never looked at this code before today. | 00:16 |
| dims__ | dansmith: i am trying with removing just the 2 failing tests and see if the rest of the jobs are ok. So far i have not seen evidence that the other jobs fail yet | 00:17 |
| jlvillal | dansmith, dims__ : Appreciate all the help | 00:17 |
| dansmith | dims__: I can promise you that it's broken for real because of the to_primitive fail | 00:17 |
| jlvillal | dims__: That might be the quick fix to unbreak the gate. | 00:17 |
| jlvillal | :( | 00:17 |
| dims__ | dansmith: ack | 00:17 |
| dims__ | dansmith: let me put it another way, if those 2 tests weren't there, we would not even have noticed it :) | 00:18 |
| *** tjones has left #openstack-nova | 00:19 | |
| dansmith | heh | 00:19 |
| jlvillal | dansmith: dims__ So what if we submitted a patch to global-requirements to say don't use 0.8.0 ? | 00:19 |
| jlvillal | That would give us some time to fix our issues. | 00:19 |
| dansmith | jlvillal: doesn't help nova much | 00:19 |
| dims__ | jlvillal: that will terribly constrain nova and cinder :( | 00:19 |
| dansmith | man, there is something weird here too | 00:20 |
| dansmith | objects.fields depends on objects.base :( | 00:20 |
| * jlvillal wonders if ironic only could not use 0.8.0 somehow. | 00:20 | |
| dims__ | jlvillal: they have been waiting 2 weeks for this release | 00:20 |
| jlvillal | dims__: Understood. It just seems like Ironic is impacted badly | 00:21 |
| jlvillal | devananda: ^^^^ If you are around. | 00:21 |
| *** tellesnobrega_ has joined #openstack-nova | 00:22 | |
| *** tellesnobrega_ has quit IRC | 00:23 | |
| *** tellesnobrega_ has joined #openstack-nova | 00:23 | |
| *** tellesnobrega_ has quit IRC | 00:24 | |
| dansmith | okay, I think I got a fix, hang on | 00:26 |
| dansmith | mrda: dims__: jlvillal: https://review.openstack.org/#/c/216472/ | 00:27 |
| jlvillal | dansmith: Thanks! | 00:27 |
| dansmith | passes unit tests for me locally | 00:27 |
| dims__ | dansmith: will this work with both older and newer o.vo? | 00:27 |
| *** asettle-afk is now known as asettle | 00:28 | |
| dims__ | (curious) | 00:28 |
| dansmith | dims__: yeah, this just avoids using some code from o.vo altogether, so should be fine | 00:28 |
| dims__ | aha | 00:28 |
| dansmith | dims__: the problem is the original patch uses code from o.vo that they can't be using without having converted to the new registry | 00:28 |
| *** ccarmack has joined #openstack-nova | 00:28 | |
| dims__ | ack. sounds like the right thing to do | 00:29 |
| dansmith | yeah, I think so | 00:29 |
| jlvillal | dansmith: Thanks. Also passes tests locally for me. | 00:29 |
| dansmith | do I get an ironic shirt in tokyo (if this makes it in)? | 00:29 |
| jlvillal | I hope so :) I wish I had an Ironic shirt.... | 00:30 |
| dansmith | haha | 00:30 |
| mrda | dansmith: sounds fine by me :) | 00:30 |
| dansmith | well, if anyone is listening, I'm a L :P | 00:30 |
| *** mtanino has quit IRC | 00:30 | |
| dims__ | haha | 00:32 |
| *** zul has joined #openstack-nova | 00:34 | |
| *** thorst has quit IRC | 00:36 | |
| eliqiao | hi ,Good eveining cores, can you kindly help to check this https://review.openstack.org/#/c/216123/ patch, I think this need to be merged soon because we can not delete instance in some cases. | 00:38 |
| *** baoli has joined #openstack-nova | 00:38 | |
| *** shoutm_ has joined #openstack-nova | 00:43 | |
| *** zhenguo has joined #openstack-nova | 00:43 | |
| *** shoutm has quit IRC | 00:45 | |
| gmann | alex_xu: johnthetubaguy : regarding .ini tempest failure. with current https://review.openstack.org/#/c/215436/6 we do not strip out extra param | 00:46 |
| *** BadCub has quit IRC | 00:46 | |
| *** takedakn has quit IRC | 00:47 | |
| gmann | so extra param goes to v21 and raise NotImplemented error instead of 400 | 00:47 |
| *** BadCub has joined #openstack-nova | 00:48 | |
| *** tjones has joined #openstack-nova | 00:50 | |
| *** tjones has left #openstack-nova | 00:50 | |
| *** liusheng has quit IRC | 00:54 | |
| *** Zhang has joined #openstack-nova | 00:58 | |
| *** salv-orlando has joined #openstack-nova | 01:00 | |
| alex_xu | gmann: which failure return NotImplemented? | 01:01 |
| gmann | alex_xu: update host with extra param | 01:03 |
| *** shoutm_ has quit IRC | 01:03 | |
| gmann | alex_xu: actually in current https://review.openstack.org/#/c/215436/6 we are not strip out extra param from request. we only handle patterproty thing there | 01:04 |
| *** yanxb_ has quit IRC | 01:04 | |
| *** yanxb has joined #openstack-nova | 01:04 | |
| gmann | alex_xu: right? | 01:04 |
| *** yanxb_ has joined #openstack-nova | 01:04 | |
| *** shaohe_feng has joined #openstack-nova | 01:04 | |
| alex_xu | gmann: yea | 01:04 |
| *** yanxb__ has quit IRC | 01:04 | |
| alex_xu | gmann: maybe we should discuss this in the meeting today | 01:04 |
| gmann | alex_xu: i think your logic to delete extra param from instance was correct | 01:04 |
| gmann | alex_xu: i remember our discussion to strip out those but may be i missed some discussion? | 01:05 |
| alex_xu | gmann: there is another discussion between me and johnthetubaguy, let me find it | 01:05 |
| gmann | alex_xu: ohhk | 01:05 |
| gmann | on irc? | 01:06 |
| alex_xu | gmann: here http://eavesdrop.openstack.org/irclogs/%23openstack-nova/%23openstack-nova.2015-08-21.log.html#t2015-08-21T09:22:59 | 01:06 |
| garyk | sdague: would it be possible that you can please take a look at https://review.openstack.org/#/c/212786/ | 01:06 |
| *** browne has quit IRC | 01:07 | |
| gmann | alex_xu: ahh i see. | 01:07 |
| *** shoutm has joined #openstack-nova | 01:08 | |
| gmann | alex_xu: actually if we do not strip out extra param then, we fall back to situation where 400 gets converted to 5xx or something | 01:09 |
| gmann | alex_xu: we can discuss that in today meeting | 01:09 |
| alex_xu | gmann: yea, I also have such concern | 01:09 |
| alex_xu | gmann: let talk about that with john again in the meeting, then let us reach agreement | 01:09 |
| *** xyang2 has joined #openstack-nova | 01:10 | |
| gmann | alex_xu: sure | 01:11 |
| *** xyang2 has quit IRC | 01:12 | |
| *** 64MADOHP4 has joined #openstack-nova | 01:12 | |
| *** wanghua has joined #openstack-nova | 01:14 | |
| *** ssurana has quit IRC | 01:18 | |
| *** shaohe_feng1 has joined #openstack-nova | 01:20 | |
| *** liusheng has joined #openstack-nova | 01:20 | |
| *** dprince has quit IRC | 01:21 | |
| *** shaohe_feng has quit IRC | 01:22 | |
| *** 64MADOHP4 has quit IRC | 01:24 | |
| *** Apoorva has quit IRC | 01:25 | |
| *** salv-orlando has quit IRC | 01:30 | |
| *** salv-orlando has joined #openstack-nova | 01:31 | |
| *** oomichi has joined #openstack-nova | 01:33 | |
| *** asd112z has joined #openstack-nova | 01:35 | |
| *** salv-orlando has quit IRC | 01:36 | |
| *** mdrabe has joined #openstack-nova | 01:38 | |
| *** tpeoples has joined #openstack-nova | 01:40 | |
| *** shaohe_feng has joined #openstack-nova | 01:41 | |
| *** asettle is now known as asettle-gym | 01:42 | |
| *** shaohe_feng1 has quit IRC | 01:43 | |
| *** tsekiyam_ has joined #openstack-nova | 01:47 | |
| *** rotbeard has quit IRC | 01:48 | |
| *** tsekiyama has quit IRC | 01:50 | |
| *** tsekiyam_ has quit IRC | 01:51 | |
| *** dims__ has quit IRC | 01:55 | |
| *** rotbeard has joined #openstack-nova | 01:56 | |
| *** dave-mccowan has quit IRC | 01:56 | |
| *** humble_ has joined #openstack-nova | 01:56 | |
| *** cfarquhar has quit IRC | 01:57 | |
| *** wanghua has quit IRC | 02:00 | |
| *** shoutm_ has joined #openstack-nova | 02:01 | |
| *** nic has quit IRC | 02:02 | |
| *** shoutm has quit IRC | 02:03 | |
| *** haomaiwang has joined #openstack-nova | 02:05 | |
| openstackgerrit | Michael Still proposed openstack/nova: Don't allow instance to overcommit against itself https://review.openstack.org/215764 | 02:05 |
| *** lyanchih has joined #openstack-nova | 02:06 | |
| openstackgerrit | Ian Wienand proposed openstack/nova: Add [pci] group and whitelist_sections to configure PCI passthrough https://review.openstack.org/133138 | 02:10 |
| *** achanda has joined #openstack-nova | 02:11 | |
| openstackgerrit | Ian Wienand proposed openstack/nova: Add [pci] group and whitelist_sections to configure PCI passthrough https://review.openstack.org/133138 | 02:12 |
| *** obondarev_ has quit IRC | 02:12 | |
| *** obondarev has joined #openstack-nova | 02:13 | |
| *** Marga_ has joined #openstack-nova | 02:17 | |
| *** baoli has quit IRC | 02:18 | |
| *** ijw has quit IRC | 02:19 | |
| *** vilobhmm11 has quit IRC | 02:19 | |
| *** ijw has joined #openstack-nova | 02:19 | |
| *** xiaohui has quit IRC | 02:19 | |
| *** zhihai_s has joined #openstack-nova | 02:19 | |
| *** xiaohui has joined #openstack-nova | 02:19 | |
| *** jcoufal has quit IRC | 02:20 | |
| *** baoli has joined #openstack-nova | 02:21 | |
| *** ijw has quit IRC | 02:21 | |
| *** jwang has joined #openstack-nova | 02:21 | |
| openstackgerrit | Ian Wienand proposed openstack/nova: Add [pci] group and whitelist_sections to configure PCI passthrough https://review.openstack.org/133138 | 02:21 |
| *** ijw has joined #openstack-nova | 02:21 | |
| *** ijw has quit IRC | 02:23 | |
| *** ijw has joined #openstack-nova | 02:23 | |
| *** hemna has joined #openstack-nova | 02:34 | |
| *** angdraug has quit IRC | 02:34 | |
| *** bkopilov has quit IRC | 02:34 | |
| *** vladikr has quit IRC | 02:35 | |
| openstackgerrit | Rui Chen proposed openstack/nova-specs: Split network plane for live migration https://review.openstack.org/194990 | 02:35 |
| *** mdrabe has quit IRC | 02:41 | |
| *** venkat_p has joined #openstack-nova | 02:44 | |
| *** liusheng has quit IRC | 02:44 | |
| *** liusheng has joined #openstack-nova | 02:45 | |
| *** browne has joined #openstack-nova | 02:49 | |
| *** achanda has quit IRC | 02:50 | |
| *** vilobhmm1 has joined #openstack-nova | 02:50 | |
| *** hakimo_ has joined #openstack-nova | 02:52 | |
| *** hakimo has quit IRC | 02:55 | |
| *** shuquan has joined #openstack-nova | 02:55 | |
| *** vilobhmm1 has quit IRC | 02:57 | |
| *** vilobhmm1 has joined #openstack-nova | 02:58 | |
| *** haomaiwang has quit IRC | 03:01 | |
| *** achanda has joined #openstack-nova | 03:01 | |
| *** haomaiwang has joined #openstack-nova | 03:02 | |
| *** vilobhmm1 has quit IRC | 03:02 | |
| *** garyk has quit IRC | 03:05 | |
| *** unicell has quit IRC | 03:08 | |
| *** asettle-gym is now known as asettle | 03:08 | |
| *** rubasov has quit IRC | 03:10 | |
| *** shuquan has quit IRC | 03:21 | |
| *** links has joined #openstack-nova | 03:24 | |
| *** shuquan has joined #openstack-nova | 03:24 | |
| *** zhihai_song has joined #openstack-nova | 03:25 | |
| *** unicell has joined #openstack-nova | 03:27 | |
| *** zhihai_s has quit IRC | 03:29 | |
| *** VW_ has joined #openstack-nova | 03:30 | |
| *** asd112z has quit IRC | 03:31 | |
| *** shuquan has quit IRC | 03:31 | |
| *** VW_ has quit IRC | 03:31 | |
| *** hemna has quit IRC | 03:31 | |
| *** VW_ has joined #openstack-nova | 03:32 | |
| *** davideagnello has quit IRC | 03:33 | |
| *** shuquan has joined #openstack-nova | 03:34 | |
| *** diogogmt has quit IRC | 03:34 | |
| *** shuquan has quit IRC | 03:38 | |
| *** ssurana has joined #openstack-nova | 03:39 | |
| *** humble__ has joined #openstack-nova | 03:40 | |
| *** deepthi has joined #openstack-nova | 03:42 | |
| *** ijw has quit IRC | 03:42 | |
| *** ijw has joined #openstack-nova | 03:43 | |
| *** humble_ has quit IRC | 03:44 | |
| *** shoutm_ has quit IRC | 03:48 | |
| *** shuquan has joined #openstack-nova | 03:49 | |
| *** shoutm has joined #openstack-nova | 03:59 | |
| *** haomaiwang has quit IRC | 04:01 | |
| *** 17WAAILMU has joined #openstack-nova | 04:01 | |
| *** lyanchih has quit IRC | 04:03 | |
| *** VikasC has joined #openstack-nova | 04:14 | |
| *** VikasC has quit IRC | 04:18 | |
| *** tojuvone has joined #openstack-nova | 04:18 | |
| *** humble_ has joined #openstack-nova | 04:18 | |
| *** bkopilov has joined #openstack-nova | 04:18 | |
| *** shuquan has quit IRC | 04:19 | |
| *** humble__ has quit IRC | 04:20 | |
| *** garyk has joined #openstack-nova | 04:22 | |
| *** sudipto has joined #openstack-nova | 04:23 | |
| *** zh_song has joined #openstack-nova | 04:23 | |
| *** shuquan has joined #openstack-nova | 04:24 | |
| *** garyk has quit IRC | 04:24 | |
| *** garyk has joined #openstack-nova | 04:25 | |
| *** zhihai_song has quit IRC | 04:26 | |
| *** vilobhmm1 has joined #openstack-nova | 04:32 | |
| openstackgerrit | Zhenguo Niu proposed openstack/nova: [Ironic]Not count available resources of deployed ironic node https://review.openstack.org/204088 | 04:34 |
| *** baoli has quit IRC | 04:35 | |
| *** abhirc has quit IRC | 04:37 | |
| *** abhirc has joined #openstack-nova | 04:38 | |
| *** shuquan has quit IRC | 04:39 | |
| *** emagana has joined #openstack-nova | 04:40 | |
| *** bnemec has quit IRC | 04:42 | |
| *** Sukhdev has joined #openstack-nova | 04:42 | |
| *** abhirc has quit IRC | 04:42 | |
| *** Zhang has quit IRC | 04:47 | |
| *** abhirc has joined #openstack-nova | 04:47 | |
| *** shyama has joined #openstack-nova | 04:47 | |
| *** Zhang has joined #openstack-nova | 04:48 | |
| *** abhirc has quit IRC | 04:48 | |
| *** abhirc has joined #openstack-nova | 04:48 | |
| *** lyanchih has joined #openstack-nova | 04:49 | |
| *** Zhang has quit IRC | 04:51 | |
| *** Zhang has joined #openstack-nova | 04:51 | |
| *** sudipto has quit IRC | 04:52 | |
| *** unicell1 has joined #openstack-nova | 04:55 | |
| *** unicell has quit IRC | 04:56 | |
| *** tjones has joined #openstack-nova | 04:57 | |
| *** fawadkhaliq has joined #openstack-nova | 04:57 | |
| *** ildikov has quit IRC | 04:59 | |
| *** lyanchih has quit IRC | 04:59 | |
| *** shyama has quit IRC | 05:00 | |
| *** 17WAAILMU has quit IRC | 05:01 | |
| *** haomaiwang has joined #openstack-nova | 05:02 | |
| *** abhirc has quit IRC | 05:04 | |
| *** tjones has quit IRC | 05:07 | |
| *** tjones has joined #openstack-nova | 05:07 | |
| *** tjones has left #openstack-nova | 05:07 | |
| *** liusheng has quit IRC | 05:09 | |
| *** liusheng has joined #openstack-nova | 05:09 | |
| *** VikasC has joined #openstack-nova | 05:15 | |
| *** nkrinner has joined #openstack-nova | 05:16 | |
| *** hemna has joined #openstack-nova | 05:17 | |
| *** VW_ has quit IRC | 05:17 | |
| *** VW_ has joined #openstack-nova | 05:18 | |
| *** VW_ has quit IRC | 05:18 | |
| *** VW_ has joined #openstack-nova | 05:18 | |
| *** VikasC has quit IRC | 05:20 | |
| *** hemna has quit IRC | 05:22 | |
| *** dannywilson has joined #openstack-nova | 05:23 | |
| *** asd112z has joined #openstack-nova | 05:32 | |
| *** dannywilson has quit IRC | 05:34 | |
| *** dannywilson has joined #openstack-nova | 05:35 | |
| *** davideagnello has joined #openstack-nova | 05:36 | |
| *** Sree has joined #openstack-nova | 05:39 | |
| *** Sree has quit IRC | 05:40 | |
| *** ildikov has joined #openstack-nova | 05:40 | |
| *** Sree has joined #openstack-nova | 05:40 | |
| *** Sree has quit IRC | 05:40 | |
| *** Sree has joined #openstack-nova | 05:41 | |
| *** VW_ has quit IRC | 05:41 | |
| *** vilobhmm1 has quit IRC | 05:41 | |
| *** eliqiao has quit IRC | 05:42 | |
| *** eliqiao has joined #openstack-nova | 05:47 | |
| *** deepakcs has joined #openstack-nova | 05:48 | |
| *** ajayaa has joined #openstack-nova | 05:48 | |
| *** asd112z has quit IRC | 05:51 | |
| *** zhenguo has quit IRC | 05:53 | |
| *** jichen has joined #openstack-nova | 05:55 | |
| *** shuquan has joined #openstack-nova | 05:57 | |
| *** links has quit IRC | 05:57 | |
| *** exploreshaifali has joined #openstack-nova | 05:58 | |
| *** sudipto has joined #openstack-nova | 05:59 | |
| *** shuquan has quit IRC | 05:59 | |
| *** lyanchih has joined #openstack-nova | 06:00 | |
| *** shaohe_feng has quit IRC | 06:00 | |
| *** links has joined #openstack-nova | 06:00 | |
| *** moshele has joined #openstack-nova | 06:00 | |
| *** haomaiwang has quit IRC | 06:01 | |
| *** haomaiwang has joined #openstack-nova | 06:01 | |
| *** shuquan has joined #openstack-nova | 06:05 | |
| *** baoli has joined #openstack-nova | 06:05 | |
| *** zhenguo has joined #openstack-nova | 06:09 | |
| *** irenab has quit IRC | 06:11 | |
| *** BadCub has quit IRC | 06:12 | |
| *** boris-42 has quit IRC | 06:12 | |
| *** serverascode has quit IRC | 06:12 | |
| *** ramishra has quit IRC | 06:12 | |
| *** DuncanT has quit IRC | 06:12 | |
| *** patrickeast has quit IRC | 06:12 | |
| *** vivekd has joined #openstack-nova | 06:13 | |
| *** BadCub has joined #openstack-nova | 06:14 | |
| *** serverascode has joined #openstack-nova | 06:14 | |
| *** DuncanT has joined #openstack-nova | 06:15 | |
| *** ramishra has joined #openstack-nova | 06:15 | |
| *** vilobhmm1 has joined #openstack-nova | 06:16 | |
| *** armax has quit IRC | 06:16 | |
| *** asettle has quit IRC | 06:17 | |
| *** patrickeast has joined #openstack-nova | 06:17 | |
| *** boris-42 has joined #openstack-nova | 06:17 | |
| openstackgerrit | IWAMOTO Toshihiro proposed openstack/nova: Omit RescheduledException in instance_fault.message https://review.openstack.org/165751 | 06:19 |
| *** baoli has quit IRC | 06:20 | |
| *** Murali has joined #openstack-nova | 06:21 | |
| *** shuquan has quit IRC | 06:22 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Imported Translations from Transifex https://review.openstack.org/212367 | 06:23 |
| *** unicell1 has quit IRC | 06:24 | |
| *** unicell has joined #openstack-nova | 06:25 | |
| *** shuquan has joined #openstack-nova | 06:30 | |
| *** shaohe_feng has joined #openstack-nova | 06:31 | |
| *** unicell has quit IRC | 06:32 | |
| *** unicell1 has joined #openstack-nova | 06:32 | |
| openstackgerrit | Deepak C Shetty (deepakcs) proposed openstack/nova: libvirt: Fix snapshot delete for network disk type for blockRebase op https://review.openstack.org/212518 | 06:32 |
| *** sudipto has quit IRC | 06:32 | |
| *** Sukhdev has quit IRC | 06:34 | |
| *** fawadk has joined #openstack-nova | 06:34 | |
| *** fawadkhaliq has quit IRC | 06:35 | |
| *** irenab has joined #openstack-nova | 06:38 | |
| *** exploreshaifali has quit IRC | 06:40 | |
| *** SourabhP has joined #openstack-nova | 06:40 | |
| *** SourabhP_ has joined #openstack-nova | 06:41 | |
| *** lykinsbd has quit IRC | 06:42 | |
| *** lykinsbd has joined #openstack-nova | 06:43 | |
| *** evgenyf has joined #openstack-nova | 06:44 | |
| *** SourabhP has quit IRC | 06:45 | |
| *** SourabhP_ is now known as SourabhP | 06:45 | |
| *** rubasov has joined #openstack-nova | 06:46 | |
| *** sudipto has joined #openstack-nova | 06:48 | |
| *** shoutm has quit IRC | 06:49 | |
| *** vivekd_ has joined #openstack-nova | 06:50 | |
| *** shoutm has joined #openstack-nova | 06:51 | |
| *** vivekd__ has joined #openstack-nova | 06:51 | |
| *** SourabhP has quit IRC | 06:52 | |
| *** SourabhP has joined #openstack-nova | 06:53 | |
| *** vivekd has quit IRC | 06:53 | |
| *** vivekd__ is now known as vivekd | 06:53 | |
| *** alexschm has joined #openstack-nova | 06:53 | |
| *** ankit_ag has joined #openstack-nova | 06:53 | |
| *** vivekd_ has quit IRC | 06:55 | |
| *** shaohe_feng has quit IRC | 06:56 | |
| *** davideagnello has quit IRC | 06:57 | |
| *** vilobhmm1 has quit IRC | 06:58 | |
| *** garyk has quit IRC | 06:58 | |
| *** garyk has joined #openstack-nova | 06:59 | |
| *** haomaiwang has quit IRC | 07:01 | |
| *** haomaiwang has joined #openstack-nova | 07:02 | |
| *** SourabhP has quit IRC | 07:02 | |
| *** nkrinner has quit IRC | 07:05 | |
| *** Sree has quit IRC | 07:07 | |
| *** Sree has joined #openstack-nova | 07:08 | |
| *** nkrinner has joined #openstack-nova | 07:09 | |
| *** achanda has quit IRC | 07:10 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/216568 | 07:10 |
| *** Sree has quit IRC | 07:12 | |
| *** evgenyf has quit IRC | 07:13 | |
| *** achanda has joined #openstack-nova | 07:14 | |
| gmann | alex_xu: hi | 07:14 |
| alex_xu | gmann: hi | 07:14 |
| *** rajesht has joined #openstack-nova | 07:14 | |
| gmann | alex_xu: I just updated api meeting agend | 07:15 |
| gmann | alex_xu: mainly added patches up for making v21 default on gate | 07:15 |
| alex_xu | gmann: cool, thanks | 07:15 |
| alex_xu | gmann: that is worth move under the v2.0 on v2.1 section | 07:16 |
| gmann | alex_xu: good idea. ll do that | 07:16 |
| alex_xu | thanks | 07:16 |
| rajesht | hi cores, | 07:17 |
| rajesht | could anybody please review it | 07:17 |
| rajesht | https://review.openstack.org/#/c/170118/ | 07:17 |
| rajesht | thanks | 07:17 |
| *** sahid has joined #openstack-nova | 07:17 | |
| gmann | alex_xu: done added under API infrastructure | 07:18 |
| alex_xu | gmann: thanks | 07:18 |
| *** browne has quit IRC | 07:21 | |
| openstackgerrit | Rui Chen proposed openstack/nova-specs: Task states improvement https://review.openstack.org/216578 | 07:26 |
