| *** annashen has quit IRC | 00:00 | |
| openstackgerrit | Tracy Jones proposed openstack/nova: Cleanup of compute api reboot method https://review.openstack.org/127427 | 00:02 |
|---|---|---|
| *** annashen has joined #openstack-nova | 00:02 | |
| *** oro has joined #openstack-nova | 00:03 | |
| *** annashen has quit IRC | 00:07 | |
| *** annashen has joined #openstack-nova | 00:08 | |
| *** sdake has quit IRC | 00:09 | |
| *** annashen has quit IRC | 00:09 | |
| *** doude has quit IRC | 00:11 | |
| *** otter768 has joined #openstack-nova | 00:12 | |
| *** smccully has quit IRC | 00:15 | |
| *** angdraug has quit IRC | 00:15 | |
| *** otter768 has quit IRC | 00:17 | |
| *** smccully has joined #openstack-nova | 00:18 | |
| *** baoli has joined #openstack-nova | 00:21 | |
| *** zzzeek has quit IRC | 00:22 | |
| *** Marga_ has joined #openstack-nova | 00:29 | |
| *** oro has quit IRC | 00:30 | |
| *** mtanino has quit IRC | 00:33 | |
| *** yingjun has joined #openstack-nova | 00:36 | |
| *** packet has joined #openstack-nova | 00:36 | |
| openstackgerrit | Eli Qiao proposed openstack/nova-specs: Add more properties to limits and used_limits api https://review.openstack.org/145689 | 00:40 |
| *** achanda has quit IRC | 00:41 | |
| *** promulo__ has joined #openstack-nova | 00:41 | |
| *** promulo_ has quit IRC | 00:46 | |
| *** mwagner_lap has quit IRC | 00:46 | |
| *** packet has quit IRC | 00:46 | |
| *** BadCub is now known as BadCub_Hiding | 00:48 | |
| *** sdake has joined #openstack-nova | 00:48 | |
| *** joefides_ has quit IRC | 00:50 | |
| *** joefides has joined #openstack-nova | 00:51 | |
| *** zhenguo has joined #openstack-nova | 00:51 | |
| jhesketh | dansmith, jogo: okay, rebuilt th images and tested them, so now running against 174480 which should work now | 00:53 |
| *** sdake has quit IRC | 00:53 | |
| dansmith | jhesketh: cool | 00:56 |
| *** packet has joined #openstack-nova | 00:57 | |
| jhesketh | dansmith: okay it's working except that your new migration reads more rows that turbo-hipster allows | 00:59 |
| dansmith | jhesketh: probably not surprising though, right? | 00:59 |
| jhesketh | so I'll need to create an exception in the dataset which is another image rebuild | 00:59 |
| jhesketh | yeah, and an exception seems reasonable | 00:59 |
| dansmith | jhesketh: sorry about that | 01:00 |
| jhesketh | oh that's not a problem | 01:00 |
| jhesketh | it is important to note though as it means that migration may take time in a real deployment | 01:00 |
| *** browne has quit IRC | 01:00 | |
| dansmith | jhesketh: so, hmm, actually, what rows is it reading? | 01:00 |
| dansmith | they should all be converted because you've ran the migration, right? | 01:01 |
| jhesketh | unless there is a way to optimise the query | 01:01 |
| jhesketh | dansmith: https://ssl.rcbops.com/turbo_hipster/logviewer/?q=/turbo_hipster/results/80/174480/14/check/real-db-upgrade_nova_mysql_user_001:th-mysql/f2b1417ca83f4b6f9c92dcbb1fc2c252/user_001.log | 01:01 |
| jhesketh | dansmith: actually, it could be the reads between running the migrate flavors | 01:02 |
| dansmith | jhesketh: wait, you mean the migrate_flavor_data is | 01:02 |
| dansmith | right, okay | 01:02 |
| jhesketh | so it's a bug in my code | 01:02 |
| jhesketh | sorry | 01:02 |
| dansmith | hah, no problem | 01:02 |
| jhesketh | nevertheless, it's working, so if you wanted to merge your code I'd say go for it | 01:03 |
| jhesketh | otherwise I'll fix up my code and rebuild and have a positive vote in ~1hr | 01:03 |
| dansmith | jhesketh: cool, can you comment on there that you think it's good save that issue or something? | 01:03 |
| dansmith | ah, okay | 01:03 |
| dansmith | if it's an hours thing, I'm cool waiting for a vote | 01:03 |
| *** mwagner_lap has joined #openstack-nova | 01:03 | |
| dansmith | if it won't be ready by my morning, then a comment by you on there would be appreciated and we'll unblock everyone waiting on this | 01:04 |
| jhesketh | okay, will do | 01:04 |
| jhesketh | what time is it for you currently | 01:04 |
| dansmith | 1804 | 01:04 |
| dansmith | "my morning" can be pretty fluid :) | 01:04 |
| *** zhangtralon has joined #openstack-nova | 01:05 | |
| jhesketh | heh | 01:06 |
| dansmith | jhesketh: I gotta run off to dinner.. thanks a ton for your help with all this! | 01:07 |
| jhesketh | no trouble at all | 01:07 |
| jhesketh | have a good evening | 01:07 |
| *** Jeffrey4l has quit IRC | 01:09 | |
| *** markvoelker has joined #openstack-nova | 01:11 | |
| *** rfolco has quit IRC | 01:14 | |
| *** baoli has quit IRC | 01:15 | |
| *** baoli has joined #openstack-nova | 01:16 | |
| *** mdbooth has quit IRC | 01:16 | |
| *** mmedvede has quit IRC | 01:19 | |
| *** achanda has joined #openstack-nova | 01:19 | |
| *** eliqiao has quit IRC | 01:19 | |
| *** popw has joined #openstack-nova | 01:20 | |
| *** eliqiao has joined #openstack-nova | 01:20 | |
| *** mdbooth has joined #openstack-nova | 01:22 | |
| *** eliqiao has left #openstack-nova | 01:26 | |
| *** eliqiao has joined #openstack-nova | 01:26 | |
| openstackgerrit | Vincent Hou proposed openstack/nova: Correct the order of the parameters to call swap_volume https://review.openstack.org/174307 | 01:32 |
| *** unicell has quit IRC | 01:38 | |
| *** Jeffrey4l has joined #openstack-nova | 01:42 | |
| *** alex_xu has quit IRC | 01:43 | |
| *** IanGovett has joined #openstack-nova | 01:47 | |
| *** vilobhmm1 has quit IRC | 01:48 | |
| *** dims has joined #openstack-nova | 01:48 | |
| *** Marga_ has quit IRC | 01:50 | |
| tonyb | dims: ping? | 01:50 |
| dims | tonyb: pong | 01:50 |
| *** Marga_ has joined #openstack-nova | 01:51 | |
| tonyb | dims: I have a RPC/oslo_messaging query for you (or your delegate) | 01:51 |
| tonyb | I'll state right now RPC has alwasy just workd for me so I don't really know a whole lot about it | 01:52 |
| *** patrickeast has quit IRC | 01:52 | |
| tonyb | I'm seeing an RPC call produced (and I can see it in rabbit) but the message times out so I suspect that somehow there isn't a comsumer subscribed/registered | 01:53 |
| tonyb | dims: is that something you've see before? or any ideas how to track down why the message isn't reaching the consumer? | 01:53 |
| *** Marga_ has quit IRC | 01:55 | |
| dims | tonyb: no errors, stack traces? | 01:55 |
| tonyb | dims: Yeah I get the generic 500 (something didn't work) error message and the RPC timeout stacktrace. | 01:56 |
| tonyb | dims: let me regenerate both and paste | 01:58 |
| dims | cool, also where do you see it? nova api log or conductor log? | 01:58 |
| tonyb | api log | 01:58 |
| *** packet has quit IRC | 01:58 | |
| tonyb | dims: I don't want to derail your plans so feel free to say 'file a bug' at any point | 01:59 |
| tonyb | http://paste.openstack.org/show/LiifCkMJGjK3ZHUqG12R/ | 02:01 |
| tonyb | dims: I don't have tracinf turned on in rabbit right now but in the past I see rabbit publish the message and no reply | 02:01 |
| dims | any hits in the log for the message id? dce43cb688214328845a4b24acc1ed8f | 02:02 |
| *** Marga_ has joined #openstack-nova | 02:02 | |
| *** yingjun has quit IRC | 02:03 | |
| tonyb | dims a grep just shows 2 lines both in the paste above | 02:03 |
| dims | some sort of HA in your environment or plain ol' devstack? | 02:04 |
| *** yingjun has joined #openstack-nova | 02:05 | |
| *** smccully has quit IRC | 02:06 | |
| tonyb | dims: plain old devstack in a rackspace VM devstack and nova are from ~48hours agao | 02:06 |
| dims | fedora or ubuntu? | 02:06 |
| tonyb | dims: (aslo 48hours ago) I did a brand new install of 10.04 so all the pip modules should be current | 02:07 |
| *** baoli has quit IRC | 02:07 | |
| tonyb | sorry ubuntu 14.04 NOT 10.04 that'd be nuts ;P | 02:07 |
| *** baoli has joined #openstack-nova | 02:07 | |
| dims | grab me "rabbitmqctl status"? | 02:08 |
| dims | so this sounds like the nova compute that is supposed to pick up the message is not picking it up | 02:08 |
| tonyb | http://paste.openstack.org/show/213020/ | 02:08 |
| tonyb | dims: yeah that's what I was thinking but I don't have any idea on how to track that down | 02:09 |
| *** Jeffrey4l has quit IRC | 02:10 | |
| dims | one thing to try quickly is to restart just the nova compute process | 02:11 |
| dims | after cleaning up the compute logs and making sure there are no errors after it starts up | 02:12 |
| *** yamahata has quit IRC | 02:12 | |
| *** otter768 has joined #openstack-nova | 02:13 | |
| dims | if that does not work, please log a bug and we can request sileht to help. there are various tricks with rabbitmq (like rabbitmqctl force_reset) etc. we may have to try after we capture a bit more logs etc (including rabbitmq logs) | 02:14 |
| tonyb | dims: So when I killed nova-compute I get "2015-04-30 02:15:11.224 ERROR oslo_messaging._drivers.impl_rabbit [-] Failed to consume message from queue:" | 02:16 |
| *** Marga_ has quit IRC | 02:16 | |
| dims | with a fresh restart? | 02:17 |
| dims | of nova-compute | 02:17 |
| *** Marga_ has joined #openstack-nova | 02:17 | |
| *** signed8bit has joined #openstack-nova | 02:17 | |
| *** otter768 has quit IRC | 02:18 | |
| tonyb | That was killing nova-compute after It'd run the rpc client at least one. | 02:19 |
| tonyb | and after killing and restarting nova-compute the message still times out. | 02:20 |
| *** Marga_ has quit IRC | 02:21 | |
| tonyb | okay I'll open a bug | 02:21 |
| dims | weird, thanks! | 02:24 |
| eliqiao | hi dims: still around? | 02:25 |
| tonyb | dims: is there an easy way to incraes the log level of oslo_messaging to DEBUG? | 02:25 |
| dims | eliqiao: pong | 02:25 |
| *** IanGovett has quit IRC | 02:25 | |
| *** yingjun has quit IRC | 02:26 | |
| eliqiao | dims: I have a question about compute_api's lock/unlock/get_lock method, I see that in nova-api we have alreate use isntances object instead of instance_uuid to call nova_compute_api, my question is since we have get an instance object in nova-api, we can get/change the instance's status in nova-api, why we still need to call nova_compute api to do lock/unlock/get_lock? | 02:27 |
| dims | tonyb: quickest it to touch up the level in openstack/nova/nova/config.py | 02:27 |
| *** yingjun has joined #openstack-nova | 02:28 | |
| dims | eliqiao: not sure about that, may need dansmith or mriedem i think | 02:29 |
| tonyb | dims: thanks | 02:29 |
| *** Jeffrey4l has joined #openstack-nova | 02:29 | |
| eliqiao | dims: ah sorry , my mistake, I thought dansmith is dims, my fault... | 02:29 |
| eliqiao | dims: I am quite strange that I still see dansmith at this time(he should be in bed) | 02:30 |
| eliqiao | :) | 02:30 |
| *** VW_ has joined #openstack-nova | 02:30 | |
| dims | eliqiao: some folks have a setup to they can log and catch up on the irc chatter overnight :) | 02:31 |
| dims | i am signing off now :) | 02:31 |
| *** dims has quit IRC | 02:31 | |
| *** VW_ has quit IRC | 02:32 | |
| *** VW_ has joined #openstack-nova | 02:32 | |
| *** bkopilov has quit IRC | 02:36 | |
| *** fifieldt has joined #openstack-nova | 02:36 | |
| *** yamahata has joined #openstack-nova | 02:43 | |
| *** alex_xu has joined #openstack-nova | 02:44 | |
| *** browne has joined #openstack-nova | 02:45 | |
| *** ir2ivps3_ has quit IRC | 02:46 | |
| *** ir2ivps3 has joined #openstack-nova | 02:48 | |
| *** heyongli has quit IRC | 02:49 | |
| *** otter768 has joined #openstack-nova | 02:50 | |
| *** baoli has quit IRC | 02:51 | |
| *** heyongli has joined #openstack-nova | 02:51 | |
| *** baoli has joined #openstack-nova | 02:52 | |
| *** baoli_ has joined #openstack-nova | 02:53 | |
| *** baoli has quit IRC | 02:53 | |
| *** dikonoor has joined #openstack-nova | 02:54 | |
| *** harlowja has quit IRC | 02:56 | |
| *** harlowja has joined #openstack-nova | 02:56 | |
| *** harlowja has quit IRC | 02:57 | |
| *** sdake has joined #openstack-nova | 02:59 | |
| *** sdake_ has joined #openstack-nova | 03:02 | |
| *** sdake has quit IRC | 03:02 | |
| *** kaufer has joined #openstack-nova | 03:02 | |
| *** zhangtralon has quit IRC | 03:09 | |
| *** erlon has quit IRC | 03:11 | |
| *** penick has joined #openstack-nova | 03:13 | |
| *** vilobhmm1 has joined #openstack-nova | 03:18 | |
| *** packet has joined #openstack-nova | 03:20 | |
| *** vilobhmm11 has joined #openstack-nova | 03:21 | |
| *** vilobhmm1 has quit IRC | 03:24 | |
| *** Jeffrey4l has quit IRC | 03:24 | |
| *** harlowja has joined #openstack-nova | 03:27 | |
| *** gtt116__ has joined #openstack-nova | 03:31 | |
| *** popw1 has joined #openstack-nova | 03:33 | |
| *** popw has quit IRC | 03:34 | |
| *** gtt116_ has quit IRC | 03:35 | |
| *** baoli_ has quit IRC | 03:36 | |
| *** dave-mccowan has quit IRC | 03:40 | |
| *** penick_ has joined #openstack-nova | 03:41 | |
| *** penick has quit IRC | 03:42 | |
| *** penick_ is now known as penick | 03:42 | |
| *** tjones1 has quit IRC | 03:44 | |
| *** otter768 has quit IRC | 03:44 | |
| *** tjones1 has joined #openstack-nova | 03:45 | |
| *** tjones1 has left #openstack-nova | 03:46 | |
| *** penick has quit IRC | 03:52 | |
| *** sdake_ has quit IRC | 03:52 | |
| *** yingjun has quit IRC | 03:54 | |
| *** isd has joined #openstack-nova | 04:00 | |
| *** vladikr has quit IRC | 04:03 | |
| *** unicell has joined #openstack-nova | 04:05 | |
| *** vilobhmm11 has quit IRC | 04:09 | |
| *** achanda has quit IRC | 04:11 | |
| *** rajesht has joined #openstack-nova | 04:12 | |
| *** unicell1 has joined #openstack-nova | 04:14 | |
| *** unicell has quit IRC | 04:15 | |
| *** VW_ has quit IRC | 04:22 | |
| *** sdake has joined #openstack-nova | 04:24 | |
| jhesketh | dansmith: sorry it took longer than expected, but th now passes on https://review.openstack.org/#/c/174480/ | 04:25 |
| *** bkopilov has joined #openstack-nova | 04:25 | |
| *** packet has quit IRC | 04:26 | |
| *** yingjun has joined #openstack-nova | 04:45 | |
| *** yingjun has quit IRC | 04:46 | |
| *** ajo has joined #openstack-nova | 04:49 | |
| *** ildikov has quit IRC | 04:54 | |
| *** signed8bit has quit IRC | 05:02 | |
| *** flwang1 has quit IRC | 05:05 | |
| *** ildikov has joined #openstack-nova | 05:07 | |
| *** sudipto has joined #openstack-nova | 05:09 | |
| *** _gryf|afk is now known as _gryf | 05:14 | |
| *** achanda has joined #openstack-nova | 05:16 | |
| *** josecastroleon has joined #openstack-nova | 05:17 | |
| *** Nikolay_St has quit IRC | 05:17 | |
| *** emagana has joined #openstack-nova | 05:19 | |
| *** josecastroleon has quit IRC | 05:19 | |
| *** openstackgerrit has quit IRC | 05:22 | |
| *** openstackgerrit has joined #openstack-nova | 05:22 | |
| *** harlowja is now known as harlowja_away | 05:23 | |
| *** fazal_ has quit IRC | 05:33 | |
| *** cfriesen has quit IRC | 05:34 | |
| *** kaufer has quit IRC | 05:37 | |
| openstackgerrit | Mh Raies proposed openstack/python-novaclient: Adding missing nova read only CLI test https://review.openstack.org/171148 | 05:40 |
| *** kiran-r has joined #openstack-nova | 05:40 | |
| *** otter768 has joined #openstack-nova | 05:45 | |
| *** ildikov has quit IRC | 05:45 | |
| *** ajayaa has joined #openstack-nova | 05:46 | |
| *** claudiub has joined #openstack-nova | 05:49 | |
| *** garyk has joined #openstack-nova | 05:50 | |
| *** otter768 has quit IRC | 05:50 | |
| *** yingjun has joined #openstack-nova | 05:56 | |
| *** isd has quit IRC | 05:57 | |
| *** Murali has joined #openstack-nova | 06:00 | |
| *** josecastroleon has joined #openstack-nova | 06:00 | |
| *** tojuvone has joined #openstack-nova | 06:03 | |
| *** Hosam has joined #openstack-nova | 06:05 | |
| *** oro has joined #openstack-nova | 06:06 | |
| *** armax has quit IRC | 06:08 | |
| sudipto | johnthetubaguy, bauzas Hi good morning! | 06:10 |
| *** inhumanitas has joined #openstack-nova | 06:10 | |
| *** kiran-r has quit IRC | 06:14 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Imported Translations from Transifex https://review.openstack.org/178055 | 06:18 |
| openstackgerrit | melanie witt proposed openstack/nova: Update self with db result in InstanceInfoCache.save https://review.openstack.org/178942 | 06:18 |
| *** doron_afk has joined #openstack-nova | 06:20 | |
| *** ildikov has joined #openstack-nova | 06:22 | |
| *** dikonoor has quit IRC | 06:22 | |
| *** dikonoor has joined #openstack-nova | 06:24 | |
| *** rajesht has quit IRC | 06:24 | |
| *** rajesht has joined #openstack-nova | 06:24 | |
| *** fawadkhaliq has joined #openstack-nova | 06:27 | |
| *** oro has quit IRC | 06:38 | |
| *** kiran-r has joined #openstack-nova | 06:40 | |
| *** kiran-r has quit IRC | 06:46 | |
| *** romainh has joined #openstack-nova | 06:48 | |
| *** kevinbenton has quit IRC | 06:49 | |
| *** kevinbenton has joined #openstack-nova | 06:50 | |
| *** inhumanitas has quit IRC | 06:52 | |
| *** Longgeek has joined #openstack-nova | 06:52 | |
| *** dulek has joined #openstack-nova | 06:53 | |
| *** zul has joined #openstack-nova | 06:53 | |
| *** kiran-r has joined #openstack-nova | 06:55 | |
| *** rushiagr_away is now known as rushiagr | 06:57 | |
| *** danieru_ has quit IRC | 06:58 | |
| *** danieru has joined #openstack-nova | 06:58 | |
| *** inhumanitas has joined #openstack-nova | 06:59 | |
| *** doron_afk is now known as doron | 07:00 | |
| openstackgerrit | Eli Qiao proposed openstack/nova: API: remove admin require from compute_node_get_all_by_* from db layer https://review.openstack.org/175334 | 07:02 |
| openstackgerrit | Eli Qiao proposed openstack/nova: API: Add policy enforcement test cases for pci API https://review.openstack.org/175745 | 07:02 |
| openstackgerrit | Eli Qiao proposed openstack/nova: API: remove instance_get_all_by_host hard-code admin check from db https://review.openstack.org/177673 | 07:02 |
| openstackgerrit | Eli Qiao proposed openstack/nova: API: remove admin require for compute_node(get_all/search_by_hyperviso) from db https://review.openstack.org/175744 | 07:02 |
| openstackgerrit | Eli Qiao proposed openstack/nova: API: remove admin require for compute_node_create/update/delete from db layer https://review.openstack.org/175693 | 07:02 |
| *** jcoufal has joined #openstack-nova | 07:03 | |
| *** doron is now known as doron_afk | 07:03 | |
| *** Longgeek has quit IRC | 07:07 | |
| *** sahid has joined #openstack-nova | 07:08 | |
| *** mpaolino has joined #openstack-nova | 07:09 | |
| *** browne has quit IRC | 07:09 | |
| *** mpaolino has quit IRC | 07:10 | |
| *** bfic has joined #openstack-nova | 07:10 | |
| *** dulek has quit IRC | 07:11 | |
| *** zul has quit IRC | 07:12 | |
| *** lpetrut has joined #openstack-nova | 07:12 | |
| *** zul has joined #openstack-nova | 07:12 | |
| *** subscope has quit IRC | 07:12 | |
| *** ociuhandu has joined #openstack-nova | 07:14 | |
| *** subscope has joined #openstack-nova | 07:19 | |
| *** yingjun has quit IRC | 07:22 | |
| *** Longgeek has joined #openstack-nova | 07:22 | |
| *** alexpilotti has joined #openstack-nova | 07:23 | |
| *** dikonoor has quit IRC | 07:25 | |
| *** dikonoor has joined #openstack-nova | 07:26 | |
| *** nkrinner has joined #openstack-nova | 07:27 | |
| *** gtt116_ has joined #openstack-nova | 07:28 | |
| *** gtt116__ has quit IRC | 07:30 | |
| *** oro has joined #openstack-nova | 07:33 | |
| *** oro has quit IRC | 07:39 | |
| *** jlanoux has joined #openstack-nova | 07:40 | |
| *** mpavone has joined #openstack-nova | 07:40 | |
| *** Hosam_ has joined #openstack-nova | 07:42 | |
| *** Hosam has quit IRC | 07:45 | |
| *** otter768 has joined #openstack-nova | 07:46 | |
| *** pixelb has joined #openstack-nova | 07:48 | |
| *** irenab_ has joined #openstack-nova | 07:48 | |
| *** irenab has quit IRC | 07:49 | |
| *** irenab_ is now known as irenab | 07:49 | |
| *** otter768 has quit IRC | 07:51 | |
| *** stefanb has quit IRC | 07:51 | |
| *** fawadkhaliq has quit IRC | 07:52 | |
| *** doron_afk is now known as doron | 08:02 | |
| *** kiran-r has quit IRC | 08:03 | |
| *** jistr has joined #openstack-nova | 08:04 | |
| *** stefanb has joined #openstack-nova | 08:04 | |
| *** jichen has joined #openstack-nova | 08:05 | |
| *** yassine_ has joined #openstack-nova | 08:05 | |
| *** derekh has joined #openstack-nova | 08:10 | |
| *** doron is now known as doron_afk | 08:13 | |
| *** yingjun has joined #openstack-nova | 08:15 | |
| openstackgerrit | Radoslav Gerganov proposed openstack/nova-specs: VMware: Native HTML5 consoles for VMware https://review.openstack.org/127283 | 08:15 |
| *** moshele has joined #openstack-nova | 08:15 | |
| openstackgerrit | Alvaro Lopez Garcia proposed openstack/nova: scheduler: weight nodes based on their cached images https://review.openstack.org/173304 | 08:16 |
| openstackgerrit | Alvaro Lopez Garcia proposed openstack/nova: Add new image cache monitor in compute nodes https://review.openstack.org/173303 | 08:16 |
