| *** ferest has joined #openstack-nova | 00:00 | |
| *** jwang has quit IRC | 00:00 | |
| *** jwang has joined #openstack-nova | 00:00 | |
| *** ferest has quit IRC | 00:00 | |
| *** penick has quit IRC | 00:01 | |
| *** ferest has joined #openstack-nova | 00:01 | |
| *** Sukhdev has quit IRC | 00:01 | |
| *** thorst has quit IRC | 00:04 | |
| *** vishwanathj has quit IRC | 00:05 | |
| *** ferest has quit IRC | 00:05 | |
| *** ferest has joined #openstack-nova | 00:06 | |
| *** ferest has quit IRC | 00:06 | |
| *** penick has joined #openstack-nova | 00:06 | |
| *** ferest has joined #openstack-nova | 00:07 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/nova: Add total_vms field in API https://review.openstack.org/203382 | 00:08 |
|---|---|---|
| openstackgerrit | Jeffrey Zhang proposed openstack/nova: Add total_vms field to ComputeNode https://review.openstack.org/202472 | 00:08 |
| *** markvoelker has joined #openstack-nova | 00:09 | |
| dims | edleafe: ping | 00:11 |
| *** ferest has quit IRC | 00:11 | |
| *** ferest has joined #openstack-nova | 00:12 | |
| openstackgerrit | Nicolas Simonds proposed openstack/nova: libvirt: Split out resize_image logic from create_image https://review.openstack.org/187857 | 00:12 |
| openstackgerrit | Nicolas Simonds proposed openstack/nova: libvirt: Fix/implement revert-resize for RBD-backed images https://review.openstack.org/187395 | 00:12 |
| *** ferest has quit IRC | 00:12 | |
| dims | edleafe: (re: https://review.openstack.org/#/c/187968/), i've responded to your question, please reconsider the -1 | 00:12 |
| *** ferest has joined #openstack-nova | 00:13 | |
| *** andreykurilin__ has quit IRC | 00:14 | |
| *** penick has quit IRC | 00:19 | |
| *** ferest has quit IRC | 00:19 | |
| *** ferest has joined #openstack-nova | 00:20 | |
| *** ferest has quit IRC | 00:21 | |
| *** jasondotstar has quit IRC | 00:21 | |
| *** ferest has joined #openstack-nova | 00:21 | |
| *** ferest has quit IRC | 00:21 | |
| *** jasondotstar has joined #openstack-nova | 00:22 | |
| *** dims has quit IRC | 00:22 | |
| *** ferest has joined #openstack-nova | 00:23 | |
| *** ddaskal has quit IRC | 00:23 | |
| *** ferest has quit IRC | 00:24 | |
| *** mtanino_ has joined #openstack-nova | 00:25 | |
| *** ferest has joined #openstack-nova | 00:25 | |
| *** ferest has quit IRC | 00:26 | |
| openstackgerrit | Meghal Gosalia proposed openstack/nova-specs: Quota Management for Dynamic Quota Resources https://review.openstack.org/206160 | 00:26 |
| *** mtanino has quit IRC | 00:26 | |
| *** ferest has joined #openstack-nova | 00:27 | |
| *** ferest has quit IRC | 00:27 | |
| *** tsekiyam_ has joined #openstack-nova | 00:27 | |
| *** tonytan4ever has joined #openstack-nova | 00:28 | |
| *** ferest has joined #openstack-nova | 00:28 | |
| *** Mike_D_laptop has quit IRC | 00:30 | |
| *** ferest has quit IRC | 00:30 | |
| *** xyang2 has joined #openstack-nova | 00:30 | |
| *** ferest has joined #openstack-nova | 00:31 | |
| *** tsekiyama has quit IRC | 00:31 | |
| *** dims has joined #openstack-nova | 00:32 | |
| *** tonytan4ever has quit IRC | 00:32 | |
| openstackgerrit | Meghal Gosalia proposed openstack/nova-specs: Quota Management for Dynamic Quota Resources https://review.openstack.org/206160 | 00:33 |