| openstackgerrit | Simona Iuliana Toader proposed openstack/nova: Hyper-V: Adds Hyper-V UEFI Secure Boot https://review.openstack.org/209581 | 07:26 |
| openstackgerrit | Simona Iuliana Toader proposed openstack/nova: Hyper-V: refactors utilsfactory https://review.openstack.org/214527 | 07:26 |
| *** humble__ has joined #openstack-nova | 07:28 | |
| *** humble_ has quit IRC | 07:30 | |
| *** achanda has quit IRC | 07:30 | |
| *** ociuhandu has quit IRC | 07:35 | |
| *** shaohe_feng has joined #openstack-nova | 07:35 | |
| *** Nikolay_St has joined #openstack-nova | 07:38 | |
| *** ndipanov has quit IRC | 07:38 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/216568 | 07:38 |
| *** pbrooko has joined #openstack-nova | 07:40 | |
| *** lyanchih has quit IRC | 07:40 | |
| openstackgerrit | Rui Chen proposed openstack/nova: Set task_state=None when booting instance failed https://review.openstack.org/187145 | 07:41 |
| *** matrohon has joined #openstack-nova | 07:42 | |
| *** fawadk has quit IRC | 07:43 | |
| *** hparekh has joined #openstack-nova | 07:44 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: add support for muti_statuses_query from migration https://review.openstack.org/169631 | 07:46 |
| *** salv-orlando has joined #openstack-nova | 07:47 | |
| *** lyanchih has joined #openstack-nova | 07:48 | |
| *** ajayaa has quit IRC | 07:50 | |
| *** ajayaa has joined #openstack-nova | 07:51 | |
| *** lpetrut has joined #openstack-nova | 07:51 | |
| *** IanGovett has joined #openstack-nova | 07:52 | |
| *** bkopilov has quit IRC | 07:54 | |
| *** alex_klimov has joined #openstack-nova | 07:54 | |
| *** summerWendy has joined #openstack-nova | 07:56 | |
| *** baoli has joined #openstack-nova | 07:58 | |
| *** Zhang has quit IRC | 08:00 | |
| *** haomaiwang has quit IRC | 08:01 | |
| *** haomaiwang has joined #openstack-nova | 08:01 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: Add project-id when list server-groups https://review.openstack.org/209018 | 08:02 |
| *** baoli has quit IRC | 08:02 | |
| *** xgsa has joined #openstack-nova | 08:03 | |
| *** shoutm has quit IRC | 08:03 | |
| *** asd112z has joined #openstack-nova | 08:03 | |
| *** shoutm has joined #openstack-nova | 08:05 | |
| *** oomichi has quit IRC | 08:06 | |
| *** zhihai_song has joined #openstack-nova | 08:06 | |
| *** ssurana has quit IRC | 08:07 | |
| *** asd112z has quit IRC | 08:08 | |
| *** zh_song has quit IRC | 08:09 | |
| *** exploreshaifali has joined #openstack-nova | 08:10 | |
| *** danpb has joined #openstack-nova | 08:12 | |
| *** jistr has joined #openstack-nova | 08:12 | |
| *** tpeoples has quit IRC | 08:13 | |
| *** Zhang has joined #openstack-nova | 08:14 | |
| *** Sree has joined #openstack-nova | 08:15 | |
| *** lpetrut has quit IRC | 08:17 | |
| johnthetubaguy | alex_xu: gmann: yeah, I wanted to try without stripping those out | 08:19 |
| *** deepthi has quit IRC | 08:19 | |
| *** shuquan has quit IRC | 08:20 | |
| johnthetubaguy | alex_xu: gmann: I thought we got a different tempest error when we stripped out all the params? | 08:20 |
| *** markus_z has joined #openstack-nova | 08:21 | |
| johnthetubaguy | alex_xu: gmann: my mistake, it almost looks like we get the same error whether we strip out the params or not | 08:22 |
| *** derekh has joined #openstack-nova | 08:24 | |
| *** romainh has joined #openstack-nova | 08:24 | |
| alex_xu | johnthetubaguy: without strip out, host update will return notimplement | 08:25 |
| alex_xu | johnthetubaguy: with strip out, host update will return 200/202 | 08:25 |
| *** karimb has joined #openstack-nova | 08:25 | |
| johnthetubaguy | alex_xu: I didn't see the tempest tests pass when we stripped out the extra bits though | 08:25 |
| alex_xu | johnthetubaguy: that tempest for host update still will failed, because it assert 404 | 08:26 |
| gmann | johnthetubaguy: alex_xu : with strip out 4xx will return 2xx | 08:26 |
| johnthetubaguy | OK, I am clearly looking at the wrong patchset then | 08:26 |
| alex_xu | johnthetubaguy: it's fine, I forget host update also when we discussion | 08:26 |
| gmann | johnthetubaguy: it passed - PS12 | 08:27 |
| alex_xu | johnthetubaguy: gmann so let us fall back to strip out extra params? | 08:27 |
| gmann | alex_xu: i think yes | 08:27 |
| johnthetubaguy | alex_xu: gmann: looks like we don't have an option, lets go back to stripping out the extras | 08:27 |
| alex_xu | ok, no problem, will update patch later | 08:28 |
| johnthetubaguy | alex_xu: gmann: I think we might need to relax that, for things like the scheduler, but lets deal with that on a case by case basis | 08:28 |
| gmann | johnthetubaguy: alex_xu : 4xx->2xx is fine right? | 08:28 |
| johnthetubaguy | gmann: its not good, but I can live with it, I think | 08:28 |
| alex_xu | gmann: at least better than 4xx->notimpelement... | 08:28 |
| gmann | johnthetubaguy: humm | 08:28 |
| gmann | alex_xu: yea | 08:29 |
| johnthetubaguy | gmann: alex_xu: I think this is worth an email to the mailing list, to explain the reasoning | 08:29 |
| johnthetubaguy | its the opposite of what was agreed during the spec review | 08:29 |
| gmann | johnthetubaguy: yea. do we need spec update also? | 08:30 |
| alex_xu | johnthetubaguy: yea | 08:31 |
| *** lucasagomes has joined #openstack-nova | 08:31 | |
| eliqiao | johnthetubaguy: alex_xu: gmann: regarding microversion changes, I am not sure if this change requires a microversion bump even I'v read the bump guide, can you help to check ? https://review.openstack.org/#/c/214434/ | 08:32 |
| *** shaohe_feng has quit IRC | 08:34 | |
| alex_xu | johnthetubaguy: eliqiao, about that patch, I just think there is big problem is about we have different API behavior between different hypervisor. libvirt will failed live-migration with bdm, but other hypervisor may not. | 08:34 |
| johnthetubaguy | eliqiao: are you sure it always fails? | 08:35 |
| johnthetubaguy | alex_xu: yeah, we need ways to discover these sorts of limitations, ideally | 08:35 |
| johnthetubaguy | eliqiao: I am just wondering if its true for all storage backends and all volume drivers? | 08:36 |
| eliqiao | johnthetubaguy: yes, libvirt dirver failed in that case. | 08:36 |
| * eliqiao checking code.. | 08:36 | |
| eliqiao | #link https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L6191 | 08:37 |
| *** evgenyf has joined #openstack-nova | 08:37 | |
| alex_xu | johnthetubaguy: eliqiao's patch change the API contract I think, before return 2xx, with his patch return fault. and if this is only libvirt driver, before we have way discover that, we shouldn't change the api contract | 08:38 |
| gmann | johnthetubaguy: alex_xu: eliqiao : even in such case we can catch migration error in API and raise respective error (for hypervisor does not raise this should be ok) | 08:38 |
| *** lyanchih has quit IRC | 08:38 | |
| eliqiao | this MigrationError will raise after calling _disconnect_volume called when doing roll_back | 08:38 |
| *** deepthi has joined #openstack-nova | 08:39 | |
| johnthetubaguy | eliqiao: in your patch, you should really consider removing the check you just linked to, or at least adding a comment saying its a redundant check | 08:39 |
| *** alex_klimov has quit IRC | 08:40 | |
| eliqiao | so, these is potential rask when doing roll back, _disconnect_volume may fail.. | 08:40 |
| johnthetubaguy | alex_xu: this feels OK to me, its just returning an existing error earlier, where we already return some errors | 08:40 |
| eliqiao | johnthetubaguy: yeah, agreed, my bad. | 08:40 |
| johnthetubaguy | alex_xu: now I also wish there was an easier way to discover these differences, and we list them, for me it makes block migration useless. | 08:41 |
| *** Sree_ has joined #openstack-nova | 08:41 | |
| *** haomaiwang has quit IRC | 08:41 | |
| alex_xu | johnthetubaguy: the original api didn't return some errors, it return 2xx, and rollback the instance back to active status. But there is migration recoder in os-migrations API for this migration failed | 08:41 |
| *** hop has joined #openstack-nova | 08:41 | |
| *** Sree_ is now known as Guest29053 | 08:42 | |
| johnthetubaguy | alex_xu: no, the original API already returns errors for some cases | 08:42 |
| *** haomaiwang has joined #openstack-nova | 08:42 | |
| *** zhihai_song has quit IRC | 08:42 | |
| johnthetubaguy | alex_xu: this is just an extra error case to trigger the existing error, I believe | 08:42 |
| alex_xu | johnthetubaguy: hummm | 08:43 |
| *** zhihai_s has joined #openstack-nova | 08:43 | |
| alex_xu | johnthetubaguy: the contract part isn't the return error, the contract is whether we allow live-migration with bdm. | 08:44 |
| alex_xu | johnthetubaguy: but we can return not impelement for this case? | 08:44 |
| eliqiao | johnthetubaguy: yeah, thanks for your comments on the patch, I think you are right, I need to check block_device_info and block_device_mapping | 08:44 |
| *** Sree has quit IRC | 08:44 | |
| johnthetubaguy | alex_xu: I thought the API WG said we should stop using the 501 return code? | 08:44 |
| *** vivekd has quit IRC | 08:45 | |
| alex_xu | johnthetubaguy: yea, sdague propose use 404 I remember at https://review.openstack.org/#/c/183456/, but it isn't merge yet...another confuse | 08:45 |
| johnthetubaguy | alex_xu: its a mute point thought, because it would just fail later in the process anyways, so its making the failure more user friendly, its hard to block that | 08:46 |
| *** dannywilson has quit IRC | 08:46 | |
| alex_xu | johnthetubaguy: emm... the mute point make sense | 08:46 |
| eliqiao | johnthetubaguy: my bad, I think my patch is correct, get the bdm from bdi, and it won't include swap and local disk. | 08:47 |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Switch to oslo.cache lib https://review.openstack.org/203049 | 08:47 |
| eliqiao | besides, failed early will bring more benifits, it avoid roll back | 08:48 |
| johnthetubaguy | alex_xu: agreed I don't like that difference, and it should be discoverable, the bit that worries me more is making clear that it doesn't work, so people know its something thats worth looking to fix (I managed to push hard enough to get that working with XenServer, in the end) | 08:48 |
| *** vivekd has joined #openstack-nova | 08:48 | |
| johnthetubaguy | eliqiao: agreed with you on the concept, my take is that it doesn't need an API microversion bump, but I often get those wrong, so I would like others to take a look | 08:48 |
| snikitin | dims, pkholkin: Dims, you asked "can you confirm that if [DEFAULT]/memcached_servers is not set then the configuration from oslo.cache will get picked up?" | 08:49 |
| openstackgerrit | John Garbutt proposed openstack/nova: libvirt: refactor _create_domain_setup_lxc to use Image.get_model https://review.openstack.org/215310 | 08:50 |
| snikitin | dims, pkholkin: I can't confirm it. oslo.cache conf will be ignored. Also in oslo.cache opts we have opt 'memcache_servers' and in nova we have 'memcacheD_servers' | 08:50 |
| alex_xu | johnthetubaguy: yea, the mute is not good | 08:51 |
| eliqiao | johnthetubaguy: alex_xu: hmm okay , thanks for your advice, seems we need further discuession | 08:51 |
| *** alex_klimov has joined #openstack-nova | 08:52 | |
| openstackgerrit | Alexis Lee proposed openstack/nova: Walkthrough from API to virt driver https://review.openstack.org/210467 | 08:52 |
| *** exploreshaifali has quit IRC | 08:52 | |
| snikitin | dims, pkholkin: Moreover opt 'memcache_servers' from oslo_cache is needed for oslo.cache.memcahe_pool backend. And in nova we use dogpile.cache.memcached backend | 08:53 |
| *** lyanchih has joined #openstack-nova | 08:54 | |
| gmann | eliqiao: alex_xu : even we fail on same in pre_live_migration. does it make rollback? | 08:55 |
| eliqiao | gmann: fail on pre_live_migration will do rollback. | 08:56 |
| eliqiao | gmann: #link https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L4983 | 08:57 |
| tdurakov | eliqiao, hi | 08:59 |
| eliqiao | tdurakov: hi | 08:59 |
| tdurakov | about rollback on pre-live-migration | 09:00 |
| tdurakov | deployed multinode devstack and try to l-m volume-backed instance | 09:00 |
| *** claudiub has joined #openstack-nova | 09:01 | |
| tdurakov | i've got error from glance, and when it happened no cleanup on destination happens | 09:02 |
| gmann | eliqiao: i see. Thanks. we should have some checking which can be moved earlier from pre_live_migration | 09:02 |
| gmann | eliqiao: anyways thats different thing we can do later | 09:02 |
| tdurakov | instance directory on target compute node doesn't removed | 09:02 |
| eliqiao | gmann: yeah, thanks for the remiding. good idea. | 09:03 |
| johnthetubaguy | bauzas: just wondering if you spotted some rebases being required on these patches: https://review.openstack.org/#/c/215471/3 | 09:03 |
| tdurakov | eliqiao, working on fix for that | 09:03 |
| eliqiao | tdurakov: any bug link ? | 09:03 |
| tdurakov | nope, just find that | 09:04 |
| tdurakov | https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L6147 | 09:04 |
| tdurakov | this raise happens | 09:04 |
| eliqiao | tdurakov: I think it's better to report a bug to track it. actually I didn't get your words too much. | 09:04 |
| *** ociuhandu has joined #openstack-nova | 09:04 | |
| eliqiao | tdurakov: why glance report an exception? | 09:05 |
| tdurakov | eliqiao, this is another story... | 09:05 |
| tdurakov | debugging that | 09:05 |
| eliqiao | tdurakov: ops... much more confused , it would be appreciate if you can report a bug , I'd love to try your step | 09:06 |
| tdurakov | eliqiao, sure, I'll file bug for that, it seems that live-migration cleanup_flags is the root cause | 09:07 |
| eliqiao | tdurakov: okay , cool thanks. | 09:07 |
| tdurakov | anyway, rollback on destination is overcomplecated | 09:08 |
| markus_z | tdurakov: Please post the link of that bug here when you reported it. | 09:08 |
| tdurakov | there is rpc messaging hell during live-migration | 09:08 |
| johnthetubaguy | bauzas: actually, have some questions about this one: https://review.openstack.org/#/c/215471/3 | 09:08 |
| *** fawadkhaliq has joined #openstack-nova | 09:09 | |
| markus_z | tdurakov: It's a pretty complicated think, yes. More live-migrate related bugs: https://bugs.launchpad.net/nova/+bugs?field.tag=live-migrate | 09:09 |
| *** haomaiwang has quit IRC | 09:09 | |
| *** haomaiwang has joined #openstack-nova | 09:10 | |
| eliqiao | tdurakov: agreed with markus_z: | 09:10 |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova-specs: Add project id information in os-server-group API https://review.openstack.org/209917 | 09:11 |
| markus_z | s/think/thing ... | 09:11 |
| *** humble_ has joined #openstack-nova | 09:12 | |
| *** VW_ has joined #openstack-nova | 09:12 | |
| tdurakov | complecated?) ok, simple example: if pre_live_migration fails on destination it returns fail to source which than sends request then to destination again.... | 09:15 |
| tdurakov | markus_z, eliqiao ^ | 09:15 |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: Add project-id when list server-groups https://review.openstack.org/209018 | 09:15 |
| *** humble__ has quit IRC | 09:15 | |
| johnthetubaguy | claudiub: if you get chance, could you please take a look at this hyper-v change: https://review.openstack.org/#/c/189302/23 | 09:16 |
| johnthetubaguy | claudiub: I am blind, now I see your comments on there | 09:16 |
| *** VW_ has quit IRC | 09:16 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: resize-revert can break anti-affinity policy https://review.openstack.org/160110 | 09:17 |
| claudiub | johnthetubaguy: yeah, i reviewed it, still expecting the ci to pass though. | 09:19 |
| johnthetubaguy | claudiub: +1 | 09:20 |
| *** marzif has joined #openstack-nova | 09:21 | |
| markus_z | tdurakov: The complexity of (live-)migration is IMO mainly driven by the accumulation of different use cases and their pre-reqs and post-reqs. Right now, I assume that a specific code path makes sense for a use case which is unknown to me. | 09:22 |
| markus_z | tdurakov: On the other hand it could be just a mistake :) | 09:22 |
| claudiub | johnthetubaguy: anyways, any suggestions for the hyper-v snapshotting race thingy? https://review.openstack.org/#/c/214560/ | 09:22 |
| tdurakov | markus_z, as for me, we should split different types of live-migration in code | 09:24 |
| markus_z | tdurakov: I guess this would reduce the amount of IFs and clarify use cases? | 09:25 |
| tdurakov | yep | 09:25 |
| markus_z | tdurakov: Sounds desirable :) | 09:25 |
| tdurakov | give me few minutes | 09:25 |
| tdurakov | filing bug in lp, I've talked earlier | 09:26 |
| *** lpetrut has joined #openstack-nova | 09:28 | |
| *** Zhang has quit IRC | 09:30 | |
| *** rdopiera has joined #openstack-nova | 09:30 | |
| *** sbalakri has joined #openstack-nova | 09:30 | |
| *** rotbeard has quit IRC | 09:30 | |
| *** sbalakri has left #openstack-nova | 09:30 | |
| *** marzif has quit IRC | 09:31 | |
| *** marzif has joined #openstack-nova | 09:31 | |
| eliqiao | tdurakov: what the plan of spliting different types of live-migration? | 09:32 |
| tdurakov | markus_z, eliqiao https://bugs.launchpad.net/nova/+bug/1488435 | 09:33 |
| openstack | Launchpad bug 1488435 in OpenStack Compute (nova) "Resources on destination node doesn't cleanup if live-migration fails" [Undecided,New] | 09:33 |
| tdurakov | eliqiao, i'm going to file blueprint and spec for that | 09:34 |
| *** yanxb has quit IRC | 09:34 | |
| eliqiao | tdurakov: looking forward for that. | 09:34 |
| *** yanxb has joined #openstack-nova | 09:35 | |
| tdurakov | eliqiao, markus_z, could you triage bug I've filed | 09:35 |
| johnthetubaguy | sdague: some API test tidy ups if you are in the mood for some reviews: https://review.openstack.org/#/c/174215/10 | 09:35 |
| *** zhihai_s has quit IRC | 09:36 | |
| johnthetubaguy | eliqiao: tdurakov: my big ask for folks helping with live-migration is to help make the multinode devstack tests reliably pass, once we have that in place, I think we can do more | 09:37 |
| johnthetubaguy | (until that point, I don't really want to see any big changes, really) | 09:37 |
| tdurakov | johnthetubaguy, one more thing | 09:37 |
| tdurakov | are we going to have ci for multinode? | 09:37 |
| markus_z | tdurakov: Could you add a bit more details to the bug, please? https://wiki.openstack.org/wiki/BugFilingRecommendations | 09:38 |
| eliqiao | tdurakov: if I understand your bug correctly , _create_images_and_backing failed to download image from glance ,right? | 09:38 |
| tdurakov | eliqiao, yep | 09:38 |
| johnthetubaguy | tdurakov: thats basically what I am talking about, there are people trying to make that work, but they need help, I believe: see gate-tempest-dsvm-multinode-full | 09:38 |
| markus_z | tdurakov: The commit IDs of devstack and Nova would be interesting as well as steps to reproduce. | 09:39 |
| tdurakov | markus_z, sure | 09:39 |
| tdurakov | johnthetubaguy, ++ | 09:39 |
| tdurakov | it would be really cool thing!:) | 09:39 |
| johnthetubaguy | tdurakov: yeah, I think thats the most important thing here, lets get folks to help make that reliably pass, then we can have a hope of maintaining live-migrate :) | 09:39 |
| markus_z | tdurakov: This is important for the reviews later on, to decide if the patch set matches the expectation mentioned in the bug report. | 09:40 |
| tdurakov | johnthetubaguy, so first priority for reliable multinode devstack, right? | 09:40 |
| johnthetubaguy | tdurakov: I think so | 09:41 |
| tdurakov | johnthetubaguy, ack) | 09:41 |
| johnthetubaguy | tdurakov: I would love to have that before the end of liberty | 09:41 |
| *** shoutm has quit IRC | 09:41 | |
| johnthetubaguy | if we can identify the key blockers, chase them down, fix them, etc, that would be amazingly good news for everyone! | 09:42 |
| tdurakov | i'll be back in an hour. | 09:43 |
| eliqiao | markus_z, tdurakov get a exception when calling _create_images_and_backing , so we can fake an exception around that method. | 09:44 |
| eliqiao | _rollback_live_migration doesn't clean up the instance_dir which is made in dest node. | 09:46 |
| openstackgerrit | lyanchih proposed openstack/nova: libvirt - Flavor's quota instance resource not observed for non-ephemeral storage https://review.openstack.org/201019 | 09:48 |
| *** dims__ has joined #openstack-nova | 09:48 | |
| *** dims__ has quit IRC | 09:48 | |
| *** dims__ has joined #openstack-nova | 09:49 | |
| *** BobBall has quit IRC | 09:50 | |
| *** aix has joined #openstack-nova | 09:50 | |
| *** ndipanov has joined #openstack-nova | 09:51 | |
| markus_z | eliqiao: I posted a comment in the bug report. | 09:52 |
| markus_z | eliqiao: I think this combines three different issues. | 09:52 |
| *** aix has quit IRC | 09:52 | |
| eliqiao | markus_z: agreed | 09:53 |
| *** aix has joined #openstack-nova | 09:53 | |