| openstackgerrit | Alvaro Lopez Garcia proposed openstack/nova: compute monitors: set update delta from base class https://review.openstack.org/173987 | 08:16 |
| *** Nikolay_St has joined #openstack-nova | 08:18 | |
| *** jcoufal_ has joined #openstack-nova | 08:21 | |
| *** lucasagomes has joined #openstack-nova | 08:24 | |
| *** nkrinner has quit IRC | 08:25 | |
| *** jcoufal has quit IRC | 08:25 | |
| *** nkrinner has joined #openstack-nova | 08:27 | |
| *** redbeard has joined #openstack-nova | 08:28 | |
| *** inara` has quit IRC | 08:31 | |
| *** garyk has quit IRC | 08:31 | |
| *** garyk has joined #openstack-nova | 08:32 | |
| *** ajo has quit IRC | 08:34 | |
| *** ajo_ has joined #openstack-nova | 08:34 | |
| *** popw1 has quit IRC | 08:36 | |
| *** popw has joined #openstack-nova | 08:36 | |
| *** zul has quit IRC | 08:38 | |
| *** lpetrut has quit IRC | 08:41 | |
| *** matrohon has joined #openstack-nova | 08:43 | |
| *** salv-orl_ has joined #openstack-nova | 08:45 | |
| *** salv-orlando has quit IRC | 08:46 | |
| *** belmoreira has joined #openstack-nova | 08:46 | |
| *** pask81 has joined #openstack-nova | 08:48 | |
| *** inara has joined #openstack-nova | 08:50 | |
| *** gszasz has joined #openstack-nova | 08:52 | |
| *** pkoniszewski has joined #openstack-nova | 08:53 | |
| *** kiran-r has joined #openstack-nova | 08:56 | |
| *** zhangtralon has joined #openstack-nova | 08:57 | |
| gilliard | Hello Nova! | 08:59 |
| haypo | dansmith: hi. i tried to reply to your comments on my python 3 spec, the problem is that in some places, you ask me the opposite of other developers :-) https://review.openstack.org/176868 | 09:00 |
| eliqiao | gilliard: hello, GM! | 09:01 |
| gilliard | :) | 09:01 |
| *** achanda has quit IRC | 09:01 | |
| *** hightall has joined #openstack-nova | 09:05 | |
| *** doude has joined #openstack-nova | 09:05 | |
| johnthetubaguy | haypo: we don't always all agree, thats the fun bit, I will take a look at how thats going | 09:10 |
| haypo | johnthetubaguy: thanks :) | 09:10 |
| haypo | johnthetubaguy: i don't have a strong opposition on points raised by dansmith, except for the alternative: i want to keep this paragraph | 09:11 |
| haypo | i want to explain why we don't drop python 2 immediatly, because different developers asked for this option | 09:11 |
| haypo | (don't understand why it's not possible in practice) | 09:11 |
| johnthetubaguy | haypo: we deploy >20k hosts that have a disto on them doesn't support python2.7, if that helps with the scale of the problem? | 09:12 |
| tojuvone | johnthetubaguy: Do you think the review got to agreement what to implement or should continue the discussion on that: https://review.openstack.org//169836/ | 09:13 |
| johnthetubaguy | haypo: now we compiled python2.7 and got that out to them all, but its a massive faf, but python3.4 is the way forward, so I agree we need to work on both | 09:13 |
| johnthetubaguy | haypo: at least thats kinda where I am coming from, does that make any more sense? | 09:14 |
| *** eglynn has joined #openstack-nova | 09:14 | |
| *** yingjun has quit IRC | 09:14 | |
| johnthetubaguy | tojuvone: honestly, the whole evacuate instance HA thing is looking like it needs a summit session, and some deep thought, we keep getting it very wrong, so its probably time to take a step back | 09:15 |
| haypo | johnthetubaguy: which linux distro doesn't come with prebuilt python2? | 09:16 |
| *** oro has joined #openstack-nova | 09:20 | |
| johnthetubaguy | haypo: that would be telling, but think of distros that are released before python27 was released, its one of those | 09:21 |
| johnthetubaguy | (and think debian, not something that has access to EPEL) | 09:22 |
| *** doron_afk has quit IRC | 09:22 | |
| johnthetubaguy | haypo: the main point here, is changing the python version is a big deal, in lots of ways | 09:22 |
| haypo | johnthetubaguy: "distros that are released before python27 was released" these ones are not really the target of the spec. anyway, if you need to compile your own python binary, compiling python 2.7 or 3.4 should be the same :) | 09:23 |
| haypo | johnthetubaguy: my spec doesn't change the python version, it's written explicitly, no? | 09:23 |
| johnthetubaguy | haypo: right, thats why I support the spec | 09:23 |
| haypo | johnthetubaguy: my spec adds python3 support and asks to not use it :-D | 09:23 |
| johnthetubaguy | don't get me wrong, I want us to run on python3.4, its just I need time to get there | 09:24 |
| haypo | johnthetubaguy: sure | 09:24 |
| *** doron_afk has joined #openstack-nova | 09:25 | |
| tojuvone | johnthetubaguy: Ok :) This topic would also be interesting inside OPNFV session (monday), but if there is slot for Nova, I guess it would be that and also as this is more openstack internal? | 09:26 |
| *** hightall has quit IRC | 09:29 | |
| tojuvone | johnthetubaguy: Ok, I read the summit etherpad, so there would be already something on Wednesday 3:30 - 4:10. | 09:31 |
| *** fawadkhaliq has joined #openstack-nova | 09:34 | |
| *** oro has quit IRC | 09:35 | |
| *** claudiub has quit IRC | 09:36 | |
| johnthetubaguy | tojuvone: whats the OPNFV session on Monday? is that in the conference or design summit? | 09:36 |
| *** zul has joined #openstack-nova | 09:38 | |
| *** kiwik-chenrui has quit IRC | 09:38 | |
| *** ndipanov has joined #openstack-nova | 09:39 | |
| *** Hosam_ has quit IRC | 09:40 | |
| *** emagana has quit IRC | 09:40 | |
| *** ndipanov has quit IRC | 09:40 | |
| *** ndipanov has joined #openstack-nova | 09:41 | |
| openstackgerrit | Matthew Gilliard proposed openstack/nova-specs: Updates links on tags spec https://review.openstack.org/178998 | 09:43 |
| *** otter768 has joined #openstack-nova | 09:47 | |
| *** lpetrut has joined #openstack-nova | 09:49 | |
| tojuvone | <johnthetubaguy>: Seems to be still more open that OPNFV session about what to bring up. Mostly I guess for OPNFV projects to introduce the needs against openstack. Here is the etherpad: https://etherpad.opnfv.org/p/Vancouver_OpenStack | 09:49 |
| *** otter768 has quit IRC | 09:52 | |
| johnthetubaguy | tojuvone: ah, interesting: https://openstacksummitmay2015vancouver.sched.org/event/dbd2188f366a132cd38bd3e3811cb338#.VUH7561VhBc | 09:55 |
| johnthetubaguy | thanks | 09:55 |
| *** bfic has quit IRC | 09:56 | |
| *** dims has joined #openstack-nova | 09:56 | |
| *** jlanoux has quit IRC | 09:59 | |
| openstackgerrit | zhangtralon proposed openstack/nova: DB:support for migration objects multi-statues https://review.openstack.org/169604 | 10:01 |
| sudipto | bauzas, I understood the object model to an extent...wanted to ask you a few things. | 10:01 |
| *** bfic has joined #openstack-nova | 10:03 | |
| _gryf | johnthetubaguy, I'm looking on the summit-ideas etherpd, and under "Performance of unshelving instance booted from image is too slow" topic there is Make evacuate not bad | 10:03 |
| tojuvone | <johnthetubaguy> So at least on OPNFV sessions there might be NFV side perspective shown more to OpenStack guys :) For "Evacuate" there is at least Doctor (fault and maintenance) project and HA project. | 10:03 |
| *** IanGovett has joined #openstack-nova | 10:03 | |
| _gryf | johnthetubaguy, so maybe indeed there is a good chance to make a session for the evacuation state | 10:03 |
| _gryf | sooner the better | 10:03 |
| openstackgerrit | zhangtralon proposed openstack/nova: DB:support for migration objects multi-statuses https://review.openstack.org/169604 | 10:06 |
| *** fawadkhaliq has quit IRC | 10:08 | |
| *** mwagner_lap has quit IRC | 10:13 | |
| *** jlanoux has joined #openstack-nova | 10:13 | |
| heyongli | johnthetubaguy, i put some pci topic to nova summit etherpad, hope there are some time slot avaliable. | 10:14 |
| johnthetubaguy | heyongli: what line did you add it in? | 10:14 |
| johnthetubaguy | ah, 267? | 10:15 |
| openstackgerrit | sahid proposed openstack/nova: console: clean tokens do not happen for all kind of consoles https://review.openstack.org/179014 | 10:15 |
| heyongli | johnthetubaguy, at the ehthpad's bottom | 10:15 |
| *** openstackgerrit_ has joined #openstack-nova | 10:16 | |
| heyongli | johnthetubaguy, above "List of slots available", is there ok? | 10:16 |
| johnthetubaguy | heyongli: for those, I think its best to get some specs up first, we could discuss those in the un-conference if needed, there is a session on the resource tracker already, so those bits will hopefully be covered in there, with any luck | 10:17 |
| heyongli | johnthetubaguy, sure, there is some bp, i will try to get that spec or bug ready. | 10:18 |
| openstackgerrit | zhangtralon proposed openstack/nova: DB:support for migration objects multi-statuses https://review.openstack.org/169604 | 10:18 |
| openstackgerrit | zhangtralon proposed openstack/nova: add support for muti_statuses_query from migration https://review.openstack.org/169631 | 10:19 |
| *** bfic has quit IRC | 10:21 | |
| sudipto | johnthetubaguy, if I drop the scheduled_at column from the instance table and there are no references in the code that actually updates or acts on that field, can you tell me what kinda things I should be taking care of? | 10:21 |
| johnthetubaguy | sudipto: well, its an object version issue here, the old version that has that field need to have it dropped before its saved into the new DB | 10:24 |
| sudipto | johnthetubaguy, it seems, I'd have to change the object version | 10:24 |
| sudipto | johnthetubaguy, yeah, ok - so I was wondering about how the upgrade actually takes place. | 10:25 |
| johnthetubaguy | sudipto: its an API change, so yes, all API change mean a version change | 10:25 |
| sudipto | johnthetubaguy, but none of the APIs seem to be using this field at all...in such a case too, it's an API change? As in the references have been removed I feel. | 10:26 |
| johnthetubaguy | sudipto: so normally its a case of converting back to older objects (in this case it might be just add the extra null value), I am not certainly myself without digging in the code and reminding myself | 10:26 |
| johnthetubaguy | sudipto: the filed its-self is part of the API | 10:26 |
| heyongli | johnthetubaguy, beside pci resource management, there are othere thing about pci most of is make it stable, get a trusted ci, or minor feature, hope that could get some time slot also. | 10:26 |
| *** tjones1 has joined #openstack-nova | 10:27 | |
| johnthetubaguy | heyongli: its unlikely at this point, as we have all the slots mostly full, by we can consider it for sure, can I answer the questions you have about that here? I thought we agreed direction on the CI and were just waiting for someone to create one, and it sounded like the mellanox one was very close to that | 10:29 |
| *** flwang1 has joined #openstack-nova | 10:29 | |
| heyongli | johnthetubaguy, we had CI, you might not know, so i just want to said, we have, and we should make it stable. | 10:30 |
| *** sudipto_ has joined #openstack-nova | 10:30 | |
| *** sudipto has quit IRC | 10:30 | |
| *** mwagner_lap has joined #openstack-nova | 10:30 | |
| sudipto_ | johnthetubaguy, sorry, i got disconnected. Basically converting back to older objects would happen when you downgrade your controller? | 10:31 |
| heyongli | Intel PCI CI/SRIOV CI cover most of PCI testing, there might some gap, we can discuss on that at any time slot, then hope that will make pci get some progress on feature and stabibility. | 10:31 |
| *** openstackgerrit_ has quit IRC | 10:31 | |
| johnthetubaguy | sudipto_: compute nodes are older than the control plane, nova-conductor will downgrade | 10:31 |
| johnthetubaguy | heyongli: that was the one where it didn't run the upstream tempest tests I guess | 10:31 |
| *** apuimedo has joined #openstack-nova | 10:32 | |
| johnthetubaguy | heyongli: so my question is, what do you need to discuss with the rest of the Nova community, what feedback/answers are you wanting from the session? it seems like the direction you need to go is already agreed, I guess I am missing something? | 10:32 |
| heyongli | johnthetubaguy, yes, but we can run that also, and also hope pci testing case running now could be into tempest. | 10:33 |
| johnthetubaguy | heyongli: I suspect they would accept good tests if they were proposed, but thats a QA session really, not a Nova session as such, and probably best just to get the tests up for review in tempest | 10:33 |
| sudipto_ | johnthetubaguy, Ok so it means that the compute is sending us an object over RPC and we don't have a way to handle it in the control plane... | 10:34 |
| johnthetubaguy | sudipto_: sorry, I am being unclear, we really need to get this documented somewhere | 10:34 |
| johnthetubaguy | sudipto_: the new control place sends new objects to the compute node | 10:35 |
| heyongli | johnthetubaguy, ok got it, i might need move the CI part to QA team session. | 10:35 |
| johnthetubaguy | heyongli: I would just submit your patches, or join one of their meetings, I don't think a summit session will be needed, but I could be wrong | 10:35 |
| *** zul has quit IRC | 10:35 | |
| johnthetubaguy | sudipto_: then compute rejects that as its on old code | 10:36 |
| johnthetubaguy | sudipto_: nova-conductor and then convert it into the version that code can accept | 10:36 |
| *** yamahata has quit IRC | 10:36 | |
| *** sudipto has joined #openstack-nova | 10:37 | |
| johnthetubaguy | sudipto_: in your case, it might just be adding the null extra property: https://github.com/openstack/nova/blob/master/nova/objects/instance.py#L323 | 10:37 |
| sudipto | johnthetubaguy, got it, basically - if you get a compute data that has that field, we would just null it in the instance.py | 10:38 |
| heyongli | johnthetubaguy, maybe not deserve a summit meeting, if pci featues/bug fix won't be blocke by ci. | 10:38 |
| *** bfic has joined #openstack-nova | 10:39 | |
| johnthetubaguy | sudipto: yeah, its possible the base logic already adds in the defaults, I can't remember, but thats the stuff that needs checking, its backporting the object for older compute nodes, I hope that helps a bit, the spec doesn't need to be over detailed | 10:39 |
| *** ociuhandu has quit IRC | 10:40 | |
| *** sudipto_ has quit IRC | 10:40 | |
| heyongli | and cause pci is not a priority task so far, so it's progressed slow, i'like seek way to change this situation if possible. | 10:40 |
| johnthetubaguy | heyongli: yeah, clean ups we would welcome, but yeah, we need a CI before we add too many more features in that area really, I am just not sure what questions you need answers in the summit session, we need discussion in there, not present of plans, if that makes sense | 10:41 |
| johnthetubaguy | heyongli: CI is the key bit, lets get it well tested, and that will need stability bug fixes, and they are high priority at the moment | 10:41 |
| johnthetubaguy | (or were, we haven't decided the priorities for liberty yet) | 10:41 |
| sudipto | johnthetubaguy, yeah this actually helps. If the instance object is being sent by the compute, and it has sent an older version that has the scheduled_at field, we would have to make sure we don't call a 'save' on that, since in reality that column wouldn't exist. | 10:43 |
| *** josecastroleon has quit IRC | 10:44 | |
| *** exploreshaifali has joined #openstack-nova | 10:45 | |
| *** pask81 has quit IRC | 10:45 | |
| johnthetubaguy | sudipto: hmm, thats a good point, we don't have online migrations yet, which would help with that, hmm… I am guessing the object experts could answer this one quickly, I would just get the spec up with notes on where you are not sure | 10:45 |
| sudipto | johnthetubaguy, ok.. | 10:45 |
| johnthetubaguy | sudipto: I was thinking about the opposite issues actually, new objects sent to the old compute nodes, the plan in my head says we don't drop the old collumn till the old computes are upgraded | 10:46 |
| johnthetubaguy | sudipto: its possible we could just drop the column a release after we drop the object property, but I suspect there is a better way to do that | 10:46 |
| heyongli | johnthetubaguy, i agree, i try to fix pci issue now, and working on CI.we had ci ruuning thirdpart testcases, seem this is not enough. question is what a ci meet the testing requiemnts, all testcases tempest and be well reviewed? any more? | 10:47 |
| sudipto | johnthetubaguy, hmm yeah the other way is also a very legit case, where the compute expects to work on that field but then that's not there... | 10:47 |
| sudipto | johnthetubaguy, bauzas seems to be away today...any other object version expert...maybe Dan Smith? | 10:47 |
| johnthetubaguy | sudipto: as discussed in that summit session, basically we expect the same as any other third party CI, to start with, that includes saying which tempest tests you skip, and give the logs of the tempest tests you have run: http://ci.openstack.org/third_party.html | 10:49 |
| heyongli | or this is qa team's problem, but i would like get your suggestion, thansk, johnthetubaguy | 10:49 |
| sudipto | johnthetubaguy, ok, sorry i wasn't a part of the summit session :( ok...thankx... | 10:49 |
| johnthetubaguy | sudipto: ideally all tempest tests, but some might not make sense or might not work, and thats fine, if you publish the list, then we can work on from that | 10:50 |
| *** bfic has quit IRC | 10:50 | |
| sudipto | johnthetubaguy, ok.,.. | 10:50 |
| johnthetubaguy | sudipto: its possible that wasn't clear, sorry for that, but thats what I though we agreed, running the tempest tests in the usual way | 10:50 |
| johnthetubaguy | sudipto: I agree you might need to add some tests to tempest, to get extra coverage you need, and you might need to skip some tempest tests as they don't make sense, but thats the general idea anyways | 10:51 |
| johnthetubaguy | sudipto: sorry, I cross the streams here | 10:52 |
| johnthetubaguy | heyongli: sorry those messages were for you about this: http://ci.openstack.org/third_party.html | 10:52 |
| johnthetubaguy | sudipto: yeah, dansmith should be able to help you with that when he wakes up and comes online | 10:52 |
| *** alexpilotti has quit IRC | 10:52 | |
| johnthetubaguy | sudipto: I would get your draft spec uploaded though, that will make it easier to talk about | 10:53 |
| sudipto | johnthetubaguy, ok, so this is the way i understood. If we remove this field, some tempest CIs may fail...and hence we need to list all the failures? | 10:53 |
| johnthetubaguy | heyongli: what do you mean by the "qa team" here, not sure I undertood? | 10:53 |
| *** salv-orlando has joined #openstack-nova | 10:53 | |
| johnthetubaguy | sudipto: sorry, my bad, the existing test frameworks should check you code quite nicely here | 10:53 |
| *** doude has quit IRC | 10:53 | |
| *** bfic has joined #openstack-nova | 10:53 | |
| sudipto | johnthetubaguy, ok, sorry for asking too many questions, but once we submit this spec, should we also ensure ourselves that the tempest use cases pass? | 10:54 |
| johnthetubaguy | heyongli: the qa team maintains the tempest code, feature developers who want extra coverage should help write tempest tests for them, ideally, but the vendor specific CI should be maintained by the vendor, and ideally run all the tempest tests | 10:54 |
| johnthetubaguy | sudipto: yes, you need tempest passing, but thats done for you when you upload your patches for reviews, sorry the other conversation is about testing some hardware specific features in a new test framework | 10:55 |
| johnthetubaguy | sudipto: I meant to send the tempest things to heyongli but my fingers didn't do that | 10:56 |
| sudipto | johnthetubaguy, oh :D Ok np... :) | 10:56 |
| sudipto | johnthetubaguy, I think I will get the spec up first and then go from there... | 10:56 |
| *** salv-orl_ has quit IRC | 10:56 | |
| sudipto | johnthetubaguy, but thanks for your consultation, I know the version miss matches that we are dealing with... | 10:56 |
| johnthetubaguy | sudipto: sounds good, I think you have all the key bits now, the spec doesn't need too much detail, just enough so we agree with the approach | 10:56 |
| sudipto | johnthetubaguy, one last thing, if i submit the patch, tempest runs, it fails gives a -1 - then do I need to get in touch with someone to have it fixed? | 10:57 |
| sudipto | johnthetubaguy, this is a more generic question, i suppose than being related to this spec :) | 10:58 |
| sudipto | ...and a bit novice as well. | 10:58 |
| johnthetubaguy | sudipto: it almost always means its your code thats broken, but we can work through that with you | 11:00 |
| sudipto | johnthetubaguy, yeah ok... | 11:00 |
| *** Hosam has joined #openstack-nova | 11:00 | |
| sudipto | johnthetubaguy, yeah i meant a case, where my code has broken the tests. | 11:00 |
| johnthetubaguy | sudipto: that means you broke the API, so you need to not do that, basically | 11:00 |