| *** xyang2 has quit IRC | 00:33 | |
| *** salv-orlando has joined #openstack-nova | 00:33 | |
| *** tpeoples has joined #openstack-nova | 00:34 | |
| *** ferest has quit IRC | 00:34 | |
| *** ferest has joined #openstack-nova | 00:36 | |
| *** ferest has quit IRC | 00:37 | |
| *** salv-orlando has quit IRC | 00:37 | |
| *** jasondotstar is now known as jasondotstar|afk | 00:43 | |
| openstackgerrit | Ed Leafe proposed openstack/nova: Add non-negative integer and float fields https://review.openstack.org/206249 | 00:43 |
| *** pixelbeat_ has joined #openstack-nova | 00:44 | |
| *** ildikov has joined #openstack-nova | 00:46 | |
| *** openstackgerrit has quit IRC | 00:46 | |
| *** openstackgerrit has joined #openstack-nova | 00:47 | |
| openstackgerrit | sajeesh cs proposed openstack/nova: Nested Quota Driver: Adding v2 and v3(v2.1) support https://review.openstack.org/149828 | 00:52 |
| *** dave-mcc_ has joined #openstack-nova | 00:53 | |
| *** markmcclain has quit IRC | 00:53 | |
| *** diogogmt has joined #openstack-nova | 00:54 | |
| *** annashen has joined #openstack-nova | 00:54 | |
| *** kevinbenton has quit IRC | 00:54 | |
| *** jbernard has quit IRC | 00:56 | |
| *** jbernard has joined #openstack-nova | 00:56 | |
| *** kevinbenton has joined #openstack-nova | 00:57 | |
| *** markmcclain has joined #openstack-nova | 00:57 | |
| *** dave-mccowan has quit IRC | 00:57 | |
| *** mdrabe has quit IRC | 00:58 | |
| *** mdrabe has joined #openstack-nova | 00:59 | |
| *** annashen has quit IRC | 00:59 | |
| *** Jeffrey4l has quit IRC | 01:00 | |
| *** Jeffrey4l has joined #openstack-nova | 01:12 | |
| openstackgerrit | Shiina, Hironori proposed openstack/nova: Add a compute RPC API to inject an NMI https://review.openstack.org/202605 | 01:12 |
| openstackgerrit | Eli Qiao proposed openstack/nova: Object: Fix KeyError when loading instance from db https://review.openstack.org/205917 | 01:13 |
| *** Sukhdev has joined #openstack-nova | 01:14 | |
| *** jasondotstar|afk has quit IRC | 01:17 | |
| *** haomaiwang has quit IRC | 01:22 | |
| *** haomaiwang has joined #openstack-nova | 01:23 | |
| *** pixelbeat_ has quit IRC | 01:24 | |
| *** meghal has quit IRC | 01:24 | |
| *** pixelbeat_ has joined #openstack-nova | 01:25 | |
| *** Sukhdev has quit IRC | 01:25 | |
| *** vilobhmm has quit IRC | 01:26 | |
| *** jwcroppe has quit IRC | 01:30 | |
| *** jwcroppe has joined #openstack-nova | 01:30 | |
| *** jwcroppe has quit IRC | 01:35 | |
| *** shuquan has joined #openstack-nova | 01:35 | |
| *** davideagnello has quit IRC | 01:36 | |
| *** salv-orlando has joined #openstack-nova | 01:44 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Enable python34 tests for nova/tests/unit/scheduler/test*.py https://review.openstack.org/187968 | 01:44 |
| *** shuquan has quit IRC | 01:46 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Get py34 subunit.run test discovery to work https://review.openstack.org/204723 | 01:47 |
| *** salv-orlando has quit IRC | 01:48 | |
| *** shuquan has joined #openstack-nova | 01:48 | |
| *** MentalRay has joined #openstack-nova | 01:49 | |
| *** ssurana has joined #openstack-nova | 01:54 | |
| *** annashen has joined #openstack-nova | 01:55 | |
| *** bauzas has quit IRC | 01:58 | |
| *** Piet has quit IRC | 01:58 | |