| *** sahid has quit IRC | 09:54 | |
| *** sahid has joined #openstack-nova | 09:54 | |
| eliqiao | hi , can anyone help to review this patch https://review.openstack.org/#/c/216123/ , I think it is an emergency one. | 09:55 |
| markus_z | It's an important catch I think. And it's good to know people who are willing to dig into the migration topic. | 09:55 |
| markus_z | eliqiao: I'll look at it after the lunch in 1h. | 09:56 |
| eliqiao | markus_z: thanks, but that one isn't LM related. appreciate if you could help to review. | 09:57 |
| *** abhishekk has joined #openstack-nova | 09:59 | |
| openstackgerrit | lyanchih proposed openstack/nova: libvirt: serial console ports count upper limit needs to be checked https://review.openstack.org/207373 | 09:59 |
| *** romainh has quit IRC | 10:00 | |
| openstackgerrit | John Garbutt proposed openstack/nova: api: tidy up paste.ini https://review.openstack.org/214085 | 10:03 |
| *** exploreshaifali has joined #openstack-nova | 10:03 | |
| openstackgerrit | zhangfeng proposed openstack/nova: Fix the error of duplicate words. https://review.openstack.org/216646 | 10:04 |
| *** Zhang has joined #openstack-nova | 10:05 | |
| *** humble__ has joined #openstack-nova | 10:06 | |
| *** haomaiwang has quit IRC | 10:09 | |
| *** humble_ has quit IRC | 10:09 | |
| *** haomaiwa_ has joined #openstack-nova | 10:10 | |
| *** Zhang has quit IRC | 10:11 | |
| *** shaohe_feng has joined #openstack-nova | 10:12 | |
| *** Zhang has joined #openstack-nova | 10:13 | |
| *** lpetrut has quit IRC | 10:13 | |
| *** summerWendy has quit IRC | 10:13 | |
| *** Zhang has quit IRC | 10:15 | |
| *** exploreshaifali has quit IRC | 10:15 | |
| *** exploreshaifali has joined #openstack-nova | 10:16 | |
| *** lpetrut has joined #openstack-nova | 10:17 | |
| *** irenab has quit IRC | 10:18 | |
| *** salv-orlando has quit IRC | 10:20 | |
| *** vivekd has quit IRC | 10:22 | |
| *** vivekd_ has joined #openstack-nova | 10:22 | |
| *** vivekd_ is now known as vivekd | 10:22 | |
| *** dave-mcc_ has joined #openstack-nova | 10:23 | |
| *** Zhang has joined #openstack-nova | 10:26 | |
| *** Zhang has quit IRC | 10:27 | |
| openstackgerrit | Adelina Tuvenie proposed openstack/nova: Fixes Bug "destroy_vm fails with HyperVException" https://review.openstack.org/204007 | 10:27 |
| *** Zhang has joined #openstack-nova | 10:38 | |
| *** Zhang has quit IRC | 10:39 | |
| tdurakov | johnthetubaguy, about ci jobs for live-migration: we need more than one environment for that | 10:40 |
| openstackgerrit | George Peristerakis proposed openstack/python-novaclient: Search the list of servers with unicode text. https://review.openstack.org/200649 | 10:40 |
| johnthetubaguy | tdurakov: agreed, but need to get one working reliably first | 10:40 |
| tdurakov | johnthetubaguy, we couldn't say then that we have reliable live-migration: we need at least one with shares storage, say nfs as easiest to config(to check ephemerals) one with rbd | 10:41 |
| johnthetubaguy | tdurakov: totally agree, but first things first | 10:42 |
| *** Zhang has joined #openstack-nova | 10:42 | |
| tdurakov | so, what would be the first? maybe no shared storage and check volume-backed instances and block-migration? | 10:43 |
| *** salv-orlando has joined #openstack-nova | 10:43 | |
| tdurakov | johnthetubaguy^ | 10:43 |
| johnthetubaguy | tdurakov: the first is getting gate-tempest-dsvm-multinode-full voting, I am unsure if that already includes one of the options | 10:43 |
| johnthetubaguy | tdurakov: once we know we have a base line working, we can look to add the other systems (I am tempted to say RBD multinode first) | 10:43 |
| tdurakov | johnthetubaguy, ok, where I can check it's code? | 10:44 |
| tdurakov | ci job i mean | 10:44 |
| johnthetubaguy | thats the name of the ci job | 10:44 |
| johnthetubaguy | gate-tempest-dsvm-multinode-full | 10:44 |
| tdurakov | is there any git repo for that? | 10:45 |
| johnthetubaguy | the folks in infra can help with find the config for that, its in the usual places: http://docs.openstack.org/infra/ | 10:46 |
| johnthetubaguy | tdurakov: if you look at one of the running ones, you can see the devstack and tempest config that was used: http://logs.openstack.org/06/201606/10/check/gate-tempest-dsvm-multinode-full/6247a47/logs/ | 10:47 |
| *** lyanchih has quit IRC | 10:47 | |
| tdurakov | johnthetubaguy, thanks! | 10:47 |
| johnthetubaguy | you should be able to start your own 8GB VM, and use that config, to reproduce the test, I belive | 10:47 |
| johnthetubaguy | mriedem, when he is online, might be a good person to ask about its current state | 10:48 |
| * johnthetubaguy goes for some lunch | 10:48 | |
| *** smatzek has joined #openstack-nova | 10:52 | |
| *** pbrooko has quit IRC | 10:55 | |
| *** Zhang has quit IRC | 10:56 | |
| tdurakov | markus_z, I've updated bug description | 10:59 |
| tdurakov | johnthetubaguy, markus_z, eliqiao: I've also made a patch to live_migration_cleanup_flags, could you review that: https://review.openstack.org/#/c/216329/ | 11:01 |
| *** asd112z has joined #openstack-nova | 11:03 | |
| *** Zhang has joined #openstack-nova | 11:03 | |
| *** asd112z has quit IRC | 11:08 | |
| *** vivekd has quit IRC | 11:08 | |
| *** ajayaa has quit IRC | 11:09 | |
| *** haomaiwa_ has quit IRC | 11:09 | |
| bauzas | johnthetubaguy: heya, missed your pings since I'm still a bit jetlagged | 11:09 |
| *** Zhang has quit IRC | 11:09 | |
| *** Marga_ has quit IRC | 11:10 | |
| *** Zhang has joined #openstack-nova | 11:10 | |
| *** haomaiwa_ has joined #openstack-nova | 11:10 | |
| *** adam_g has quit IRC | 11:12 | |
| markus_z | tdurakov: thanks for the update on that bug | 11:15 |
| *** ftersin has joined #openstack-nova | 11:16 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: default_schedule_zone should be a list of AZ(s) https://review.openstack.org/207221 | 11:17 |
| *** dprince has joined #openstack-nova | 11:19 | |
| *** shaohe_feng has quit IRC | 11:20 | |
| *** fawadkhaliq has quit IRC | 11:20 | |
| *** lucasagomes has quit IRC | 11:21 | |
| *** Guest29053 has quit IRC | 11:22 | |
| *** Sree has joined #openstack-nova | 11:22 | |
| *** lucasagomes has joined #openstack-nova | 11:22 | |
| *** ajayaa has joined #openstack-nova | 11:23 | |
| *** ccesario has quit IRC | 11:24 | |
| *** Sree has quit IRC | 11:27 | |
| johnthetubaguy | bauzas: no worries, I really hate jet lag, it hurts :( | 11:29 |
| bauzas | johnthetubaguy: thanks for commenting my change anyway, really good point | 11:30 |
| bauzas | johnthetubaguy: I made 2 answers, because I misuderstood at the first time :D | 11:31 |
| bauzas | johnthetubaguy: so, tl;dr: using the stats field is smart, but... https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L335 | 11:31 |
| *** sudipto has quit IRC | 11:32 | |
| bauzas | johnthetubaguy: which means that Ironic is writing a totally different string than the other drivers for that field | 11:32 |
| *** pbrooko has joined #openstack-nova | 11:32 | |
| bauzas | johnthetubaguy: that's because when writing the ComputeNode object, we left that ugly Text fieldtype | 11:33 |
| bauzas | johnthetubaguy: which is room for problems | 11:33 |
| *** vivekd has joined #openstack-nova | 11:35 | |
| *** thorst has joined #openstack-nova | 11:35 | |
| *** ccesario has joined #openstack-nova | 11:36 | |
| *** exploreshaifali has quit IRC | 11:36 | |
| *** Zhang has quit IRC | 11:37 | |
| *** apuimedo has joined #openstack-nova | 11:38 | |
| *** Zhang has joined #openstack-nova | 11:39 | |
| johnthetubaguy | bauzas: yeah, I forgot how nasty that ironic bit was | 11:39 |
| *** irenab has joined #openstack-nova | 11:40 | |
| bauzas | johnthetubaguy: so yeah, you're totally right, the spec was wrong :/ | 11:40 |
| *** vivekd has quit IRC | 11:40 | |
| *** Marga_ has joined #openstack-nova | 11:40 | |
| johnthetubaguy | bauzas: no worries, that happens | 11:40 |
| bauzas | johnthetubaguy: I'm about to write a Depends-On which would ensure that if that change merges, then the spec would be identical | 11:40 |
| *** exploreshaifali has joined #openstack-nova | 11:41 | |
| *** gszasz has joined #openstack-nova | 11:41 | |
| *** lucasagomes is now known as lucas-hungry | 11:45 | |
| johnthetubaguy | bauzas: I wouldn't worry about that too much at this point, lets just get the code sorted, we only have a few days left before all these patches will get a procedural -2 applied | 11:47 |
| *** vivekd_ has joined #openstack-nova | 11:47 | |
| bauzas | ack, FYI just working on the last change (scheduler side | 11:47 |
| bauzas | johnthetubaguy: ^ | 11:47 |
| bauzas | johnthetubaguy: I'll rebase at the same time | 11:48 |
| *** erlon has joined #openstack-nova | 11:48 | |
| johnthetubaguy | bauzas: OK, cool | 11:48 |
| johnthetubaguy | bauzas: it shouldn't really be needscodereview until all the patches are up for review, but thats a minor process nit | 11:48 |
| claudiub | johnthetubaguy: it would be great to have this in asap: https://review.openstack.org/#/c/216568/ | 11:49 |
| claudiub | johnthetubaguy: we need oslo.service 0.7.0 on hyper-v. | 11:49 |
| claudiub | for * | 11:50 |
| bauzas | johnthetubaguy: tbh, I said NeedsCodeReview because I wanted to have the compute related patches to be merged before L-3 | 11:50 |
| bauzas | johnthetubaguy: the sched side can be waiting for Mitaka | 11:50 |
| bauzas | johnthetubaguy: because since we need to wait for all compute nodes to be having the same fields, waiting 1 more cycle would be not good | 11:51 |
| *** Kevin_Zheng_ has joined #openstack-nova | 11:51 | |
| johnthetubaguy | bauzas: true | 11:51 |
| *** vivekd_ is now known as vivekd | 11:51 | |
| johnthetubaguy | claudiub: yeah, thats a good heads up, +2 from me | 11:51 |
| *** marcusvrn_ has joined #openstack-nova | 11:53 | |
| claudiub | ty :) | 11:54 |
| *** zhenguo has quit IRC | 11:55 | |
| alex_xu | nova api meeting in 5 mins | 11:56 |
| *** gmann_ has joined #openstack-nova | 11:58 | |
| *** pbrooko has quit IRC | 11:58 | |
| *** oomichi has joined #openstack-nova | 11:58 | |
| alex_xu | nova api meeting is running | 12:01 |
| *** shoutm has joined #openstack-nova | 12:02 | |
| johnthetubaguy | oomichi: I have a +2 on most of these now, I think most had very small updates since your last +2 https://blueprints.launchpad.net/nova/+spec/test-collapse-v2-and-v21 | 12:03 |
| *** pbrooko has joined #openstack-nova | 12:05 | |
| *** venkat_p has quit IRC | 12:08 | |
| *** haomaiwa_ has quit IRC | 12:09 | |
| *** fawadkhaliq has joined #openstack-nova | 12:10 | |
| *** haomaiwang has joined #openstack-nova | 12:10 | |
| ftersin | ndipanov: hi. are you objections still actulal in https://review.openstack.org/#/c/170243/ | 12:10 |
| *** fawadkhaliq has quit IRC | 12:14 | |
| *** deepakcs has quit IRC | 12:14 | |
| *** tojuvone has quit IRC | 12:14 | |
| rajesht | johnthetubaguy: need your help, if you have a little bit of time | 12:14 |
| johnthetubaguy | rajesht: sorry, in the API meeting, should be free for a bit after that | 12:15 |
| rajesht | no problem :) | 12:15 |
| *** Zhang has quit IRC | 12:16 | |
| *** ftersin has quit IRC | 12:19 | |
| *** Zhang has joined #openstack-nova | 12:19 | |
| *** fawadkhaliq has joined #openstack-nova | 12:20 | |
| *** romainh has joined #openstack-nova | 12:20 | |
| *** ftersin has joined #openstack-nova | 12:21 | |
| ndipanov | ftersin, ah sorry about that - | 12:23 |
| ndipanov | will look at it in a few minutes | 12:25 |
| bauzas | ndipanov: around for a question ? | 12:27 |
| bauzas | ndipanov: I'm just wondering how the cpu/ram ratios are related to the NUMA filters | 12:28 |
| *** annegentle has joined #openstack-nova | 12:28 | |
| *** rotbeard has joined #openstack-nova | 12:32 | |
| *** VW_ has joined #openstack-nova | 12:33 | |
| *** tojuvone has joined #openstack-nova | 12:33 | |
| openstackgerrit | jichenjc proposed openstack/nova: Refactor test cases for live-migrate error case https://review.openstack.org/179048 | 12:34 |
| *** smatzek has quit IRC | 12:35 | |
| *** ajayaa has quit IRC | 12:35 | |
| *** edmondsw has joined #openstack-nova | 12:36 | |
| *** takedakn has joined #openstack-nova | 12:36 | |
| *** deepthi has quit IRC | 12:37 | |
| *** aysyd has joined #openstack-nova | 12:38 | |
| *** scheuran has joined #openstack-nova | 12:39 | |
| bauzas | ndipanov: just testing it and it seems a bit buggy... | 12:40 |
| openstackgerrit | George Peristerakis proposed openstack/python-novaclient: Search the list of servers with unicode text. https://review.openstack.org/200649 | 12:42 |
| *** baoli has joined #openstack-nova | 12:44 | |
| *** lucas-hungry is now known as lucasagomes | 12:46 | |
| *** pbrooko_ has joined #openstack-nova | 12:46 | |
| *** VW_ has quit IRC | 12:47 | |
| *** vivekd has quit IRC | 12:48 | |
| *** VW_ has joined #openstack-nova | 12:48 | |
| *** pbrooko has quit IRC | 12:48 | |
| *** baoli has quit IRC | 12:48 | |
| *** annegentle has quit IRC | 12:50 | |
| openstackgerrit | sahid proposed openstack/nova: libvirt: make guest to return power state https://review.openstack.org/216263 | 12:51 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: move domain info to guest https://review.openstack.org/191046 | 12:51 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: clean code of NOTE tags https://review.openstack.org/212580 | 12:51 |
| *** dustins has joined #openstack-nova | 12:52 | |
| *** VW_ has quit IRC | 12:52 | |
| dims__ | ftersin: want to pick up https://review.openstack.org/#/c/207221/ ? | 12:53 |
| *** ccarmack has left #openstack-nova | 12:53 | |
| *** smatzek has joined #openstack-nova | 12:54 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/216568 | 12:54 |
| *** apuimedo has quit IRC | 12:57 | |
| *** burgerk has joined #openstack-nova | 12:58 | |
| *** zhihai_s has joined #openstack-nova | 12:58 | |
| *** zhihai_song has joined #openstack-nova | 12:59 | |
| *** apuimedo has joined #openstack-nova | 12:59 | |
| *** burgerk_ has joined #openstack-nova | 13:00 | |
| *** burgerk has quit IRC | 13:00 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Add cpu_allocation_ratio and ram_allocation_ratio to ComputeNode https://review.openstack.org/215471 | 13:01 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Update ComputeNode values with allocation ratios in the RT https://review.openstack.org/216362 | 13:01 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Update HostManager and filters to use RT ratios https://review.openstack.org/216696 | 13:01 |
| bauzas | johnthetubaguy: ^ I know you're in meeting, but here are the latest changes, incl. rebases and scheduler side | 13:01 |
| Kevin_Zheng_ | So we will do that in M? | 13:01 |
| Kevin_Zheng_ | The project_id thin | 13:02 |
| Kevin_Zheng_ | The project_id thing | 13:02 |
| alex_xu | Kevin_Zheng_: I guess, the time is tight | 13:02 |
| johnthetubaguy | bauzas: OK, thanks, might not get to those right away I am afraid | 13:02 |
| bauzas | johnthetubaguy: no worries | 13:03 |
| *** zhihai_s has quit IRC | 13:03 | |
| ftersin | dims__: hi, didn't think about it. why? did i've wrote here something special, which doesn't related to the fix? | 13:03 |
| Kevin_Zheng_ | alex_xu: so should I registe a BP for this? | 13:03 |
| *** Sree has joined #openstack-nova | 13:04 | |
| gmann_ | sdague: johnthetubaguy : not able to connect my system remotely. I will rebase https://review.openstack.org/#/c/214985/ tomorrow. | 13:04 |
| johnthetubaguy | gmann_: no worries | 13:04 |
| alex_xu | Kevin_Zheng_: let me review it, then I will try to comment on | 13:04 |
| Kevin_Zheng_ | alex_xu: Ok then, Thanks a lot. | 13:04 |
| alex_xu | Kevin_Zheng_: np | 13:05 |
| ftersin | dims__: but sure, if you ask to do there what i meant, i can... | 13:05 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: default_schedule_zone should be a list of AZ(s) https://review.openstack.org/207221 | 13:05 |
| * alex_xu need draft the meeting content, then won't typo such slow~ | 13:06 | |
| *** Sree has quit IRC | 13:06 | |
| *** Zhang has quit IRC | 13:06 | |
| *** xyang1 has joined #openstack-nova | 13:06 | |
| *** baoli has joined #openstack-nova | 13:07 | |
| dims__ | ftersin: i am not comfortable making more changes than minimum needed at this point of the cycle and you seem to have a good idea on how this should work, hence asking. | 13:07 |
| *** rlrossit has joined #openstack-nova | 13:08 | |
| *** nelsnelson has joined #openstack-nova | 13:08 | |
| dims__ | ftersin: please feel free to pick it up if you wish | 13:08 |
| *** Sree has joined #openstack-nova | 13:08 | |
| *** gmann_ has quit IRC | 13:08 | |
| ftersin | dims__: now in that case you return no az, and the launch is continued. But previously if default az is done, _validate_and_build_base_options catched this and failed | 13:08 |
| dims__ | ftersin: see that's what i am saying :) | 13:09 |
| *** pbrooko has joined #openstack-nova | 13:09 | |
| *** oomichi has quit IRC | 13:09 | |
| *** haomaiwang has quit IRC | 13:09 | |
| sdague | gmann: sounds fine, I approved the rest of the stack | 13:10 |
| bauzas | dims__: hum, just seeing https://review.openstack.org/#/c/207221/ | 13:10 |
| ftersin | dims__: ok, in a few hours | 13:10 |
| *** browne has joined #openstack-nova | 13:10 | |
| *** haomaiwang has joined #openstack-nova | 13:10 | |
| dims__ | bauzas: y, ftersin graciously accepted to help fix that review better :) | 13:11 |
| bauzas | dims__: I think that https://bugs.launchpad.net/nova/+bug/1474467 is rather a feature than a bugfix | 13:11 |
| openstack | Launchpad bug 1474467 in OpenStack Compute (nova) "default_schedule_zone should be list" [Low,In progress] - Assigned to Davanum Srinivas (DIMS) (dims-v) | 13:11 |
| *** annegentle has joined #openstack-nova | 13:11 | |
| dims__ | bauzas: sure, it can wait till M opens up, just want to get the code out there | 13:11 |
| bauzas | dims__: just putting it as Wishlist | 13:12 |
| *** pbrooko_ has quit IRC | 13:12 | |
| bauzas | dims__: also, I'm not very fond of that | 13:12 |
| dims__ | bauzas: ack | 13:12 |
| *** annegentle has quit IRC | 13:12 | |
| bauzas | honestly | 13:12 |
| bauzas | dims__: I need to think more about that but that's a pretty invasive change and lots of operators don't really understand what a Nova AZ is | 13:12 |
| ftersin | i think so too | 13:13 |
| bauzas | dims__: I was about to write an ML thread about that btw. because some discussion with the Trove PTL made me think that lots of us also think that AZs == failure domains | 13:13 |
| bauzas | which is *not* the case | 13:13 |
| *** raildo-afk is now known as raildo | 13:13 | |
| *** annegentle has joined #openstack-nova | 13:14 | |
| *** baoli_ has joined #openstack-nova | 13:14 | |
| *** emagana has quit IRC | 13:14 | |
| *** zhenguo has joined #openstack-nova | 13:16 | |
| markus_z | bauzas: jaypipes referenced the AZ during the midcycle as exactly that. | 13:16 |
| bauzas | dims__: rereading the bugfile makes me think it's just a problem between what's a default AZ and what's not | 13:17 |
| bauzas | markus_z: as exactly what ? failure domains ? | 13:17 |
| *** baoli has quit IRC | 13:17 | |
| bauzas | markus_z: they aren't | 13:17 |
| markus_z | bauzas: yes AZs == failure domains | 13:17 |
| annegentle | alex_xu: johnthetubaguy: sorry I missed the api meeting, first day of school here :) | 13:18 |
| bauzas | markus_z: and I know that jay is keen to have AZs == failure domains, but I just think it's not needed as cells v2 are somehow those failure domains we need | 13:18 |
| markus_z | bauzas: That's what jay said. I haven't yet digged into the AZs | 13:18 |
| *** markmcclain has quit IRC | 13:18 | |
| bauzas | markus_z: AZs are just metadata applied to aggregates + specific silly rules | 13:18 |
| *** kylek3h has joined #openstack-nova | 13:18 | |
| johnthetubaguy | annegentle: no worries, not much news right now | 13:19 |
| *** davidlenwell has quit IRC | 13:19 | |
| markus_z | bauzas: I mentioned it as a hint that a ML discussion would be helpful to clarify the assumptions about AZs. | 13:19 |
| *** Zhang has joined #openstack-nova | 13:19 | |
| bauzas | markus_z: yeah, I owe that discussion | 13:19 |
| bauzas | markus_z: lots of people are confused by the term | 13:19 |
| *** ccarmack has joined #openstack-nova | 13:20 | |
| markus_z | bauzas: I'm one of them :) | 13:20 |
| bauzas | markus_z: https://twitter.com/sylvainbauza/status/634425145281503232 | 13:20 |
| *** thangp has joined #openstack-nova | 13:20 | |
| *** davidlenwell has joined #openstack-nova | 13:20 | |
| markus_z | bauzas: No chars left for more "O"s? ;) | 13:20 |
| annegentle | johnthetubaguy: ok, good. alex_xu 's notes are good to have | 13:20 |
| bauzas | markus_z: I should have said more OOOOs | 13:20 |
| *** markmcclain has joined #openstack-nova | 13:21 | |
| bauzas | each time I'm hearing such phrase, it just pisses me off | 13:21 |