| sudipto | johnthetubaguy, yeah makes sense. Got it! | 11:01 |
| sudipto | johnthetubaguy, thanks for your time. There's a huge power cut that has happened here post thunderstorms, so i am running on battery and praying that the power would be back soon :D | 11:02 |
| johnthetubaguy | sudipto: eeek, thats happened to me before | 11:02 |
| johnthetubaguy | sudipto: time to turn down the screen brightness, etc! | 11:02 |
| heyongli | johnthetubaguy, got some valueable information from you, thansk. seems we need run tempest tests also. and i tried to put tests to tempest, but not get approved. | 11:02 |
| sudipto | johnthetubaguy, yeah i will turn off the laptop till the power is back. thanks again for your time! | 11:02 |
| johnthetubaguy | heyongli: yes, basically the idea is that you want as much of normal nova things to work when running your particular combination of components | 11:03 |
| johnthetubaguy | sudipto: no problem, happy to help | 11:03 |
| *** sudipto has quit IRC | 11:03 | |
| heyongli | johnthetubaguy, from your point of view, before pci could move on, the third party ci must get vote right? thanks your suggestion about runing all tempest cases, i now know why we need that -:) | 11:07 |
| johnthetubaguy | heyongli: we would welcome bug fixes and stability work, thats fine, features, I really worry before we get some CI in place | 11:07 |
| *** rushiagr is now known as rushiagr_away | 11:08 | |
| heyongli | johnthetubaguy, get CI in place, could you talk about more about this? get all tempest cases running on third party ci, i get it, and is thre more concern, it's valuable for pci, thanks. | 11:10 |
| *** zhangtralon has quit IRC | 11:13 | |
| *** kiran-r has quit IRC | 11:15 | |
| ttx | johnthetubaguy: about to retag Nova final | 11:18 |
| *** markvoelker has quit IRC | 11:18 | |
| johnthetubaguy | ttx: awesome, thank you | 11:18 |
| *** josecastroleon has joined #openstack-nova | 11:19 | |
| *** jamielennox is now known as jamielennox|away | 11:20 | |
| *** aix has joined #openstack-nova | 11:21 | |
| johnthetubaguy | heyongli: it has to be running and stable, and eventually voting, if that makes sense | 11:22 |
| *** danieru has quit IRC | 11:23 | |
| *** danieru has joined #openstack-nova | 11:23 | |
| *** mwagner_lap has quit IRC | 11:24 | |
| *** yassine_ has quit IRC | 11:25 | |
| openstackgerrit | jichenjc proposed openstack/nova: Do not delete volumes during reschedule https://review.openstack.org/169097 | 11:26 |
| *** yassine_ has joined #openstack-nova | 11:26 | |
| *** bfic has quit IRC | 11:29 | |
| heyongli | johnthetubaguy, ok, i know your concern now, thanks, really. that CI will running all tempest code soon, and we will post testing statics somewhere to proofe stability, and working on to get vote. thanks so much of your time. now time to offline for me. see you next day,and summit. | 11:30 |
| *** yassine_ has quit IRC | 11:30 | |
| johnthetubaguy | heyongli: thanks for you hard work pushing on this! talk soon | 11:30 |
| *** Longgeek_ has joined #openstack-nova | 11:31 | |
| *** Longgeek has quit IRC | 11:32 | |
| *** ociuhandu has joined #openstack-nova | 11:32 | |
| ndipanov | very interesting bug I found yesterday: https://bugs.launchpad.net/nova/+bug/1450438 | 11:33 |
| openstack | Launchpad bug 1450438 in oslo-incubator "loopingcall: if a time drift to the future occurs, all timers will be blocked" [Undecided,New] | 11:33 |
| *** ferest has joined #openstack-nova | 11:33 | |
| openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: convert imagebackend to support nova.virt.image.model classes https://review.openstack.org/132024 | 11:35 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: virt: convert VFS API to use nova.virt.image.model https://review.openstack.org/132022 | 11:35 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: virt: convert disk API over to use nova.virt.image.model https://review.openstack.org/132023 | 11:35 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: virt: introduce model for describing local image metadata https://review.openstack.org/132020 | 11:35 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: virt: convert disk mount API to use nova.virt.image.model https://review.openstack.org/132021 | 11:35 |
| *** alexpilotti has joined #openstack-nova | 11:36 | |
| heyongli | johnthetubaguy, sure, bye | 11:37 |
| *** kiran-r has joined #openstack-nova | 11:37 | |
| *** ajayaa has quit IRC | 11:38 | |
| *** Kevin_Zheng has quit IRC | 11:40 | |
| *** liusheng has quit IRC | 11:40 | |
| *** hanzhang has quit IRC | 11:41 | |
| openstackgerrit | jichenjc proposed openstack/nova: Restore shelved instance original state https://review.openstack.org/144889 | 11:42 |
| *** doron_afk has quit IRC | 11:43 | |
| *** derekh has quit IRC | 11:44 | |
| *** derekh has joined #openstack-nova | 11:45 | |
| *** doron_afk has joined #openstack-nova | 11:46 | |
| *** popw has quit IRC | 11:47 | |
| *** rfolco has joined #openstack-nova | 11:48 | |
| *** otter768 has joined #openstack-nova | 11:48 | |
| *** yassine_ has joined #openstack-nova | 11:48 | |
| *** alexpilotti has quit IRC | 11:48 | |
| *** pask81 has joined #openstack-nova | 11:49 | |
| *** zul has joined #openstack-nova | 11:50 | |
| *** jcoufal_ is now known as jcoufal | 11:51 | |
| *** otter768 has quit IRC | 11:53 | |
| *** kiran-r has quit IRC | 11:54 | |
| *** kiranr has joined #openstack-nova | 11:54 | |
| *** baoli has joined #openstack-nova | 11:56 | |
| *** raildo has joined #openstack-nova | 11:58 | |
| *** VW_ has joined #openstack-nova | 11:58 | |
| *** kiranr has quit IRC | 11:59 | |
| *** ildikov has quit IRC | 11:59 | |
| *** VW_ has quit IRC | 12:01 | |
| *** VW__ has joined #openstack-nova | 12:01 | |
| openstackgerrit | Daniel Berrange proposed openstack/nova: objects: define the ImageMeta & ImageMetaProps objects https://review.openstack.org/76234 | 12:01 |
| gilliard | ndipanov: nice! | 12:01 |
| *** dikonoor has quit IRC | 12:02 | |
| *** ZZelle has quit IRC | 12:02 | |
| *** ZZelle has joined #openstack-nova | 12:02 | |
| openstackgerrit | Petrut Lucian proposed openstack/nova-specs: Hyper-V: Add storage QoS support https://review.openstack.org/175422 | 12:03 |
| ndipanov | gilliard, isn't it :) | 12:05 |
| *** claudiub has joined #openstack-nova | 12:05 | |
| gilliard | ndipanov: is the drift to the future necessary? Just from reading the bug report, it seems like any clock correction backwards would halt events until the clock catches up to where it was before., | 12:06 |
| *** tojuvone_ has joined #openstack-nova | 12:06 | |
| sdague | gilliard: does it recover? | 12:07 |
| ndipanov | gilliard, right | 12:07 |
| sdague | ndipanov: I guess that's the question, does it actually recover once it's gotten back to the future? | 12:07 |
| ndipanov | sdague, it does but it can be days | 12:07 |
| gilliard | I guess if the clock jumps back 2 hours, you have 2 hours until the processing restarts, right? | 12:07 |
| ndipanov | y | 12:08 |
| sdague | ndipanov: right, but sudo won't work until then right? | 12:08 |
| *** chuck__ has joined #openstack-nova | 12:08 | |
| sdague | if sudo goes back in time, I think it stops working until it gets back to the future | 12:08 |
| ndipanov | hmmm | 12:09 |
| *** mwagner_lap has joined #openstack-nova | 12:09 | |
| sdague | which means nova compute would be boned anyway | 12:09 |
| sdague | I've at least seen that in the past, maybe sudo doesn't do it any more | 12:09 |
| *** dikonoor has joined #openstack-nova | 12:10 | |
| *** zul has quit IRC | 12:10 | |
| sdague | but reverse time drift is basically going to make all kinds of things faceplant | 12:10 |
| ndipanov | gilliard, the thing with forward time drift | 12:10 |
| sdague | cache computations, etc. | 12:10 |
| gilliard | sudo works alright. | 12:10 |
| ndipanov | is that it is sure to set all the times off | 12:10 |
| *** ferest has quit IRC | 12:11 | |
| *** doron_afk is now known as doron | 12:11 | |
| *** markvoelker has joined #openstack-nova | 12:11 | |
| *** bfic has joined #openstack-nova | 12:11 | |
| *** dulek has joined #openstack-nova | 12:11 | |
| ndipanov | sdague, anyway - the solution is to use monotonic clocks for timers - not really controversial :) | 12:12 |
| *** heyongli has quit IRC | 12:12 | |
| sdague | ndipanov: sure that's fine, it's not a bad change | 12:12 |
| gilliard | oh no I changed my system clock and now all https certs are invalid and I can't browse the internet ;( | 12:12 |
| sdague | I'm just saying that the chance that your system is going to keep working if you set the clock back like that remains low | 12:13 |
| *** heyongli has joined #openstack-nova | 12:13 | |
| *** romainh has quit IRC | 12:13 | |
| *** bfic has quit IRC | 12:13 | |
| *** romainh has joined #openstack-nova | 12:13 | |
| ndipanov | sdague, yeah - NTP is assumed | 12:13 |
| sdague | ndipanov: right, but ntp doesn't set time back like that | 12:14 |
| *** bfic has joined #openstack-nova | 12:14 | |
| sdague | it does it with fast slow seconds to drift back | 12:14 |
| *** ildikov has joined #openstack-nova | 12:15 | |
| ndipanov | sdague, also bouncing a service will fix it | 12:15 |
| ndipanov | it's not by any means critical | 12:15 |
| ndipanov | just interesting... | 12:16 |
| sdague | yep | 12:16 |
| sdague | a git grep of time.time in nova does show how many other places we'll go off the rails here | 12:17 |
| sdague | I'd +2 the monotonic add though, seems sensible | 12:18 |
| *** VW__ has quit IRC | 12:19 | |
| *** pask81 has quit IRC | 12:19 | |
| ndipanov | sdague, well anything using loopingcalls is affected | 12:19 |
| ndipanov | meaning periodic tasks and heartbeat | 12:19 |
| sdague | yeh | 12:20 |
| sdague | so is that fix going to make all time.time calls monotonic? | 12:21 |
| ndipanov | well we have a now() in oslo.timeutils | 12:21 |
| ndipanov | that is set to a monotonic clock if available | 12:22 |
| ndipanov | so we should just use that I guess | 12:22 |
| sdague | yeh, that seems reasonable | 12:22 |
| ndipanov | the thing is - in py2.7 | 12:22 |
| ndipanov | monotonic is not part of stdlib | 12:22 |
| sdague | I kind of wonder what the impact is for when we write that value to disk | 12:22 |
| ndipanov | sdague, you mean in the db | 12:23 |
| sdague | ndipanov: it seems like a reasonable thing to add to requirements | 12:23 |
| ndipanov | sdague, for sure - it's like a simple wrapper around a syscall | 12:23 |
| *** Hosam has quit IRC | 12:23 | |
| sdague | the image cache writes timestamp values into checksums, right? | 12:24 |
| sdague | or am I reading this wrong | 12:24 |
| ndipanov | sdague, if it does and has no code to handle negative values (like servicegroup kinda does) | 12:24 |
| ndipanov | then it's gonna be a problem | 12:24 |
| sdague | it looks like negative values are probably ok | 12:25 |
| sdague | just stuff will live a lot longer | 12:25 |
| sdague | I guess that's where this breaks down, if you are using this value off box, it ends up causing new problems | 12:26 |
| sdague | if it's just used in a single box to ensure we don't go backwards, it's ok | 12:26 |
| ndipanov | not sure if we should use it there - basically anything that is supposed to survive restarts should use a utcnow() | 12:26 |
| sdague | but you don't want to interchange with that | 12:26 |
| ndipanov | that is synced | 12:26 |
| ndipanov | all timers should use monotonic | 12:26 |
| sdague | yeh, that seems right | 12:27 |
| sdague | that's just why I wanted to make sure that the monkey patch wasn't going to monkey patch monotonic into time.time() in other uses | 12:27 |
| ndipanov | sdague, no - it's basically what clock should the hub use | 12:27 |
| sdague | ok | 12:27 |
| ndipanov | oddly enough - oslo is not affected because py-amqp does use monotonic for heartbeats | 12:28 |
| *** bkopilov has quit IRC | 12:31 | |
| openstackgerrit | Neil Jerram proposed openstack/nova: Add new VIF type VIF_TYPE_TAP https://review.openstack.org/146914 | 12:32 |
| *** Longgeek has joined #openstack-nova | 12:32 | |
| *** Longgeek_ has quit IRC | 12:33 | |
| *** joefides has quit IRC | 12:34 | |
| *** achanda has joined #openstack-nova | 12:36 | |
| *** jistr has quit IRC | 12:39 | |
| *** achanda has quit IRC | 12:41 | |
| openstackgerrit | Neil Jerram proposed openstack/nova: Add new VIF type VIF_TYPE_TAP https://review.openstack.org/146914 | 12:41 |
| *** alexpilotti has joined #openstack-nova | 12:43 | |
| *** Longgeek has quit IRC | 12:44 | |
| *** dprince has joined #openstack-nova | 12:44 | |
| openstackgerrit | jichenjc proposed openstack/nova: Refactory test cases for live-migrate error case https://review.openstack.org/179048 | 12:44 |
| *** kaufer has joined #openstack-nova | 12:47 | |
| *** doron is now known as doron_afk | 12:48 | |
| *** Longgeek_ has joined #openstack-nova | 12:48 | |
| *** ivasev has joined #openstack-nova | 12:50 | |
| *** rgcs has joined #openstack-nova | 12:52 | |
| *** chuck__ has quit IRC | 12:53 | |
| *** jistr has joined #openstack-nova | 12:54 | |
| *** jistr is now known as jistr|biab | 12:55 | |
| *** dave-mccowan has joined #openstack-nova | 12:55 | |
| *** ivasev has quit IRC | 12:55 | |
| rgcs | Hello. Is it known that live-migration in kilo rc is broken on shared storage even when patches https://review.openstack.org/174307 and https://review.openstack.org/174059 which fix other broken issues with live-migration are applied? I'm getting a "Migration error: Cannot block migrate instance xxxxxx with mapped volumes" when attempting a normal live-migration explicitly without block-migration set. | 12:57 |
| *** joefides has joined #openstack-nova | 12:59 | |
| claudiub | hello folks, I have a little bit of a problem... can you tell me why in nova, in requirements.txt we have oslo.config>=1.9.3,<1.10.0 and on neutron, there is oslo.config>=1.11.0? | 12:59 |
| sdague | claudiub: in which branches? | 12:59 |
| claudiub | master | 12:59 |
| gilliard | rgcs: thanks. looking. | 13:00 |
| *** thomasem has quit IRC | 13:00 | |
| gilliard | rgcs: do you know if there's an open bug or anything? That error is thrown from code guarded by 'if is_block_migration ...' | 13:01 |
| rgcs | gilliard: I haven't been able to find any bug, no. That's why I came here | 13:02 |
| sdague | claudiub: it looks like nova hasn't landed the requirements update because it hasn't passed unit tests | 13:02 |
| sdague | claudiub: this is why devstack does the requirements override | 13:02 |
| gilliard | rgcs: suggest you raise a bug, please. | 13:02 |
| sdague | gilliard: I thought someone was throwing patches on this the other day | 13:02 |
| sdague | or at least I saw a tempest test I think | 13:03 |
| gilliard | sdague: the patch that raises this error was being backported to stable/juno | 13:03 |
| claudiub | sdague: I see. can you point me to the commit please? | 13:03 |
| sdague | claudiub: https://review.openstack.org/#/c/176844/9 | 13:03 |
| gilliard | sdague: am just rebasing https://review.openstack.org/#/c/157054/ so I'll look once I've finished that. | 13:04 |
| *** Longgeek_ has quit IRC | 13:04 | |
| claudiub | sdague: thanks! | 13:04 |
| *** fifieldt has quit IRC | 13:04 | |
| *** Longgeek has joined #openstack-nova | 13:05 | |
| *** bfic has left #openstack-nova | 13:07 | |
| openstackgerrit | jichenjc proposed openstack/nova-specs: flavor access create should check public/private https://review.openstack.org/172388 | 13:07 |
| haypo | i see tests/unit/ and tests/functional/ in nova/, is tests/functional/ related to Tempest? | 13:08 |
| *** artom has joined #openstack-nova | 13:08 | |
| haypo | or does nova have 3 kinds of tests: unit, functional and tempest? | 13:08 |
| openstackgerrit | jichenjc proposed openstack/nova-specs: flavor access create should check public/private https://review.openstack.org/172388 | 13:09 |
| *** xyang1 has joined #openstack-nova | 13:10 | |
| sdague | haypo: 3 kinds | 13:11 |
| haypo | sdague: ok. i asked to decide the scope of my python3 spec :-) | 13:11 |
| *** dboik has quit IRC | 13:12 | |
| haypo | it's unclear what "nova tests" stands for | 13:12 |
| *** ivasev has joined #openstack-nova | 13:12 | |
| sdague | I think it's unit tests first, then working tempest run | 13:14 |
| sdague | the functional tests are something we don't intend to run on multiple python versions | 13:14 |
| openstackgerrit | jichenjc proposed openstack/nova-specs: flavor access create should check public/private https://review.openstack.org/172388 | 13:15 |
| haypo | sdague: do you mean that it would be acceptable to restrict the spec to unit tests? | 13:16 |
| rgcs | gilliard: Ok opened https://bugs.launchpad.net/nova/+bug/1450471 sorry should it be a duplicate | 13:18 |
| openstack | Launchpad bug 1450471 in OpenStack Compute (nova) "live-migration fails on shared storage with "Cannot block migrate"" [Undecided,New] | 13:18 |
| gilliard | rgcs: thanks. There's definitely something off in the code there. | 13:19 |
| rgcs | as far as I understand it, it should never even attempt to do block migration in this setup (and it didn't in juno) | 13:20 |
| rgcs | thanks for looking into it | 13:20 |
| jaypipes | gilliard: mornin! :) | 13:21 |
| gilliard | Hi jaypipes :) | 13:21 |
| jaypipes | gilliard: so, are you OK with those steps I outlined for tasks the API WG liaison would be responsible for? | 13:21 |
| gilliard | jaypipes: Yeah! I think that's sound. | 13:22 |
| *** VW_ has joined #openstack-nova | 13:22 | |
| jaypipes | gilliard: OK, cool. The API WG guys had a suggestion for #4 to be "find an API WG person to write a guideline spec if not comfortable doing so", which I thought was a good suggestion. | 13:23 |
| *** thomasem has joined #openstack-nova | 13:23 | |
| gilliard | jaypipes: I think that's what informally happens at the moment, but if it's written down somewhere like you said it, then we can share the work away from just being oomichi | 13:23 |
| jaypipes | righty.. | 13:23 |
| *** eharney has quit IRC | 13:24 | |
| gilliard | jaypipes: are you OK with the suggested meeting time 1200UTC/Fridays? | 13:24 |
| jaypipes | gilliard: coolio. I'll make a wiki page listing you and alex_xu being the current API WG liaisons along with those 4 tasks. | 13:25 |
| gilliard | jaypipes: awesome :) | 13:25 |
| jaypipes | gilliard: sure, I'm fine with that, though I'll prolly miss a few meetings. | 13:25 |
| alex_xu | jaypipes: cool | 13:25 |
| gilliard | Hey alex_xu :) | 13:25 |
| jaypipes | alex_xu: hey :) | 13:25 |
| alex_xu | gilliard: jaypipes hey folks :) | 13:25 |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Added server tags support in nova-api https://review.openstack.org/128940 | 13:25 |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Added method exists to the Tag object https://review.openstack.org/178103 | 13:25 |
| jaypipes | alex_xu: you cool with those 4 tasks I listed in the email, yes? | 13:25 |
| alex_xu | jaypipes: I didn't get time go into the detail, I just saw the first, and first one is cool | 13:26 |
| alex_xu | let me quick check others | 13:26 |
| jaypipes | alex_xu: ok, if you wouldn't mind, pls have a detailed read and lemme know if cool. | 13:26 |
| *** mriedem has joined #openstack-nova | 13:26 | |
| gilliard | rgcs: I suspect the problem may be that the block_migration parameter in this function is not used: https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L5174-L5202 | 13:26 |
| *** thangp has joined #openstack-nova | 13:27 | |
| gilliard | rgcs: The libvirt driver looks for it in a the migrate_data dict instead. | 13:27 |
| alex_xu | jaypipes: ok, got it, I will repiled before the end of day | 13:27 |
| *** zhenguo has quit IRC | 13:27 | |
| rgcs | gilliard: hm is there a simple hack I could apply for a quick test? | 13:29 |
| jaypipes | alex_xu: cheers | 13:29 |
| *** lsmola_ is now known as lsmola | 13:29 | |
| *** dboik has joined #openstack-nova | 13:30 | |
| *** dboik has quit IRC | 13:30 | |
| alex_xu | jaypipes: would you like give some suggestion on this guideline https://review.openstack.org/177778 when you have time? Hope to hear your opinion :) | 13:31 |
| openstackgerrit | garyk proposed openstack/nova: VMware: create commom object for instance Limits https://review.openstack.org/179059 | 13:31 |
| *** dboik has joined #openstack-nova | 13:31 | |
| jaypipes | alex_xu: will do today. had that on my todo list for a while now. sorry for the delay. | 13:31 |
| openstackgerrit | garyk proposed openstack/nova: VMware: Resource limits for memory https://review.openstack.org/179060 | 13:31 |
| alex_xu | jaypipes: thanks :) no hurry | 13:31 |
| openstackgerrit | garyk proposed openstack/nova: VMware: store extra_specs object https://review.openstack.org/179061 | 13:32 |