| *** jlvillal has quit IRC | 01:58 | |
| *** jlvillal has joined #openstack-nova | 01:59 | |
| *** annashen has quit IRC | 02:00 | |
| *** bauzas has joined #openstack-nova | 02:00 | |
| *** nic has quit IRC | 02:03 | |
| *** tsekiyam_ has quit IRC | 02:03 | |
| *** wanghua has joined #openstack-nova | 02:04 | |
| *** unicell has quit IRC | 02:06 | |
| *** mtanino_ has quit IRC | 02:06 | |
| *** d34dh0r53 has quit IRC | 02:07 | |
| *** d34dh0r53 has joined #openstack-nova | 02:08 | |
| *** d34dh0r53 has joined #openstack-nova | 02:08 | |
| *** d34dh0r53 has quit IRC | 02:09 | |
| *** d34dh0r53 has joined #openstack-nova | 02:10 | |
| *** humble_ has joined #openstack-nova | 02:11 | |
| *** jasondotstar|afk has joined #openstack-nova | 02:12 | |
| *** angdraug has quit IRC | 02:12 | |
| *** tjones1 has quit IRC | 02:12 | |
| *** wanghua has quit IRC | 02:15 | |
| *** mdrabe has quit IRC | 02:17 | |
| *** annashen has joined #openstack-nova | 02:17 | |
| *** mdrabe has joined #openstack-nova | 02:18 | |
| *** zhenguo has joined #openstack-nova | 02:19 | |
| *** artom has quit IRC | 02:29 | |
| *** shyu has joined #openstack-nova | 02:33 | |
| *** bkopilov has quit IRC | 02:35 | |
| openstackgerrit | sajeesh cs proposed openstack/nova: Nested Quota Driver: Adding v2 and v3 support https://review.openstack.org/149828 | 02:35 |
| *** dims has quit IRC | 02:35 | |
| ildikov | alex_xu: hi. are you around? | 02:37 |
| *** changbl has quit IRC | 02:37 | |
| alex_xu | ildikov: hi, I'm here | 02:37 |
| *** Murali has joined #openstack-nova | 02:38 | |
| ildikov | alex_xu: can I disturb you with some dumb microversioning questions? | 02:38 |
| alex_xu | ildikov: sure, glad to help | 02:39 |
| ildikov | alex_xu: great, thanks | 02:39 |
| *** mragupat has joined #openstack-nova | 02:40 | |
| ildikov | alex_xu: so here is this patch: https://review.openstack.org/#/c/147520/48 | 02:40 |
| *** suresh12 has quit IRC | 02:40 | |
| ildikov | alex_xu: what we change in the API is the list that the jsonschema validates against | 02:40 |
| alex_xu | ildikov: yea | 02:41 |
| ildikov | alex_xu: ... and in the controller class I don't see any version info at the moment | 02:41 |
| ildikov | alex_xu: so I got a little confused with what I should bump now and how | 02:41 |
| alex_xu | ildikov: let me find some link for you | 02:41 |
| ildikov | alex_xu: as the schema itself is not changed | 02:41 |
| ildikov | alex_xu: I read the devref, but it mentioned schema changes as much as I saw | 02:42 |
| *** davideagnello has joined #openstack-nova | 02:42 | |
| alex_xu | ildikov: your patch is similar with this https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/plugins/v3/remote_consoles.py#L158 | 02:43 |
| ildikov | alex_xu: so then I should add version info here: https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/plugins/v3/server_groups.py#L129 | 02:44 |
| ildikov | alex_xu: and also add numbering to the schema | 02:45 |
| alex_xu | ildikov: yes | 02:45 |
| alex_xu | ildikov: you should have different schema for different version | 02:45 |
| ildikov | alex_xu: so I need to change the schema itself too to use the "old" list and the newer version should use the extended list, right? | 02:46 |
| ildikov | alex_xu: the old API should remain unchanged now? I mean emove the changes under contrib? | 02:47 |