| *** ccarmack has left #openstack-nova | 13:21 | |
| *** mdrabe has joined #openstack-nova | 13:21 | |
| *** changbl has quit IRC | 13:21 | |
| *** Zhang has quit IRC | 13:22 | |
| *** rlrossit has left #openstack-nova | 13:22 | |
| johnthetubaguy | annegentle: we are thinking in two weeks time, we can push hard on the docs, once we know where everything has landed, its a way too late really, but its our best bet right now | 13:22 |
| bauzas | markus_z: (and others) http://blog.russellbryant.net/2013/05/21/availability-zones-and-host-aggregates-in-openstack-compute-nova/ | 13:22 |
| annegentle | johnthetubaguy: yeah | 13:22 |
| johnthetubaguy | annegentle: prototype for swagger auto generation (non WADL) by tokyo, seems a good aim too | 13:22 |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Added server tags support in nova-api https://review.openstack.org/128940 | 13:23 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/216568 | 13:23 |
| *** links has quit IRC | 13:23 | |
| markus_z | bauzas: Thanks, I'll read it soon | 13:23 |
| *** mtanino has joined #openstack-nova | 13:23 | |
| *** lyanchih has joined #openstack-nova | 13:24 | |
| *** dims__ has quit IRC | 13:24 | |
| bauzas | markus_z: so will do my ML thread | 13:24 |
| *** dims__ has joined #openstack-nova | 13:25 | |
| *** ajayaa has joined #openstack-nova | 13:25 | |
| alex_xu | sdague: hope you can take look at this also https://review.openstack.org/#/c/215100/ | 13:26 |
| *** ferest has joined #openstack-nova | 13:26 | |
| sdague | alex_xu: looking now | 13:26 |
| alex_xu | sdague: thanks | 13:27 |
| *** jwcroppe has quit IRC | 13:27 | |
| *** IanGovett1 has joined #openstack-nova | 13:27 | |
| annegentle | johnthetubaguy: would love that | 13:28 |
| *** davideagnello has joined #openstack-nova | 13:28 | |
| *** davideagnello has quit IRC | 13:28 | |
| johnthetubaguy | sdague: are you sure its worth renaming osapi_v3 group, given we are deprecating them anyway? | 13:28 |
| *** ctrath has joined #openstack-nova | 13:28 | |
| *** shuquan has joined #openstack-nova | 13:28 | |
| sdague | alex_xu: right, so reverting that isn't an option. We probably should just chase the queue state with a microversion to signal folks that it's something they can get after a certain point | 13:29 |
| sdague | johnthetubaguy: so, I think because we don't know exactly when we are deleting them, yes | 13:30 |
| *** jwcroppe has joined #openstack-nova | 13:30 | |
| alex_xu | sdague: update the doc of microversions? | 13:30 |
| *** diogogmt has joined #openstack-nova | 13:30 | |
| *** IanGovett has quit IRC | 13:30 | |
| *** rlrossit has joined #openstack-nova | 13:30 | |
| openstackgerrit | John Garbutt proposed openstack/nova: api: deprecate the concept of extensions in v2.1 https://review.openstack.org/214592 | 13:30 |
| johnthetubaguy | sdague: true | 13:30 |
| *** ctrath1 has joined #openstack-nova | 13:30 | |
| *** ferest has quit IRC | 13:31 | |
| alex_xu | dansmith: we are just talking about https://review.openstack.org/#/c/215100/ ^ | 13:31 |
| *** IanGovett has joined #openstack-nova | 13:31 | |
| sdague | alex_xu: yeh, I think we just need a bump to indicate this is out there and part of the contract. | 13:31 |
| *** Zhang has joined #openstack-nova | 13:31 | |
| alex_xu | sdague: emm...that sounds a smart way | 13:31 |
| sdague | it means there will be a small number of changes where it's there and not advertised, but it should be fine for the majority of folks | 13:32 |
| *** IanGovett1 has quit IRC | 13:32 | |
| dansmith | so, just to be clear, | 13:32 |
| dansmith | you know that we previously added more statuses than just in that patch, | 13:33 |
| dansmith | and that ndipanov is about to potentially add another terminal status for live migration (and maybe evacuate) | 13:33 |
| *** IanGovett1 has joined #openstack-nova | 13:33 | |
| dansmith | I don't understand why the values were ever enumerated in schema, but I think it was a mistake | 13:33 |
| sdague | I did not, however, if those get to the user, we do need to version that data | 13:34 |
| sdague | because otherwise we'll have people assuming they can test for that state, and it's not even possible in their code | 13:34 |
| *** claudiub has quit IRC | 13:34 | |
| *** jecarey has joined #openstack-nova | 13:34 | |
| dansmith | my point is, I disagree that it's a state | 13:34 |
| *** jwcroppe has quit IRC | 13:34 | |
| dansmith | and it's only visible to admins anyway | 13:34 |
| *** ctrath has quit IRC | 13:34 | |
| ndipanov | sdague, dansmith we are nowhere near ready to make it part of the public api contract imho | 13:35 |
| sdague | is it accessible over the API? | 13:35 |
| dansmith | ndipanov: unfortunately, it is already coded in schema, but I have no idea why :( | 13:35 |
| ndipanov | sdague, why does that matter | 13:35 |
| ndipanov | it's not something we should commit to | 13:35 |
| sdague | because, then it is public contract | 13:35 |
| ndipanov | well I disagree | 13:35 |
| tdurakov | markus_z, | 13:35 |
| ndipanov | it's visible - yes | 13:35 |
| sdague | that's the whole point of the API | 13:35 |
| *** shuquan has quit IRC | 13:35 | |
| ndipanov | ok then | 13:35 |
| dansmith | sdague: we treat this like a status string | 13:36 |
| tdurakov | markus_z, are you around? | 13:36 |
| dansmith | sdague: almost free-form | 13:36 |
| *** dims has joined #openstack-nova | 13:36 | |
| markus_z | tdurakov: yes, I am | 13:36 |
| *** tpeoples has joined #openstack-nova | 13:36 | |
| *** dims__ has quit IRC | 13:36 | |
| alex_xu | at least we don't have doc or conract tell user, that isnt part of contract.... | 13:36 |
| tdurakov | left one more comment on my bug | 13:36 |
| ndipanov | as long as it's a string and not an int or an object | 13:36 |
| tdurakov | markus_z, glance problem was trivial | 13:36 |
| *** IanGovett has quit IRC | 13:36 | |
| ndipanov | we should not commit to it's value | 13:36 |
| ndipanov | that's what I'm saying | 13:36 |
| markus_z | tdurakov: What was the number again? | 13:37 |
| tdurakov | https://bugs.launchpad.net/nova/+bug/1488435 | 13:37 |
| openstack | Launchpad bug 1488435 in OpenStack Compute (nova) "Resources on destination node doesn't cleanup if live-migration fails" [Medium,In progress] - Assigned to Timofey Durakov (tdurakov) | 13:37 |
| sdague | ndipanov: so you'd be happy if that's the way the python stdlib worked? it's not an int, it can be anything? | 13:37 |
| *** shuquan has joined #openstack-nova | 13:37 | |
| tdurakov | so there is only cleanup issue | 13:37 |
| dansmith | sdague: if stdlib is just fetching us a value from the os, then yeah | 13:37 |
| *** mriedem has joined #openstack-nova | 13:38 | |
| alex_xu | ndipanov: we use neutron client like this https://github.com/openstack/nova/blob/master/nova/network/neutronv2/api.py#L1644 ... | 13:38 |
| markus_z | tdurakov: You refer there to an incorrect configuration, right? | 13:38 |
| ndipanov | I mean | 13:38 |
| sdague | if you don't want to commit this to the contract, then filter it out of what goes back to the user | 13:38 |
| tdurakov | markus_z, yep | 13:38 |
| *** Zhang has quit IRC | 13:38 | |
| tdurakov | no port on dest node were provided(( | 13:38 |
| sdague | you can't have it both ways | 13:38 |
| dansmith | sdague: should we schemaify the service disable reason field? | 13:38 |
| sdague | reason isn't a state | 13:39 |
| dansmith | this isn't either | 13:39 |
| dansmith | it's status | 13:39 |
| markus_z | tdurakov: Ah, good to hear. Add me as reviewer to the patch for 2) and 3) when you are ready, ok? | 13:39 |
| dansmith | and reason is sometimes status when we set it programmatically | 13:39 |
| dansmith | in response to auto-disabling the state | 13:39 |
| *** dims_ has joined #openstack-nova | 13:39 | |
| ndipanov | I think it's a bad idea to arbitrarily draw the line and say from now on _everything_ is treated as a contract - we are nowhere near that in terms of consistency | 13:39 |
| ndipanov | and doing so will impose a massive cost to future development | 13:40 |
| tdurakov | markus_z, sure, I'm going to finish with it, so there will be reference config for nova that enables live-migration. than we could move to enabling ci job | 13:40 |
| sdague | ndipanov: that was a bit point to microversions, we know lots of things are bad, but we can evolve them towards something good in a structured way | 13:40 |
| ndipanov | for I have to say gains that I don't understand... | 13:40 |
| dansmith | sdague: I argued the same thing when ndipanov proposed his migration state machine, that the state field should be a new one that *is* consistent, because status isn't and never was | 13:40 |
| *** rotbart has joined #openstack-nova | 13:40 | |
| ndipanov | sdague, right but if you need to keep backwards compat - it's not free either so we should do it when it makes sense | 13:40 |
| dansmith | IMHO, status is and should be considered just text.. informative | 13:40 |
| ndipanov | dansmith, the more I looked at that code the more I agree | 13:41 |
| sdague | dansmith: so then it should be a localized string | 13:41 |
| markus_z | tdurakov: Sorry, I didn't understand this. What do you mean with "reference config"? | 13:41 |
| *** dims__ has joined #openstack-nova | 13:41 | |
| sdague | if it's just descriptive, make a localized sentence | 13:41 |
| sdague | if it's mean to be consumed by software | 13:41 |
| sdague | it should be part of the contract | 13:42 |
| *** dims has quit IRC | 13:42 | |
| dansmith | sdague: yeah, except that we (wrongly) key on it internally in a variety of places | 13:42 |
| sdague | dansmith: which is indicative that it's a useful part of the contract | 13:42 |
| ndipanov | I agree in principle but we need to do some work before we do it in this case | 13:42 |
| dansmith | sdague: it's not enumerated anywhere though, it's just strings in code in random places | 13:42 |
| pkholkin | mriedem: Hello Matt! Saw your message. We made some updates to our patch 'switch to olso_cache'. Do you still have some questions? Could you please make another review, thanks! https://review.openstack.org/#/c/203049/16 | 13:42 |
| *** claudiub has joined #openstack-nova | 13:42 | |
| sdague | right, I do get that we have lots of random strings lots of places | 13:43 |
| tdurakov | markus_z, after i've fixed error with glance nova fails with File "/usr/local/lib/python2.7/dist-packages/libvirt.py", line 1586, in migrateToURI2 | 13:43 |
| tdurakov | if ret == -1: raise libvirtError ('virDomainMigrateToURI2() failed', dom=self) | 13:43 |
| mriedem | pkholkin: i'm in class all day so probably can't get to it in detail | 13:43 |
| pkholkin | mriedem: ok, please look if you have time | 13:43 |
| *** rotbeard has quit IRC | 13:44 | |
| sdague | it seems like a very useful thing for access via the api to understand where things stand, useful enough that we use it internally (rightly or wrongly) | 13:44 |
| openstackgerrit | John Garbutt proposed openstack/nova: api: deprecate the concept of extensions in v2.1 https://review.openstack.org/214592 | 13:44 |
| tdurakov | markus_z, I think problem is in libvirt_migration_url and migration flag in nova.conf | 13:44 |
| *** mriedem is now known as mriedem_class | 13:44 | |
| *** bapalm has quit IRC | 13:44 | |
| ndipanov | sdague, it is - that's why we want to add more of it and formalize it | 13:44 |
| sdague | not making it part of the contract just means people are going to code regex to it and get mad about when things aren't there | 13:44 |
| *** Zhang has joined #openstack-nova | 13:45 | |
| tdurakov | also need to check why no port were provided on destination compute in config | 13:45 |
| *** VW_ has joined #openstack-nova | 13:45 | |
| *** dims_ has quit IRC | 13:45 | |
| dansmith | sdague: right now, the only sane way to do it is look for values you recognize and ignore the ones you don't | 13:45 |
| sdague | ndipanov: ok, so do that. Make it formal. Don't keep patching it adhoc | 13:45 |
| dansmith | sdague: which is why I don't think it should be codified | 13:45 |
| sdague | dansmith: which is *terrible* | 13:45 |
| *** karimb has quit IRC | 13:46 | |
| * dansmith gives up | 13:46 | |
| sdague | with microversions we get to say "ok, things were adhoc before, here's where they stand" | 13:46 |
| sdague | then add the more states when they come | 13:46 |
| dansmith | sdague: you should wait to codify this random silliness until ndipanov's patch merges which adds more | 13:46 |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: DO NOT MERGE: Example of config options reshuffle https://review.openstack.org/214581 | 13:46 |
| ndipanov | yeah I'd say we want to wait on that whole thing if we agree that the work we started doing | 13:47 |
| ndipanov | which is | 13:47 |
| *** jwcroppe has joined #openstack-nova | 13:47 | |
| ndipanov | make Migration a staus holder for moving instances around | 13:47 |
| ndipanov | is the right way to do it | 13:47 |
| dansmith | we haven't even merged the patch that exposes the migration type field, | 13:48 |
| sdague | dansmith: ok, the benefit of that is what? | 13:48 |
| openstackgerrit | John Garbutt proposed openstack/nova: api: deprecate the api v2 extension configuration https://review.openstack.org/214601 | 13:48 |
| sdague | I guess I don't understand what benefit we get by not declaring this bit of contract change | 13:48 |
| dansmith | which means anyone that is looking for specific "states" will be utterly confused since a live migration shares none of the "states" with a normal migration | 13:48 |
| dansmith | and they can't tell the difference from the api side | 13:48 |
| sdague | ok, so how do we get there? | 13:49 |
| markus_z | tdurakov: Why do you see the problem there? | 13:49 |
| dansmith | sdague: because ndipanov is just going to change it again and a new microversion in between just means more silly code pollution to filter out (or whatever you think should happen) the new "states" | 13:49 |
| *** BobBall has joined #openstack-nova | 13:49 | |
| sdague | ok, so what's the plan to get this into a contractable state? | 13:50 |
| dansmith | sdague: ndipanov has a full state machine implementation for all migrations | 13:50 |
| ndipanov | well I had defined a state machine | 13:50 |
| openstackgerrit | John Garbutt proposed openstack/nova: api: deprecate the concept of extensions in v2.1 https://review.openstack.org/214592 | 13:50 |
| ndipanov | but then I abandoned it in hopes of getting the actual code merged | 13:51 |
| ndipanov | as there was some back and forth about the implementation | 13:51 |
| openstackgerrit | John Garbutt proposed openstack/nova: api: deprecate the api v2 extension configuration https://review.openstack.org/214601 | 13:51 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova-specs: libvirt: support pvpanic device https://review.openstack.org/216719 | 13:51 |
| ndipanov | eventually josh added some support for it in automaton | 13:51 |
| ndipanov | so I'll use some of that | 13:51 |
| tdurakov | markus_z, in my config there no flags at all http://docs.openstack.org/admin-guide-cloud/compute-configuring-migrations.html | 13:51 |
| *** bapalm_ is now known as bapalm | 13:51 | |
| dansmith | ndipanov: abandoned in gerrit, or just gave up expecting it to merge in liberty? | 13:51 |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Remove get_all method from servicegroup API https://review.openstack.org/214629 | 13:52 |
| bauzas | dims__: so heavy painy -1 here https://review.openstack.org/#/c/207221/8 | 13:52 |
| alex_xu | ndipanov: bump microverwion when your statemachine finshed? tell user don't trust status field before this version? | 13:52 |
| ndipanov | dansmith, well I can merge the patches for migration resource tracking | 13:52 |
| sdague | can we hide that info from the user until the whole state machine is ready? because my concern is we keep changing stuff adhoc here for a long time, people just get ragey that not of this is something you can program to. | 13:52 |
| dansmith | ndipanov: yeah | 13:52 |
| ndipanov | so I abandoned it | 13:52 |
| ndipanov | but they can be added later | 13:52 |
| ndipanov | (with more states) | 13:52 |
| *** karimb has joined #openstack-nova | 13:52 | |
| *** thumpba has joined #openstack-nova | 13:53 | |
| ndipanov | alex_xu, yeah that is a good way to go imho | 13:53 |
| dansmith | sdague: IMHO, we should remove the schema on the status field right now, not filter the values, add a new microversion when we expose the actual state field | 13:53 |
| dansmith | which is what alex_xu just said | 13:53 |
| ndipanov | at that point we'll likely introduce the new field | 13:53 |
| dansmith | yeah +1 | 13:53 |
| ndipanov | yes | 13:53 |
| *** evgenyf has quit IRC | 13:53 | |
| *** ferest has joined #openstack-nova | 13:53 | |
| alex_xu | we better have another version to tell user, from this version don't trust the status.... | 13:53 |
| dansmith | that microversion will do nothing though, right? | 13:54 |
| dansmith | except signaling? unless we're going to filter before that point, I'm not sure that buys us anything | 13:54 |
| dansmith | but, if the compromise is to filter before that, | 13:54 |
| sdague | yeh, it won't. Is this API call documented at all? If so we'd update that, but our docs are pretty bad | 13:54 |
| dansmith | sdague: it's admin-only and very cryptic already | 13:54 |
| sdague | honestly, probably should at least document the API samples | 13:54 |
| dansmith | not sure it is | 13:54 |
| alex_xu | update this https://github.com/openstack/nova/blob/master/nova/api/openstack/rest_api_version_history.rst? | 13:55 |
| *** BobBall has quit IRC | 13:55 | |
| sdague | admin-only doesn't really matter, people can open this stuff up. And the admin api doesn't get to punt on contracts | 13:55 |
| alex_xu | the only doc for microversion for now | 13:55 |
| dansmith | if the compromise is to land a microversion, filter before that, and make it "just a string" after that, then I'll agree to it as long as it fully lands in liberty | 13:55 |
| dansmith | sdague: I meant it's been ignored, re: docs | 13:55 |
| *** Sree has quit IRC | 13:55 | |
| sdague | dansmith: no, I agree, a microversion isn't useful | 13:55 |
| markus_z | tdurakov: Do you assume that there could be another configuration issue which could make the bug invalid? | 13:55 |
| dansmith | okay | 13:55 |
| sdague | dansmith: we still have samples for these though, right | 13:55 |
| *** GnomeKing has joined #openstack-nova | 13:56 | |
| sdague | which people go and read some times to figure out request semantics | 13:56 |
| sdague | we could stick a comment in there maybe | 13:56 |
| *** Sree has joined #openstack-nova | 13:56 | |
| dansmith | I assume we have to | 13:56 |
| sdague | so what's the story on when the "whole thing" lands? | 13:56 |
| sdague | the contract state machine | 13:57 |
| dansmith | wanna hear something awesome? | 13:57 |
| dansmith | the status field in the api sample is not one that we could/would ever have :P | 13:58 |
| tdurakov | markus_z, bug i've filed is valid:) but i haven't managed yet to live-migrate volume-backed instance on multinode devstack | 13:58 |
| *** jaypipes has joined #openstack-nova | 13:58 | |
| markus_z | tdurakov: I see, OK, I didn't understand this in the first place. | 13:58 |
| dansmith | https://github.com/openstack/nova/blob/master/nova/tests/functional/api_sample_tests/api_samples/os-migrations/migrations-get.json.tpl#L14 | 13:59 |
| *** emagana has joined #openstack-nova | 13:59 | |
| ndipanov | dansmith, the version could switch over to the new field | 13:59 |
| dansmith | capital first letter, and "done" is never used, AFAIK | 13:59 |
| ndipanov | and then revert back to the old that can change | 13:59 |
| ndipanov | when asked fro a lower one | 13:59 |
| tdurakov | hm... libvirtError: operation failed: Failed to connect to remote libvirt URI qemu+ssh://stack@node2/system: Cannot recv data: Host key verification failed.: Connection reset by peer | 13:59 |
| dansmith | ndipanov: IMHO, the version should just add the new field, leave the old one as it was | 13:59 |
| *** yamahata has joined #openstack-nova | 14:00 | |
| ndipanov | yes or that | 14:00 |
| *** Sree has quit IRC | 14:00 | |
| sdague | dansmith: yeh, the samples are often like that | 14:00 |
| sdague | which is a whole other issue | 14:01 |
| alex_xu | if we finally think status isn't part of contract, +1 for dansmith | 14:01 |
| tdurakov | johnthetubaguy, markus_z ndipanov, danpb what live_migration_url should be used by default? | 14:01 |
| dansmith | sdague: well, it means (to me) that the sample isn't conveying that it's an actual state (since it looks like text) | 14:02 |
| alex_xu | +1 for add new field, sounds make thing easy | 14:02 |
| sdague | dansmith: there is tons of contract content that's like that in samples, but anyway | 14:02 |
