| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-01-16 | |||
| 22:18:38 | mriedem | i could | |
| 22:18:52 | mriedem | and get less cute with the list_cells awk | |
| 22:19:36 | mriedem | yeah let me separate that quick, basically just if n-cpu; then discover_hosts | |
| 22:19:46 | dansmith | yeah | |
| 22:21:13 | openstackgerrit | Dan Smith proposed openstack/nova: WIP Move instance creation to conductor https://review.openstack.org/319379 | |
| 22:28:24 | mriedem | ok done | |
| 22:28:39 | mriedem | dansmith: here is the change you'll need https://review.openstack.org/#/c/420981/1 | |
| 22:29:00 | dansmith | thanks | |
| 22:36:54 | mriedem | jaypipes: +2 on https://review.openstack.org/#/c/398473/ now | |
| 22:36:55 | mriedem | thanks | |
| 22:50:45 | mriedem | alright apparently i'm going to die in an ice storm and should head home, so doing that now, will be back online later | |
| 23:07:53 | openstackgerrit | Dan Smith proposed openstack/nova: WIP Move instance creation to conductor https://review.openstack.org/319379 | |
| 23:10:07 | openstackgerrit | Gage Hugo proposed openstack/nova: Fixed warning messages when building docs https://review.openstack.org/420990 | |
| 23:35:28 | openstackgerrit | Takashi NATSUME proposed openstack/nova: api-ref: Fix deprecated proxy API parameters https://review.openstack.org/416882 | |
| 23:44:59 | openstackgerrit | Dan Smith proposed openstack/nova: Multicell support for instance listing https://review.openstack.org/396775 | |
| 23:45:00 | openstackgerrit | Dan Smith proposed openstack/nova: WIP Move instance creation to conductor https://review.openstack.org/319379 | |
| 23:45:00 | openstackgerrit | Dan Smith proposed openstack/nova: Make api_samples tests use simple cell environment https://review.openstack.org/410473 | |
| 23:53:03 | mriedem | hmm, wtf is the difference between gate-tempest-dsvm-multinode-full-ubuntu-xenial-nv and gate-tempest-dsvm-neutron-multinode-full-ubuntu-xenial-nv now that both are running neutron | |
| 23:56:14 | mriedem | heh, turns out, not a gd thing | |
| #openstack-nova - 2017-01-17 | |||
| 00:04:16 | mriedem | i farqed up the discover_hosts devstack patch, update coming | |
| 00:05:09 | mikal | mriedem: what does it mean to have an instance mapping in a cell with no hosts? | |
| 00:05:27 | mriedem | mikal: mayhem is what | |
| 00:05:56 | mriedem | mikal: context? | |
| 00:05:59 | mikal | mriedem: see comment on https://review.openstack.org/#/c/420451 | |
| 00:06:13 | mikal | mriedem: you check if there are any hosts, and only if there are none do you check if there are any instances | |
| 00:06:18 | mikal | mriedem: which seems... redundant to me | |
| 00:06:24 | mikal | mriedem: but also not the end of the world | |
| 00:06:43 | mriedem | good question | |
| 00:06:49 | dansmith | they're different things, and if we had a host_map delete command, | |
| 00:06:56 | dansmith | then you'd still need to check that there aren't any instances | |
| 00:07:06 | dansmith | this delete command is to let you delete cells you created and didn't setup | |
| 00:07:25 | dansmith | since you can't really have multiple real cells yet, we don't likely need more of the delete commands just yet | |
| 00:07:31 | mikal | dansmith: but doesn't no hosts imply no instances? | |
| 00:07:36 | dansmith | no | |
| 00:07:45 | dansmith | I mean, I know what you're saying, | |
| 00:07:49 | mikal | dansmith: right, explain that bit in small words for me | |
| 00:07:56 | dansmith | but if you had deleted a host you would still have some instance mappings | |
| 00:08:16 | mikal | So deleting the host which has the instances on it doesn't cascade the delete to the instances? | |
| 00:08:17 | dansmith | there is no foreign key constraint on that stuff, | |
| 00:08:22 | dansmith | so you could have nuked some things | |
| 00:08:33 | dansmith | you can't delete a host yet, so .. no :) | |
| 00:09:03 | mikal | Heh | |
| 00:09:30 | mriedem | like today you can delete a compute node in the db by deleting the nova-compute service for that host | |
| 00:09:44 | mriedem | and i'm pretty sure that doesn't mean you can't still have instnaces in the db pointing at that node | |
| 00:09:48 | dansmith | right | |
| 00:09:52 | mikal | That's a fair point | |
| 00:09:56 | dansmith | just sanity checking | |
| 00:10:01 | mikal | I don't feel we need to hold the new thing to a higher standard than what we have now | |
| 00:14:18 | mriedem | mikal: i'm failing to pick up on sarcasm there | |
| 00:19:00 | mikal | mriedem: Oh, sorry | |
| 00:19:19 | mikal | mriedem: I am glad the new code is consistent in its level of terrible with the old | |
| 00:19:23 | mikal | mgood enough? | |