| alex_xu | ildikov: yea, https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/schemas/v3/remote_consoles.py#L91-L113 | 02:47 |
| ildikov | alex_xu: yeap, this is what I meant | 02:48 |
| alex_xu | ildikov: cool :) | 02:48 |
| ildikov | alex_xu: should this file still be changed: https://review.openstack.org/#/c/147520/48/nova/api/openstack/compute/contrib/server_groups.py ? | 02:49 |
| ildikov | alex_xu: or just plugins/v3/... from now on? | 02:50 |
| alex_xu | ildikov: I think no, the v2 API is freeze now | 02:50 |
| alex_xu | ildikov: yes | 02:50 |
| *** mragupat has quit IRC | 02:51 | |
| ildikov | alex_xu: ok, then I remove that change | 02:51 |
| ildikov | alex_xu: thanks for your help! | 02:51 |
| alex_xu | ildikov: np :) | 02:51 |
| ildikov | alex_xu: and sorry for the kinda obvious questions :) | 02:52 |
| *** hakimo_ has joined #openstack-nova | 02:52 | |
| alex_xu | ildikov: it's fine, microversion is new stuff, everyone need time to get fimilar with it | 02:53 |
| *** pixelbeat_ has quit IRC | 02:53 | |
| *** patrickeast has quit IRC | 02:54 | |
| *** hakimo has quit IRC | 02:54 | |
| ildikov | alex_xu: yeap, it looks complicated enough to deserve some 'let's go for sure' questions :) | 02:54 |
| alex_xu | ildikov: yea | 02:55 |
| *** suresh12 has joined #openstack-nova | 02:55 | |
| ildikov | alex_xu: well, hopefully it will work out well, then it worth it to deal with the competence buildup phase | 02:56 |
| *** annashen_ has joined #openstack-nova | 02:56 | |
| *** vladikr has joined #openstack-nova | 03:00 | |
| *** annashen_ has quit IRC | 03:01 | |
| openstackgerrit | wangxiyuan proposed openstack/python-novaclient: Add 'deleted' status check in _poll_for_status https://review.openstack.org/197457 | 03:05 |
| *** Murali has quit IRC | 03:07 | |
| *** baoli has joined #openstack-nova | 03:08 | |
| *** sajeesh has quit IRC | 03:10 | |
| *** baoli has quit IRC | 03:13 | |
| *** otter768 has quit IRC | 03:14 | |
| *** shuquan has quit IRC | 03:19 | |
| *** shuquan has joined #openstack-nova | 03:20 | |
| *** Kennan2 has quit IRC | 03:20 | |
| *** Kennan has joined #openstack-nova | 03:20 | |
| *** Sukhdev has joined #openstack-nova | 03:21 | |
| *** jasondotstar|afk has quit IRC | 03:24 | |
| *** diogogmt has quit IRC | 03:24 | |
| *** VW_ has joined #openstack-nova | 03:29 | |
| *** jwcroppe has joined #openstack-nova | 03:30 | |
| *** shuquan has quit IRC | 03:32 | |
| *** coolsvap|away is now known as coolsvap | 03:32 | |
| *** kiran-r has joined #openstack-nova | 03:34 | |
| *** suresh12 has quit IRC | 03:36 | |
| *** Jeffrey4l has quit IRC | 03:37 | |
| *** kiran-r has quit IRC | 03:40 | |
| *** davideagnello has quit IRC | 03:40 | |
| *** suresh12 has joined #openstack-nova | 03:43 | |
| *** Nastooh_ has quit IRC | 03:44 | |
| *** annashen has quit IRC | 03:45 | |
| openstackgerrit | Zaina Afoulki proposed openstack/nova: XenAPI Driver: set capabilites explicitly https://review.openstack.org/206288 | 03:47 |
| *** Jeffrey4l has joined #openstack-nova | 03:51 | |
| *** meghal has joined #openstack-nova | 03:53 | |
| *** vilobhmm has joined #openstack-nova | 03:55 | |
| *** Jeffrey4l has quit IRC | 03:57 | |
| *** annashen has joined #openstack-nova | 03:57 | |