| sdague | what's the time frame for the whole ball of wax? | 14:02 |
| dansmith | sdague: what and where do you want an annotation in the api samples that it's a text field? like just hand-edit the doc/ one? | 14:02 |
| sdague | dansmith: yeh, so honestly, let's skip that bit for now, and try to figure out what and when is end game | 14:03 |
| dansmith | okay, did you find api docs for this? | 14:03 |
| sdague | dansmith: no, I assume it's not yet documented | 14:03 |
| dansmith | okay | 14:03 |
| *** diogogmt has quit IRC | 14:04 | |
| lxsli | bauzas: ping | 14:04 |
| dims__ | bauzas: no worries, i am just going through really old reviews/in progress bugs and making sure we don't lose anything valuable | 14:04 |
| alex_xu | the only I need to do is abandon https://review.openstack.org/#/c/215100/ for now, right? | 14:04 |
| sdague | so, what changes are landing in Liberty? and when during Mitaka can we declare this a strong contract? | 14:04 |
| dansmith | so ndipanov if you want to resurrect your patch, make the changes you have planned, I'll look at it | 14:05 |
| *** mnestratov|2 has joined #openstack-nova | 14:05 | |
| dansmith | sdague: I think drop the schema for L and see how much progress we can make on liberty with ndipanov's stuff | 14:05 |
| lxsli | bauzas: when we request destination for Ironic, that's by nodename right? which gives me a problem because there's no objects.compute_node method to lookup by nodename. Is it OK to add one do you think? | 14:05 |
| dansmith | if he has time to work on it in the short term it might be mergeable | 14:05 |
| *** geekinutah has joined #openstack-nova | 14:06 | |
| bauzas | lxsli: I don't get your question | 14:06 |
| sdague | ok | 14:06 |
| alex_xu | dansmith: the schema you point to tempest? | 14:07 |
| sdague | it still seems like if this is outside of the schema, it should be in the i18n catalog | 14:07 |
| bauzas | lxsli: I mean, Ironic is particular because of the 1:N relationship, but the boot req is similar | 14:07 |
| *** abhishekk has quit IRC | 14:07 | |
| dansmith | alex_xu: no, I thought you said that our schema in nova enumerated the possible values | 14:07 |
| sdague | because it's meant for description, not for programming | 14:07 |
| dansmith | alex_xu: you said that in your revert no,? | 14:07 |
| ndipanov | dansmith, resurrect the state machine patch? | 14:07 |
| alex_xu | dansmith: we don't have schema for response | 14:07 |
| lxsli | bauzas: I'm asking whether the requested_destination field we're adding to RequestSpec has the hostname or nodename (or both?) as its primitive form | 14:08 |
| bauzas | lxsli: since the scheduler gives back a tuple (host, node, limits), then the conductor calls the right service manager | 14:08 |
| *** mnestratov has quit IRC | 14:08 | |
| dansmith | alex_xu: okay | 14:08 |
| dansmith | ndipanov: yeah | 14:08 |
| bauzas | lxsli: ah that | 14:08 |
| alex_xu | so let me abandon https://review.openstack.org/#/c/215100/ | 14:08 |
| bauzas | lxsli: so, there is no way to pass the node in the API atm | 14:09 |
| bauzas | lxsli: sorry, I mean for evacuate and live-migrate | 14:09 |
| *** haomaiwang has quit IRC | 14:09 | |
| ndipanov | dansmith, OK but it's kind of sort of dependent on https://review.openstack.org/#/c/203284/ | 14:09 |
| bauzas | lxsli: which makes sense | 14:09 |
| ndipanov | not really though | 14:09 |
| dansmith | ndipanov: okay, well, maybe we can poke him today | 14:10 |
| bauzas | lxsli: because Ironic doesn't support both http://docs.openstack.org/developer/nova/support-matrix.html | 14:10 |
| lxsli | bauzas: right, thinking about LM would be hard for Ironic | 14:10 |
| dansmith | ndipanov: we also shouldn't rush it because we want the machine to be right | 14:10 |
| *** haomaiwang has joined #openstack-nova | 14:10 | |
| ndipanov | dansmith, ok let me resurrect the patch and then we can talk | 14:10 |
| lxsli | bauzas: ok I'll assume hostname == nodename then, thanks | 14:10 |
| bauzas | lxsli: well, that's a good point | 14:11 |
| ndipanov | lxsli, not sure that's a good idea | 14:11 |
| bauzas | lxsli: I'd rather see how force_hosts is done | 14:11 |
| *** tonytan4ever has joined #openstack-nova | 14:11 | |
| *** ferest has quit IRC | 14:11 | |
| bauzas | lxsli: yeah, I +1 ndipanov | 14:11 |
| ndipanov | thre is code that does different things | 14:11 |
| ndipanov | I am removing it though | 14:11 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova-specs: libvirt: support pvpanic device https://review.openstack.org/216719 | 14:11 |
| alex_xu | dansmith: ndipanov sdague thanks for the talk, abandoned https://review.openstack.org/#/c/215100/ | 14:11 |
| *** Piet has joined #openstack-nova | 14:12 | |
| dansmith | alex_xu: thanks | 14:12 |
| ndipanov | https://review.openstack.org/#/c/214713/ | 14:12 |
| ndipanov | lxsli, bauzas ^^ | 14:12 |
| *** Zhang has quit IRC | 14:12 | |
| bauzas | lxsli: FYI here is what happens with a boot request | 14:12 |
| bauzas | https://github.com/openstack/nova/blob/master/nova/compute/api.py#L563-L585 | 14:12 |
| bauzas | lxsli: I hate that code | 14:13 |
| bauzas | lxsli: but that's how we pass dest hosts when booting | 14:13 |
| bauzas | ndipanov: yeah that makes sense | 14:13 |
| *** vladikr has joined #openstack-nova | 14:13 | |
| bauzas | ndipanov: because select_dests gives back a (host, node, limits) which are passed over RPC to compute | 14:14 |
| ndipanov | yes | 14:14 |
| *** mtanino has quit IRC | 14:14 | |
| bauzas | ndipanov: so +1 to you | 14:14 |
| ndipanov | so we can't assume anything about the node really | 14:14 |
| bauzas | at least for that specific bit | 14:14 |
| ndipanov | if you need it - pass it | 14:14 |
| ndipanov | or dig it up | 14:14 |
| bauzas | right | 14:14 |
| lxsli | so should we split req_dest into req_host and req_node? | 14:15 |
| ndipanov | imho no we should be going the other way in our APIs | 14:15 |
| bauzas | lxsli: I'd rather say that req_dest is a ComputeNode obkect | 14:15 |
| bauzas | lxsli: and I wouldn't be happy by a req_host:req_dest stuff | 14:16 |
| openstackgerrit | Andrew Melton proposed openstack/nova: virt: allow for direct mounting of LocalBlockImages https://review.openstack.org/215311 | 14:16 |
| *** raildo is now known as raildo-afk | 14:16 | |
| lxsli | So it's a ComputeNode and its primitive is a (host, node) tuple? | 14:16 |
| bauzas | lxsli: okay, lemme rephrase | 14:17 |
| *** raildo-afk is now known as raildo | 14:17 | |
| bauzas | lxsli: I don't see much benefits of adding a node information for live-migrate and evacuate | 14:17 |
| bauzas | lxsli: so when you'd get the string, you'd figure out the related ComputeNode | 14:17 |
| bauzas | but if two are possible, then it would be an API error IMHO | 14:18 |
| bauzas | because Conflict | 14:18 |
| sdague | alex_xu: so I'm actually really confused why sending an extra param in v21 triggers a 501 - http://logs.openstack.org/85/214085/18/check/gate-tempest-dsvm-full/f7c7244/console.html#_2015-08-25_10_52_06_012 | 14:18 |
| lxsli | I can do that | 14:19 |
| bauzas | so that if Ironic would like to extend the possibility to specifically ask for a node, then it would microversion a new change which would accept a new body arg (or whatever) | 14:19 |
| bauzas | that would defer the discussion to when it would be needed | 14:19 |
| bauzas | here, no reasons to complicate stuff because X | 14:19 |
| lxsli | yep ok | 14:20 |
| *** xgsa has quit IRC | 14:21 | |
| *** agireud has joined #openstack-nova | 14:22 | |
| *** shuquan has quit IRC | 14:23 | |
| *** y_sawai has joined #openstack-nova | 14:24 | |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: Allow for migration object to be passed to _move_claim https://review.openstack.org/199679 | 14:25 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: rt: generalize claim code to be useful for other move actions https://review.openstack.org/197138 | 14:25 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: rt: move filtering of migration by type lower in the call stack https://review.openstack.org/199678 | 14:25 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: Make rebuild/evacuate use resource claims https://review.openstack.org/214770 | 14:25 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: compute: Make rebuild consider the node that was passed in https://review.openstack.org/214713 | 14:26 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: compute: Make rebuild_instance RPC take additional args https://review.openstack.org/200485 | 14:26 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: compute: Split the rebuild_instance method https://review.openstack.org/213461 | 14:26 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: db: Add the migration_context to the instance_extra table https://review.openstack.org/216731 | 14:26 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: objects: Add MigrationContext object https://review.openstack.org/216732 | 14:26 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: objects: Hook migration object into Instance https://review.openstack.org/216733 | 14:26 |
| *** venkat_p has joined #openstack-nova | 14:26 | |
| *** nkrinner has quit IRC | 14:26 | |
| *** shuquan has joined #openstack-nova | 14:26 | |
| dansmith | dammit ndipanov | 14:27 |
| ndipanov | as in dammit now I have so much awesome stuff to review I will probably have to work longer? :) | 14:28 |
| dansmith | I was working through that set | 14:28 |
| *** mdrabe has quit IRC | 14:28 | |
| dansmith | and had uncommitted things so now I don't get my -1 credit | 14:28 |
| ndipanov | ah so it's only adding patches and rebased the lower ones | 14:28 |
| ndipanov | no changes in the ones that were posted | 14:29 |
| ndipanov | I would normally not rebase | 14:29 |
| dansmith | I had comments on the bottom on | 14:29 |
| ndipanov | but there were conflicts :( | 14:29 |
| mriedem_class | you can still -1 | 14:30 |
| mriedem_class | "-1, see ps x' | 14:30 |
| dansmith | heh, yeah | 14:30 |
| dansmith | if I was really concerned | 14:30 |
| sdague | hmmm... - http://logs.openstack.org/85/214085/18/check/gate-tempest-dsvm-full/f7c7244/logs/screen-n-api.txt.gz#_2015-08-25_10_28_45_546 - what's up with the service group driver spam | 14:30 |
| *** ankit_ag has quit IRC | 14:31 | |
| dansmith | nice | 14:31 |
| *** tsekiyama has joined #openstack-nova | 14:32 | |
| mriedem_class | sdague: https://github.com/openstack/nova/commit/f05eecbc3fe3a93057e6ab6adb6743dc0f2cf328 | 14:32 |
| *** asd112z has joined #openstack-nova | 14:32 | |
| mtreinish | sdague: hah, what is the ServiceGroup driver defined as? | 14:33 |
| *** asd112z has quit IRC | 14:33 | |
| *** asd112z has joined #openstack-nova | 14:33 | |
| *** asd112z has quit IRC | 14:34 | |
| sdague | hey, jaypipes ^^^ | 14:34 |
| sdague | something else must have changed, because I don't remember those popping so often | 14:34 |
| *** asd112z has joined #openstack-nova | 14:34 | |
| *** xgsa has joined #openstack-nova | 14:35 | |
| sdague | ok, really trying to figure out why schema validation is getting translated to a 501 | 14:35 |
| lxsli | bauzas: re: the walkthrough, how do you expect a sample HTTP request to help enlighten the reader about the API entrypoint? | 14:35 |
| bauzas | lxsli: I was just thinking of all the if branches that we have | 14:35 |
| bauzas | lxsli: in general, I'm considering that giving a sample is always a good pattern for documentation | 14:36 |
| *** zhihai_song has quit IRC | 14:36 | |
| bauzas | lxsli: like for example, you can put very long text for describing a CLI, but giving examples is good enough | 14:36 |
| lxsli | bauzas: where it helps enlighten I'd agree, but I'm worried this would obscure instead, by putting attention on parameters instead of the call stack | 14:37 |
| jaypipes | mtreinish: what was yourquestion? | 14:37 |
| bauzas | lxsli: well, I'd rather it's the starting point - btw. did you made also the keystonemiddleware walkthru ? | 14:37 |
| mtreinish | jaypipes: I was reading that nova log and I just couldn't figure out what the ServiceGroup driver was defined as :) | 14:37 |
| lxsli | bauzas: nope, not me | 14:37 |
| jaypipes | oh, lol | 14:38 |
| jaypipes | mtreinish: I'll put in a quick patch to remove that line... | 14:38 |
| bauzas | lxsli: while I really like the idea, I think it needs support from all of us | 14:38 |
| *** ijuwang has joined #openstack-nova | 14:38 | |
| bauzas | lxsli: because it would be a pain for maintaining that | 14:38 |
| mtreinish | jaypipes: heh, thanks | 14:38 |
| bauzas | lxsli: so, it's all about the aim that we have for that document | 14:38 |
| bauzas | lxsli: and I defer that to johnthetubaguy | 14:38 |
| sdague | jaypipes: yeh, in moving from a singleton pattern to instances, the logging didn't translate so well :) | 14:38 |
| jaypipes | sdague: sorry mate...fixing now. | 14:39 |
| bauzas | lxsli: I mean, it's basically code walkthru | 14:39 |
| johnthetubaguy | lxsli: whats the context here? can I help? | 14:39 |
| lxsli | bauzas: I personally am happy for it to age disgracefully until it gets deleted/updated | 14:39 |
| lxsli | bauzas: better out-of-date than nothing and the big themes should remain true even as the code drifts | 14:39 |
| bauzas | johnthetubaguy: well, I just made you blinking because lxsli is documenting a code review | 14:39 |
| lxsli | it's about this I showed you the other day: https://review.openstack.org/#/c/210467/ | 14:40 |
| bauzas | johnthetubaguy: by saying 'code review', I'm not speaking of Gerrit, a real code review like in other companies :) | 14:40 |
| lxsli | bauzas is concerned about it getting out of date | 14:40 |
| bauzas | lxsli: I'm more concerned about the target we aim | 14:40 |
| bauzas | lxsli: and how we can maintain that | 14:41 |
| *** annegentle has quit IRC | 14:41 | |
| openstackgerrit | Jay Pipes proposed openstack/nova: Remove debug log message in SG API constructor https://review.openstack.org/216741 | 14:41 |
| bauzas | lxsli: not purely a stale stuff because yeah, it's fair enough to see it's worth it than nothing | 14:41 |
| jaypipes | mtreinish, sdague: feel free to fast-track ^^ | 14:41 |
| lxsli | bauzas: so what target would you like it to aim for? | 14:41 |
| johnthetubaguy | lxsli: honestly, I am a fan of longer term docs that point to the overall flow, so you can go find the current details | 14:41 |
| *** annegentle has joined #openstack-nova | 14:41 | |
| johnthetubaguy | lxsli: if its too hard to understand we got the code wrong (hence cells_v2 I guess) | 14:42 |
| bauzas | lxsli: I'd rather wonder how we could improve http://docs.openstack.org/developer/nova/devref/architecture.html (given jogo's changes) | 14:42 |
| lxsli | johnthetubaguy: if Python had a good IDE I might agree - when you have to grep, I think this sort of thing is more useful | 14:42 |
| bauzas | lxsli: well, I do have a good IDE | 14:43 |
| bauzas | (it's not Friday :p) | 14:43 |
| bauzas | so here is the thing | 14:43 |
| lxsli | can it find definitions through a "self.driver = CONF.sched_driver" ? | 14:43 |
| *** links has joined #openstack-nova | 14:43 | |
| lxsli | also through "rpc.cast(cctxt, 'blah')" | 14:44 |
| bauzas | lxsli: well, I don't think it's far different from any other project - you need to read code to understand | 14:44 |
| lxsli | Sure you can say that, but it took me maybe 3 days by now putting that together | 14:45 |
| lxsli | if someone can read it in 15mins, checking the links, that's a big win for them | 14:45 |
| bauzas | lxsli: hey, I'm working in Nova for 2 years now, and I'm still discovering bits | 14:45 |
| johnthetubaguy | lxsli: I am just thinking about the middle ground, or at least a way to make it easier to update, I guess | 14:46 |
| bauzas | johnthetubaguy: zactlyu | 14:46 |
| *** mdrabe has joined #openstack-nova | 14:46 | |
| lxsli | I'm worried that if you make it too easy to maintain, it'll lose the concreteness that makes it simple for a newcomer to grasp | 14:46 |
| bauzas | my take on that is "hell yeah, you get all my support for that, but I just don't want that to be just a code walk" | 14:46 |
| bauzas | lxsli: like for example, the links you provided will be outdated just by the next change impacting the file | 14:47 |
| bauzas | lxsli: so, by some extend, the reader has to find those methods by his own | 14:47 |
| lxsli | I can put SHAs in those, and a warning at the top saying it was written for Liberty and may be inaccurate for future versions | 14:47 |
| johnthetubaguy | lxsli: thats true, but a new comer having a list of broken links is bad too | 14:47 |
| johnthetubaguy | lxsli: +1 for SHA in the links, thats a good start | 14:48 |
| bauzas | lxsli: I don't think it really helps here, because the guy couldn't understand why this specific method | 14:48 |
| markus_z | dims__: Are you available? | 14:48 |
| johnthetubaguy | lxsli: what about, here is the overall idea, now lets look through this example of how it worked for this particular point in time, as a handy starting point? | 14:49 |
| dims__ | markus_z: pong | 14:49 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova: Add functional test for server group https://review.openstack.org/147514 | 14:49 |
| bauzas | lxsli: johnthetubaguy: I just feel we need to write some design docs, exactly like any other proprietary software could have | 14:49 |
| lxsli | bauzas: yes but I'm not volunteering | 14:49 |
| openstackgerrit | Radomir Dopieralski proposed openstack/nova: Gracefully handle 404 responses from neutron when polling for floating IPs https://review.openstack.org/216746 | 14:50 |
| markus_z | dims__: You do a lot in oslo, am I right? I don't find a description how their blueprint process looks like. | 14:50 |
| lxsli | bauzas: I wrote a walkthrough because I was reading the code myself so it was convenient | 14:50 |
| lxsli | bauzas: if you don't like the result, I can abandon it | 14:50 |
| lxsli | johnthetubaguy's suggestion seems reasonable but I'm not writing something completely different | 14:51 |
| dims__ | markus_z: don't really have anything written down, any specific angle or information you need? | 14:51 |
| bauzas | lxsli: here is what I would love to have it updated http://bodenr.blogspot.fr/2014/03/openstack-nova-boot-server-call-diagram.html | 14:52 |
| lxsli | bauzas: I look forward to reviewing your patch :) | 14:52 |
| *** Murali has quit IRC | 14:52 | |
| *** shoutm has quit IRC | 14:52 | |
| dims__ | markus_z: here's one thing we have - http://specs.openstack.org/openstack/oslo-specs/specs/policy/spec-approval.html | 14:53 |
| *** diogogmt has joined #openstack-nova | 14:53 | |
| bauzas | lxsli: I don't get it, you don't prefer to have diagrams instead of long texts ? | 14:53 |
| *** rotbart has quit IRC | 14:53 | |
| lxsli | bauzas: the perfect has killed the good | 14:54 |
| *** amotoki has joined #openstack-nova | 14:54 | |
| bauzas | lxsli: okay, what I could say ? update your change, put your SHA1s, and leave others comment on it | 14:54 |
| *** angdraug has joined #openstack-nova | 14:54 | |
| *** achanda has joined #openstack-nova | 14:54 | |
| markus_z | dims__: I think that pointer covers what I need, thanks! | 14:54 |
| lxsli | bauzas: I didn't understand that | 14:56 |
| lxsli | bauzas: I can restore it or leave it, I'm easy | 14:56 |
| *** Nikolay_St has quit IRC | 14:56 | |
| lxsli | bauzas: if you want me to restore it, you can patch on top of it | 14:56 |
| *** venkat_p has quit IRC | 14:57 | |
| bauzas | lxsli: well, it's opensource, so leave your change, make sure your links are not wrong, and leave it to be reviewed | 14:57 |
| *** venkat_p has joined #openstack-nova | 14:57 | |
| bauzas | lxsli: if you're telling me what I should do, I appreciate | 14:57 |
| *** venkat_p has quit IRC | 14:57 | |
| lxsli | OK I'll do that | 14:57 |
| *** xgsa has quit IRC | 14:57 | |
| *** rotbeard has joined #openstack-nova | 14:58 | |
| *** exploreshaifali has quit IRC | 14:59 | |
| *** zzzeek has joined #openstack-nova | 14:59 | |
| *** vivekd has joined #openstack-nova | 15:00 | |
| openstackgerrit | lyanchih proposed openstack/nova: libvirt - Unrescue will not remove rescue disk when image_type=rbd https://review.openstack.org/205766 | 15:00 |
| *** armax has joined #openstack-nova | 15:01 | |
| *** shuquan has quit IRC | 15:02 | |
| *** alexschm has quit IRC | 15:02 | |
| *** mtanino has joined #openstack-nova | 15:02 | |
| *** hemnafk is now known as hemna | 15:03 | |
| *** achanda has quit IRC | 15:04 | |
| *** rushil has joined #openstack-nova | 15:04 | |
| *** thedodd has joined #openstack-nova | 15:04 | |
| *** xgsa has joined #openstack-nova | 15:04 | |