| openstackgerrit | garyk proposed openstack/nova: VMware: add resource limits for disk https://review.openstack.org/179062 | 13:32 |
| *** sdake_ has joined #openstack-nova | 13:34 | |
| *** vladikr has joined #openstack-nova | 13:35 | |
| *** zul has joined #openstack-nova | 13:35 | |
| *** kaufer has quit IRC | 13:35 | |
| sdague | hmmm so this unit test failure is interesting in that we now get some odd coupling with oslo.utils | 13:36 |
| *** sdake has quit IRC | 13:38 | |
| *** exploreshaifali has quit IRC | 13:38 | |
| openstackgerrit | garyk proposed openstack/nova: VMware: set create_virtual_disk_spec method as local https://review.openstack.org/179066 | 13:38 |
| *** packet has joined #openstack-nova | 13:39 | |
| *** zul has quit IRC | 13:40 | |
| *** ajo_ is now known as ajo | 13:40 | |
| *** rajesht has quit IRC | 13:41 | |
| *** neelashah has joined #openstack-nova | 13:41 | |
| gilliard | rgcs: Well, that method calls into the libvirt driver to eventually cause your error here: https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L5174-L5202 | 13:43 |
| gilliard | rgcs: So I guess that inspecting the contents of migrate_data in that function will be useful. Although it looks like it's put in the debug log anyway on #5775 so perhaps you already have it. | 13:44 |
| gilliard | rgcs: and you can compare that to the value for block_migration I linked above, which is not used. | 13:44 |
| gilliard | rgcs: sorry wrong link. Error is raised here https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L5849 | 13:45 |
| *** dave-mccowan has quit IRC | 13:46 | |
| *** burt has quit IRC | 13:46 | |
| *** sdake has joined #openstack-nova | 13:47 | |
| sdague | dansmith: you up yet? | 13:48 |
| dansmith | sdague: depends, what's broken? | 13:48 |
| gilliard | ha | 13:48 |
| *** kaufer has joined #openstack-nova | 13:48 | |
| sdague | oh, wait, mriedem wrote this line | 13:49 |
| *** otter768 has joined #openstack-nova | 13:49 | |
| sdague | mriedem: https://review.openstack.org/#/c/132408/ | 13:49 |
| rgcs | gilliard: I'm not really fluent in python, so I'm a bit lost here. From the error in libvirt it seems it is convinced to have is_block_migration set to true in migrate_data while it probably should be false. Although I don't see where I can influence this. | 13:49 |
| gilliard | rgcs: I can try to replicate it but it will take some time. You could just try hard-coding block_migration=False there and seeing if the migration works. | 13:50 |
| *** sdake_ has quit IRC | 13:51 | |
| rgcs | gilliard: Yeah just got the same idea, am trying that and report back in 5 mins | 13:51 |
| sdague | mriedem: so you overloaded an unrelated test with safe encode testing, but that seems to be varying in the current oslo set of patchs | 13:52 |
| mriedem | sdague: so we can't mock encodeutils? | 13:52 |
| sdague | mriedem: well... it's returning 4 calls if you run the whole test suite | 13:52 |
| mriedem | sdague: this is a new thing? | 13:52 |
| *** jistr|biab is now known as jistr | 13:53 | |
| sdague | yes, that's blocking our requirements update | 13:53 |
| *** otter768 has quit IRC | 13:53 | |
| sdague | I'm guessing it's part of magical i18n | 13:54 |
| mriedem | sdague: link? | 13:54 |
| mriedem | self.assertEqual(2, mock_safe_decode.call_count) fails b/c call_count is 4? | 13:55 |
| sdague | https://review.openstack.org/#/c/176844/9 | 13:55 |
| sdague | yes | 13:55 |
| sdague | if you run the whole module | 13:55 |
| sdague | if you run the test, it's 2 | 13:55 |
| mriedem | we've been running the whole module since that change was merged, right? | 13:55 |
| mriedem | which was last november | 13:55 |
| sdague | mriedem: this is with a set of updated oslo libraries | 13:55 |
| sdague | which is blocking us updating any libraries | 13:56 |
| mriedem | great | 13:56 |
| *** garyk has quit IRC | 13:56 | |
| *** jecarey has joined #openstack-nova | 13:56 | |
| *** erlon has joined #openstack-nova | 13:56 | |
| sdague | something has to be leaking global state here, but I've been poking at this for the last hour and can't figure out what | 13:56 |
| mriedem | yeah, was going to say, something else must be mocking encodeutils | 13:56 |
| mriedem | i can poke at this but on a call until 9:30 | 13:57 |
| sdague | ok | 13:57 |
| rgcs | gilliard: no it doesn't work. I get the same error still MigrationError: Migration error: Cannot block migrate instance 53d225ac-1915-4cff-8f15-54b5c66c20a3 with mapped volumes. had nova-compute restarted after changing it. Still happens in pre_live_migration 5826 (since I commented the existing line out). | 13:57 |
| *** eharney has joined #openstack-nova | 13:57 | |
| openstackgerrit | Victor Stinner proposed openstack/nova-specs: Adding Python 3.4 support to Nova https://review.openstack.org/176868 | 13:58 |
| *** irenab has quit IRC | 13:59 | |
| *** armax has joined #openstack-nova | 13:59 | |
| mriedem | nova meeting? | 14:00 |
| johnthetubaguy | abhishekk mikal tjones cburgess jgrimm adrian_otto funzo mjturek jcookekhugen irina_pov krtaylor danpb alexpilotti flip214 jaypipes gilliard garyk edleafe dims moshele anteaya Nisha sileht claudiub lxsli neiljerram markus_z swamireddy alevine tonyb andreykurilin ndipanov sc68cal akuriata time for nova meeting #openstack-meeting | 14:00 |
| *** irenab has joined #openstack-nova | 14:00 | |
| *** zul has joined #openstack-nova | 14:00 | |
| *** signed8bit has joined #openstack-nova | 14:01 | |
| rgcs | gilliard: wait, I changed the is_block_migration in the if migration_data block and that didn't take. So I changed the one on 5773 to false | 14:01 |
| rgcs | gilliard: now I get something new entirely. It tells me: Migration failure: Unable to pre-create chardev file '/var/lib/nova/instances/53d225ac-1915-4cff-8f15-54b5c66c20a3/console.log': No such file or directory | 14:01 |
| rgcs | gilliard: I think I read about that one somewhere while investigating | 14:02 |
| *** dulek has quit IRC | 14:02 | |
| *** doron_afk has quit IRC | 14:03 | |
| *** dave-mccowan has joined #openstack-nova | 14:04 | |
| *** nkrinner has quit IRC | 14:04 | |
| *** browne has joined #openstack-nova | 14:04 | |
| *** irenab has quit IRC | 14:04 | |
| *** mmedvede has joined #openstack-nova | 14:06 | |
| gilliard | rgcs: thanks - update the bug if yuo find anything, I'll look at it in a couple hours | 14:06 |
| *** dave-mccowan has quit IRC | 14:06 | |
| *** mpaolino has joined #openstack-nova | 14:07 | |
| *** iamjarvo has joined #openstack-nova | 14:11 | |
| *** iamjarvo has quit IRC | 14:11 | |
| openstackgerrit | dave-mcnally proposed openstack/nova: Create a reserved aggregate filter https://review.openstack.org/178750 | 14:11 |
| *** iamjarvo has joined #openstack-nova | 14:11 | |
| *** lsmola has quit IRC | 14:12 | |
| *** unicell has joined #openstack-nova | 14:12 | |
| *** unicell1 has quit IRC | 14:13 | |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: introduce libosinfo library to set hardware policy https://review.openstack.org/149625 | 14:14 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: use osinfo when configuring network model https://review.openstack.org/149627 | 14:14 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: adding libosinfo configuration https://review.openstack.org/149626 | 14:14 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: use osinfo when configuring the disk bus https://review.openstack.org/149628 | 14:14 |
| *** cfriesen has joined #openstack-nova | 14:18 | |
| *** heyongli has quit IRC | 14:18 | |
| *** heyongli has joined #openstack-nova | 14:18 | |
| *** mpaolino has quit IRC | 14:18 | |
| *** BadCub_Hiding is now known as BadCub | 14:20 | |
| *** dave-mccowan has joined #openstack-nova | 14:20 | |
| *** tonytan4ever has joined #openstack-nova | 14:21 | |
| *** jgrimm is now known as zz_jgrimm | 14:22 | |
| *** doude has joined #openstack-nova | 14:23 | |
| *** zul has quit IRC | 14:26 | |
| *** dikonoor has quit IRC | 14:26 | |
| *** VW_ has quit IRC | 14:29 | |
| openstackgerrit | Tobias Engelbert proposed openstack/nova-specs: Add support for shared volumes between guests https://review.openstack.org/179104 | 14:29 |
| *** irenab has joined #openstack-nova | 14:31 | |
| *** eharney has quit IRC | 14:32 | |
| *** VW_ has joined #openstack-nova | 14:32 | |
| *** josecastroleon has quit IRC | 14:35 | |
| *** irenab has quit IRC | 14:35 | |
| *** dprince has quit IRC | 14:36 | |
| johnthetubaguy | beagles: let me know if I can help push any process forward for that neutron refactor stuff | 14:40 |
| bauzas | johnthetubaguy: my question was about reviewing all of that, how can I track all the stuff re: the neutron refactor, since it's not (yet?) a prio ? | 14:40 |
| bauzas | sounds hairy | 14:41 |
| johnthetubaguy | bauzas: not sure it needs tracking more than anything else at this point, sounds like there is a natural ordering of the specs, but I could be wrong | 14:41 |
| bauzas | mmmm | 14:42 |
| bauzas | okay, let's see that when the patches will come | 14:42 |
| jaypipes | johnthetubaguy: so, about the instance-tagging blueprint, if you could find another nova-driver to review, that would be great to approve that this week. :) | 14:42 |
| sdague | dansmith: so https://review.openstack.org/#/c/178942/1/nova/objects/instance_info_cache.py,cm is just a refresh from the db? | 14:43 |
| sdague | that call signature just seems weird to me, but I guess thats how objects roll? | 14:43 |
| dansmith | sdague: which seems weird? | 14:43 |
| dansmith | sdague: that makes that object inline with every other object | 14:43 |
| *** dims has quit IRC | 14:44 | |
| sdague | self.foo(...., self, ....) | 14:44 |
| dansmith | sdague: it's a staticmethod because it's called before we have an instance of that object at times | 14:44 |
| sdague | so... why isn't it called statically | 14:44 |
| dansmith | like when we're hydrating after rpc | 14:44 |
| johnthetubaguy | jaypipes: yeah, I can ask around, people have started to follow up on other +2s now, so hopefully that will get moving soon | 14:44 |
| *** dims has joined #openstack-nova | 14:44 | |
| dansmith | sdague: because of inheritance | 14:44 |
| dansmith | sdague: so if you subclassed that, you'd want the right hydration to run, not just that of your parent | 14:45 |
| *** emagana has joined #openstack-nova | 14:45 | |
| jaypipes | johnthetubaguy: coolio. thx again! | 14:45 |
| sdague | dansmith: so... if it always is going to be called on an instance then, why not just make it an instance method? | 14:45 |
| dansmith | sdague: because sometimes we call it from a class after looking up the class in the registry by name | 14:45 |
| sdague | anyway, if that's how it is, that's fine, just commenting that it makes for some awkwardness | 14:45 |
| dansmith | sdague: which still honors the potential for a subclass, because name | 14:45 |
| dansmith | sdague: okay :) | 14:46 |
| sdague | anyway, that's why I was asking "this is how this is supposed to work?" :) | 14:46 |
| sdague | +2 on the patch | 14:46 |
| dansmith | sure, no complaints | 14:46 |
| *** eharney has joined #openstack-nova | 14:47 | |
| *** dulek has joined #openstack-nova | 14:47 | |
| beagles | johnthetubaguy, yep and thanks | 14:48 |
| *** zul has joined #openstack-nova | 14:49 | |
| beagles | johnthetubaguy, a new consideration that probably needs covering here is that I believe there is a lot of related discussion on the neutron side of things for L. sc68cal might have some insight/thoughts on that | 14:49 |
| *** dulek has quit IRC | 14:53 | |
| jaypipes | alex_xu: https://wiki.openstack.org/wiki/Nova/APIWGLiaisons | 14:54 |
| *** nelsnelson has joined #openstack-nova | 14:56 | |
| gilliard | jaypipes: +1 thanks | 14:58 |
| *** zzzeek has joined #openstack-nova | 14:58 | |
| jaypipes | gilliard: hmm, sorry, I put your IRC nick as mgilliard... feel free to correct! | 14:58 |
| gilliard | no problem mjaypipes | 14:59 |
| *** dave-mccowan has quit IRC | 14:59 | |
| *** Murali has quit IRC | 15:00 | |
| *** dboik_ has joined #openstack-nova | 15:00 | |
| bauzas | alex_xu: FYI I noticed yesterday that /os-services wasn't present in the official docs for v2.1 http://developer.openstack.org/api-ref-compute-v2.1.html | 15:01 |
| bauzas | alex_xu: where should I provide a change for adding it ? | 15:01 |
| *** emagana has quit IRC | 15:02 | |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Added server tags support in nova-api https://review.openstack.org/128940 | 15:02 |
| bauzas | man, it's just midnight for him... | 15:02 |
| gilliard | alex_xu operates in every timeznoe. | 15:02 |
| *** annegentle has joined #openstack-nova | 15:02 | |
| dims | bauzas: git://git.openstack.org/openstack/api-site i believe | 15:03 |
| *** dboik has quit IRC | 15:03 | |
| bauzas | jaypipes: ^? | 15:03 |
| * bauzas jumping off for 15 mins | 15:03 | |
| *** baoli has quit IRC | 15:05 | |
| *** baoli has joined #openstack-nova | 15:07 | |
| *** hightall has joined #openstack-nova | 15:08 | |
| *** yassine_ has quit IRC | 15:08 | |
| *** oro has joined #openstack-nova | 15:10 | |
| *** apuimedo has quit IRC | 15:10 | |
| *** Marga_ has joined #openstack-nova | 15:11 | |
| *** krtaylor has quit IRC | 15:12 | |
| *** Marga_ has quit IRC | 15:15 | |
| *** zz_jgrimm is now known as jgrimm | 15:17 | |
| *** krtaylor has joined #openstack-nova | 15:18 | |
| *** _gryf is now known as _gryf|afk | 15:19 | |
| *** neelashah has quit IRC | 15:19 | |
| *** neelashah has joined #openstack-nova | 15:19 | |
| johnthetubaguy | jaypipes: there is a spec up for json-home by the way, not sure if that came up at the API-WG at all: https://review.openstack.org/#/c/130715/ | 15:20 |
| jaypipes | johnthetubaguy: oh, cool, didn't see that. thanks! | 15:21 |
| *** oro has quit IRC | 15:22 | |
| openstackgerrit | Andrew Laski proposed openstack/nova: devref: update the nova architecture doc https://review.openstack.org/177374 | 15:23 |
| *** obedmr has joined #openstack-nova | 15:23 | |
| *** belmoreira has quit IRC | 15:24 | |
| *** neelashah1 has joined #openstack-nova | 15:24 | |
| *** neelashah has quit IRC | 15:26 | |
| *** neelashah1 has quit IRC | 15:30 | |
| *** marun has joined #openstack-nova | 15:30 | |
| *** garyk has joined #openstack-nova | 15:31 | |
| *** irenab has joined #openstack-nova | 15:31 | |
| *** hemna_ has joined #openstack-nova | 15:32 | |
| *** jichen has quit IRC | 15:35 | |
| *** deepakcs has joined #openstack-nova | 15:35 | |
| deepakcs | ttx, hi, whats the reason for removal of kilo-rc-potential tag in https://bugs.launchpad.net/bugs/1438027 ? | 15:36 |
| openstack | Launchpad bug 1438027 in OpenStack Compute (nova) "backing file path isn't relative path for file-based volume drivers using qcow2 snapshots" [Low,In progress] - Assigned to Deepak C Shetty (dpkshetty) | 15:36 |
| *** irenab has quit IRC | 15:36 | |
| *** neelashah has joined #openstack-nova | 15:36 | |
| dansmith | deepakcs: because kilo is out | 15:37 |
| russellb | dansmith: can't go back in time and apply a fix to an rc? | 15:38 |
| russellb | disappointing | 15:38 |
| dansmith | russellb: delorean is in the shop :/ | 15:38 |
| *** hightall has quit IRC | 15:39 | |
| *** achanda has joined #openstack-nova | 15:39 | |
| jaypipes | russellb: disappointing indeed. | 15:39 |
| *** emagana has joined #openstack-nova | 15:39 | |
| mtreinish | dansmith: did you trade your jeep in for a delorean? | 15:39 |
| dansmith | mtreinish: heh | 15:40 |
| russellb | i need a jeep. | 15:40 |
| russellb | need! | 15:40 |
| jaypipes | mtreinish: no, for a VM cabriolet. | 15:40 |
| jaypipes | VW.. | 15:40 |
| dansmith | jaypipes: let's not get into comparing cars, mkay? :D | 15:40 |
| jaypipes | damn I've been working in cloud for too long. | 15:40 |
| jaypipes | dansmith: :) | 15:40 |
| mtreinish | jaypipes: haha | 15:40 |
| jaypipes | dansmith: Old Blue is in the garage. I know you're jealous. | 15:41 |
| *** pkoniszewski has quit IRC | 15:41 | |
| dansmith | jaypipes: hah, nice | 15:41 |
| ttx | With enough git karma, everything is possible | 15:42 |
| *** READ10 has joined #openstack-nova | 15:42 | |
| dansmith | ttx: sadly I have neither a delorean nor that amount of git karma.. sad. | 15:43 |
| *** achanda has quit IRC | 15:44 | |
| ttx | dansmith: on the good side, Kilo is the first release I'll have a real party to celebrate | 15:45 |
| ttx | meaning more than just me and myself | 15:45 |
| dansmith | ttx: oh? | 15:45 |
| *** jcoufal has quit IRC | 15:45 | |
| russellb | ooh, sounds fun | 15:45 |
| ttx | jeblair coming tonight | 15:45 |
| *** hightall has joined #openstack-nova | 15:45 | |
| *** hightall has quit IRC | 15:46 | |
| openstackgerrit | Brooks Kaminski proposed openstack/nova: Xenapi: Change generate-configdrive to use cdrom https://review.openstack.org/179135 | 15:48 |
| *** kaufer has quit IRC | 15:49 | |
| *** iamjarvo has quit IRC | 15:49 | |
| *** yamahata has joined #openstack-nova | 15:49 | |
| *** otter768 has joined #openstack-nova | 15:50 | |
| dansmith | woot! | 15:51 |
| dansmith | wedge landed | 15:51 |
| *** neelashah has quit IRC | 15:52 | |
| *** kaufer has joined #openstack-nova | 15:52 | |
| *** annegentle has quit IRC | 15:53 | |
| *** otter768 has quit IRC | 15:55 | |
| openstackgerrit | dave-mcnally proposed openstack/nova: Create a reserved aggregate filter https://review.openstack.org/178750 | 15:55 |
| *** EmilienM is now known as EmilienM|afk | 15:55 | |
| *** romainh has left #openstack-nova | 15:55 | |
| *** emagana has quit IRC | 15:57 | |
| *** angdraug has joined #openstack-nova | 15:57 | |
| cfriesen | Could one of the cores take a look at https://review.openstack.org/#/c/170780/ ? It's been basically ready to go for a couple weeks, mriedem has +2'd it three times now | 15:57 |
| *** emagana has joined #openstack-nova | 15:59 | |
| *** dave-mccowan has joined #openstack-nova | 16:01 | |
| *** baoli has quit IRC | 16:02 | |
| *** xgsa has quit IRC | 16:03 | |
| *** baoli_ has joined #openstack-nova | 16:03 | |
| *** deepakcs has quit IRC | 16:03 | |
| *** dboik_ has quit IRC | 16:04 | |
| *** rgcs has quit IRC | 16:04 | |
| *** dboik has joined #openstack-nova | 16:04 | |
| *** jlanoux has quit IRC | 16:07 | |
| *** moshele has quit IRC | 16:07 | |
| sdague | ttx: so, can you un -2 this now - https://review.openstack.org/#/c/176806/ ? | 16:08 |
| ttx | sdague: need taht in first: https://review.openstack.org/#/c/179083/ | 16:09 |
| ttx | first post-release stable/kilo must be the preversion bump so taht we generate correct version metadata | 16:10 |
| ttx | I'll un-2 all my stable/kilo once those merge | 16:10 |
| sdague | ttx: great | 16:11 |
| ttx | also fungi will revert the ACL | 16:11 |
| ttx | soi that it behaves like stable/* | 16:11 |
| *** busterswt has joined #openstack-nova | 16:14 | |
| *** salv-orlando has quit IRC | 16:17 | |
| *** Nic has joined #openstack-nova | 16:18 | |
| *** salv-orlando has joined #openstack-nova | 16:19 | |
| dansmith | sdague: okay melwitt's patch is in, keep your eyes peeled for ssh timeouts | 16:19 |
| dansmith | honestly, I really don't think this is related, but it's close enough to watch anyway | 16:19 |
| *** ijw has joined #openstack-nova | 16:21 | |
| *** Nikolay_St has quit IRC | 16:21 | |
| sdague | ok, will do | 16:21 |
| openstackgerrit | Matthew Gilliard proposed openstack/nova-specs: Cancel an ongoing live migration https://review.openstack.org/179149 | 16:23 |
| *** pankaj__ has joined #openstack-nova | 16:23 | |
| *** krtaylor has quit IRC | 16:25 | |
| *** matrohon has quit IRC | 16:25 | |
| mriedem | ttx: i'm not +2 on stable/kilo for nova for some reason? | 16:25 |
| ttx | mriedem: because against all appearances it's not stable/kilo yet | 16:26 |
| mriedem | ok | 16:26 |
| ttx | mriedem, you will stable/* rights on kilo as soon as https://review.openstack.org/179045 merges | 16:26 |
| ttx | we need https://review.openstack.org/#/q/branch:stable/kilo+topic:next-kilo,n,z in first | 16:27 |
| openstackgerrit | Marian Horban proposed openstack/nova: WIP: Graceful shutdown WSGI server https://review.openstack.org/129167 | 16:27 |
| ttx | mriedem: hopefully later today | 16:27 |
| mriedem | ok | 16:27 |
| *** tjones2 has joined #openstack-nova | 16:31 | |
| ndipanov | mriedem, jaypipes dansmith - you all love neutron api https://review.openstack.org/#/c/175742/ | 16:31 |
| *** sdake_ has joined #openstack-nova | 16:31 | |
| dansmith | ndipanov: false. | 16:32 |
| ndipanov | denial... sounds like love to me | 16:33 |
| dansmith | heh | 16:33 |
| mriedem | that change scares me | 16:33 |
| ndipanov | mriedem, why | 16:33 |
| dansmith | because it touches even a single line in neutronapi? | 16:33 |
| mriedem | that's the kind of change that will completely bust some corner case in how neutron api works | 16:33 |
| ndipanov | mriedem, what do you mean | 16:33 |