| *** annashen has quit IRC | 04:01 | |
| *** bkopilov has joined #openstack-nova | 04:02 | |
| *** jwcroppe has quit IRC | 04:02 | |
| *** patrickeast has joined #openstack-nova | 04:03 | |
| *** jwcroppe has joined #openstack-nova | 04:03 | |
| *** VW_ has quit IRC | 04:04 | |
| *** markvoelker has quit IRC | 04:04 | |
| *** VW_ has joined #openstack-nova | 04:04 | |
| *** jwcroppe_ has joined #openstack-nova | 04:06 | |
| *** jwcroppe has quit IRC | 04:08 | |
| *** VW_ has quit IRC | 04:09 | |
| *** Jeffrey4l has joined #openstack-nova | 04:09 | |
| openstackgerrit | lyanchih proposed openstack/nova: Remove fixed test in tempest-dsvm-cells-rc https://review.openstack.org/205512 | 04:09 |
| openstackgerrit | lyanchih proposed openstack/nova: Creating instance fail when inject ssh key in cells mode https://review.openstack.org/197514 | 04:09 |
| *** nelsnelson has quit IRC | 04:13 | |
| *** meghal has quit IRC | 04:14 | |
| *** jwcroppe_ has quit IRC | 04:16 | |
| *** jwcroppe has joined #openstack-nova | 04:17 | |
| *** jwcroppe has quit IRC | 04:22 | |
| *** ssurana has quit IRC | 04:35 | |
| *** suresh12 has quit IRC | 04:37 | |
| *** chhavi has joined #openstack-nova | 04:39 | |
| *** mdrabe has quit IRC | 04:41 | |
| *** Mike_D_laptop has joined #openstack-nova | 04:49 | |
| *** deepakcs has joined #openstack-nova | 04:51 | |
| *** tjones has joined #openstack-nova | 04:51 | |
| *** patrickeast_ has joined #openstack-nova | 04:54 | |
| *** jwcroppe has joined #openstack-nova | 04:56 | |
| *** patrickeast has quit IRC | 04:56 | |
| *** patrickeast_ is now known as patrickeast | 04:56 | |
| *** Mike_D_laptop has left #openstack-nova | 04:57 | |
| *** jwcroppe has quit IRC | 04:58 | |
| *** annashen has joined #openstack-nova | 04:58 | |
| *** jwcroppe has joined #openstack-nova | 04:58 | |
| *** yamahata has joined #openstack-nova | 04:59 | |
| *** _gryf has quit IRC | 05:00 | |
| *** jwcroppe_ has joined #openstack-nova | 05:01 | |
| *** annashen has quit IRC | 05:02 | |
| *** Nikolay_St has joined #openstack-nova | 05:03 | |
| *** jwcroppe has quit IRC | 05:03 | |
| openstackgerrit | lyanchih proposed openstack/nova: Remove fixed test in tempest-dsvm-cells-rc https://review.openstack.org/205512 | 05:05 |
| openstackgerrit | lyanchih proposed openstack/nova: Creating instance fail when inject ssh key in cells mode https://review.openstack.org/197514 | 05:05 |
| *** markvoelker has joined #openstack-nova | 05:05 | |
| *** markvoelker has quit IRC | 05:10 | |
| *** doude has quit IRC | 05:13 | |
| openstackgerrit | lyanchih proposed openstack/nova: libvirt - Unrescue will not remove rescue disk in ceph when image_type=rbd https://review.openstack.org/205766 | 05:13 |
| *** tjones has quit IRC | 05:14 | |
| *** unicell has joined #openstack-nova | 05:14 | |
| *** otter768 has joined #openstack-nova | 05:15 | |
| *** otter768 has quit IRC | 05:19 | |
| *** Sukhdev has quit IRC | 05:20 | |
| openstackgerrit | yuntongjin proposed openstack/nova: Add missing policy for limits extention https://review.openstack.org/173178 | 05:28 |
| openstackgerrit | Alex Xu proposed openstack/python-novaclient: Adds support for x509 certificates as keypairs https://review.openstack.org/136458 | 05:31 |