| openstackgerrit | Ed Leafe proposed openstack/nova: Remove 'v3' from extension code https://review.openstack.org/214311 | 15:05 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Remove v3 references in unit test 'contrib' https://review.openstack.org/212851 | 15:05 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Remove more 'v3' references from the code https://review.openstack.org/214290 | 15:05 |
| *** vivekd has quit IRC | 15:07 | |
| *** dustins has quit IRC | 15:08 | |
| *** sudipto has joined #openstack-nova | 15:08 | |
| *** haomaiwang has quit IRC | 15:09 | |
| *** mnestratov|2 has quit IRC | 15:10 | |
| sudipto | dansmith, Hi, getting errors from Jenkins on removal of the NovaObjectDictCompat from the agent object. Something that suggests that certain properties are not initialized....http://logs.openstack.org/71/216371/1/check/gate-nova-python27/6975550/console.html | 15:10 |
| *** haomaiwang has joined #openstack-nova | 15:10 | |
| *** mnestratov has joined #openstack-nova | 15:10 | |
| sudipto | dansmith, suggestions on the best way to deal with this? | 15:10 |
| *** gszasz has quit IRC | 15:12 | |
| ndipanov | dansmith, so are you cool with leaving the migration type default6ing as it is? | 15:12 |
| *** dave-mcc_ has quit IRC | 15:12 | |
| openstackgerrit | lalit dagre proposed openstack/nova: WIP: Flavor create and Flavor get_* methods support both DB https://review.openstack.org/213041 | 15:13 |
| *** ahilan has quit IRC | 15:13 | |
| dansmith | sudipto: just a sec | 15:13 |
| dansmith | ndipanov: did you reply to me? | 15:13 |
| ndipanov | dansmith, probably didn't post it :/ | 15:14 |
| *** fawadkhaliq has quit IRC | 15:15 | |
| *** Kevin_Zheng_ has quit IRC | 15:15 | |
| *** xgsa has quit IRC | 15:16 | |
| dansmith | sudipto: you have a bug on line 39 of your agent object | 15:16 |
| dansmith | sudipto: you're only ever setting the name attribute, so all the others are unset and thus trying to lazy-load when they're used | 15:16 |
| openstackgerrit | jichenjc proposed openstack/nova: Check host status before scheduler https://review.openstack.org/216759 | 15:17 |
| openstackgerrit | Tomoki Sekiyama proposed openstack/nova: Ironic: Support boot from Cinder volume https://review.openstack.org/215385 | 15:17 |
| sudipto | dansmith, darn, thanks figured! | 15:17 |
| sudipto | dansmith, thought name was the 'name' of the property - didn't notice the loop before that. Thanks! | 15:18 |
| dansmith | sudipto: my bug finding service is $30/hr, minimum 15 minute increments | 15:18 |
| sudipto | dansmith, ahem ok :D | 15:18 |
| dansmith | heh | 15:18 |
| ndipanov | dansmith, I did now | 15:18 |
| dansmith | ndipanov: okay, reading your comment is prioritized just under "get breakfast" | 15:19 |
| ndipanov | lol | 15:19 |
| *** geekinutah has quit IRC | 15:19 | |
| *** venkat_p has joined #openstack-nova | 15:21 | |
| openstackgerrit | Alexis Lee proposed openstack/nova: Walkthrough from API to virt driver https://review.openstack.org/210467 | 15:21 |
| *** emagana has quit IRC | 15:22 | |
| *** dannywilson has joined #openstack-nova | 15:22 | |
| lxsli | dansmith: do you take credit cards? | 15:23 |
| dansmith | I take paypal | 15:23 |
| *** rajesht has quit IRC | 15:24 | |
| *** Zhang has joined #openstack-nova | 15:25 | |
| *** scheuran has quit IRC | 15:26 | |
| *** dave-mcc_ has joined #openstack-nova | 15:26 | |
| openstackgerrit | Sudipta Biswas proposed openstack/nova: cleanup NovaObjectDictCompat from agent https://review.openstack.org/216371 | 15:27 |
| *** fawadkhaliq has joined #openstack-nova | 15:27 | |
| *** pbrooko_ has joined #openstack-nova | 15:27 | |
| *** markus_z has quit IRC | 15:28 | |
| *** moshele has quit IRC | 15:28 | |
| *** davideagnello has joined #openstack-nova | 15:29 | |
| *** xgsa has joined #openstack-nova | 15:29 | |
| *** pbrooko has quit IRC | 15:29 | |
| dansmith | ndipanov: I replied, but ... just to be clear: is the plan to leave move_type=None, but always set it based on the _determine helper if it's =None? | 15:30 |
| mhorban | sdague: Hello, could you please re-examine https://review.openstack.org/#/c/195303/13 | 15:30 |
| ndipanov | dansmith, yeah | 15:30 |
| ndipanov | see a couple of patches up | 15:30 |
| ndipanov | I was calling _determine in the RT | 15:31 |
| *** thangp has quit IRC | 15:31 | |
| *** garyk has quit IRC | 15:31 | |
| dansmith | right, right, cool | 15:31 |
| mriedem_class | apmelton: heh, only 57 test failures with libvirt lxc + lvm in tempest dsvm | 15:31 |
| mriedem_class | http://logs.openstack.org/33/215733/3/check/gate-tempest-dsvm-full/50241c8/console.html#_2015-08-24_20_14_23_489 | 15:31 |
| ndipanov | as for adding the usage of those params and the RPC bump in one patch - that's gonna be tough | 15:32 |
| ndipanov | adding the funcionality is like 3 patches to split up the method etc | 15:32 |
| apmelton | mriedem_class: that job is running libvirt-lxc as the libvirt-type? | 15:32 |
| ndipanov | I think I'll just shuffle those around dansmith | 15:32 |
| mriedem_class | apmelton: yeah | 15:32 |
| apmelton | neat | 15:32 |
| dansmith | ndipanov: well, I think that adding the rpc version before we honor them is not a good thing to do | 15:32 |
| *** Zhang has quit IRC | 15:32 | |
| mriedem_class | http://logs.openstack.org/33/215733/3/check/gate-tempest-dsvm-full/50241c8/logs/etc/nova/nova.conf.txt.gz | 15:32 |
| mriedem_class | apmelton: ^ | 15:33 |
| mriedem_class | virt_type = lxc images_volume_group = stack-volumes-default images_type = lvm | 15:33 |
| dansmith | ndipanov: how about you pull that out to a helper that takes the args, add all the functionality, tests, etc, and then one patch after that to expose them over rpc? | 15:33 |
| ndipanov | dansmith, yeah I guess I agree | 15:33 |
| *** ildikov has quit IRC | 15:33 | |
| ndipanov | that's what I meant by shuffle them around | 15:33 |
| dansmith | okay cool | 15:33 |
| *** davideagnello has quit IRC | 15:34 | |
| mriedem_class | apmelton: now it's just a matter of sorting through the failures and toggling tempest to not run those, either via config options in tempest or blacklisting them with a regex | 15:34 |
| *** zhenguo has quit IRC | 15:36 | |
| *** Zhang has joined #openstack-nova | 15:36 | |
| *** rushil has quit IRC | 15:36 | |
| *** rushil has joined #openstack-nova | 15:37 | |
| *** geekinutah has joined #openstack-nova | 15:37 | |
| mriedem_class | apmelton: looks like mostly snapshots and network stuff, not surprising | 15:37 |
| apmelton | yea, I don't think snapshots work at all | 15:37 |
| mriedem_class | it also looks like tempest was timing out on deleting servers so we eventually go overquota | 15:38 |
| apmelton | mriedem_class: I did some digging on what was causing the network to die | 15:38 |
| mriedem_class | so there might actually be fewer failures than i'm seeing here | 15:38 |
| mriedem_class | orly | 15:38 |
| apmelton | mriedem_class: I can talk in a little bit, in a meeting at the moment | 15:38 |
| mriedem_class | apmelton: sure, or post in the etherpad under notes or whatever | 15:38 |
| mriedem_class | https://etherpad.openstack.org/p/nova-lxc-ci | 15:38 |
| *** thangp has joined #openstack-nova | 15:39 | |
| apmelton | actually, just wrapped up | 15:39 |
| dansmith | sudipto: looks better :) | 15:40 |
| apmelton | mriedem_class: so, what we noticed was the br100 bridge that gets spun up by nova-network starts with eth0's mac address, like it should | 15:40 |
| apmelton | mriedem_class: and then, as libvirt starts the instance, some how the mac on br100 gets changed to the mac of the instance's port on the bridge | 15:41 |
| apmelton | and thus, the mac spoofing rules on the underlying hypervisor start blocking traffic | 15:41 |
| mriedem_class | weird, libvirt issue or nova issue? | 15:41 |
| apmelton | haven't figured it out yet | 15:41 |
| sudipto | dansmith, :) | 15:42 |
| apmelton | mriedem_class: are you actually seeing the network issue in jenkins? | 15:42 |
| *** Marga_ has quit IRC | 15:42 | |
| *** Marga_ has joined #openstack-nova | 15:43 | |
| mriedem_class | apmelton: i haven't gotten far enough yet, just saw floating ips tests failing | 15:43 |
| apmelton | gotcha | 15:43 |
| *** meghal has joined #openstack-nova | 15:43 | |
| mriedem_class | in that devstack change i had to basically neuter cells and neutron runs | 15:43 |
| *** lsmola has quit IRC | 15:44 | |
| mriedem_class | https://review.openstack.org/#/c/215733/ | 15:44 |
| *** links has quit IRC | 15:44 | |
| mriedem_class | because with neutron, jenkins can't even post results devstack fails so hard | 15:44 |
| mriedem_class | so it was in a perpetual retry loop all weekend | 15:44 |
| apmelton | dang | 15:44 |
| mriedem_class | so if you guys are using lxc + neutron-something at rax and have special patches, push them up :) | 15:45 |
| *** thangp has quit IRC | 15:46 | |
| apmelton | mriedem_class: I don't think we have any patches to nova to get get neutron working | 15:47 |
| mriedem_class | do you guys only run nova + lxc with nova-network at rax? | 15:48 |
| *** meghal has left #openstack-nova | 15:48 | |
| *** meghal has joined #openstack-nova | 15:49 | |
| apmelton | no, so to hook into our custom quark stuff, we're setting up the interfaces outside of nova and libvirt via libvirt hooks | 15:49 |
| apmelton | it's definitely not ideal | 15:49 |
| *** VW__ has joined #openstack-nova | 15:50 | |
| apmelton | I don't think we're doing anything really special though, as long as the neutron agent is handling that, it should work | 15:50 |
| mriedem_class | apmelton: hmm, DiskFilter might be wonking things in the scheduler with the lvm-backend | 15:51 |
| mriedem_class | http://logs.openstack.org/33/215733/3/check/gate-tempest-dsvm-full/50241c8/logs/screen-n-sch.txt.gz#_2015-08-24_20_06_44_105 | 15:51 |
| *** vivekd has joined #openstack-nova | 15:52 | |
| mriedem_class | maybe not, it's the ComputeFilter that fails | 15:52 |
| mriedem_class | edleafe: btw, nice filtering logs added there ^ | 15:52 |
| mriedem_class | to see which filter fails | 15:52 |
| annegentle | Any one know why doc/source/api is in the .gitignore file for nova? | 15:52 |
| *** VW___ has joined #openstack-nova | 15:53 | |
| openstackgerrit | jichenjc proposed openstack/nova: Catch OverQuota in volume create function https://review.openstack.org/173985 | 15:53 |
| *** VW__ has quit IRC | 15:53 | |
| apmelton | mriedem_class: is that running with the BlockMount patch as well? | 15:53 |
| mriedem_class | yeah | 15:53 |
| mriedem_class | via depends-on | 15:53 |
| apmelton | ah gotcha | 15:53 |
| *** VW___ has quit IRC | 15:53 | |
| mriedem_class | annegentle: because that's automatically generated by the docs build | 15:53 |
| *** VW_ has quit IRC | 15:54 | |
| annegentle | mriedem_class: ah it's "internal" api docs for python modules? | 15:54 |
| mriedem_class | tox -e docs will create doc/source/api with all of the autoindex modules | 15:54 |
| mriedem_class | yeah | 15:54 |
| *** VW_ has joined #openstack-nova | 15:54 | |
| annegentle | got it. wonder why swift has doc/source/api? | 15:54 |
| * annegentle investigates further | 15:54 | |
| mriedem_class | i think some projects actually have their autoindex docs in the modules in source | 15:54 |
| mriedem_class | rather than generated automatically | 15:54 |
| annegentle | yeah that makes sense for swift actually | 15:54 |
| mriedem_class | apmelton: this is really weird, i'm seeing novalidhost scheduling failures, but n-cpu logs are clean | 15:55 |
| mriedem_class | not sure where things are actually piling up and tipping over | 15:55 |
| *** changbl has joined #openstack-nova | 15:55 | |
| *** cfarquhar has joined #openstack-nova | 15:57 | |
| *** cfarquhar has joined #openstack-nova | 15:57 | |
| openstackgerrit | lyanchih proposed openstack/nova: libvirt - Flavor's quota instance resource not observed for non-ephemeral storage https://review.openstack.org/201019 | 15:57 |
| apmelton | mriedem_class: if i'd like to replicate this tempest run against my devstack, how'd I go about that? | 15:58 |
| mriedem_class | apmelton: well i guess start with this https://review.openstack.org/#/c/215733/ | 15:58 |
| mriedem_class | and set NOVA_BRANCH=refs/changes/11/215311/4 in stackrc | 15:58 |
| mriedem_class | to pull down that dependent change | 15:58 |
| mriedem_class | i think you have to enable the tempest service in local.conf | 15:59 |
| *** Nikolay_St has joined #openstack-nova | 15:59 | |
| *** davideagnello has joined #openstack-nova | 16:00 | |
| *** apuimedo is now known as apuimedo|away | 16:02 | |
| *** takedakn has quit IRC | 16:03 | |
| *** jichen has quit IRC | 16:03 | |
| sdague | dansmith: so, any idea why johnthetubaguy' | 16:04 |
| *** Apoorva has joined #openstack-nova | 16:04 | |
| dansmith | no I have no idea | 16:04 |
| sdague | s patch on paste changes is tickling this in nova-compute - http://logs.openstack.org/85/214085/18/check/gate-tempest-dsvm-full/f7c7244/logs/screen-n-cpu.txt.gz?level=TRACE#_2015-08-25_10_31_23_394 | 16:04 |
| sdague | yeh, sorry, enter key fail | 16:04 |
| dansmith | just a sec | 16:05 |
| sdague | we're getting an unexpected 501 back when passing an extra param to the API, however, the fail is all the way down at the compute host | 16:05 |
| mriedem_class | apmelton: http://logs.openstack.org/33/215733/3/check/gate-tempest-dsvm-full/50241c8/logs/screen-n-cpu.txt.gz#_2015-08-24_20_03_20_485 shows it's using your lvm patch | 16:05 |
| apmelton | nice | 16:05 |
| *** hop has quit IRC | 16:05 | |
| johnthetubaguy | sdague: oh, that looks a bit like the magic maintenance mode stuff... | 16:06 |
| sdague | johnthetubaguy: that's the only reason your patch is failing | 16:06 |
| sdague | which looks like a legit bug somewhere that's gotten exposed now | 16:06 |
| johnthetubaguy | sdague: oh, freaky, maybe the 'corrected' invalid input triggered that somehow, yeah, needs a dig | 16:07 |
| sdague | yeh, so I think we failed early before | 16:07 |
| sdague | now we relaxed a thing | 16:07 |
| sdague | and it got to the compute, but exploded | 16:07 |
| johnthetubaguy | sdague: thats only implemented in XenAPI, or only in a few drivers anyways | 16:08 |
| johnthetubaguy | just wondering why it go there | 16:08 |
| *** aix has quit IRC | 16:08 | |
| sdague | johnthetubaguy: because before we were rejecting the request very early | 16:08 |
| *** xgsa has quit IRC | 16:08 | |
| *** Zhang has quit IRC | 16:09 | |
| apmelton | mriedem_class: running tempest, it looks like the n-cpu process has simply frozen | 16:09 |
| sdague | because of some half assed validation on the v2.0 side | 16:09 |
| sdague | with the 2.0 on 2.1 we dropped a bunch of validation | 16:09 |
| sdague | so in this case the request gets further | 16:09 |
| apmelton | and then n-sch stops sending instances because it's heartbeat is too old | 16:09 |
| *** haomaiwang has quit IRC | 16:09 | |
| johnthetubaguy | sdague: yeah, nasty | 16:09 |
| sdague | ok, so, I guess nova is doing the right thing | 16:10 |
| sdague | because this is a xen only feature | 16:10 |
| *** haomaiwa_ has joined #openstack-nova | 16:10 | |
| dansmith | yeah | 16:10 |
| edleafe | mriedem_class: glad the filter logging helped :) | 16:10 |
| johnthetubaguy | sdague: have you got a link to the request tempest is making | 16:11 |
| sdague | ok, so I'll work with mtreinish to figure out the right disable path here | 16:11 |
| sdague | yeh, one sec | 16:11 |
| sdague | johnthetubaguy: http://logs.openstack.org/85/214085/18/check/gate-tempest-dsvm-full/f7c7244/logs/screen-n-api.txt.gz#_2015-08-25_10_31_23_374 | 16:12 |
| *** alex_klimov has quit IRC | 16:12 | |
| sdague | slightly cryptic | 16:12 |
| sdague | but that's the request body | 16:12 |
| *** marzif has quit IRC | 16:12 | |
| mriedem_class | apmelton: yeah, n-cpu just dies off after deleting a couple of instances | 16:14 |
| johnthetubaguy | sdague: it might just be we need to exclude that test for libvirt that doesn't implement the API, but it seems odd we had that test on for so long before | 16:14 |
| *** pbrooko_ has quit IRC | 16:14 | |
| *** yamahata has quit IRC | 16:14 | |
| sdague | johnthetubaguy: well, it was effectively hitting our old validation code early enough that it failed on all platforms | 16:14 |
| johnthetubaguy | sdague: yeah, I guess its odd we only have a negative test, but maybe thats all we have | 16:15 |
| sdague | so, possitive tests here are really hard to put into tempest | 16:15 |
| *** SourabhP has joined #openstack-nova | 16:15 | |
| sdague | because they actually turn off parts of your infrastructure | 16:15 |
| sdague | and then all kinds of weirdness happens | 16:15 |
| sdague | which is why this is only negative tested | 16:15 |
| *** vivekd has quit IRC | 16:16 | |
| johnthetubaguy | sdague: oh right, good point | 16:16 |
| sdague | I agree, it's a very weird set of edge cases here | 16:16 |
| sdague | I'll work with mtreinish to figure out the right fixes | 16:16 |
| openstackgerrit | Simona Iuliana Toader proposed openstack/nova: Hyper-V: Adds Hyper-V UEFI Secure Boot https://review.openstack.org/209581 | 16:16 |
| openstackgerrit | Simona Iuliana Toader proposed openstack/nova: Hyper-V: refactors utilsfactory https://review.openstack.org/214527 | 16:16 |
| johnthetubaguy | sdague: cool, thank you, yeah, I think its just not supported by libvirt, due to the "maintenance_mode": "enable" being present | 16:17 |
| *** jistr has quit IRC | 16:17 | |
| johnthetubaguy | sdague: like you say, in the past we would reject due to "param": "XXX" | 16:17 |
| johnthetubaguy | sdague: one other, crazy, option is to turn on strict checking for this API when using the v2.1 legacy mode? | 16:18 |
| *** davideagnello has quit IRC | 16:18 | |
| *** jlk has left #openstack-nova | 16:18 | |
| mriedem_class | apmelton: looks like there is supposed to be a guru meditation report in the n-cpu logs but i don't see one | 16:18 |
| mriedem_class | that might have been helpful | 16:18 |
| johnthetubaguy | sdague: like add legacy_is_strict: True into the schema or something, but that feels really dirty | 16:18 |
| sdague | yeh, honestly, this is way off in the deep woods | 16:19 |
| johnthetubaguy | sdague: fair point | 16:19 |
| *** tonytan4ever has quit IRC | 16:20 | |
| *** SourabhP has quit IRC | 16:20 | |
| *** takedakn has joined #openstack-nova | 16:20 | |
| mriedem_class | apmelton: should devstack setup with lxc be running nova-idmapshift? | 16:20 |
| apmelton | mriedem_class: no | 16:20 |
| apmelton | mriedem_class: that's only necessary if you're running with user namespaces | 16:21 |
| *** takedakn has quit IRC | 16:22 | |
| *** smatzek has quit IRC | 16:22 | |
| *** vladikr has quit IRC | 16:23 | |
| apmelton | mriedem_class: it looks like libvirt my even be frozen | 16:23 |
| *** diogogmt has quit IRC | 16:24 | |
| mriedem_class | yeah | 16:24 |
| mriedem_class | http://logs.openstack.org/33/215733/3/check/gate-tempest-dsvm-full/50241c8/logs/libvirt/libvirtd.txt.gz#_2015-08-24_20_03_28_365 | 16:24 |
| mriedem_class | it looks like it just hangs up trying to delete instance 2 | 16:24 |
| *** vivekd has joined #openstack-nova | 16:25 | |
| *** diogogmt has joined #openstack-nova | 16:25 | |
| *** bnemec has joined #openstack-nova | 16:25 | |
| *** tonytan4ever has joined #openstack-nova | 16:26 | |
| apmelton | mriedem_class: I wonder if we've got too much going on in the volume group | 16:26 |
| apmelton | with nova trying to zero out instance disks | 16:26 |
| apmelton | I've noticed that when disks get overloaded, instances don't die fast enough for libvirt and things end up in a bad state | 16:26 |
| *** exploreshaifali has joined #openstack-nova | 16:27 | |
| *** marzif has joined #openstack-nova | 16:28 | |
| *** ByPasS has quit IRC | 16:28 | |
| mriedem_class | apmelton: ok, i can rebase my devstack change on top of ccarmack's that adds the NOVA_VOLUME_CLEAR flag | 16:28 |
| *** unicell has joined #openstack-nova | 16:30 | |
| *** moshele has joined #openstack-nova | 16:30 | |
| *** sachin has joined #openstack-nova | 16:30 | |
| *** unicell1 has quit IRC | 16:31 | |
| *** ByPasS has joined #openstack-nova | 16:31 | |
| *** unicell has quit IRC | 16:31 | |
| *** matrohon has quit IRC | 16:32 | |
| *** bkopilov has joined #openstack-nova | 16:32 | |
| *** VW__ has joined #openstack-nova | 16:33 | |
| *** nic has joined #openstack-nova | 16:33 | |
| *** eglynn has joined #openstack-nova | 16:35 | |
| dansmith | sdague: sorry :( https://review.openstack.org/#/c/147514/47 | 16:35 |
| *** ftersin has quit IRC | 16:35 | |
| *** ftersin has joined #openstack-nova | 16:35 | |
| *** rushil has quit IRC | 16:37 | |
| *** VW_ has quit IRC | 16:37 | |
| *** Apoorva has quit IRC | 16:37 | |
| *** moshele has quit IRC | 16:39 | |
| *** yamahata has joined #openstack-nova | 16:40 | |
| *** __TheDodd__ has joined #openstack-nova | 16:43 | |
| *** thedodd has quit IRC | 16:43 | |
| sdague | dansmith: ok, well I'm going to -1 bauzas patch with my explaination from irc yesterday | 16:45 |