| ndipanov | it just gets rid of even considering to check neutron quotas because it makes no sense to do it | 16:34 |
| ndipanov | it doesn nothing more | 16:34 |
| ndipanov | basically it removes code that is racy by design | 16:34 |
| *** unicell has quit IRC | 16:35 | |
| *** sdake has quit IRC | 16:35 | |
| ndipanov | mriedem, it's only called from the API | 16:35 |
| ndipanov | one place | 16:35 |
| ndipanov | seemed like a win to me | 16:35 |
| mriedem | i have to dig into the change deeper | 16:36 |
| ndipanov | read my comments as well | 16:36 |
| mriedem | yes i saw that | 16:36 |
| mriedem | it's kind of what i'm talking about | 16:36 |
| *** dprince has joined #openstack-nova | 16:36 | |
| ndipanov | lol | 16:36 |
| ndipanov | mriedem, <3 | 16:38 |
| openstackgerrit | Matthew Gilliard proposed openstack/nova-specs: Cancel an ongoing live migration https://review.openstack.org/179149 | 16:38 |
| pankaj__ | is ther a way to set instance name = hostname in openstack . where hostname is generated by some external REST service base on some values of tenant/project | 16:38 |
| *** krtaylor has joined #openstack-nova | 16:38 | |
| ndipanov | name is generated from a template isn't it? | 16:39 |
| bauzas | johnthetubaguy: procedural question on https://review.openstack.org/#/c/76234/ | 16:39 |
| mriedem | ndipanov: yeah, template | 16:39 |
| *** oro has joined #openstack-nova | 16:39 | |
| mriedem | pankaj__: http://docs.openstack.org/kilo/config-reference/content/list-of-compute-config-options.html | 16:39 |
| mriedem | instance_name_template = instance-%08x | 16:40 |
| mriedem | (StrOpt) Template string to be used to generate instance names | 16:40 |
| *** zul has quit IRC | 16:40 | |
| pankaj__ | i have REST service whcih takes OS+appcode+location and generates next available hostname . can i have openstack use that for instance naming | 16:40 |
| *** mpavone has quit IRC | 16:40 | |
| *** sdake has joined #openstack-nova | 16:40 | |
| *** mmedvede has quit IRC | 16:41 | |
| *** iamjarvo has joined #openstack-nova | 16:41 | |
| mriedem | pankaj__: you can name the instance on boot, but instance_name_template is for the hypervisor i believe | 16:41 |
| dansmith | yep | 16:41 |
| bauzas | mriedem: yup | 16:41 |
| ndipanov | so there's 3 names for an instance really :) | 16:41 |
| ndipanov | simple | 16:42 |
| bauzas | not really | 16:42 |
| *** ociuhandu has quit IRC | 16:42 | |
| pankaj__ | these would be instance names in openstack . would it be same as hostname inside the VM | 16:42 |
| ndipanov | bauzas, you are right there's 4 | 16:44 |
| ndipanov | name, display_name, hostname, uuid | 16:44 |
| *** sdake_ has quit IRC | 16:44 | |
| ndipanov | pankaj__, I think nova sets it to display_name if not otherwise provided but would have to check | 16:45 |
| dansmith | bauzas: ndipanov: am I off base here? https://review.openstack.org/#/c/163440/3 | 16:46 |
| *** alexpilotti has quit IRC | 16:46 | |
| *** igordcard_ has quit IRC | 16:46 | |
| ndipanov | dansmith, well | 16:46 |
| bauzas | ndipanov: I double-checked and it seems that display_name is what the user provides | 16:46 |
| ndipanov | no | 16:46 |
| ndipanov | but... | 16:46 |
| bauzas | dansmith: looking... | 16:47 |
| ndipanov | so my but | 16:47 |
| ndipanov | (lol) | 16:47 |
| bauzas | eh | 16:47 |
| ndipanov | is with regards to timelines | 16:47 |
| ndipanov | so you are right | 16:47 |
| ndipanov | once we have the request spec that is persisted | 16:47 |
| ndipanov | and migrate the resource tracking to use it - this can go away | 16:48 |
| *** derekh has quit IRC | 16:48 | |
| *** dims has quit IRC | 16:48 | |
| ndipanov | dansmith, but it's not a trivial amount of work - so the question is whether we want to hold up a bugfix on it... | 16:49 |
| *** dims has joined #openstack-nova | 16:49 | |
| dansmith | if the bugfix is a schema change and a bunch of code to do things in one place just to move them to another place, I say: "please no" | 16:49 |
| bauzas | ndipanov: dansmith: well, you're right, it's all about paying a debt for fixing that now | 16:49 |
| ndipanov | dansmith, that might be the correct answer | 16:49 |
| dansmith | this is debt that we introduced with numa, I'd rather not make it worse with more debt, especially since we have the right fix on the near horizon | 16:50 |
| ndipanov | well | 16:50 |
| bauzas | dansmith: agreed, it's a migration... | 16:50 |
| ndipanov | it's not introduced with numa | 16:50 |
| ndipanov | it's introduced with pci | 16:51 |
| bauzas | that's the main problem IMHO, because that's persisting more debt | 16:51 |
| dansmith | um, what? | 16:51 |
| ndipanov | but numa is the same | 16:51 |
| dansmith | this is specifically about numa_topoligy | 16:51 |
| dansmith | er topology | 16:51 |
| ndipanov | right but we need the same thing for pci | 16:51 |
| dansmith | even more reason to do it right :) | 16:51 |
| dansmith | and not have *two* hacks :) | 16:51 |
| ndipanov | agreed | 16:51 |
| ndipanov | I've been meaning to write an email explaining why these 2 break a lot of asusmptions in the scheduler | 16:52 |
| openstackgerrit | Matthew Gilliard proposed openstack/nova-specs: Cancel an ongoing live migration https://review.openstack.org/179149 | 16:52 |
| ndipanov | and resource tracker | 16:52 |
| openstackgerrit | Shilla Saebi proposed openstack/nova: minor edit to policy_enforcement.rst https://review.openstack.org/178463 | 16:52 |
| bauzas | could we just focus on persisting the spec for Liberty ? | 16:53 |
| ndipanov | bauzas, well sadly it's not how it works | 16:53 |
| bauzas | I think it would be bad just to add more on the migration table | 16:53 |
| ndipanov | ah from that pov | 16:53 |
| bauzas | for that purpose | 16:53 |
| ndipanov | yes | 16:53 |
| ndipanov | I agree with you guys | 16:53 |
| ndipanov | We need a generic solution to this | 16:53 |
| johnthetubaguy | bauzas: looking at that patch | 16:53 |
| dansmith | bauzas: yes, that's what I want, to focus on the right spec | 16:54 |
| bauzas | so I would rather give an optimistic -1 saying "I disagree with your change but I'm fine if you find another solution that doesn't require persisting states in DB" | 16:54 |
| ndipanov | bauzas, not sure you can do that | 16:54 |
| bauzas | hence the word "optimistic" | 16:54 |
| ndipanov | bauzas, bare in mind that means | 16:54 |
| ndipanov | add crap to a json blob | 16:54 |
| bauzas | lol | 16:54 |
| *** melwitt has joined #openstack-nova | 16:55 | |
| bauzas | I thought we called the Canadian customs to specifically ask them to deny visas for blobs ? | 16:55 |
| dansmith | bauzas: it's still debt, even if we stuff it into something else | 16:55 |
| dansmith | probably worse debt even | 16:55 |
| ndipanov | yes - I agree | 16:55 |
| ndipanov | we should really fix it | 16:55 |
| ndipanov | but _fix_ it | 16:55 |
| bauzas | dansmith: agreed, my point is to say "there are probably other ways to do that, keep looking" | 16:56 |
| dansmith | bauzas: I don't want to send him off on a goose chase | 16:56 |
| ndipanov | bauzas, there are if you don't care about races ;) | 16:56 |
| dansmith | bauzas: we know how to fix it, let's not pretend | 16:56 |
| ndipanov | yes | 16:56 |
| *** patrickeast has joined #openstack-nova | 16:56 | |
| bauzas | dansmith: I just remember how russellb found a good plan for checking on migrations the affinity rules | 16:56 |
| *** igordcard_ has joined #openstack-nova | 16:56 | |
| bauzas | then I need more time to get a final answers than just "reqspec persistence is the plan" | 16:57 |
| bauzas | answer even | 16:57 |
| *** dsanders has joined #openstack-nova | 16:57 | |
| bauzas | dansmith: but don't get me wrong, yes it needs to be done | 16:57 |
| *** gszasz has quit IRC | 16:57 | |
| *** eglynn has quit IRC | 16:57 | |
| bauzas | and I'm hassling the guy who is just working on the first step, ie. creating a ReqSpec object | 16:57 |
| bauzas | that guy is dumb | 16:58 |
| *** igordcard_ has joined #openstack-nova | 16:58 | |
| dansmith | not dumb, just slow :) | 16:58 |
| *** exploreshaifali has joined #openstack-nova | 16:58 | |
| *** dikonoor has joined #openstack-nova | 16:58 | |
| *** igordcard_ has quit IRC | 16:58 | |
| bauzas | dansmith: he's French, you know, these dudes have lots of holidays | 16:58 |
| bauzas | starting by tomorrow, even | 16:59 |
| ndipanov | anyway bauzas - having code up we could look at would also help | 16:59 |
| *** dikonoor has quit IRC | 16:59 | |
| bauzas | ndipanov: I had a first dep on https://review.openstack.org/#/c/76234/ | 17:00 |
| bauzas | ndipanov: since that's now in a pretty good shape, I can continue to work on | 17:00 |
| *** emagana has quit IRC | 17:00 | |
| *** mtanino has joined #openstack-nova | 17:01 | |
| *** irenab has joined #openstack-nova | 17:01 | |
| *** achanda has joined #openstack-nova | 17:02 | |
| openstackgerrit | Dan Smith proposed openstack/nova-specs: Add robustify-evacuate spec https://review.openstack.org/161444 | 17:02 |
| *** Sukhdev has joined #openstack-nova | 17:02 | |
| dansmith | bauzas: ndipanov can one or both of you comment on that patch? it's a new contributor so I want to make sure it doesn't look like I'm just being a dick | 17:02 |
| dansmith | basically, I want him to have to learn that slowly like everyone else did | 17:02 |
| bauzas | dansmith: sure, will do | 17:02 |
| dansmith | thanks | 17:02 |
| ndipanov | dansmith, I talked to them on irc | 17:03 |
| ndipanov | they reported a bunch of bugs around numa | 17:03 |
| dansmith | ndipanov: in the last few minutes? | 17:03 |
| ndipanov | and actually sent an email about this | 17:03 |
| *** Nic has quit IRC | 17:03 | |
| ndipanov | in the last few years! | 17:04 |
| ndipanov | :D | 17:04 |
| *** patrickeast has quit IRC | 17:04 | |
| ndipanov | will comment no worries | 17:04 |
| ndipanov | first reading bauzas spec | 17:04 |
| dansmith | okay | 17:05 |
| *** unicell has joined #openstack-nova | 17:05 | |
| *** unicell has quit IRC | 17:05 | |
| *** vilobhmm1 has joined #openstack-nova | 17:05 | |
| *** unicell has joined #openstack-nova | 17:05 | |
| *** irenab has quit IRC | 17:06 | |
| *** unicell has quit IRC | 17:06 | |
| *** unicell has joined #openstack-nova | 17:06 | |
| *** emagana has joined #openstack-nova | 17:06 | |
| *** lucasagomes has quit IRC | 17:06 | |
| *** yamahata has quit IRC | 17:07 | |
| *** haypo has left #openstack-nova | 17:07 | |
| *** subscope has quit IRC | 17:08 | |
| *** beagles is now known as beagles_brb | 17:11 | |
| *** harlowja_away is now known as harlowja | 17:11 | |
| *** sahid has quit IRC | 17:12 | |
| *** salv-orlando has quit IRC | 17:16 | |
| *** salv-orlando has joined #openstack-nova | 17:16 | |
| *** Mike_D_laptop has joined #openstack-nova | 17:16 | |
| *** Nic has joined #openstack-nova | 17:17 | |
| *** mmedvede has joined #openstack-nova | 17:18 | |
| *** beagles_brb is now known as beagles | 17:18 | |
| *** salv-orlando has quit IRC | 17:21 | |
| *** salv-orlando has joined #openstack-nova | 17:21 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Remove orphaned tables - iscsi_targets, volumes https://review.openstack.org/166214 | 17:23 |
| *** yamahata has joined #openstack-nova | 17:28 | |
| *** tjones2 has quit IRC | 17:30 | |
| *** tjones2 has joined #openstack-nova | 17:31 | |
| *** redbeard has quit IRC | 17:32 | |
| *** emagana has quit IRC | 17:35 | |
| *** ijw has quit IRC | 17:36 | |
| *** isd has joined #openstack-nova | 17:36 | |
| *** thangp_ has joined #openstack-nova | 17:37 | |
| *** dims_ has joined #openstack-nova | 17:37 | |
| *** dboik_ has joined #openstack-nova | 17:37 | |
| *** signed8b_ has joined #openstack-nova | 17:38 | |
| *** ptm has joined #openstack-nova | 17:39 | |
| *** sdake_ has joined #openstack-nova | 17:39 | |
| *** asselin__ has joined #openstack-nova | 17:40 | |
| *** VW_ has quit IRC | 17:40 | |
| artom | Maybe someone in here knows - can I get devstack to install certain things form pip as opposed to RPMs? | 17:41 |
| *** Nikolay_St has joined #openstack-nova | 17:41 | |
| *** VW_ has joined #openstack-nova | 17:41 | |
| artom | Specifically, greenlet has disappeared from rhel7 repos... | 17:41 |
| *** anish_ has joined #openstack-nova | 17:41 | |
| *** lpetrut has left #openstack-nova | 17:41 | |
| artom | (And yes, in this case it *has* to be rhel7) | 17:41 |
| *** haomaiw__ has joined #openstack-nova | 17:41 | |
| *** jamespd has joined #openstack-nova | 17:42 | |
| *** jeblair_ has joined #openstack-nova | 17:42 | |
| *** VW__ has joined #openstack-nova | 17:42 | |
| *** VW__ has quit IRC | 17:42 | |
| *** gzhai has joined #openstack-nova | 17:43 | |
| *** VW__ has joined #openstack-nova | 17:43 | |
| *** ildikov has quit IRC | 17:45 | |
| *** coolsvapl has joined #openstack-nova | 17:45 | |
| *** cburgess_ has joined #openstack-nova | 17:45 | |
| *** rpodolyaka_ has joined #openstack-nova | 17:45 | |
| *** sdague_ has joined #openstack-nova | 17:45 | |
| *** keekz_ has joined #openstack-nova | 17:45 | |
| dansmith | artom: ask in -qa | 17:46 |
| dansmith | artom: higher density of people that would know that in there | 17:46 |
| *** VW_ has quit IRC | 17:46 | |
| dansmith | my guess is "not" | 17:46 |
| *** dims has quit IRC | 17:46 | |
| *** pankaj__ has quit IRC | 17:46 | |
| *** heyongli has quit IRC | 17:47 | |
| *** ZZelle has quit IRC | 17:47 | |
| *** haomaiwang has quit IRC | 17:47 | |
| *** kbyrne has quit IRC | 17:47 | |
| *** jamespd_ has quit IRC | 17:47 | |
| *** anish has quit IRC | 17:47 | |
| *** greghaynes has quit IRC | 17:47 | |
| *** cburgess has quit IRC | 17:47 | |
| *** rpodolyaka has quit IRC | 17:47 | |
| *** sdake has quit IRC | 17:47 | |
| *** dboik has quit IRC | 17:47 | |
| *** signed8bit has quit IRC | 17:47 | |
| *** thangp has quit IRC | 17:47 | |
| *** ivasev has quit IRC | 17:47 | |
| *** jistr has quit IRC | 17:47 | |
| *** pixelb has quit IRC | 17:47 | |
| *** jeblair has quit IRC | 17:47 | |
| *** PaulMurray has quit IRC | 17:47 | |
| *** asselin_ has quit IRC | 17:47 | |
| *** sdague has quit IRC | 17:47 | |
| *** gabriel-bezerra has quit IRC | 17:47 | |
| *** coolsvap|afk has quit IRC | 17:47 | |
| *** keekz has quit IRC | 17:47 | |
| *** mjturek1 has quit IRC | 17:47 | |
| *** gzhai1 has quit IRC | 17:47 | |
| *** salv-orlando has quit IRC | 17:47 | |
| *** bnemec has quit IRC | 17:47 | |
| *** jogo has quit IRC | 17:47 | |
| *** mkoderer has quit IRC | 17:47 | |
| *** morganfainberg has quit IRC | 17:47 | |
| *** triggerz_ has quit IRC | 17:47 | |
| *** gibi has quit IRC | 17:47 | |
| *** jesusaurus has quit IRC | 17:47 | |
| *** vipul has quit IRC | 17:47 | |
| *** devananda has quit IRC | 17:47 | |
| *** tanee has quit IRC | 17:47 | |
| *** rpodolyaka_ is now known as rpodolyaka | 17:47 | |
| *** VW_ has joined #openstack-nova | 17:47 | |
| *** sdague_ is now known as sdague | 17:47 | |
| *** tjones3 has joined #openstack-nova | 17:48 | |
| *** raildo has quit IRC | 17:48 | |
| *** tjones2 has quit IRC | 17:48 | |
| *** raildo has joined #openstack-nova | 17:49 | |
| *** salv-orlando has joined #openstack-nova | 17:49 | |
| *** bnemec has joined #openstack-nova | 17:49 | |
| *** jogo has joined #openstack-nova | 17:49 | |
| *** mkoderer has joined #openstack-nova | 17:49 | |
| *** morganfainberg has joined #openstack-nova | 17:49 | |
| *** triggerz_ has joined #openstack-nova | 17:49 | |
| *** gibi has joined #openstack-nova | 17:49 | |
| *** jesusaurus has joined #openstack-nova | 17:49 | |
| *** vipul has joined #openstack-nova | 17:49 | |
| *** devananda has joined #openstack-nova | 17:49 | |
| *** tanee has joined #openstack-nova | 17:49 | |
| *** VW__ has quit IRC | 17:51 | |
| *** otter768 has joined #openstack-nova | 17:51 | |
| *** Mike_D_laptop has quit IRC | 17:51 | |
| *** ijw has joined #openstack-nova | 17:51 | |
| *** ZZelle has joined #openstack-nova | 17:52 | |
| dansmith | bauzas: I'm continuing to spout off numa words in responses to questions in that review.. hope I'm getting it right :) | 17:53 |
| *** Nic has quit IRC | 17:53 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Move our ObjectListBase to subclass from the Oslo one https://review.openstack.org/173995 | 17:53 |
| openstackgerrit | Dan Smith proposed openstack/nova: Start the conversion to oslo.versionedobjects https://review.openstack.org/173479 | 17:53 |
| *** pixelb has joined #openstack-nova | 17:53 | |
| *** ndipanov has quit IRC | 17:53 | |
| *** sdake has joined #openstack-nova | 17:53 | |
| *** greghaynes has joined #openstack-nova | 17:54 | |
| *** gabriel-bezerra has joined #openstack-nova | 17:54 | |
| *** kbyrne has joined #openstack-nova | 17:54 | |
| *** bkopilov has joined #openstack-nova | 17:54 | |
| *** otter768 has quit IRC | 17:55 | |
| *** mjturek1 has joined #openstack-nova | 17:55 | |
| *** tjones2 has joined #openstack-nova | 17:55 | |
| *** sdake_ has quit IRC | 17:56 | |
| *** igordcard has joined #openstack-nova | 17:56 | |
| *** tjones3 has quit IRC | 17:56 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Remove orphaned tables - iscsi_targets, volumes https://review.openstack.org/166214 | 17:57 |
| zzzeek | dansmith: can you give me a 5 second answer what i need to do with https://review.openstack.org/#/c/154243/ ? change the migration number and rebase ? | 17:58 |
| *** harlowja has quit IRC | 17:58 | |
| * dansmith aims for extreme brevity | 17:59 | |
| dansmith | zzzeek: y | 17:59 |
| zzzeek | T | 17:59 |
| dansmith | zzzeek: n | 18:00 |
| dims_ | zzzeek: you and me have to duke it out for migration number 292 :) | 18:00 |
| zzzeek | dims_: i want to eradicate migration numbers | 18:00 |
| zzzeek | dims_ dansmith is there a specific nova session where we can talk about migrations? | 18:00 |
| dims_ | ++++ | 18:00 |
| dansmith | then I guess dims_ wins | 18:00 |
| dims_ | yay! | 18:01 |
| * zzzeek should have number 292 becaue that is his apartment building number | 18:01 | |
| dims_ | :) | 18:02 |
| *** irenab has joined #openstack-nova | 18:02 | |
| *** Nic has joined #openstack-nova | 18:03 | |
| *** angdraug has quit IRC | 18:05 | |
| *** annegentle has joined #openstack-nova | 18:05 | |
| *** ssurana has joined #openstack-nova | 18:07 | |
| *** aix has quit IRC | 18:07 | |
| *** iamjarvo has quit IRC | 18:08 | |
| *** lpetrut has joined #openstack-nova | 18:08 | |
| *** ijw_ has joined #openstack-nova | 18:08 | |
| *** sdake_ has joined #openstack-nova | 18:08 | |
| *** irenab has quit IRC | 18:09 | |
| *** lpetrut has left #openstack-nova | 18:09 | |
| *** isd has quit IRC | 18:09 | |
| openstackgerrit | Deepak C Shetty proposed openstack/nova: libvirt: Use 'relative' flag for online snapshot's commit/rebase operations https://review.openstack.org/168805 | 18:09 |
| *** sdake has quit IRC | 18:10 | |
| *** sdake has joined #openstack-nova | 18:10 | |
| *** EmilienM|afk is now known as EmilienM | 18:11 | |
| *** ijw has quit IRC | 18:12 | |
| *** emagana has joined #openstack-nova | 18:12 | |
| *** sdake_ has quit IRC | 18:14 | |
| *** iamjarvo has joined #openstack-nova | 18:14 | |
| *** VW_ has quit IRC | 18:18 | |
| *** Sukhdev has quit IRC | 18:20 | |
| *** baoli_ has quit IRC | 18:20 | |
| *** Sukhdev has joined #openstack-nova | 18:20 | |
| *** baoli_ has joined #openstack-nova | 18:21 | |
| sc68cal | Hi, for this spec http://specs.openstack.org/openstack/nova-specs/specs/liberty/approved/wrap-neutronclient.html - Could an alternative be to start using python-openstacksdk | 18:22 |
| *** bkopilov has quit IRC | 18:23 | |
| sc68cal | beagles: ^ | 18:23 |
| *** baoli_ has quit IRC | 18:25 | |
| *** ajo has quit IRC | 18:26 | |
| mriedem | sc68cal: what would the advantage of that be? | 18:26 |
| *** bkopilov has joined #openstack-nova | 18:27 | |
| sc68cal | mriedem: IIRC, at the summit in paris we had a talk from amotoki about the poor state of the python-neutronclient API, and I think the conclusion was to move towards python-openstacksdk which had a more developer friendly API | 18:27 |
| sc68cal | mriedem: let me dig up the etherpad from the session | 18:28 |
| *** jichen has joined #openstack-nova | 18:28 | |
| mriedem | sc68cal: but but but, | 18:29 |
| mriedem | what about python-openstackclient? | 18:29 |
| mriedem | how many clients and SDKs can we have? won't there be one to rule them all? | 18:29 |
| mriedem | why is the neutronclient API in a poor state? people making backwards incompatible changes? | 18:30 |
| mriedem | e.g. https://review.openstack.org/#/c/172646/ | 18:30 |
| mriedem | having a NeutronClientException -> NovaException translation shim would go a long way | 18:34 |
| mriedem | that's the big part of the wrap-neutronclient spec for me | 18:34 |
| beagles | sc68cal, I guess the thing is whatever client library is used it should be the one that gets precedence with tracking neutron changes. | 18:34 |