| 00:19:49 | Kevin_Zheng | Hi cores, any particular reason that we have limited the length of instance tag to be 60 but created a db model with max length of 80? | |
| 00:22:22 | Kevin_Zheng | http://git.openstack.org/cgit/openstack/nova/tree/nova/db/sqlalchemy/models.py#n1456 db model | |
| 00:22:56 | mriedem | Kevin_Zheng: hmm, probably a question for jaypipes | |
| 00:23:31 | Kevin_Zheng | mriedm: hmm ok | |
| 00:24:10 | mriedem | my guess is the 60 number in the API comes from the spec https://specs.openstack.org/openstack/nova-specs/specs/liberty/approved/tag-instances.html#data-model-impact | |
| 00:24:37 | Kevin_Zheng | http://git.openstack.org/cgit/openstack/nova/tree/nova/objects/tag.py#n18 FYI | |
| 00:24:50 | Kevin_Zheng | yeah, but why 80 in db model :) | |
| 00:25:00 | jaypipes | Kevin_Zheng: no idea... probably just an oversight. | |
| 00:25:22 | mriedem | snikitin: ^ | |
| 00:27:20 | mriedem | the db model change was https://review.openstack.org/#/c/97168/ | |
| 00:27:25 | mriedem | which was a long long slog | |
| 00:27:32 | mriedem | so my guess would be, the db model stuff merged, | |
| 00:27:40 | mriedem | and then a release or more later the REST API change merged | |
| 00:27:45 | Kevin_Zheng | maybe we can expand limit to 80? as we already created 80 in db | |
| 00:27:49 | mriedem | at which point we'd totally forgotten what we were doing | |
| 00:27:58 | Kevin_Zheng | haha | |
| 00:30:52 | mriedem | yeah even looking at patch set 1 of https://review.openstack.org/#/c/128940/127/nova/objects/tag.py | |
| 00:30:57 | Kevin_Zheng | looks like it start with 80 in the initial patch and nobody asked about it | |
| 00:30:59 | Kevin_Zheng | yeah | |
| 00:31:00 | mriedem | https://review.openstack.org/#/c/128940/1/nova/api/openstack/compute/schemas/v2/instance_tags.py | |
| 00:31:02 | mriedem | it was 60 | |
| 00:31:10 | mriedem | ^ had 127 patch sets | |
| 00:31:34 | mriedem | i don't know if the 60 vs 80 has anything to do with unicode? | |
| 00:32:02 | Kevin_Zheng | ah maybe, he said something about converting in spec | |
| 00:32:10 | mriedem | Kevin_Zheng: i assume you're asking because you need to know the schema for booting an instance with tags right? | |
| 00:32:41 | mriedem | snikitin would be the person to ask if he remembers a good reason for it | |
| 00:34:17 | Kevin_Zheng | ah, good, thanks, will try to catch him | |
| 00:35:09 | openstackgerrit | Claudiu Belu proposed openstack/nova: Adds CPUFeaturesField to ImageMetaProps https://review.openstack.org/420960 | |
| 00:35:47 | jaypipes | mriedem: no, 60 vs. 80 has nothing to do with unicode. | |
| 00:39:21 | Kevin_Zheng | jaypipes: so is it just an oversight or do we have any particular reason? | |
| 00:48:19 | mriedem | i think it's probably just an oversight | |
| 00:49:18 | Kevin_Zheng | lets sync them :) | |
| 00:50:56 | mriedem | well, | |
| 00:50:59 | mriedem | we could | |
| 00:51:13 | mriedem | Kevin_Zheng: is there a technical reason you need tags in the REST API to allow up to 80? | |
| 00:52:17 | mriedem | Kevin_Zheng: normally changing something like this in the REST API schema would be a microversion change, but this isn't making things more restrictive so it might be ok to change it without a microversion | |
| 00:52:29 | Kevin_Zheng | yeah | |
| 00:52:37 | mriedem | Kevin_Zheng: i think we should start with a post to the ML thread about it so snikitin has a chance to respond (and alex_xu) | |
| 00:52:43 | mriedem | Kevin_Zheng: can you start that thread? | |
| 00:52:48 | Kevin_Zheng | yeah sure | |
| 00:52:54 | mriedem | thanks | |
| 00:54:31 | openstackgerrit | Zhenyu Zheng proposed openstack/nova: Expand tag length limit to 80 https://review.openstack.org/421004 | |
| 00:55:05 | Kevin_Zheng | ^ was already doing this, but we can block it until have a clean idea from ML | |
| 01:02:13 | mriedem | something still failing in discover_hosts on the subnode http://logs.openstack.org/76/420976/4/check/gate-tempest-dsvm-neutron-multinode-full-ubuntu-xenial-nv/1e4a55e/logs/subnode-2/devstacklog.txt.gz#_2017-01-17_00_35_47_248 | |
| 01:13:37 | openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/420844 | |
| 01:43:49 | openstackgerrit | Tuan Luong-Anh proposed openstack/nova: Support Ironic interface attach/detach in nova virt https://review.openstack.org/364413 | |
| 01:45:04 | openstackgerrit | Sivasathurappan Radhakrishnan proposed openstack/nova: Initiate network changes from source during live migration https://review.openstack.org/404607 | |
| 01:57:00 | openstackgerrit | Anh Tran proposed openstack/nova: Remove mox stubs in api/openstack/compute/test_server_reset_state.py https://review.openstack.org/320826 | |