| bauzas | sdague: heya, which one ? :) | 16:45 |
| dansmith | I think tab fail | 16:45 |
| dansmith | sdague: I commented that you would | 16:45 |
| sdague | bauzas: https://review.openstack.org/#/c/147514/47 | 16:45 |
| dansmith | sdague: that's not bauzas, it's gibi | 16:46 |
| bauzas | sdague: oh that one, I need to review it but it's not my son :) | 16:46 |
| bauzas | the servergroup soft affinity thing is interesting | 16:46 |
| sdague | oh, sorry | 16:46 |
| *** amotoki has quit IRC | 16:46 | |
| sdague | you know, pattern matching on Ba.* | 16:46 |
| bauzas | hah | 16:46 |
| *** sahid has quit IRC | 16:47 | |
| * bauzas looking for creating a patent on Ba* | 16:47 | |
| mriedem_class | apmelton: should the glance image have the vm_mode property set to 'exe' for lxc? | 16:47 |
| bauzas | would prevent guys to reproduce it | 16:47 |
| sudipto | dansmith, one more doubt, https://github.com/openstack/nova/blob/master/nova/virt/firewall.py#L547 - this line is trying to add a fw rule - assuming every field in the security_group_rule obj is a string - however, the cidr field is a IPNetwork obj - so i guess this place would need a change? | 16:47 |
| dansmith | sudipto: str(ipnetwork) works out to the proper cidr representation, so it should be okay, right? | 16:48 |
| mriedem_class | that kind of str(network object) thing happens all over the nova api before returning a response | 16:49 |
| *** adam_g has joined #openstack-nova | 16:49 | |
| *** adam_g has quit IRC | 16:49 | |
| *** adam_g has joined #openstack-nova | 16:49 | |
| sudipto | dansmith, I am getting a TypeError though...on that same line - sequence item 3: expected string or Unicode, IPNetwork found | 16:50 |
| dansmith | sudipto: well, might need to coerce it then, depending on what is doing the check | 16:50 |
| dansmith | sudipto: is this related to a dict conversion? | 16:50 |
| sudipto | dansmith, yeah | 16:50 |
| *** lyanchih has quit IRC | 16:51 | |
| sudipto | dansmith, tests were using the cidr as a string value before... | 16:51 |
| dansmith | sudipto: I'm surprised that such a change would trigger a behavioral difference like this.. did you remove some special handling? | 16:51 |
| *** romainh has left #openstack-nova | 16:51 | |
| mriedem_class | apmelton: nevermind, i guess vm_mode='exe' wouldn't matter | 16:51 |
| mriedem_class | i remember that being needed for the parallels virt type | 16:51 |
| sudipto | dansmith, https://review.openstack.org/#/c/201282/3/nova/objects/security_group_rule.py | 16:52 |
| *** rushil has joined #openstack-nova | 16:53 | |
| *** ssurana has joined #openstack-nova | 16:53 | |
| dansmith | sudipto: L324 | 16:53 |
| dansmith | well, | 16:53 |
| dansmith | hmm | 16:53 |
| *** Apoorva has joined #openstack-nova | 16:53 | |
| *** rushil has quit IRC | 16:54 | |
| sudipto | dansmith, you meant the firewall file? | 16:54 |
| *** rushil has joined #openstack-nova | 16:54 | |
| apmelton | mriedem_class: yea, I think the switches in the libvirt driver hard code the exe setup in the xml | 16:55 |
| sudipto | dansmith, https://github.com/openstack/nova/blob/master/nova/virt/firewall.py#L392 | 16:55 |
| dansmith | sudipto: L519, put str() around rule.cidr | 16:56 |
| sudipto | dansmith, yeah sure... | 16:56 |
| dansmith | I dunno why it matters now though | 16:56 |
| mriedem_class | apmelton: yup, just found that | 16:57 |
| sudipto | dansmith, yeah i was a bit irked... | 16:57 |
| mriedem_class | apmelton: not sure if this helps you with the libvirt network stuff with lxc https://libvirt.org/drvlxc.html#securenetworking | 16:57 |
| dansmith | sudipto: yeah, seems like it should be failing before your change if it's going to fail | 16:58 |
| dansmith | sudipto: but assuming join() is being pedantic about what you're joining it with, it makes sense to have to coerce it before the join | 16:58 |
| sudipto | dansmith, yeah - these dict compat changes are also requiring conversion of tests to use objects instead of returning dicts. | 16:59 |
| dansmith | yeah | 16:59 |
| *** meghal has quit IRC | 16:59 | |
| *** derekh has quit IRC | 16:59 | |
| *** meghal has joined #openstack-nova | 17:00 | |
| apmelton | mriedem_class: interesting, we may want to expose that through nova.conf, or simply require it if lxc is enabled | 17:00 |
| apmelton | I can't imagine anything deploying nova would ever want lxc instances in the host's network namespace | 17:00 |
| *** venkat_p has quit IRC | 17:00 | |
| *** meghal has quit IRC | 17:01 | |
| *** meghal has joined #openstack-nova | 17:01 | |
| mriedem_class | apmelton: right :) | 17:01 |
| mriedem_class | same with filesystem isolation | 17:01 |
| mriedem_class | i think the openstack docs do say you don't want to use lxc + nova in production b/c of security | 17:02 |
| apmelton | mriedem_class: not sure that actually makes sense in the nova use case of libvirt... unless you can some how create an instance without a backing image/volume | 17:02 |
| dansmith | apmelton: yeah, because then you could potentially tcpdump on the host interface, which would be like bad and stuff | 17:03 |
| bauzas | jaypipes: food for thoughts before going to dinner https://review.openstack.org/#/c/213590/4 | 17:03 |
| bauzas | jaypipes: any idea why we shouldn't claim against vcpus ? | 17:03 |
| apmelton | mriedem_class: I honestly think the network issue I'm hitting is on libvirt or nova's side of things | 17:04 |
| bauzas | jaypipes: you actually hitted something that I missed before and now I'm just... torn | 17:04 |
| apmelton | if the mac from the instance's namespace was landing on br100, that would be an issue | 17:04 |
| *** karimb has quit IRC | 17:04 | |
| apmelton | but it's the mac from the hosts vnet0 that lands on br100 | 17:04 |
| apmelton | what's also odd is that if I force the mac on br100 back to eth0's mac, it stays that way | 17:05 |
| *** ftersin has left #openstack-nova | 17:06 | |
| bauzas | dansmith: good points on reviewing https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/allocation-ratio-to-resource-tracker,n,z but I'll need to think about that during dinner | 17:07 |
| dansmith | bauzas: I'll allow it | 17:07 |
| bauzas | dansmith: I just wonder if it's good to hide that if logic by having the fields defaulted to the conf values | 17:07 |
| mriedem_class | apmelton: interesting https://review.opencontrail.org/#/c/2811/1 | 17:07 |
| dansmith | bauzas: you're defaulting the values now, except you have to default them in every filter until the end of time | 17:08 |
| dansmith | bauzas: the whole point of the object facade is to hide changing schema underneath the rest of nova | 17:08 |
| *** haomaiwa_ has quit IRC | 17:09 | |
| *** xgsa has joined #openstack-nova | 17:10 | |
| *** haomaiwang has joined #openstack-nova | 17:10 | |
| *** unicell has joined #openstack-nova | 17:11 | |
| *** unicell1 has joined #openstack-nova | 17:13 | |
| *** unicell has quit IRC | 17:13 | |
| *** mestery has quit IRC | 17:14 | |
| *** nic1 has joined #openstack-nova | 17:14 | |
| *** nic has quit IRC | 17:15 | |
| *** achanda has joined #openstack-nova | 17:15 | |
| *** yanxb has quit IRC | 17:15 | |
| *** yanxb has joined #openstack-nova | 17:16 | |
| *** ferest has joined #openstack-nova | 17:19 | |
| GnomeKing | johnthetubaguy: ping? | 17:20 |
| GnomeKing | johnthetubaguy: Regarding your reviews (thanks) - got a moment so I can advise the team properly tomorrow? | 17:20 |
| *** sudipto has quit IRC | 17:21 | |
| *** sudipto has joined #openstack-nova | 17:22 | |
| johnthetubaguy | GnomeKing: sure, I can try, which review? | 17:22 |
| GnomeKing | Well - two questions... 1) Deprecating options: If the option is nonsensical should we just leave it in and tell people it's ignored? | 17:23 |
| GnomeKing | In this case the option cannot be used in the new implementation and it will automatically use the correct behaviour | 17:23 |
| *** mestery has joined #openstack-nova | 17:23 | |
| johnthetubaguy | GnomeKing: yeah, that could work, deprecated for removal, and well, just ignored because they are crazy | 17:24 |
| *** liusheng has quit IRC | 17:24 | |
| GnomeKing | That q was for https://review.openstack.org/#/c/213112/ | 17:24 |
| GnomeKing | OK - so we'll leave it present, mark as deprecated and have the comment saying that it's nonsensical. | 17:25 |
| *** davideagnello has joined #openstack-nova | 17:25 | |
| *** liusheng has joined #openstack-nova | 17:25 | |
| johnthetubaguy | GnomeKing: I think thats the best way, we basically say you should be able to upgrade without changing your config, thats where the rule comes from | 17:26 |
| GnomeKing | For https://review.openstack.org/#/c/205831/ you commented that we should track it somehow but didn't vote :) Do you have a suggestion for this? How are other minor changes that don't need a BP tracked? | 17:26 |
| johnthetubaguy | GnomeKing: its a boarderline feature, but you could claim its a wishlist bug, images with only one vhd with non standard name should be accepted, or something | 17:27 |
| johnthetubaguy | GnomeKing: its more so it gets the correct review attention | 17:27 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova-specs: libvirt: support pvpanic device https://review.openstack.org/216719 | 17:27 |
| johnthetubaguy | GnomeKing: this one looks like its missing its blueprint, so putting a bug on there might convince me otherwise | 17:27 |
| GnomeKing | OK - so we'll raise a wishlist bug so it can be tracked in LP through that and explain on the bug why it's not a BP | 17:27 |
| GnomeKing | Understood, thanks. | 17:28 |
| *** ferest has quit IRC | 17:28 | |
| johnthetubaguy | GnomeKing: yeah, that does track when the support was added at least, as an added extra | 17:28 |
| *** VW__ has quit IRC | 17:28 | |
| GnomeKing | I don't want to circumvent the BP process of course - just hoping this falls into the category where BPs aren't needed for very minor features :) | 17:28 |
| *** VW_ has joined #openstack-nova | 17:29 | |
| *** vladikr has joined #openstack-nova | 17:29 | |
| GnomeKing | I'll let you get back to your evening now! | 17:29 |
| *** jecarey has quit IRC | 17:29 | |
| *** meghal has quit IRC | 17:30 | |
| *** meghal has joined #openstack-nova | 17:30 | |
| *** Apoorva has quit IRC | 17:33 | |
| openstackgerrit | Sudipta Biswas proposed openstack/nova: cleanup NovaObjectDictCompat from security_group_rule https://review.openstack.org/201282 | 17:33 |
| *** VW_ has quit IRC | 17:33 | |
| sudipto | dansmith, this one was an easy one : https://review.openstack.org/#/c/215078/ | 17:35 |
| *** meghal has quit IRC | 17:35 | |
| *** meghal has joined #openstack-nova | 17:35 | |
| dansmith | sudipto: nice :) | 17:36 |
| *** ajayaa has quit IRC | 17:37 | |
| *** garyk has joined #openstack-nova | 17:39 | |
| *** exploreshaifali has quit IRC | 17:40 | |
| *** dustins has joined #openstack-nova | 17:43 | |
| *** rdopiera has quit IRC | 17:45 | |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Add 'marker' argument to server list cli https://review.openstack.org/216815 | 17:46 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Add "limit" option to servers list cli https://review.openstack.org/216816 | 17:46 |
| *** lucasagomes is now known as lucas-dinner | 17:47 | |
| *** lpetrut has quit IRC | 17:48 | |
| *** Apoorva has joined #openstack-nova | 17:48 | |
| *** Apoorva has quit IRC | 17:49 | |
| *** dustins has quit IRC | 17:50 | |
| *** tpeoples has quit IRC | 17:50 | |
| *** Apoorva has joined #openstack-nova | 17:51 | |
| openstackgerrit | Sudipta Biswas proposed openstack/nova: cleanup NovaObjectDictCompat from external_event https://review.openstack.org/216822 | 17:56 |
| *** tojuvone has quit IRC | 17:56 | |
| *** tpeoples has joined #openstack-nova | 18:00 | |
| *** VW_ has joined #openstack-nova | 18:01 | |
| *** dannywilson has quit IRC | 18:02 | |
| *** dannywilson has joined #openstack-nova | 18:02 | |
| *** ildikov has joined #openstack-nova | 18:03 | |
| openstackgerrit | Sudipta Biswas proposed openstack/nova: cleanup NovaObjectDictCompat from instance_fault https://review.openstack.org/216825 | 18:04 |
| *** browne has quit IRC | 18:06 | |
| *** sudipto has quit IRC | 18:09 | |
| *** tjones has joined #openstack-nova | 18:09 | |
| *** tjones has left #openstack-nova | 18:09 | |
| *** haomaiwang has quit IRC | 18:09 | |
| openstackgerrit | Sean Dague proposed openstack/nova: api: tidy up paste.ini https://review.openstack.org/214085 | 18:09 |
| sdague | johnthetubaguy: ^^^ that now depends on the tempest test drop that should make it work | 18:10 |
| *** haomaiwang has joined #openstack-nova | 18:10 | |
| *** lpetrut has joined #openstack-nova | 18:10 | |
| *** marzif has quit IRC | 18:11 | |
| *** rushil has quit IRC | 18:11 | |
| *** vilobhmm1 has joined #openstack-nova | 18:11 | |
| *** garyk has quit IRC | 18:12 | |
| *** tonytan4ever has quit IRC | 18:14 | |
| *** nithyag_ has joined #openstack-nova | 18:16 | |
| *** andrearo1 has joined #openstack-nova | 18:16 | |
| *** meghal has quit IRC | 18:17 | |
| *** NithyaG has quit IRC | 18:17 | |
| *** andrearosa has quit IRC | 18:17 | |
| *** rushil has joined #openstack-nova | 18:20 | |
| *** jlk has joined #openstack-nova | 18:21 | |
| *** burgerk_ has quit IRC | 18:22 | |
| *** jecarey has joined #openstack-nova | 18:27 | |
| *** ociuhandu_ has joined #openstack-nova | 18:28 | |
| *** VW__ has joined #openstack-nova | 18:28 | |
| *** ociuhandu has quit IRC | 18:29 | |
| *** ociuhandu_ is now known as ociuhandu | 18:29 | |
| *** VW__ has quit IRC | 18:29 | |
| *** VW__ has joined #openstack-nova | 18:30 | |
| *** VW__ has quit IRC | 18:30 | |
| *** VW__ has joined #openstack-nova | 18:31 | |
| *** danpb has quit IRC | 18:31 | |
| *** tonytan4ever has joined #openstack-nova | 18:31 | |
| *** VW_ has quit IRC | 18:31 | |
| *** tpeoples has quit IRC | 18:31 | |
| *** tdp has joined #openstack-nova | 18:31 | |
| *** mtanino_ has joined #openstack-nova | 18:32 | |
| *** meghal has joined #openstack-nova | 18:32 | |
| mriedem_class | rlrossit: https://review.openstack.org/#/c/159890/ | 18:32 |
| *** dustins has joined #openstack-nova | 18:33 | |
| *** mtanino has quit IRC | 18:34 | |
| *** haomaiwang has quit IRC | 18:42 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Pre-load expected attrs that the view builder needs for server details https://review.openstack.org/215859 | 18:43 |
| *** Sukhdev has joined #openstack-nova | 18:43 | |
| *** haomaiwang has joined #openstack-nova | 18:45 | |
| bauzas | dansmith: heya | 18:46 |
| * dansmith runs | 18:46 | |
| dansmith | I mean.. wassup? | 18:46 |
| bauzas | dansmith: so my stomach is done | 18:46 |
| dansmith | heh | 18:46 |
| *** kylek3h_ has joined #openstack-nova | 18:47 | |
| bauzas | dansmith: so, re: https://review.openstack.org/#/c/215471/4 I thought about that (I mean defaulting to something) before doing that | 18:47 |
| bauzas | dansmith: I thought that having nullable fields was okay to know which computes were not updated by the RT | 18:48 |
| dansmith | but is knowing that useful? | 18:48 |
| dansmith | to me, it seems like knowing just means we consult the CONF, | 18:48 |
| dansmith | which we will have to continue doing forever, as long as the field is nullable | 18:48 |
| dansmith | we did a similar thing here: https://github.com/openstack/nova/blob/master/nova/objects/migration.py#L59-L62 | 18:48 |
| *** changbl has quit IRC | 18:49 | |
| bauzas | dansmith: well, my point was to remove the if branch by the next 2nd cycle | 18:49 |
| dansmith | which means everything can now just look at migration.migration_type without having to consider "has this been upgraded?" | 18:49 |
| *** kylek3h has quit IRC | 18:49 | |
| bauzas | I see | 18:49 |
| * bauzas looking | 18:49 | |
| dansmith | bauzas: you can't do that until the next major version really | 18:49 |
| bauzas | meh okay | 18:49 |
| dansmith | bauzas: and if you rely on it, you'd need a DB migration to make sure that they had filled in all their computes before we move | 18:49 |
| dansmith | bauzas: what's the reason to *not* just convert on load? | 18:49 |
| bauzas | dansmith: well, since we don't support N-2 cycle, I was saying that it was okay to wait until Nxxx | 18:50 |
| bauzas | dansmith: that's fine | 18:50 |
| bauzas | dansmith: I mean, I now understand your poinut | 18:50 |
| dansmith | yeah, you can't say that for DB stuff | 18:50 |
| bauzas | okay, looks you are right | 18:50 |
| dansmith | unless you put a block in place, and even still, at N+2 the object is still nullable, which means you still need to check, and can't change the nullability until we bump | 18:50 |
| bauzas | gotcha | 18:51 |
| bauzas | well, I remember some convo we had for the host/service_id implementation and yeah I understand | 18:51 |
| dansmith | http://github.org/openstack/nova/tree/master/nova/db/sqlalchemy/migrate_repo/versions/291_enforce_flavors_migrated.py#L29-L35 | 18:51 |
| dansmith | that is how we're doing it for flavors ^ | 18:51 |
| dansmith | which is more complicated because we're not actually removing anything, but moving where it's stored | 18:52 |
| dansmith | and even still, it's all hidden from the user inside the object | 18:52 |
| bauzas | okay, I see | 18:52 |
| dansmith | and I guarantee you that some operator is going to shoot me for that when they start deploying liberty :D | 18:52 |
| *** marzif has joined #openstack-nova | 18:52 | |
| bauzas | 1/ so if the compute node is Liberty, then the RT would update the fields | 18:52 |
| bauzas | 2/ if the compute node is Kilo, then the object will backwards updates those fields using the opts | 18:53 |
| bauzas | dansmith: ^ | 18:53 |
| dansmith | yeah, which gives the same behavior as before, which is that the config determines the ratios, right? | 18:53 |
| bauzas | yup, and I said that for Liberty, you first need to have the same ratios for all of your nodes | 18:54 |
| dansmith | yeah | 18:54 |
| mriedem_class | sdague: probably need you to weigh in here https://review.openstack.org/#/c/173985/ | 18:54 |
| mriedem_class | alex_xu: ^ | 18:54 |
| bauzas | dansmith: then, once your computes are all upgraded to Liberty, you can just have different values per compute if you wish | 18:54 |
| dansmith | bauzas: they can start having different ratios for the upgraded compute nodes before all are upgraded, right? | 18:55 |
| dansmith | and the unupgraded ones will honor the config on the controller nodes, right? | 18:55 |
| bauzas | dansmith: zactly | 18:55 |
| dansmith | yeah, so that seems ideal to me | 18:55 |
| dansmith | almost like you planned it | 18:56 |
| bauzas | dansmith: if running Kilo, the object will backwards update the fields using the opt values | 18:56 |
| bauzas | (which are on the controller) | 18:56 |
| dansmith | right | 18:56 |
| bauzas | dansmith: but if a new host is updated with Liberty, then the RT would update those fields with the changed opt values | 18:56 |
| bauzas | okay, I need to work then :) | 18:57 |
| bauzas | and I'll update the ReleaseNotes to make it explicit | 18:57 |
| dansmith | cool | 18:57 |
| bauzas | Kilo == controller opt values, Liberty == per compute opt values | 18:57 |
| *** mtanino has joined #openstack-nova | 18:58 | |
| bauzas | dansmith: thanks a lot for your help ! | 18:58 |
| sdague | mriedem_class: looking | 18:58 |
| dansmith | np ;) | 18:58 |
| dansmith | it's my stomach's turn now | 18:58 |
| bauzas | hah | 18:58 |
| bauzas | never discuss when your stomach is empty | 18:58 |
| *** apuimedo|away is now known as apuimedo | 18:58 | |
| dansmith | well, it's not empty yet, but the "empty soon" warning light is flashing | 18:58 |
| *** changbl has joined #openstack-nova | 18:59 | |
| sdague | mriedem_class: on https://review.openstack.org/#/c/173985 I'll accept that 403 is already in the allowed set because of keystone | 18:59 |
| mriedem_class | sdague: yeah, it was a good point from ftersin | 19:00 |
| sdague | mriedem_class: did you have any other objections to that code? | 19:00 |
| mriedem_class | sdague: no | 19:00 |
| mriedem_class | i'll +2 if you're on board | 19:00 |
| mriedem_class | i would also add a note to the microversion rules doc saying that 403 is an exceptional case | 19:00 |
| *** mtanino_ has quit IRC | 19:01 | |
| mriedem_class | my -1 on jichen's change is basically to get your input | 19:01 |
| sdague | mriedem_class: yeh, I just +2ed | 19:01 |
| sdague | mriedem_class: I'd be +2 on such a change to the microversion rule doc | 19:01 |
| mriedem_class | will work on it when i'm not being a good student | 19:02 |
| sdague | right, I forgot that you didn't know how to use elastic search :) | 19:02 |
| mriedem_class | i will say, | 19:03 |
| dansmith | sdague: pretty sure this class is at the level of "what is elasticsearch?" | 19:03 |
| mriedem_class | kibana > 2 confuses me | 19:03 |