| *** bkopilov has quit IRC | 18:35 | |
| *** esmute has quit IRC | 18:35 | |
| mriedem | beagles: +1, i wouldn't want to move to something that the neutron team isn't primarily maintaining and testing | 18:35 |
| sc68cal | mriedem: the problem is the python-neutronclient API is kind of bad. You just end up constructing dictionaries with k/vs that match the Neutron API | 18:35 |
| *** bkopilov has joined #openstack-nova | 18:35 | |
| mriedem | otherwise we'd have all been on openstackclient by now right? | 18:35 |
| sc68cal | there's no classes for any of the API resources | 18:35 |
| *** patrickeast has joined #openstack-nova | 18:35 | |
| *** esmute has joined #openstack-nova | 18:35 | |
| beagles | seems a little like the arguments for the sdks are also dicts? | 18:36 |
| beagles | sorry s/sdks/opensdk/ | 18:36 |
| mriedem | sc68cal: hrm, will neutron project be deprecating the client in favor of the sdk? | 18:36 |
| mriedem | or change the client to use the sdk under the covers? | 18:36 |
| beagles | in any case I'm not opposed to one library over the other, but if it proves an extra obstacle to facilitating interaction I'm going to have failed | 18:37 |
| *** tojuvone_ has quit IRC | 18:37 | |
| sc68cal | mriedem: unknown - what is the status of python-keystoneclient? I know dtroyer replaced python-keystoneclient with openstacksdk calls in DevStack | 18:37 |
| mriedem | i thought keystoneclient was being deprecated in favor of openstackclient | 18:37 |
| beagles | (one might argue then I should contribute also openstacksdk ;) but I digress) | 18:37 |
| mriedem | openstackclient has the keystone v3 goodies | 18:38 |
| sc68cal | So then I think Neutron would probably follow the precedent set by keystone | 18:38 |
| beagles | sc68cal, can we leave that open for the moment? keep an eye out of things that the sdk library would need? | 18:38 |
| beagles | sc68cal, who would be the person on the sdk that is working on the neutron related stuff? | 18:39 |
| briancurtin | beagles: right now myself and terry howe are the primary people working on SDK. terry did most of neutron | 18:39 |
| beagles | briancurtin, do you have Terry's IRC handle? | 18:40 |
| briancurtin | terrylhowe | 18:40 |
| briancurtin | he's in #openstack-sdks and a few others, but not in here | 18:40 |
| beagles | in the immediate, I can't see targeting the SDK - but I would think that once the wrapper is established it is a natural layer under which a switchover could be made | 18:41 |
| sc68cal | beagles: does it make sense to create yet another layer? | 18:42 |
| *** iamjarvo has quit IRC | 18:43 | |
| *** bkopilov has quit IRC | 18:43 | |
| sc68cal | I'd rather see us try and use openstacksdk. If it isn't good, then let's fix instead of papering over | 18:43 |
| sc68cal | part of my reasoning is that i've tried to use the neutronclient as a python programmer and it's terrible | 18:43 |
| sc68cal | I want to see this fixed so I can stop telling people to use ruby and the fog library\ | 18:43 |
| beagles | sc68cal, there is a mismatch between nova and neutron that isn't resolved by changing SDKs | 18:43 |
| beagles | or APIs as it were | 18:43 |
| sc68cal | when you say API mismatch, do you mean between Nova's network API and Neutron's? | 18:44 |
| beagles | sc68cal, in a sense yes.. but it might be fair to say Nova in general and the neutron API. exceptions are an example for instance. Another thing is high level abstractions and neutron implementation details | 18:46 |
| beagles | e.g. give me a list of networks that is available to my tenant | 18:46 |
| *** iamjarvo has joined #openstack-nova | 18:47 | |
| beagles | that is not simply a question of the basic form of how a request is made, but details such as shared vs non-shared, impact of various extensions, etc. | 18:47 |
| sc68cal | and this is mostly because we've gotten stuck on the nova-net vs. neutron question | 18:48 |
| sc68cal | so we have to do a lot on the nova side to smooth out those differences? | 18:48 |
| sc68cal | when replying to requests | 18:48 |
| beagles | I'm not sure about that :) allocate_for_instance would still exist in some form or the other (maybe) | 18:48 |
| beagles | sc68cal, there is a lot of smoothing out :) | 18:50 |
| *** Nikolay_St has quit IRC | 18:50 | |
| *** ijw has joined #openstack-nova | 18:51 | |
| beagles | sc68cal, somebody was recently trying to do a fail early kind of thing with a binding_type=failed response | 18:51 |
| sc68cal | beagles: interesting | 18:52 |
| beagles | sc68cal, it is crazy - port client vs. tenant client, what data is available when, etc. Completely untenable for a nova oriented developer | 18:52 |
| *** annegentle has quit IRC | 18:52 | |
| sc68cal | ah - marun was asking me a similar kind of thing a couple days ago- the scope of the credentials passed into the client. I think we found it was basically getting passed in via the magic "context" arg - so it could be either, depending, lol. | 18:53 |
| dansmith | sc68cal: for what it's worth, I think our trouble with neutronapi has precisely zero to do with what client we use | 18:53 |
| sc68cal | admin vs. tennat | 18:53 |
| *** exploreshaifali has quit IRC | 18:54 | |
| *** ijw_ has quit IRC | 18:54 | |
| sc68cal | dansmith: ah, ok, my first couple reads of beagles spec, I was not quite sure where this refactor was going to happen in the nova codebase | 18:55 |
| beagles | sc68cal, yeah it gets a little interesting. Nova can't do what it needs to all of the time with the tenant's privileges (creating and deleting ports, getting extra port details that are filtered by policies etc) | 18:55 |
| dansmith | sc68cal: neutronapi for us is the thing that bridges nova concepts and neutron concepts.. the api bit is the "internal nova network api" | 18:56 |
| *** redbeard has joined #openstack-nova | 18:56 | |
| *** bkopilov has joined #openstack-nova | 18:56 | |
| *** eharney has quit IRC | 18:56 | |
| dansmith | sc68cal: which started out as "exactly what nova-network does, but calling into neutron" | 18:56 |
| sc68cal | dansmith: ahh I see | 18:56 |
| dansmith | sc68cal: which is dumb because things that were atomic, or a single action in nova-network are not atomic and a billion actions in neutron | 18:56 |
| dansmith | sc68cal: the authors of it seemed to shrug at the differences and do the best to make it seamless | 18:56 |
| dansmith | when what should have happened is refactoring the api so that it made sense for both | 18:57 |
| bauzas | dansmith: sorry, I was afk for dinner and reading stories | 18:58 |
| dansmith | sc68cal: so what we've got now is a thing that we're completely terrified to touch because every time we fix one thing, we break ten more | 18:58 |
| *** matrohon has joined #openstack-nova | 18:58 | |
| bauzas | dansmith: re: the patch, that's just hurting a piece of code that I'm not expert, ie. the migrations | 18:59 |
| bauzas | well, not the migrations, the migrations table rather | 18:59 |
| dansmith | sc68cal: which has been my reasoning for building a replacement alongside what we have and proving that it works, instead of refactoring the current thing | 18:59 |
| dansmith | mriedem: agree with that analysis? | 18:59 |
| bauzas | dansmith: but seeing it on the top level, I think you're right - the migrations table just seems useless for any other thing than a flavor if we persist the spec | 19:00 |
| * mriedem has been sidetracked with other arguing in other channels | 19:00 | |
| mriedem | reading scrollback | 19:00 |
| dansmith | bauzas: it's not useless, it's just a record of the movement, and has nothing to do with resource claims, etc | 19:00 |
| dansmith | bauzas: it used to be that all we could change was from one flavor to the other, so we stuck the from/to flavor ids in that migration object, but things are more complex now | 19:01 |
| bauzas | dansmith: I need to read carefully your last comment because you explain how it's done re: flavors | 19:01 |
| dansmith | yeah | 19:01 |
| bauzas | dansmith: yeah, because of your bp in Kilo right? | 19:01 |
| dansmith | bauzas: not really | 19:01 |
| dansmith | bauzas: this has been in play for a long time now | 19:01 |
| *** jecarey_ has joined #openstack-nova | 19:01 | |
| *** jecarey has quit IRC | 19:01 | |
| bauzas | ok, then I need to take time for giving a view | 19:02 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Remove orphaned tables - iscsi_targets, volumes https://review.openstack.org/166214 | 19:02 |
| bauzas | but as I said, on an upper view, I think the patch is not needed if we persist the spec - just because it has nothing to see with claims IIUC | 19:02 |
| *** ijw has quit IRC | 19:02 | |
| *** iamjarvo has quit IRC | 19:03 | |
| mriedem | dansmith: i agree with your analysis :) | 19:03 |
| mriedem | formally | 19:03 |
| *** bkopilov has quit IRC | 19:03 | |
| *** irenab has joined #openstack-nova | 19:05 | |
| *** vladikr has quit IRC | 19:08 | |
| *** angdraug has joined #openstack-nova | 19:09 | |
| *** irenab has quit IRC | 19:11 | |
| dansmith | alaski: look at line 114 | 19:12 |
| dansmith | alaski: I just didn't realize I had said it elsewhere in the spec, but "data model" is where you made the comment in the first place | 19:12 |
| beagles | mriedem, and yes I will address you comments on the spec :) | 19:12 |
| *** xyang1 has quit IRC | 19:12 | |
| alaski | dansmith: yeah. it's also on line 178 | 19:12 |
| alaski | dansmith: do you get a popup when I comment on your reviews? :) | 19:13 |
| dansmith | alaski: ah right | 19:13 |
| beagles | mriedem, I wonder... can I submit a revision patch to the spec? I don't recall seeing that before | 19:13 |
| mriedem | beagles: people have done that | 19:13 |
| dansmith | alaski: I rooted your system ages ago, so I just get the event stream that way | 19:13 |
| *** iamjarvo has joined #openstack-nova | 19:13 | |
| *** ildikov has joined #openstack-nova | 19:14 | |
| alaski | dansmith: nice, bypass the middleman | 19:14 |
| beagles | mriedem, ok cool. When I saw it get approved and merged it was "Yay" then I saw your comments then it was "s#!t" | 19:14 |
| dansmith | alaski: either way, I can fix those, but since they're either/or I think they're probably okay, and the implementation will go the type way | 19:14 |
| beagles | :) | 19:14 |
| alaski | dansmith: right, that's why I didn't pick on it | 19:14 |
| dansmith | okay cool | 19:14 |
| dansmith | alaski: also, you should close that window in the NW corner of your screen, at least until the end of working hours.. c'mon.. | 19:15 |
| *** bkopilov has joined #openstack-nova | 19:15 | |
| *** Hosam has joined #openstack-nova | 19:16 | |
| alaski | dansmith: what, no... that's just puppies or some other adorable thing | 19:16 |
| *** bkopilov has quit IRC | 19:16 | |
| dansmith | oh sure, _now_ it is | 19:16 |
| *** kaufer1 has joined #openstack-nova | 19:17 | |
| *** marun has quit IRC | 19:17 | |
| *** kaufer has quit IRC | 19:18 | |
| *** VW_ has joined #openstack-nova | 19:18 | |
| *** bkopilov has joined #openstack-nova | 19:20 | |
| *** ZZelle_ has joined #openstack-nova | 19:21 | |
| *** marun has joined #openstack-nova | 19:22 | |
| *** eharney has joined #openstack-nova | 19:25 | |
| *** jlk has joined #openstack-nova | 19:25 | |
| jlk | Hey all, I'm hoping somebody might be able to help me figure out a resize issue, specifically a resize revert when the revert winds up trying to resize /down/ an instance. I'm getting failures there regarding the filesystem that I need some help with. | 19:26 |
| *** thangp_ has quit IRC | 19:26 | |
| *** bkopilov has quit IRC | 19:29 | |
| *** pixelb has quit IRC | 19:30 | |
| *** lpetrut1 has joined #openstack-nova | 19:30 | |
| *** oro has quit IRC | 19:32 | |
| *** xyang1 has joined #openstack-nova | 19:32 | |
| *** heyongli has joined #openstack-nova | 19:32 | |
| *** ajo has joined #openstack-nova | 19:33 | |
| *** kevinbenton has quit IRC | 19:33 | |
| *** ekhugen has quit IRC | 19:33 | |
| *** wolsen has quit IRC | 19:33 | |
| *** wolsen_ has joined #openstack-nova | 19:33 | |
| *** cburgess_ has quit IRC | 19:34 | |
| *** ekhugen has joined #openstack-nova | 19:34 | |
| *** cburgess has joined #openstack-nova | 19:34 | |
| *** flwang1 has quit IRC | 19:35 | |
| *** bkopilov has joined #openstack-nova | 19:35 | |
| *** kevinbenton has joined #openstack-nova | 19:35 | |
| jlk | ooooh! this is a vif plugging timeout on a resize revert. Interesting | 19:37 |
| openstackgerrit | jichenjc proposed openstack/nova: Fix typo for test cases https://review.openstack.org/179211 | 19:40 |
| mriedem | cfriesen: was fixing a bug related to evacuate with neutron and vif plugging timeouts | 19:40 |
| mriedem | evacuate != resize though | 19:40 |
| mriedem | https://review.openstack.org/#/c/169827/ | 19:41 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova-specs: Proposes check destination on migrations https://review.openstack.org/179224 | 19:41 |
| *** bkopilov has quit IRC | 19:41 | |
| *** bkopilov has joined #openstack-nova | 19:42 | |
| *** burt has joined #openstack-nova | 19:42 | |
| *** jecarey_ has quit IRC | 19:43 | |
| *** tonytan4ever has quit IRC | 19:44 | |
| openstackgerrit | Andrew Laski proposed openstack/nova-specs: Persist the RequestSpec object https://review.openstack.org/176078 | 19:44 |
| *** jecarey has joined #openstack-nova | 19:46 | |
| *** jecarey_ has joined #openstack-nova | 19:47 | |
| *** mtanino has quit IRC | 19:47 | |
| cfriesen | jlk: neutron or nova-network? | 19:49 |
| *** bkopilov has quit IRC | 19:49 | |
| jlk | neutron | 19:49 |
| jlk | still debugging, trying to make sure it does a migration to a different host | 19:49 |
| cfriesen | is there a call to self.network_api.setup_instance_network_on_host() to tell neutron that it's reverting back? (Or did we not call it yet on the new host?) | 19:50 |
| cfriesen | mriedem: you don't like the nearly-identical API calls in succession? :) | 19:51 |
| mriedem | cfriesen: love them | 19:51 |
| *** jecarey has quit IRC | 19:51 | |
| *** otter768 has joined #openstack-nova | 19:52 | |
| jlk | cfriesen: I'm trying to track down what happens in a revert. | 19:53 |
| cfriesen | when running "git review" to upload a new version, is there a way to include a comment in that command so that it gets included in as part of the "uploaded patch set X" message? | 19:53 |
| jlk | cfriesen: so a resize is successful, then its reverted, does the revert come back to the original host? or would it go to a third host, or? | 19:53 |
| dansmith | cfriesen: I don't think so | 19:54 |
| dansmith | jlk: original | 19:54 |
| mriedem | the migration record in the db has source and dest | 19:54 |
| *** marun has quit IRC | 19:54 | |
| jlk | and does it wait for the vif unplug event on the migrated to hsot? | 19:54 |
| cfriesen | you guys beat me to it | 19:54 |
| lifeless | cfriesen: no, the patch submission is literally just a git push | 19:54 |
| lifeless | cfriesen: if you can't describe it in the git content, its not included. | 19:55 |
| lifeless | cfriesen: one could imagine teaching git review' | 19:55 |
| dansmith | jlk: if you're getting a timeout, then it's waiting | 19:55 |
| lifeless | cfriesen: to make an API call post-push to add a comment. | 19:55 |
| mriedem | dansmith: jlk: but there could be a bug where it's waiting for an event that's not going to come | 19:55 |
| jlk | dansmith: yeah, it looks like it's timing out waiting for the vif plug, and then it goes to delete, and I see | 19:55 |
| dansmith | mriedem: right | 19:55 |
| jlk | 2015-04-30 19:45:43.397 23513 DEBUG nova.compute.manager [req-83390c2c-17d9-4a82-9a87-84999507ade2 None] [instance: 223a73b6-1434-4db1-a9fe-57c2f589d8d6] Events pending at deletion: network-vif-plugged-9062798b-cbbd-45c5-bc5a-0ed0334a9871 _delete_instance | 19:55 |
| *** obedmr has quit IRC | 19:55 | |
| *** Hosam has quit IRC | 19:56 | |
| *** marun has joined #openstack-nova | 19:56 | |
| cfriesen | mriedem: in https://review.openstack.org/#/c/169827/7 it wasn't getting the event because neutron thought it was on the "wrong" host | 19:56 |
| dansmith | jlk: what code are you running? | 19:56 |
| *** Mike_D_laptop has joined #openstack-nova | 19:56 | |
| jlk | juno | 19:56 |
| *** otter768 has quit IRC | 19:56 | |
| dansmith | jlk: there have been fixes to this stuff since then I think | 19:56 |
| jlk | 2015-04-30 19:45:40.681 23513 WARNING nova.virt.libvirt.driver [req-1d78fb5c-bea3-4237-b95f-f043edba0c86 None] Timeout waiting for vif plugging callback for instance 223a73b6-1434-4db1-a9fe-57c2f589d8d6 | 19:57 |
| jlk | dansmith: sent to stable/juno or...? | 19:57 |
| dansmith | jlk: I thought there was one specifically around resizes, but I'm not sure.. should have been backported, but... | 19:57 |
| dansmith | jlk: the "when neutron sends events" thing is complicated enough that I can't keep it in my head | 19:57 |
| jlk | I would hope so, because resize is required for DefCore juno :/ | 19:57 |
| dansmith | jlk: so I can confirm that it's waiting because of the timeout, but whether it should or not requires a neutron person I think | 19:58 |
| jlk | alright | 19:58 |
| dansmith | since neutron needs to do things before networking will work on the other host, | 19:58 |
| dansmith | I think we *have* to wait for an event too avoid the race we used to have | 19:58 |
| dansmith | so, I tend to think that nova is doing the right thing here, but.. | 19:59 |
| dansmith | jlk: can you look at the API to see if the event is coming in but being sent to the wrong/old host? | 19:59 |
| mriedem | dansmith: were you thinking of the stuff added in this series? https://review.openstack.org/#/q/status:merged+project:openstack/nova+branch:master+topic:bug/1323658,n,z | 19:59 |
| dansmith | jlk: also, the old compute should log the event if it goes there instead of to the new one | 20:00 |
| jlk | dansmith: I could perhaps, theres a lot of calls, so somewhat confusing. | 20:00 |
| jlk | the old compute, where the instnace was first resized to? | 20:00 |
| *** isd has joined #openstack-nova | 20:00 | |
| mriedem | resized from i think | 20:00 |
| jlk | well | 20:00 |
| jlk | instance is booted on A. Instance is resized to B. Resize to B is reverted, which sends instance back to A | 20:01 |
| mriedem | heh, well https://bugs.launchpad.net/nova/+bug/1323658 | 20:01 |
| openstack | Launchpad bug 1323658 in OpenStack Compute (nova) "Nova resize/restart results in guest ending up in inconsistent state with Neutron" [Medium,Confirmed] | 20:01 |
| mriedem | so i added some cleanup/debugging around that in kilo | 20:01 |
| dansmith | jlk: is it failing at the new host or when it tries to get reverted? | 20:01 |
| mriedem | but never actually squashed a bug | 20:01 |
| dansmith | mriedem: yeah, I recall those patches, but I'm not sure I even understand the comment at this point | 20:01 |
| jlk | dansmith: fails when the revert attempt happens | 20:02 |
| jlk | the original resize appears successful | 20:02 |
| dansmith | jlk: but fails at the original host | 20:02 |
| jlk | yes | 20:02 |
| mriedem | jlk: this happens always, or intermittent? | 20:03 |
| dansmith | jlk: so looking at the other host to see if the event shows up there would be good | 20:03 |
| dansmith | jlk: like if we've not yet set instance.host = $original | 20:03 |
| jlk | seems to happen every single time I run this test | 20:03 |
| mriedem | hmm, well in juno some neutron resize scenario tests were skipped for being too fail happy https://review.openstack.org/#/c/125150/ | 20:04 |
| mriedem | we unskipped those in kilo https://review.openstack.org/#/c/161768/ | 20:04 |
| *** baoli has joined #openstack-nova | 20:04 | |
| dansmith | mriedem: your comment can't be related since that code is skipping the wait for the event anyway | 20:04 |
| mriedem | so something is better, we just don't know what or where - could be changes in neutron | 20:04 |
| *** obedmr has joined #openstack-nova | 20:04 | |
| mriedem | dansmith: yeah, there was a series of patches there | 20:05 |
| jlk | definitely seeing Instance running successfully on the resize TO host | 20:05 |
| mriedem | while going down that rabbit hole | 20:05 |
| dansmith | jlk: if you can confirm that the event is not showing up at the other host, then we'll know if it's neutron not sending it or nova losing it | 20:05 |
| dansmith | mriedem: yeah | 20:05 |
| mriedem | fwiw bug 1323658 is still a thing | 20:05 |
| openstack | bug 1323658 in OpenStack Compute (nova) "Nova resize/restart results in guest ending up in inconsistent state with Neutron" [Medium,Confirmed] https://launchpad.net/bugs/1323658 | 20:05 |
| mriedem | 64 failures in 10 days | 20:05 |
| *** baoli has quit IRC | 20:07 | |
| jlk | hrm that definitely looks related | 20:07 |
| *** baoli has joined #openstack-nova | 20:07 | |
| cfriesen | jlk: just as a datapoint...it looks like we don't unplug the old vifs until the confirm-resize operation | 20:08 |
| dansmith | cfriesen: orly | 20:08 |
| cfriesen | so if we revert, presumably it should end up using the previously-existing vifs | 20:08 |
| jlk | hrm, wouldnt that result in the same address plugged on multiple hosts? | 20:08 |
| dansmith | cfriesen: I wonder if that means we should not be expecting the event back | 20:08 |
| dansmith | jlk: no I don't think so | 20:09 |
| dansmith | jlk: because the mac is owned inside the VM | 20:09 |
| dansmith | jlk: so unless you don't stop it on the source, nothing else will reply arp for that address | 20:09 |
| jlk | okay | 20:09 |
| *** doude has quit IRC | 20:10 | |
| cfriesen | dansmith: compute.manager.Manager._confirm_resize() calls self.network_api.setup_networks_on_host and then self.driver.confirm_migration(), which ends up calling unplug_vifs() (at least for libvirt) | 20:10 |
| dansmith | it's hard to believe that resize with neutron could have been completely broken since juno and we're just now hearing about it | 20:10 |
| *** pixelb has joined #openstack-nova | 20:10 | |
| dansmith | although I think a lot of people do disable the events because $silly reasons | 20:10 |
| cfriesen | resize works, just revert doesn't | 20:10 |
| jlk | dansmith: resize seems fine | 20:10 |
| jlk | revert not as much | 20:10 |
| dansmith | I meant resize revert | 20:10 |
| mriedem | hmmm | 20:11 |
| jlk | so what signatures can I search for in the two logs to figure out if / where events went? | 20:11 |
| mriedem | wonder if we have a resize revert test in tempest | 20:11 |
| jlk | mriedem: you do | 20:11 |
| mriedem | jlk: but with neutron? | 20:11 |
| jlk | because that's what is preventing me from passing defcore :) | 20:11 |
| dansmith | jlk: when you get the event timeout, do you get a trace? | 20:11 |
| jlk | dansmith: yes | 20:11 |
| dansmith | jlk: can haz pastebin? | 20:11 |
| dansmith | you know, after you finish your twitter hatemail :) | 20:12 |
| jlk | lol | 20:13 |
| jlk | gotta dig it out here | 20:13 |
| jlk | https://gist.github.com/j2sol/2c6432973baf1038e187 | 20:14 |
| jlk | do you need more around that? | 20:14 |
| dansmith | nope | 20:14 |
| jlk | updated to ahve the req at the top https://gist.github.com/j2sol/2c6432973baf1038e187 | 20:14 |
| openstackgerrit | Dan Smith proposed openstack/nova: WIP: Don't wait for an event on a resize-revert https://review.openstack.org/179228 | 20:16 |
| dansmith | jlk: ^ | 20:16 |
| jlk | oooh | 20:17 |
| dansmith | feel free to tweet *nice* things about the excellent service you received here today :) | 20:17 |
| jlk | let me hack it in | 20:17 |
| jlk | geez, have I been that bad on twitter? :/ | 20:17 |
| dansmith | yes | 20:17 |
| dansmith | sometimes at night, I can still hear the screaming | 20:17 |
| openstackgerrit | Andrew Laski proposed openstack/nova-specs: Fix pep8 violations in tests https://review.openstack.org/179230 | 20:18 |
| *** erlon is now known as erlon_awaY | 20:19 | |
| *** oro has joined #openstack-nova | 20:19 | |
| *** Sukhdev has quit IRC | 20:20 | |
| jlk | what the heck. I seem to already have that change here? | 20:20 |
| mriedem | fail defcore | 20:20 |
| *** harlowja has joined #openstack-nova | 20:20 | |
| jlk | nope, wrong area | 20:20 |
| dansmith | yeah, there are other places | 20:20 |
| dansmith | mriedem: sadly, that change doesn't fail any unit tests :( | 20:21 |
| mriedem | i'm checking the tempest test we have | 20:22 |
| jlk | dansmith: that was totally it! | 20:22 |
| jlk | tempest passes now | 20:22 |
| dansmith | jlk: okay, I'll backport to kilo and juno | 20:23 |
| jlk | rock1 | 20:23 |
| mriedem | jlk: you're testing trunk tempest right? | 20:23 |
| jlk | ! | 20:23 |
| dansmith | jlk: after this call you will have an opportunity to rate my service on twitter from 1-10 | 20:23 |
| jlk | wow, so much burn | 20:23 |
| dims_ | melwitt: early congrats! | 20:24 |
| *** Sukhdev has joined #openstack-nova | 20:24 | |
| mriedem | dansmith: jlk: note that the resize revert test passes in the neutron job http://logs.openstack.org/50/177450/1/check/check-tempest-dsvm-neutron-full/fca0bfd/console.html.gz#_2015-04-24_21_20_54_839 | 20:25 |
| mriedem | if this were 100% fail that would be busted | 20:25 |
| mriedem | even on juno http://logs.openstack.org/89/177889/1/check/check-tempest-dsvm-neutron-full/41adbee/console.html#_2015-04-27_18_14_05_401 | 20:26 |
| dansmith | mriedem: same host | 20:26 |
| dansmith | mriedem: I think that's the deal | 20:26 |
| mriedem | oh yes | 20:26 |
| mriedem | let me check the multinode job | 20:26 |
| mriedem | sweet sweet multinode job | 20:26 |
| dansmith | jlk: karma reset to zero :) | 20:26 |
| openstackgerrit | Dan Smith proposed openstack/nova: Don't wait for an event on a resize-revert https://review.openstack.org/179228 | 20:27 |
| *** READ10 has quit IRC | 20:27 | |
| jlk | sounds like y'all were planning a blanket party for me at summit... | 20:27 |
| *** ajo has quit IRC | 20:28 | |
| dansmith | jlk: I do have to make a few calls to cancel things now | 20:28 |
| mriedem | jlk: can you get a bug open so we can track against ^ | 20:30 |
| mriedem | need that for backports | 20:30 |
| jlk | sure | 20:30 |
| dansmith | alkready dont | 20:30 |
| dansmith | er, already done :) | 20:30 |
| dansmith | https://bugs.launchpad.net/nova/+bug/1450624 | 20:30 |
| openstack | Launchpad bug 1450624 in OpenStack Compute (nova) "Nova waits for events from neutron on resize-revert that aren't coming" [High,Confirmed] - Assigned to Dan Smith (danms) | 20:30 |
| *** emagana has quit IRC | 20:30 | |
| jlk | yeah launchpad pointed me to that | 20:31 |
| openstackgerrit | Dan Smith proposed openstack/nova: Don't wait for an event on a resize-revert https://review.openstack.org/179228 | 20:31 |
| dansmith | mriedem: should be good now I think ^ | 20:32 |
| jlk | Related to https://bugs.launchpad.net/nova/+bug/1380934 ? | 20:32 |
| openstack | Launchpad bug 1380934 in OpenStack Compute (nova) "failure rebuilding instance after resize-revert" [Low,Fix released] - Assigned to Yasuaki Nagata (yasuaki-nagata) | 20:32 |
| *** marun has quit IRC | 20:32 | |
| mriedem | jlk: certianly looks related | 20:33 |
| mriedem | https://review.openstack.org/#/c/135819/ also needs to be cherry picked | 20:34 |
| mriedem | to stable/juno | 20:34 |
| *** ajo has joined #openstack-nova | 20:34 | |
| dansmith | https://review.openstack.org/#/q/Id515137747a4b76e9b7057c95f80c8ae74017519,n,z | 20:36 |
| dansmith | backports ^ | 20:36 |
| *** Mike_D_laptop has quit IRC | 20:38 | |
| artom | That's it, I want a Canadian/English version of OpenStack that spells flavour correctly. | 20:39 |
| mriedem | it's flavor | 20:39 |
| artom | This must be the 12th time I make this "mistake" ;) | 20:39 |
| beagles | artom: are you situated in Canada? | 20:41 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Cells: Call compute api methods with instance objects https://review.openstack.org/79741 | 20:42 |
| openstackgerrit | Adelina Tuvenie proposed openstack/nova: Adds Hyper-V vNic hot-plug implementation https://review.openstack.org/138255 | 20:42 |
| openstackgerrit | Adelina Tuvenie proposed openstack/nova: Adds Hyper-V OVS ViF driver https://review.openstack.org/140045 | 20:42 |
| artom | beagles, yes. | 20:42 |
| beagles | referring to it as "Canadian" spelling made me wonder | 20:43 |
| *** achanda has quit IRC | 20:43 | |
| *** harlowja has quit IRC | 20:45 | |
| *** harlowja has joined #openstack-nova | 20:45 | |
| artom | beagles, it certainly isn't Quebecois :p | 20:46 |
| beagles | artom, non | 20:46 |
| *** htruta has quit IRC | 20:46 | |
| *** raildo has quit IRC | 20:46 | |
| *** dprince has quit IRC | 20:49 | |
| *** keekz_ is now known as keekz | 20:50 | |
| mriedem | dansmith: so the neutron multinode job must be running the resize test as single node | 20:50 |
| mriedem | http://logs.openstack.org/54/157054/11/experimental/check-tempest-dsvm-neutron-multinode-full/2521846/console.html#_2015-04-30_14_41_06_700 | 20:50 |
| mriedem | hrm | 20:50 |
| dansmith | well, | 20:51 |
| mriedem | resize_to_same_host probably takes precedence | 20:51 |
| dansmith | it might be happening to land there | 20:51 |
| dansmith | yheah | 20:51 |
| mriedem | we should change that to not resize to same host | 20:51 |
| mriedem | s/we/I/ | 20:51 |
| dansmith | certainly on the multinode one | 20:51 |
| mriedem | yeah | 20:51 |
| dansmith | I also wonder if we'll start to see that job's reliability go *down* if we swtich that | 20:51 |
| dansmith | all kinds of things we can hit with that which weren't previously in play | 20:52 |
| dansmith | but good for sure | 20:52 |
| mriedem | right, kind of the point | 20:52 |
| dansmith | yeah | 20:52 |
| mriedem | so much for that idea | 20:53 |
| mriedem | http://logs.openstack.org/54/157054/11/experimental/check-tempest-dsvm-neutron-multinode-full/2521846/logs/etc/nova/nova.conf.txt.gz | 20:53 |
| mriedem | we don't set allow_resize_to_same_host | 20:53 |
| mriedem | that gd jogo is always one step ahead | 20:53 |
| bauzas | dansmith: so I code reviewed all the resources related to the resize, and I'm +1 wiht you | 20:53 |
| dansmith | bauzas: cool | 20:54 |
| bauzas | dansmith: I'll provide a new comment on it | 20:54 |
| dansmith | mriedem: well, I don't know how to explain it otherwise | 20:54 |
| dansmith | mriedem: he clearly hit the issue and this fixed it | 20:54 |
| dansmith | mriedem: maybe something else is going on in that test that we're not seeing | 20:54 |
| *** alanf-mc has joined #openstack-nova | 20:55 | |
| mriedem | vif_plugging_is_fatal = True | 20:55 |
| mriedem | yeah, not sure | 20:55 |
| *** sdake_ has joined #openstack-nova | 20:55 | |
| cfriesen | artom: we should honour the correct spelling of flavour? | 20:56 |
| sdague | mriedem: did you ever get around to circling on - https://review.openstack.org/#/c/176844/9 ? | 20:56 |
| mriedem | sdague: not really | 20:56 |
| sdague | and the fact that the mock call count on safe_encode is apparently different now? | 20:56 |
| artom | cfriesen, totally. | 20:56 |
| mriedem | sdague: will check quick | 20:56 |
| *** isd has quit IRC | 20:56 | |
| sdague | maybe dims_ has an idea | 20:56 |
| dims_ | looking | 20:57 |
| sdague | because atm, that's blocking us updating any requirements, which is going to be a problem | 20:57 |
| dansmith | mriedem: unless the test isn't noticing the failure | 20:57 |
| artom | cfriesen, like, submit all the patches right now. I will personally +3 all of them. | 20:57 |
| sdague | dims_: my instinct is oslo.i18n global state | 20:57 |
| dansmith | jlk: when it fails, you get the instance in an error state, right? | 20:57 |
| mriedem | dansmith: the test completes in like 21 seconds | 20:58 |
| mriedem | dansmith: with a timeout of 300 | 20:58 |
| dansmith | mriedem: seems kinda quick | 20:58 |
| dansmith | mriedem: yeah, and it should be waiting for the event for a long time | 20:58 |
| dansmith | mriedem: but if it doesn't check the state of the instance after the revert or something and then just deletes it... | 20:58 |
| sdague | so... be cautious of test times, some of these tests build computes in setup | 20:59 |
| *** sdake has quit IRC | 20:59 | |
| sdague | which means it's not accounted for correctly | 20:59 |
| *** marun has joined #openstack-nova | 20:59 | |
| *** sdake has joined #openstack-nova | 20:59 | |
| mriedem | http://git.openstack.org/cgit/openstack/tempest/tree/tempest/api/compute/servers/test_server_actions.py#n233 | 20:59 |
| mtreinish | sdague: that's something I still want to look into fixing | 20:59 |
| mriedem | it does the revert and waits for active | 20:59 |
| dansmith | yeah | 20:59 |
| mtreinish | it shouldn't be too hard to add events to the subunit stream | 21:00 |
| *** annegentle has joined #openstack-nova | 21:00 | |
| mtreinish | from setupclass | 21:00 |
| *** vilobhmm1 has quit IRC | 21:00 | |
| dansmith | so | 21:00 |
| sdague | mtreinish: yeh, though server_actions also does the rebuild optimization so it's doing it on teardown or something | 21:01 |
| dansmith | IIRC, neutron doesn't do this for all plugins | 21:01 |
| sdague | jsut don't trust the timings in that test :) | 21:01 |
| dansmith | are we running on ovs in this test? | 21:01 |
| *** vilobhmm1 has joined #openstack-nova | 21:01 | |
| *** unicell has quit IRC | 21:01 | |
| sdague | I thought neutron multinode was ovs | 21:01 |
| *** angdraug has quit IRC | 21:01 | |
| *** sdake has quit IRC | 21:01 | |
| mriedem | yes | 21:01 |
| *** unicell has joined #openstack-nova | 21:01 | |
| dansmith | okay | 21:01 |
| mriedem | http://logs.openstack.org/54/157054/11/experimental/check-tempest-dsvm-neutron-multinode-full/2521846/logs/etc/neutron/neutron.conf.txt.gz | 21:01 |
| *** sdake has joined #openstack-nova | 21:01 | |
| mriedem | well http://logs.openstack.org/54/157054/11/experimental/check-tempest-dsvm-neutron-multinode-full/2521846/logs/etc/neutron/plugins/ml2/ml2_conf.ini.txt.gz | 21:02 |
| mriedem | mechanism_drivers = openvswitch,linuxbridge,l2population | 21:02 |
| mriedem | whatever that means | 21:02 |
| dansmith | mriedem: no nova admin creds there | 21:02 |
| dansmith | mriedem: so it can't be sending events, AFAIK | 21:02 |
| *** vilobhmm1 has quit IRC | 21:02 | |
| *** vilobhmm1 has joined #openstack-nova | 21:03 | |
| mriedem | dansmith: ah yup | 21:03 |
| mriedem | # nova_admin_tenant_id = | 21:03 |
| *** sdake_ has quit IRC | 21:03 | |
| dansmith | _rightg | 21:03 |
| *** patrickeast has quit IRC | 21:03 | |
| dansmith | so something is up there I think | 21:03 |
| *** vilobhmm1 has quit IRC | 21:03 | |
| *** patrickeast has joined #openstack-nova | 21:03 | |
| mriedem | wouldn't nova be waiting then for something that doesn't happen? | 21:04 |
| dansmith | that's why I was wondering about the vif driver | 21:04 |
| *** vilobhmm1 has joined #openstack-nova | 21:04 | |
| bauzas | alaski: around for https://review.openstack.org/#/c/171951/1/nova/scheduler/chance.py,cm ? | 21:04 |
| *** josecastroleon has joined #openstack-nova | 21:04 | |
| *** melwitt has quit IRC | 21:04 | |
| *** melwitt has joined #openstack-nova | 21:04 | |
| *** angdraug has joined #openstack-nova | 21:05 | |
| dims_ | sdague: possibly the oslo.log bump - https://review.openstack.org/#/c/163027/ | 21:05 |
| *** rfolco has quit IRC | 21:05 | |
| *** melwitt has quit IRC | 21:05 | |
| *** melwitt has joined #openstack-nova | 21:05 | |
| *** anish_ is now known as anish | 21:06 | |
| *** anish has quit IRC | 21:06 | |
| *** anish has joined #openstack-nova | 21:06 | |
| mriedem | dims_: sdague: checking nova tests, the only place that mocks out safe_decode is the one test that's failing | 21:06 |
| melwitt | dims_: thanks :) (late reply) | 21:06 |
| mriedem | so yeah has to be something global in oslo | 21:06 |
| *** harlowja has quit IRC | 21:07 | |
| bauzas | melwitt: kudos :) | 21:07 |
| *** josecastroleon has quit IRC | 21:07 | |
| melwitt | bauzas: thanks :) | 21:08 |
| dansmith | mriedem: and yet, events are coming in: http://logs.openstack.org/54/157054/11/experimental/check-tempest-dsvm-neutron-multinode-full/2521846/logs/screen-n-cpu.txt.gz#_2015-04-30_14_37_03_514 | 21:08 |
| bauzas | melwitt: so prepare yourself to get a huge list of changes I want you to +2 | 21:08 |
| dims_ | mriedem: that's the one spot i recall being touched recently (review 163027), should be easy enough to check, by just using the older version min/max for oslo.log, want me to try that? | 21:08 |
| bauzas | melwitt: do you know how I can bribe you ? beers at Vancouver ? | 21:08 |
| melwitt | bauzas: lol ;) | 21:08 |
| dims_ | melwitt: yes, kudos again :) | 21:09 |
| jlk | dansmith: yes, I believe so, but tempest deletes it right away. I see it getting set to ERROR state though. | 21:09 |
| mriedem | dims_: sure, that must be it though | 21:09 |
| dims_ | mriedem: i am going to use the same review if there are no objectsions | 21:09 |
| mriedem | use what for what now? | 21:09 |
| dansmith | jlk: yeah, I forgot you were using tempest | 21:09 |
| jlk | just because I haven't bothered to try it manually. I can if that'll help anything | 21:10 |
| *** VW_ has quit IRC | 21:10 | |
| dansmith | np | 21:10 |
| dansmith | jlk: we're trying to figure out why our tests aren't hitting this, because it seems like they should | 21:10 |
| jlk | okay | 21:11 |
| mriedem | dims_: https://review.openstack.org/#/c/163027/1 is in 1.1.0, which nova was already using | 21:11 |
| mriedem | https://review.openstack.org/#/c/176844/9/requirements.txt | 21:11 |
| mriedem | oh nvm | 21:11 |
| mriedem | oslo.log>=1.0.0,<1.1.0 | 21:11 |
| mriedem | so yeah, that has to be the issue | 21:11 |
| mriedem | so maybe i just change the nova test to check call_count >= 2 | 21:12 |
| dims_ | mriedem: let me test the theory and then we can change the call_count | 21:12 |
| jlk | oh crap. I'm getting failures again. | 21:12 |
| *** ildikov has quit IRC | 21:12 | |
| jlk | oh it's different | 21:14 |
| jlk | 2015-04-30 21:11:19.509 28855 ERROR oslo.messaging.rpc.dispatcher [req-b4391ef1-ccf3-4106-bb3e-09eefb57ebc5 ] Exception during message handling: Resize error: Unable to resize disk down. | 21:14 |
| *** kaufer has joined #openstack-nova | 21:14 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Updated from global requirements https://review.openstack.org/176844 | 21:15 |
| dansmith | jlk: which hypervisor? | 21:15 |
| jlk | libvirt | 21:15 |
| jlk | kvm | 21:15 |
| dansmith | most don't support downsizing things | 21:15 |
| jlk | I was running into this earlier, but it seemed to have gone away in favor of the network issue | 21:15 |
| dansmith | disks, that is | 21:15 |
| jlk | wonder if it's because the same instance was used for a number of other tests and that's making things different | 21:16 |
| *** kaufer1 has quit IRC | 21:16 | |
| alaski | bauzas: in the middle of something atm, can look in a bit | 21:17 |
| bauzas | alaski: nah, just wanted to discuss with you about force_hosts and if we should just kill that because I hate it :) | 21:18 |
| *** vilobhmm1 has quit IRC | 21:18 | |
| *** matrohon has quit IRC | 21:18 | |
| alaski | bauzas: yeah, that would be lovely | 21:18 |
| *** vilobhmm1 has joined #openstack-nova | 21:19 | |
| bauzas | alaski: because calling a scheduler for telling it not to schedule is just nonsense to me | 21:19 |
| bauzas | like "eh I love you but I don't trust you" | 21:19 |
| *** vilobhmm1 has quit IRC | 21:19 | |
| jlk | oh this is fantastic | 21:19 |
| dansmith | mriedem: well, I don't get it, but sending events when no creds are configured seems really strange to me | 21:19 |
| jlk | if I run the resize revert test in isloation, totes works | 21:19 |
| mriedem | dansmith: yeah that seems busted | 21:19 |
| dansmith | jlk: sounds like maybe a tempest race | 21:20 |
| dansmith | jlk: hassle mtreinish | 21:20 |
| jlk | not sure why that results in resize down errors but perhaps | 21:20 |
| alaski | bauzas: if the scheduler is going to do claims then it has to go through there. but a user specifying a host doesn't have a clear use case IMO | 21:20 |
| bauzas | alaski: okay, I have very small bandwidth for that, but at Ieast I should fire something in ML | 21:20 |
| dansmith | jlk: it makes sense | 21:20 |
| dansmith | jlk: tempest tries to re-use instances to avoid taking eleventy billion minutes to run | 21:20 |
| bauzas | alaski: well, it's just all about forcing a destination | 21:20 |
| dansmith | jlk: so if something is in a bad ordering, it would cause that | 21:21 |
| bauzas | alaski: like we do for live migrate or evacuate | 21:21 |
| bauzas | alaski: that doesn't require a scheduler call IMHO | 21:21 |
| alaski | bauzas: sure, but why specify a destination is my question | 21:21 |
| bauzas | alaski: because operators want to hack that | 21:21 |
| *** redbeard has quit IRC | 21:21 | |
| bauzas | alaski: what's just insane is that it's something end-user possibly | 21:22 |
| bauzas | alaski: because it's an AZ attribute | 21:22 |
| bauzas | I mean on the API side | 21:22 |
| *** mwagner_lap has quit IRC | 21:22 | |
| alaski | I understand people want the capability, I would just like a more clear use case than "I don't trust the scheduler" | 21:22 |
| dansmith | the scheduler stole a candy bar from me earlier in the week, so I'm pretty untrusting of it | 21:23 |
| *** vilobhmm1 has joined #openstack-nova | 21:23 | |
| jlk | argh, yeah. | 21:24 |
| bauzas | dansmith: we know that the scheduler is a bullier right | 21:24 |
| jlk | mtreinish: hi! race condition alert :) | 21:24 |
| bauzas | (is that the right word ?) | 21:24 |
| *** Longgeek has quit IRC | 21:24 | |
| dansmith | jlk: does it matter for your purposes? | 21:24 |