| *** dustins has quit IRC | 19:04 | |
| *** haomaiwang has quit IRC | 19:09 | |
| *** haomaiwa_ has joined #openstack-nova | 19:10 | |
| *** smatzek has joined #openstack-nova | 19:15 | |
| *** claudiub has quit IRC | 19:16 | |
| *** eglynn has quit IRC | 19:17 | |
| *** y_sawai_ has joined #openstack-nova | 19:17 | |
| *** VW_ has joined #openstack-nova | 19:18 | |
| *** VW__ has quit IRC | 19:19 | |
| *** rushil has quit IRC | 19:19 | |
| *** y_sawai has quit IRC | 19:20 | |
| mriedem_class | tox -e docs fails :( | 19:20 |
| mriedem_class | https://github.com/openstack/nova/commit/ee2ecfc68aaf52223d6491dd7a366682d06c03c6 | 19:20 |
| *** lpetrut has quit IRC | 19:24 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Add a note about the 403 response not requiring a microversion https://review.openstack.org/216846 | 19:25 |
| *** dims__ has quit IRC | 19:25 | |
| mriedem_class | sdague: ^ | 19:26 |
| sdague | mriedem_class: thank you | 19:26 |
| *** dims has joined #openstack-nova | 19:26 | |
| jaypipes | sdague: easy pickins: https://review.openstack.org/#/c/195303/ | 19:27 |
| *** ociuhandu has quit IRC | 19:27 | |
| *** exploreshaifali has joined #openstack-nova | 19:27 | |
| *** achanda has quit IRC | 19:28 | |
| *** geekinutah has quit IRC | 19:29 | |
| *** ociuhandu has joined #openstack-nova | 19:29 | |
| *** VW__ has joined #openstack-nova | 19:32 | |
| *** VW_ has quit IRC | 19:32 | |
| *** dannywilson has quit IRC | 19:32 | |
| *** VW_ has joined #openstack-nova | 19:33 | |
| *** VW__ has quit IRC | 19:35 | |
| *** davideagnello has quit IRC | 19:36 | |
| sdague | jaypipes: well, I think that's still a little odd, but it's probably better than what we have | 19:36 |
| *** davideagnello has joined #openstack-nova | 19:36 | |
| dims | sdague: ya. after that gets in only a handful of failures remain in the experimental job | 19:37 |
| sdague | we do an aweful lot of dancing around the fact that we don't use the service catalog for reflection | 19:37 |
| *** davideagnello has quit IRC | 19:37 | |
| sdague | I guess I'm fine with that moving the ball forward | 19:37 |
| sdague | and we can hopefully do a better thing down the road | 19:37 |
| sdague | dims: you made some wsgi_script patches right? are those still outstanding? | 19:38 |
| *** jichen has joined #openstack-nova | 19:38 | |
| dims | sdague: all's in | 19:38 |
| sdague | dims: woot | 19:38 |
| dims | :) thanks! | 19:38 |
| *** lpetrut has joined #openstack-nova | 19:38 | |
| jaypipes | sdague: this is so that we can deploy Nova API in Apache/nginx wsgi container all-in-one with other endpoints. | 19:39 |
| mtreinish | mriedem_class: is newer kibana too colorful or something? I've not even looked at it | 19:39 |
| sdague | jaypipes: right, I get that | 19:39 |
| sdague | I'm just kind of frustrated that to do that you have to also set os_api_prefix correctly (don't mess up) and have it match what you fed other services | 19:40 |
| mriedem_class | mtreinish: well, queries are separate from filters now, and it's weird. like i don't see a text area to just input my own query | 19:40 |
| sdague | this doesn't make that worse, but is just another issue where we've got to do a bunch of extra work because we don't really use the SC much | 19:40 |
| mriedem_class | mtreinish: i have to like build everything using toggles and then mash them together to get it to run what i want | 19:40 |
| jaypipes | mriedem_class: any opportunity you can approve this? https://review.openstack.org/#/c/216741/ | 19:40 |
| mriedem_class | jaypipes: yeah | 19:41 |
| mtreinish | mriedem_class: ugh, that's what I was afraid it was like. I guess it's their attempt to be more "user friendly" :( | 19:42 |
| sdague | mtreinish: they also open core a bunch of stuff, so their is some feature that we'd like in infra, that's only in their product version. Can't remember which one it is. | 19:43 |
| mriedem_class | mtreinish: yeah that was my comment, i'm sure it's probably easier for new people to use it (sort of) | 19:43 |
| mriedem_class | it's still not super intuitive | 19:44 |
| *** marzif_ has joined #openstack-nova | 19:44 | |
| *** lpetrut has quit IRC | 19:45 | |
| *** matrohon has joined #openstack-nova | 19:45 | |
| *** haomaiw__ has joined #openstack-nova | 19:46 | |
| *** mriedem_class is now known as mriedem | 19:46 | |
| mriedem | who am i kidding | 19:46 |
| *** haomaiwa_ has quit IRC | 19:46 | |
| mriedem | i'm not paying attention | 19:46 |
| *** marzif has quit IRC | 19:46 | |
| jaypipes | mriedem: thx | 19:48 |
| apmelton | mriedem: I took a look at that review, that's essentially what nova network is doing | 19:49 |
| apmelton | mriedem: the interface in the libvirt xml is a bridged interface | 19:49 |
| mriedem | apmelton: keep in mind that juniper/contrail patch was based on icehouse | 19:50 |
| apmelton | gotcha | 19:50 |
| *** fawadkhaliq has quit IRC | 19:50 | |
| apmelton | I also went digging through the calls nova network was making... it didn't seem like it was changing the mac on the bridge anywhere | 19:51 |
| *** dustins has joined #openstack-nova | 19:56 | |
| openstackgerrit | Tomoki Sekiyama proposed openstack/nova: libvirt: Make behavior of os_require_quiesce consistent https://review.openstack.org/167819 | 19:56 |
| mriedem | alaski: are you going to -2 or should i? https://review.openstack.org/212580 | 19:56 |
| mriedem | nvm, i did | 19:57 |
| alaski | mriedem: go for it | 19:57 |
| alaski | if he pushed another revision I was going to | 19:57 |
| openstackgerrit | Jay Pipes proposed openstack/nova: rt: Rewrite abort and update_usage tests https://review.openstack.org/213590 | 20:00 |
| openstackgerrit | Jay Pipes proposed openstack/nova: Move test of claim context manager to test_tracker https://review.openstack.org/213591 | 20:00 |
| openstackgerrit | Jay Pipes proposed openstack/nova: Cleanup RT _instance_in_resize_state() https://review.openstack.org/213589 | 20:00 |
| openstackgerrit | Jay Pipes proposed openstack/nova: Move resize claim abort test to test_tracker https://review.openstack.org/213592 | 20:00 |
| jaypipes | alaski, mriedem,, dansmith, bauzas: ^^ fixed all issues. feel free to review if have a chance. | 20:00 |
| bauzas | jaypipes: np, tomorrow | 20:01 |
| jaypipes | bauzas: cheers | 20:01 |
| bauzas | jaypipes: https://twitter.com/sylvainbauza/status/636175679348961280 | 20:02 |
| bauzas | jaypipes: but since then, I had homework given by dansmith :p | 20:02 |
| *** geekinutah has joined #openstack-nova | 20:02 | |
| jaypipes | mriedem, alaski, dansmith: also, a few folks are waiting on https://review.openstack.org/#/c/209499/ in order to propose patches that add new compute monitors... if you have a chance, would be appreciated. | 20:02 |
| *** jichen has quit IRC | 20:03 | |
| jaypipes | bauzas: :) | 20:03 |
| *** Piet has quit IRC | 20:03 | |
| openstackgerrit | Jorge Niedbalski R. proposed openstack/nova: Fixes a typo on nova.tests.unit.api.ec2.test_api.py https://review.openstack.org/216854 | 20:03 |
| *** browne has joined #openstack-nova | 20:05 | |
| apmelton | mriedem: so... I've figured out whats going on with this | 20:06 |
| apmelton | mriedem: as it turns out the default behavior for a linux bridge is to use the lowest mac address of all of the tap devices connected to it | 20:06 |
| *** rushil has joined #openstack-nova | 20:06 | |
| *** tonytan4ever has quit IRC | 20:06 | |
| apmelton | so it started with eth0 because that's the only device... then if libvirt plugs a tap that has a lower mac, the bridge's address changes | 20:07 |
| *** rushil_ has joined #openstack-nova | 20:08 | |
| *** lpetrut has joined #openstack-nova | 20:08 | |
| *** sachin has quit IRC | 20:08 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Remove doc/source/api and doc/build before building docs https://review.openstack.org/216857 | 20:08 |
| *** haomaiw__ has quit IRC | 20:09 | |
| *** sachin has joined #openstack-nova | 20:09 | |
| mriedem | apmelton: cool | 20:10 |
| *** haomaiwang has joined #openstack-nova | 20:10 | |
| *** ccarmack has joined #openstack-nova | 20:11 | |
| *** ccarmack has left #openstack-nova | 20:11 | |
| *** achanda has joined #openstack-nova | 20:11 | |
| *** rushil has quit IRC | 20:12 | |
| *** ccarmack has joined #openstack-nova | 20:14 | |
| *** lpetrut has quit IRC | 20:15 | |
| *** exploreshaifali has quit IRC | 20:17 | |
| *** sachin has quit IRC | 20:20 | |
| *** sachin has joined #openstack-nova | 20:20 | |
| *** dustins has quit IRC | 20:27 | |
| *** ociuhandu has quit IRC | 20:32 | |
| *** Apoorva has quit IRC | 20:33 | |
| *** Apoorva has joined #openstack-nova | 20:34 | |
| *** sachin has quit IRC | 20:35 | |
| *** MVenesio has joined #openstack-nova | 20:35 | |
| *** sachin has joined #openstack-nova | 20:36 | |
| *** sachin has quit IRC | 20:38 | |
| *** sachin has joined #openstack-nova | 20:38 | |
| *** ijuwang has quit IRC | 20:38 | |
| *** armax has quit IRC | 20:39 | |
| *** sachin has quit IRC | 20:39 | |
| *** sachin has joined #openstack-nova | 20:39 | |
| *** gyee has joined #openstack-nova | 20:41 | |
| *** yamahata has quit IRC | 20:41 | |
| *** smatzek_ has joined #openstack-nova | 20:42 | |
| *** kiran has joined #openstack-nova | 20:42 | |
| *** kiran has quit IRC | 20:43 | |
| *** Sukhdev has quit IRC | 20:43 | |
| *** Kiran has joined #openstack-nova | 20:44 | |
| *** ociuhandu has joined #openstack-nova | 20:45 | |
| *** smatzek has quit IRC | 20:46 | |
| *** smatzek_ has quit IRC | 20:47 | |
| *** Piet has joined #openstack-nova | 20:48 | |
| *** IanGovett1 has quit IRC | 20:49 | |
| *** changbl has quit IRC | 20:53 | |
| mriedem | apmelton: i wonder if we could use ovs instead of linuxbridge? | 20:55 |
| mriedem | or if it wouldn't matter | 20:55 |
| apmelton | mriedem: anyone else in the gate using OVS with this ubuntu image? | 20:55 |
| *** y_sawai_ has quit IRC | 20:56 | |
| mriedem | not sure | 20:56 |
| *** Kiran has quit IRC | 20:56 | |
| apmelton | mriedem: I think this is an easy fix in nova-network though | 20:57 |
| *** IanGovett has joined #openstack-nova | 20:57 | |
| clarkb | nuetron uses ovs almost everywhere | 20:57 |
| mtreinish | apmelton: yeah what clarkb said, by default we are using ovs on all the neutron jobs | 20:58 |
| mtreinish | apmelton, mriedem: sc68cal is working on getting a job with linuxbridge working and I think the plan was to migrate the gate jobs to that when its ready | 20:58 |
| clarkb | mtreinish: oh I thought we were going to keep ovs as default something something I switched multinode to ovs because neutron was all ovs happy | 20:59 |
| clarkb | :| I guess I can revert and go back to linux bridge | 20:59 |
| mtreinish | clarkb: oh, maybe not. I thought that was one of the takeaways from the multiple giant ML threads and the summit sessions on it, but I could be mistaken | 21:00 |
| sc68cal | mtreinish: no we are keeping ovs as the default | 21:00 |
| mtreinish | clarkb: sc68cal will know | 21:00 |
| mtreinish | ok I was wrong :) | 21:00 |
| sc68cal | mtreinish: but we are gating on linuxbridge and ovs | 21:00 |
| sc68cal | s/are/will be/g | 21:00 |
| clarkb | oh good | 21:00 |
| * sc68cal is busy implementing 10% of fuel to pxe boot his new shiny devstack lab | 21:03 | |
| *** raildo is now known as raildo-afk | 21:03 | |
| *** bnemec has quit IRC | 21:05 | |
| *** bnemec has joined #openstack-nova | 21:06 | |
| *** meghal has quit IRC | 21:07 | |
| *** meghal has joined #openstack-nova | 21:07 | |
| *** aysyd has quit IRC | 21:09 | |
| *** haomaiwang has quit IRC | 21:09 | |
| *** thorst has quit IRC | 21:09 | |
| *** haomaiwang has joined #openstack-nova | 21:10 | |
| *** ccarmack has quit IRC | 21:10 | |
| *** rlrossit has left #openstack-nova | 21:10 | |
| *** ctrath1 has quit IRC | 21:11 | |
| *** burgerk has joined #openstack-nova | 21:12 | |
| clarkb | sc68cal: bifrost may make that quick | 21:13 |
| sc68cal | clarkb: yeah I looked at the repo but there wasn't a lot of docs | 21:14 |
| clarkb | sc68cal: I thought the readme would get you going | 21:14 |
| *** achanda has quit IRC | 21:15 | |
| sc68cal | clarkb: so the main thing is that I don't have hardware to run ironic on | 21:16 |
| *** meghal has quit IRC | 21:16 | |
| sc68cal | clarkb: and I didn't want to rebuild my whole network - I just wanted to drop 3 intel nucs in my home rack, pxe boot them and then run devstack | 21:17 |
| clarkb | sc68cal: your laptop! but ya that would be an issue | 21:17 |
| *** achanda has joined #openstack-nova | 21:17 | |
| *** ctrath has joined #openstack-nova | 21:18 | |
| mtreinish | sc68cal: oh, you should look at sdague's amt scripts to do that stuff | 21:18 |
| *** jecarey has quit IRC | 21:18 | |
| mtreinish | sc68cal: sdague has been running a similar setup at his house | 21:18 |
| sc68cal | mtreinish: yeah I did, but the boards don't do IPMI :( | 21:18 |
| sc68cal | or amt | 21:18 |
| mtreinish | sc68cal: older nucs? | 21:18 |
| clarkb | ya only the most expensive nuc does per model year iirc | 21:18 |
| clarkb | so you have to be selective if going with amt | 21:18 |
| mtreinish | oh, I didn't realize that | 21:18 |
| sc68cal | yeah they're gigabyte GB-BXi3-4010 's | 21:19 |
| *** dprince has quit IRC | 21:19 | |
| mtreinish | sc68cal: well mirantis has been in the paper about being flush with intel cash, you should ask for free nucs with amt and ipmi :) | 21:19 |
| sc68cal | mtreinish: heh - I'm running out of ports on my switch and also running out of power outlets | 21:20 |
| sc68cal | mtreinish: so, I think 3 nucs is as big as i'll get. | 21:21 |
| *** dave-mcc_ has quit IRC | 21:21 | |
| *** jecarey has joined #openstack-nova | 21:21 | |
| mtreinish | sc68cal: sure, but you'd just replace your existing 3 with the new "free" ones | 21:22 |
| sc68cal | mtreinish: doesn't do much for me when I already paid off the credit card for them | 21:22 |
| sc68cal | mtreinish: unless they want to buy them off me and replace with the freebies | 21:22 |
| mtreinish | sc68cal: heh, you've got a plan to to talk to your boss about now :) | 21:23 |
| sc68cal | mtreinish: already tried that angle-no go | 21:24 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Add cpu_allocation_ratio and ram_allocation_ratio to ComputeNode https://review.openstack.org/215471 | 21:25 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Update HostManager and filters to use ComputeNode ratios https://review.openstack.org/216696 | 21:25 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Update ComputeNode values with allocation ratios in the RT https://review.openstack.org/216362 | 21:25 |
| bauzas | dansmith: jaypipes: ^ see my homework | 21:25 |
| bauzas | FYI of course, when you have time guys | 21:26 |
| *** exploreshaifali has joined #openstack-nova | 21:26 | |
| *** VW_ has quit IRC | 21:29 | |
| *** VW_ has joined #openstack-nova | 21:30 | |
| *** meghal has joined #openstack-nova | 21:31 | |
| *** VW_ has quit IRC | 21:35 | |
| *** davideagnello has joined #openstack-nova | 21:37 | |
| *** matrohon has quit IRC | 21:39 | |
| *** davideagnello has quit IRC | 21:43 | |
| *** sachin has quit IRC | 21:44 | |
| *** davideagnello has joined #openstack-nova | 21:44 | |
| *** sachin has joined #openstack-nova | 21:45 | |
| *** sachin has quit IRC | 21:46 | |
| *** sachin has joined #openstack-nova | 21:47 | |
| *** tonytan4ever has joined #openstack-nova | 21:47 | |
| *** thumpba has quit IRC | 21:48 | |
| *** burgerk has quit IRC | 21:48 | |
| *** mriedem has quit IRC | 21:49 | |
| *** meghal has quit IRC | 21:49 | |
| *** jecarey has quit IRC | 21:50 | |
| *** ctrath has quit IRC | 21:50 | |
| *** meghal has joined #openstack-nova | 21:54 | |
| *** jaypipes has quit IRC | 21:55 | |
| *** claudiub has joined #openstack-nova | 21:57 | |
| *** meghal has quit IRC | 21:58 | |
| *** meghal has joined #openstack-nova | 21:58 | |
| *** edmondsw has quit IRC | 21:58 | |
| *** jaypipes has joined #openstack-nova | 21:58 | |
| *** MVenesio has quit IRC | 22:00 | |
| *** agireud has quit IRC | 22:01 | |
| *** xyang1 has quit IRC | 22:04 | |
| *** humble__ has quit IRC | 22:04 | |
| *** rushil_ has quit IRC | 22:05 | |
| *** humble__ has joined #openstack-nova | 22:05 | |
| *** rushil has joined #openstack-nova | 22:05 | |
| *** aix has joined #openstack-nova | 22:07 | |
| *** bnemec has quit IRC | 22:07 | |
| *** bnemec has joined #openstack-nova | 22:09 | |
| *** angdraug has quit IRC | 22:09 | |
| *** haomaiwang has quit IRC | 22:09 | |
| *** haomaiwa_ has joined #openstack-nova | 22:10 | |
| *** eharney has quit IRC | 22:11 | |
| *** armax has joined #openstack-nova | 22:14 | |
| *** burt has quit IRC | 22:16 | |
| *** achanda has quit IRC | 22:19 | |
| *** sachin has quit IRC | 22:19 | |
| *** sachin has joined #openstack-nova | 22:21 | |
| *** achanda has joined #openstack-nova | 22:21 | |
| *** kylek3h_ has quit IRC | 22:22 | |
| *** sachin has quit IRC | 22:22 | |
| *** sachin has joined #openstack-nova | 22:23 | |
| *** sachin has quit IRC | 22:25 | |
| *** sachin has joined #openstack-nova | 22:26 | |
| *** sachin has quit IRC | 22:28 | |
| *** sachin has joined #openstack-nova | 22:28 | |
| *** sachin_ has joined #openstack-nova | 22:29 | |
| *** sachin has quit IRC | 22:29 | |
| *** tonytan4ever has quit IRC | 22:29 | |
| openstackgerrit | Jay Pipes proposed openstack/nova: rt: Rewrite abort and update_usage tests https://review.openstack.org/213590 | 22:34 |
| openstackgerrit | Jay Pipes proposed openstack/nova: Move test of claim context manager to test_tracker https://review.openstack.org/213591 | 22:34 |
| openstackgerrit | Jay Pipes proposed openstack/nova: Cleanup RT _instance_in_resize_state() https://review.openstack.org/213589 | 22:34 |
| openstackgerrit | Jay Pipes proposed openstack/nova: Move resize claim abort test to test_tracker https://review.openstack.org/213592 | 22:34 |
| *** lucas-dinner has quit IRC | 22:34 | |
| *** RaySun has joined #openstack-nova | 22:35 | |
| *** rushil has quit IRC | 22:43 | |
| *** sachin_ has quit IRC | 22:44 | |
| *** Sukhdev has joined #openstack-nova | 22:45 | |
| *** sachin has joined #openstack-nova | 22:46 | |
| *** sachin has quit IRC | 22:47 | |
| bauzas | jaypipes: meh, okay for reviewing your series tomorrow morning my time ? | 22:48 |
| bauzas | jaypipes: can see a last comment | 22:48 |
| *** jaypipes has quit IRC | 22:48 | |
| *** sachin has joined #openstack-nova | 22:48 | |
| *** sachin has quit IRC | 22:50 | |
| *** sachin has joined #openstack-nova | 22:51 | |
| *** sachin has quit IRC | 22:53 | |
| *** sachin has joined #openstack-nova | 22:54 | |
| *** mmedvede has quit IRC | 22:55 | |
| *** sachin has quit IRC | 22:55 | |
| *** sachin has joined #openstack-nova | 22:56 | |
| *** sachin has quit IRC | 22:56 | |
| *** sachin has joined #openstack-nova | 22:56 | |
| *** RaySun has quit IRC | 22:58 | |
| *** asd112z has quit IRC | 22:59 | |
| *** shoutm has joined #openstack-nova | 23:01 | |
| openstackgerrit | Andrew Melton proposed openstack/nova: net: explicitly set mac on linux bridge https://review.openstack.org/216895 | 23:04 |
| *** unicell1 has quit IRC | 23:06 | |
| *** __TheDodd__ has quit IRC | 23:08 | |
| *** haomaiwa_ has quit IRC | 23:09 | |
| *** ccarmack has joined #openstack-nova | 23:09 | |
| *** 64MADOR5X has joined #openstack-nova | 23:10 | |
| *** mmedvede has joined #openstack-nova | 23:23 | |
| *** davideagnello has quit IRC | 23:24 | |
| *** yamahata has joined #openstack-nova | 23:24 | |
| *** tdp has quit IRC | 23:25 | |
| *** bitblt has joined #openstack-nova | 23:28 | |
| *** bitblt has quit IRC | 23:28 | |
| *** sachin has quit IRC | 23:28 | |
| *** changbl has joined #openstack-nova | 23:29 | |
| *** marzif_ has quit IRC | 23:29 | |
| openstackgerrit | Pengtao Huang proposed openstack/nova: change the value "all_mappings" to "block_device_mapping" https://review.openstack.org/208511 | 23:30 |
| *** sachin has joined #openstack-nova | 23:30 | |
| *** davideagnello has joined #openstack-nova | 23:31 | |
| *** hemna is now known as hemnafk | 23:31 | |
| *** dannywilson has joined #openstack-nova | 23:33 | |
| *** vivekd_ has joined #openstack-nova | 23:34 | |
| *** unicell has joined #openstack-nova | 23:35 | |
| *** vivekd has quit IRC | 23:36 | |
| *** vivekd_ is now known as vivekd | 23:36 | |
| *** dannywilson has quit IRC | 23:38 | |
| *** diogogmt has quit IRC | 23:38 | |
| *** mdrabe has quit IRC | 23:41 | |
| *** annegent_ has joined #openstack-nova | 23:42 | |
| *** marcusvrn_ has quit IRC | 23:45 | |
| *** annegentle has quit IRC | 23:45 | |
| *** zzzeek has quit IRC | 23:45 | |
| *** erlon has quit IRC | 23:48 | |
| *** zzzeek has joined #openstack-nova | 23:49 | |
| *** dave-mccowan has joined #openstack-nova | 23:49 | |
| *** zzzeek has quit IRC | 23:53 | |
| *** takedakn has joined #openstack-nova | 23:53 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!