| dansmith | jlk: like, do you have to pass it all in one go? | 21:24 |
| jroll | alaski: testing a particular hypervisor seems like a use case | 21:24 |
| *** vilobhmm1 has quit IRC | 21:24 | |
| jroll | alaski: I *think* our ops team does that sometimes | 21:25 |
| bauzas | alaski: I actually just fired up a spec related to that and how forcing hosts is risky | 21:25 |
| bauzas | alaski: just because scheduler filters Do Something | 21:25 |
| mtreinish | jlk: heh, uh oh | 21:25 |
| *** mragupat has joined #openstack-nova | 21:25 | |
| mriedem | dims_: sdague: testing a patch now | 21:26 |
| *** ildikov has joined #openstack-nova | 21:26 | |
| dims_ | mriedem: i am waiting on https://jenkins07.openstack.org/job/gate-nova-python27/11170/console | 21:28 |
| bauzas | alaski: for example of something beasty https://bugs.launchpad.net/nova/+bug/1427772 | 21:28 |
| openstack | Launchpad bug 1427772 in OpenStack Compute (nova) "Instance that uses force-host still needs to run some filters" [Low,Confirmed] - Assigned to Sylvain Bauza (sylvain-bauza) | 21:28 |
| *** Longgeek has joined #openstack-nova | 21:28 | |
| bauzas | alaski: we have also https://bugs.launchpad.net/nova/+bug/1214943 as an exhibit on how all of this is bad https://bugs.launchpad.net/nova/+bug/1214943 | 21:29 |
| openstack | Launchpad bug 1214943 in OpenStack Compute (nova) "Live migration should use the same memory over subscription logic as instance boot" [High,Confirmed] - Assigned to Sylvain Bauza (sylvain-bauza) | 21:29 |
| *** nelsnelson has quit IRC | 21:29 | |
| *** angdraug has quit IRC | 21:29 | |
| *** VW_ has joined #openstack-nova | 21:30 | |
| *** unicell has quit IRC | 21:30 | |
| bauzas | alaski: so, re: all of that, I think we need some way to query the scheduler to know if what we do is valid, hence https://review.openstack.org/#/c/179224/ | 21:30 |
| bauzas | alaski: and we should just move the force_hosts logic down to the conductor and call that ^ (spec) | 21:30 |
| *** sdake_ has joined #openstack-nova | 21:31 | |
| *** iamjarvo has quit IRC | 21:32 | |
| *** ajo has quit IRC | 21:32 | |
| *** dboik has joined #openstack-nova | 21:33 | |
| dims_ | mriedem: indeed it was the oslo.log bump - gate-nova-python27 finished fine with older rev | 21:33 |
| mriedem | dims_: i have the fix | 21:33 |
| mriedem | will push that up and rebase the sync on top of it | 21:33 |
| *** unicell has joined #openstack-nova | 21:33 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Updated from global requirements https://review.openstack.org/176844 | 21:34 |
| *** sdake has quit IRC | 21:34 | |
| dims_ | y, i switched it back | 21:34 |
| dims_ | +1 mriedem | 21:34 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Fix assert on call count for encodeutils.safe_decode mock https://review.openstack.org/179254 | 21:34 |
| mtreinish | jlk: reading the back scroll I think there might have been a tempest bug filed about the resize tests and resize down at one point | 21:35 |
| mtreinish | but I don't really remember | 21:35 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Updated from global requirements https://review.openstack.org/176844 | 21:35 |
| mtreinish | but yeah that test probably doesn't have proper protections in place to block a resized down | 21:35 |
| mriedem | dims_: sdague: ^ rebased | 21:35 |
| dims_ | +1'ed both :) | 21:36 |
| *** signed8b_ has quit IRC | 21:36 | |
| mriedem | dansmith: proud of my use of mock.call_args_list here https://review.openstack.org/#/c/179254/1/nova/tests/unit/virt/libvirt/test_driver.py | 21:36 |
| mriedem | baoli: can someone check on the intel pci CI job? | 21:37 |
| mriedem | it's failing on everything | 21:37 |
| *** dboik_ has quit IRC | 21:37 | |
| * dansmith hands mriedem a cookie | 21:37 | |
| *** mmedvede has quit IRC | 21:37 | |
| dansmith | mriedem: it seems to be flaky actually, I saw it pass on something earlier | 21:37 |
| *** dboik has quit IRC | 21:37 | |
| dansmith | but agreed, it's less useful than noise at the moment | 21:38 |
| mriedem | if i can come away from this day with doing at least one useful thing, it might still be a win | 21:38 |
| dims_ | dansmith: cookies are dangerous! https://www.youtube.com/watch?v=Gyk55GYnGl0 | 21:39 |
| *** emagana has joined #openstack-nova | 21:39 | |
| mriedem | dims_: isn't that kind of the whole republican party's base theory? | 21:40 |
| *** angdraug has joined #openstack-nova | 21:40 | |
| mriedem | give a mouse a cookie and he'll never learn to make cookies for himself | 21:40 |
| dims_ | haha | 21:40 |
| mriedem | then eventually the mice will riot in baltimore | 21:40 |
| *** sdake_ has quit IRC | 21:40 | |
| mriedem | if only i'd listened to ted cruz | 21:41 |
| dims_ | :) | 21:41 |
| *** sdake has joined #openstack-nova | 21:42 | |
| *** annegentle has quit IRC | 21:42 | |
| *** annegentle has joined #openstack-nova | 21:42 | |
| *** jichen has quit IRC | 21:44 | |
| *** mriedem is now known as mriedem_away | 21:47 | |
| *** bkopilov has joined #openstack-nova | 21:48 | |
| *** annegentle has quit IRC | 21:48 | |
| *** lpetrut1 has quit IRC | 21:48 | |
| *** kaufer has quit IRC | 21:51 | |
| *** otter768 has joined #openstack-nova | 21:52 | |
| *** achanda has joined #openstack-nova | 21:52 | |
| jlk | mtreinish: I couldn't find anything in existing reports, so I made https://bugs.launchpad.net/tempest/+bug/1450648 | 21:53 |
| openstack | Launchpad bug 1450648 in tempest "No protection against resize down in resize testing" [Undecided,New] | 21:53 |
| *** igordcard_ has joined #openstack-nova | 21:53 | |
| *** joefides has quit IRC | 21:53 | |
| *** jgrimm is now known as zz_jgrimm | 21:53 | |
| mtreinish | jlk: ok thanks, I'll take a look | 21:54 |
| *** flwang1 has joined #openstack-nova | 21:54 | |
| mtreinish | yeah the server reuse there has always been kinda funky. It might just be easiest to move that test into a different class without server reuse | 21:55 |
| *** igordcard has quit IRC | 21:55 | |
| *** dboik has joined #openstack-nova | 21:56 | |
| *** irenab has joined #openstack-nova | 21:56 | |
| jlk | is that something I could try real quick? | 21:57 |
| *** otter768 has quit IRC | 21:57 | |
| *** dboik_ has joined #openstack-nova | 21:57 | |
| dansmith | jlk: certainly if there's an open bug for this you can "pass the test" by just showing that you work, except you have to run this one in isolation because $bug, right? | 21:57 |
| dansmith | jlk: I mean, all for helping to get it fixed, but I sensed your urgency before... | 21:58 |
| jlk | dansmith: yeah, that's right. It passes in isolation. I was hoping to do a local mod to tempest to let it run the whole gamut so I can submit to the defcore folks. | 21:58 |
| jlk | the urgency is just that once I get this done I can move on to something else fun :) | 21:59 |
| *** emagana has quit IRC | 21:59 | |
| *** oomichi has joined #openstack-nova | 22:00 | |
| *** jecarey_ has quit IRC | 22:00 | |
| dansmith | jlk: well, I dunno which is worse, hacking the verification test, or failing it but showing you pass without the bug :) | 22:00 |
| dansmith | but anyway | 22:00 |
| *** dboik has quit IRC | 22:00 | |
| jlk | haha, that's a fair point | 22:01 |
| jlk | I'm already hacking nova by putting your change in before it's merged, but I think at this point this system is something of an honor system | 22:01 |
| *** irenab has quit IRC | 22:01 | |
| alaski | jroll: yeah, testing a compute is a good use case. though force_hosts may not be the solution | 22:01 |
| dansmith | yeah, well, that's why I got the backport up for you | 22:01 |
| dansmith | so it's at least something proposed for stable/juno :) | 22:01 |
| jroll | alaski: right :) | 22:02 |
| jlk | dansmith: yeah, I do appreciate that | 22:02 |
| dansmith | jlk: so I think you're legit with your nova change, I mean | 22:02 |
| alaski | bauzas: I agree it's bad. I'm not sure we can bypass the scheduler if it's handling claims though, I thought everything would need to nominally go through the scheduler | 22:03 |
| bauzas | alaski: claims are compute objects | 22:04 |
| alaski | bauzas: now. I thought there was talk of moving them | 22:04 |
| *** oro has quit IRC | 22:04 | |
| *** dims has joined #openstack-nova | 22:04 | |
| bauzas | alaski: even if by some way we have to have a related claiming system for the scheduler | 22:04 |
| bauzas | alaski: yup | 22:04 |
| *** dims_ has quit IRC | 22:05 | |
| *** Hosam has joined #openstack-nova | 22:05 | |
| bauzas | alaski: but that's not exactly for the same reason | 22:05 |
| bauzas | alaski: that's mostly because we're loosing track for the decisions given by the scheduler once select_dest() is returning | 22:05 |
| bauzas | so that's racy | 22:05 |
| alaski | sure, but it would be racy to migrate an instance without telling the scheduler | 22:06 |
| bauzas | alaski: hence my new API | 22:06 |
| bauzas | alaski: we need to make sure the scheduler accepts that | 22:06 |
| *** thomasem has quit IRC | 22:07 | |
| bauzas | alaski: but there is a different paradigm between asking the scheduler to select from a full list of hosts and just verifying a host | 22:07 |
| *** annegentle has joined #openstack-nova | 22:08 | |
| *** vilobhmm1 has joined #openstack-nova | 22:08 | |
| *** mmedvede has joined #openstack-nova | 22:08 | |
| *** vilobhmm1 has quit IRC | 22:08 | |
| *** vilobhmm1 has joined #openstack-nova | 22:09 | |
| *** busterswt has quit IRC | 22:09 | |
| *** melwitt_ has joined #openstack-nova | 22:11 | |
| *** melwitt_ has quit IRC | 22:11 | |
| *** melwitt has quit IRC | 22:12 | |
| *** gyee has joined #openstack-nova | 22:13 | |
| *** armax has quit IRC | 22:15 | |
| *** burt has quit IRC | 22:15 | |
| *** Nic has quit IRC | 22:15 | |
| *** jamielennox|away is now known as jamielennox | 22:16 | |
| *** Sukhdev has quit IRC | 22:19 | |
| *** iamjarvo has joined #openstack-nova | 22:19 | |
| *** Sukhdev has joined #openstack-nova | 22:20 | |
| *** melwitt has joined #openstack-nova | 22:20 | |
| *** vilobhmm1 has quit IRC | 22:21 | |
| *** erlon_awaY has quit IRC | 22:21 | |
| *** vilobhmm1 has joined #openstack-nova | 22:21 | |
| *** mragupat has quit IRC | 22:22 | |
| *** alanf-mc has quit IRC | 22:28 | |
| *** packet has quit IRC | 22:30 | |
| *** alanf-mc has joined #openstack-nova | 22:30 | |
| *** Hosam has quit IRC | 22:31 | |
| openstackgerrit | melanie witt proposed openstack/nova: WIP: send object to cells instance_update_at_top https://review.openstack.org/176518 | 22:32 |
| *** VW_ has quit IRC | 22:33 | |
| *** emagana has joined #openstack-nova | 22:36 | |
| *** thomasem has joined #openstack-nova | 22:39 | |
| *** thomasem has quit IRC | 22:39 | |
| *** igordcard has joined #openstack-nova | 22:39 | |
| *** oro has joined #openstack-nova | 22:40 | |
| *** thomasem has joined #openstack-nova | 22:42 | |
| *** emagana has quit IRC | 22:42 | |
| *** annegentle has quit IRC | 22:43 | |
| *** obedmr has quit IRC | 22:45 | |
| *** emagana has joined #openstack-nova | 22:46 | |
| *** artom has quit IRC | 22:49 | |
| openstackgerrit | Fei Long Wang proposed openstack/nova: Fix nova backup for volume-backed instance https://review.openstack.org/164494 | 22:49 |
| *** jlk has left #openstack-nova | 22:51 | |
| *** annegentle has joined #openstack-nova | 22:58 | |
| *** achanda has quit IRC | 23:01 | |
| *** vilobhmm1 has quit IRC | 23:02 | |
| *** vilobhmm1 has joined #openstack-nova | 23:02 | |
| *** annegentle has quit IRC | 23:03 | |
| *** marun has quit IRC | 23:05 | |
| *** baoli has quit IRC | 23:06 | |
| *** oomichi has quit IRC | 23:06 | |
| *** marun has joined #openstack-nova | 23:06 | |
| *** melwitt has quit IRC | 23:07 | |
| *** iamjarvo has quit IRC | 23:07 | |
| *** oro has quit IRC | 23:08 | |
| openstackgerrit | Chris Friesen proposed openstack/nova: fix network setup on evacuate https://review.openstack.org/169827 | 23:08 |
| *** annegentle has joined #openstack-nova | 23:08 | |
| *** Sukhdev_ has joined #openstack-nova | 23:10 | |
| *** claudiub has quit IRC | 23:10 | |
| *** Sukhdev has quit IRC | 23:11 | |
| *** hemna_ has quit IRC | 23:11 | |
| *** marun has quit IRC | 23:11 | |
| *** vilobhmm11 has joined #openstack-nova | 23:11 | |
| *** vilobhmm11 has quit IRC | 23:12 | |
| *** vilobhmm11 has joined #openstack-nova | 23:12 | |
| *** BadCub is now known as BadCub_Hiding | 23:12 | |
| alex_xu | bauzas: it should be here https://github.com/openstack/api-site/tree/master/api-ref/src/wadls/compute-api/src/v2.1/wadl | 23:13 |
| *** vilobhmm1 has quit IRC | 23:13 | |
| dansmith | jogo: pretty nasty resize bug: https://review.openstack.org/#/c/179228/ | 23:14 |
| dansmith | jogo: apparently resize-revert has been broken with neutron since junoish | 23:14 |
| *** achanda has joined #openstack-nova | 23:14 | |
| *** salv-orl_ has joined #openstack-nova | 23:14 | |
| jogo | dansmith: not surprised | 23:15 |
| jogo | dansmith: so is there a tempest test for this somewhere that is failing? | 23:15 |
| dansmith | jogo: it should only fail on multinode, but it's not and we can't figure out why | 23:15 |
| dansmith | jogo: but that test is clearly misconfigured | 23:15 |
| dansmith | jogo: and the operator that reported the issue has confirmed this fixes it and +1d the patch | 23:16 |
| dansmith | so, we're still trying to figure out why it's not failing in our job, but so far it's not clear | 23:16 |
| *** jgriffith has quit IRC | 23:16 | |
| dansmith | i.e. the test has no nova admin creds configured, but is also successfully sending events to nova, which shouldn't be possible, as far as I know | 23:17 |
| jogo | dansmith: multinode and neutron right? | 23:17 |
| dansmith | yes | 23:17 |
| *** jgriffith has joined #openstack-nova | 23:17 | |
| jogo | ahh, hmm I recall something like this failing | 23:17 |
| *** salv-orlando has quit IRC | 23:17 | |
| jogo | but not sure now | 23:18 |
| *** vilobhmm11 has quit IRC | 23:18 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Merge tag '2015.1.0' https://review.openstack.org/179284 | 23:18 |
| *** vilobhmm1 has joined #openstack-nova | 23:18 | |
| mtreinish | dansmith: which test? | 23:19 |
| *** krak has joined #openstack-nova | 23:19 | |
| dansmith | mtreinish: the resize_revert | 23:19 |
| jogo | mtreinish: on multinode neutron | 23:19 |
| jogo | when resize is on the same node | 23:20 |
| dansmith | mtreinish: it's not a tempest problem though, because the op was hitting it with tempest and this patch makes it pass :) | 23:20 |
| dansmith | no | 23:20 |
| dansmith | jogo: when resize is *not* on the same node | 23:20 |
| jogo | err woops | 23:20 |
| openstackgerrit | Alex Xu proposed openstack/nova: Add version bump guideline in microversion devref https://review.openstack.org/177778 | 23:20 |
| mtreinish | dansmith: yeah, it doesn't look like it has admin. Is it required for resize? | 23:23 |
| dansmith | mtreinish: no, it's required for sending events back to nova | 23:23 |
| jogo | this test right ? test_resize_server_revert | 23:24 |
| dansmith | mtreinish: and nova is configured to require them | 23:24 |
| dansmith | jogo: yeah | 23:24 |
| dansmith | mtreinish: I'm talking about neutron having nova_admin_* creds in its conf file | 23:24 |
| mtreinish | dansmith: ah ok, I misunderstood | 23:24 |
| dansmith | when I said "that test" i meant "that job" | 23:24 |
| dansmith | jogo: are these using noauth? | 23:24 |
| jogo | dansmith: how do I check? | 23:25 |
| dansmith | would be nova policy I think | 23:26 |
| jogo | http://logs.openstack.org/28/179228/3/experimental/check-tempest-dsvm-neutron-multinode-full/f61ef10/logs/ | 23:26 |
| jogo | http://logs.openstack.org/28/179228/3/experimental/check-tempest-dsvm-neutron-multinode-full/f61ef10/logs/etc/nova/policy.json.txt.gz | 23:26 |
| dansmith | actually paste, | 23:27 |
| dansmith | which I'm looking at | 23:27 |
| dansmith | but I don't know how we configure to actually use one or the other :$ | 23:27 |
| jogo | I doubt we are using noauth | 23:28 |
| dansmith | I wouldn't think so | 23:28 |
| jogo | auth_strategy = keystone | 23:28 |
| jogo | http://logs.openstack.org/28/179228/3/experimental/check-tempest-dsvm-neutron-multinode-full/f61ef10/logs/etc/nova/nova.conf.txt.gz | 23:28 |
| dansmith | that's not it | 23:28 |
| dansmith | that's how we auth to neutron I think | 23:28 |
| dansmith | this is neutron calling to us | 23:29 |
| jogo | we default general auth for nova to keystone as well | 23:30 |
| dansmith | where? | 23:30 |
| jogo | nova/api/auth.py | 23:31 |
| dansmith | each endpoint in api-paste has a noauth, noauth2, and keystone | 23:31 |
| dansmith | and I don't know how we choose keystone | 23:31 |
| dansmith | ah, okay | 23:31 |
| jogo | dansmith: ugly python foo code | 23:31 |
| dansmith | I was expecting to see that default in the nova.conf | 23:31 |
| dansmith | well, I dunno, it's acting like it's just letting anything send an event with no creds | 23:32 |
| dansmith | so I dunno | 23:32 |
| jogo | so I looked at the logs from ^ and it looks like we resize across instances | 23:32 |
| jogo | can you get the users tempest.conf? | 23:32 |
| dansmith | across compute nodes you mean? | 23:32 |
| mtreinish | jogo: it's probably running with tenant isolation, so it's making a new user for each test class | 23:32 |
| *** jlk has joined #openstack-nova | 23:33 | |
| dansmith | jogo: it's jlk | 23:33 |
| jlk | ohai | 23:33 |
| dansmith | mtreinish: what do you think that has to do with this? | 23:33 |
| jogo | jlk: we are trying to reproduce the test_resize_server_revert failure on our multi node neutron job | 23:33 |
| openstackgerrit | Alex Xu proposed openstack/nova: Add the rule of separate plugin for Nova REST API in devref https://review.openstack.org/162913 | 23:34 |
| jogo | jlk: and are having trouble. Would you be able to share your tempest.conf with us for comparison? | 23:34 |
| jlk | sure | 23:34 |
| mtreinish | dansmith: nothing jogo just said users from the tempest.conf, and I just wanted to clarify that they probably weren't being used | 23:34 |
| dansmith | okay | 23:34 |
| jlk | let me grep out comments | 23:35 |
| dansmith | independent of this bug, neutron has no nova admin creds and is sending us events, because I see them in the logs | 23:35 |
| dansmith | and I don't know how that can be :) | 23:35 |
| jogo | jlk: comparing it with http://logs.openstack.org/28/179228/3/experimental/check-tempest-dsvm-neutron-multinode-full/f61ef10/logs/tempest_conf.txt.gz | 23:35 |
| *** melwitt has joined #openstack-nova | 23:36 | |
| *** annegentle has quit IRC | 23:36 | |
| *** isd has joined #openstack-nova | 23:36 | |
| jogo | dansmith: http://logs.openstack.org/28/179228/3/experimental/check-tempest-dsvm-neutron-multinode-full/f61ef10/logs/etc/neutron/neutron.conf.txt.gz | 23:37 |
| jogo | search for '[nova]' | 23:37 |
| jlk | jogo: https://gist.github.com/j2sol/4a118407aedd13d37a70 | 23:38 |
| *** emagana has quit IRC | 23:38 | |
| *** marun has joined #openstack-nova | 23:40 | |
| *** tjones2 has quit IRC | 23:41 | |
| *** emagana has joined #openstack-nova | 23:41 | |
| jogo | jlk thanks | 23:42 |
| *** Alexandra_ has joined #openstack-nova | 23:43 | |
| jogo | to be clear what test failed for you? | 23:43 |
| jogo | jlk: test_resize_server_revert? | 23:44 |
| jlk | yes | 23:44 |
| jogo | kk | 23:44 |
| jogo | btw you set run_ssh = true | 23:44 |
| jogo | then false a few lines later | 23:44 |
| jlk | yeah, the false wins | 23:44 |
| *** krtaylor has quit IRC | 23:44 | |
| mtreinish | jogo: heh, yeah false is the right choice for that one :) | 23:45 |
| *** irenab has joined #openstack-nova | 23:45 | |
| *** salv-orl_ has quit IRC | 23:46 | |
| jogo | yeah | 23:46 |
| jogo | jlk: hmm I don't see anything that stands out | 23:46 |
| jlk | I suppose it's possible that our neutron just sucks | 23:47 |
| jogo | dansmith: not sure what is going on and its too late in the day for me to dig into the nova logs | 23:47 |
| jogo | jlk: everyones neutron socks | 23:47 |
| jlk | lets try tomorrow | 23:47 |
| jogo | sucks | 23:47 |
| jogo | jlk: well the fix is going in either way | 23:47 |
| jlk | could be a false positive and the instance doesn't actually have the IP back, I didn't try to validate connectivity | 23:47 |
| *** harlowja has joined #openstack-nova | 23:49 | |
| *** irenab has quit IRC | 23:50 | |
| *** otter768 has joined #openstack-nova | 23:53 | |
| openstackgerrit | melanie witt proposed openstack/nova: Fix disconnect_volume issue when find_multipath_device returns None https://review.openstack.org/159626 | 23:54 |
| *** eliqiao1 has joined #openstack-nova | 23:55 | |
| *** krtaylor has joined #openstack-nova | 23:56 | |
| dansmith | jogo: +1, I have crap I gotta do anyway | 23:58 |
| *** otter768 has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!