| openstackgerrit | Alex Xu proposed openstack/python-novaclient: Add version discover and check in CLI https://review.openstack.org/184281 | 05:31 |
| openstackgerrit | Alex Xu proposed openstack/python-novaclient: Set "latest" as default compute api version https://review.openstack.org/184658 | 05:31 |
| *** shuquan has joined #openstack-nova | 05:35 | |
| *** abhijeetm has joined #openstack-nova | 05:39 | |
| *** Murali has joined #openstack-nova | 05:45 | |
| *** jamielennox is now known as jamielennox|away | 05:52 | |
| *** lucas__ has joined #openstack-nova | 05:52 | |
| *** rushiagr_away is now known as rushiagr | 05:55 | |
| *** garyk has joined #openstack-nova | 05:55 | |
| *** salv-orlando has joined #openstack-nova | 05:55 | |
| *** jamielennox|away is now known as jamielennox | 05:56 | |
| *** lucas__ has quit IRC | 05:57 | |
| openstackgerrit | Deepak C Shetty (deepakcs) proposed openstack/nova: libvirt: Nova part of volume snapshot improvements https://review.openstack.org/172813 | 05:57 |
| openstackgerrit | Deepak C Shetty (deepakcs) proposed openstack/nova: objects: Add support for OnlineSnapshotInfo object https://review.openstack.org/194538 | 05:57 |
| *** unicell1 has joined #openstack-nova | 05:58 | |
| *** unicell has quit IRC | 05:59 | |
| *** annashen has joined #openstack-nova | 05:59 | |
| *** annashen has quit IRC | 06:03 | |
| *** _gryf has joined #openstack-nova | 06:04 | |
| *** vilobhmm has quit IRC | 06:13 | |
| *** Jeffrey4l has quit IRC | 06:14 | |
| *** cfriesen has quit IRC | 06:16 | |
| *** deepthi has joined #openstack-nova | 06:18 | |
| *** patrickeast has quit IRC | 06:22 | |
| *** belmoreira has joined #openstack-nova | 06:25 | |
| *** Nikolay_St has quit IRC | 06:25 | |
| *** Jeffrey4l has joined #openstack-nova | 06:27 | |
| *** doude has joined #openstack-nova | 06:27 | |
| *** doude has quit IRC | 06:27 | |
| *** doude has joined #openstack-nova | 06:27 | |
| *** vishwanathj has joined #openstack-nova | 06:28 | |
| openstackgerrit | Alex Xu proposed openstack/nova: Skip additionalProperties checks when LegacyV2CompatibleWrapper enabled https://review.openstack.org/193858 | 06:36 |
| openstackgerrit | Alex Xu proposed openstack/nova: Add scenario for API sample tests with legacy v2 compatible wrapper https://review.openstack.org/203485 | 06:36 |
| openstackgerrit | Alex Xu proposed openstack/nova: Fix the wrong file name for legacy v2 compatible wrapper functional test https://review.openstack.org/206326 | 06:36 |
| *** liusheng has quit IRC | 06:36 | |
| *** liusheng has joined #openstack-nova | 06:37 | |
| *** nkrinner has joined #openstack-nova | 06:42 | |
| *** ramineni1 has joined #openstack-nova | 06:42 | |
| *** afazekas has joined #openstack-nova | 06:44 | |
| -openstackstatus- NOTICE: zuul is stuck and about to undergo an emergency restart, please be patient as job results may take a long time | 06:46 | |
| *** ChanServ changes topic to "zuul is stuck and about to undergo an emergency restart, please be patient as job results may take a long time" | 06:46 | |
| *** jwcroppe_ has quit IRC | 06:48 | |
| *** romainh has joined #openstack-nova | 06:50 | |
| *** sahid has joined #openstack-nova | 06:53 | |
| *** tpeoples has quit IRC | 06:55 | |
| *** markus_z has joined #openstack-nova | 06:59 | |
| *** apuimedo has quit IRC | 06:59 | |
| *** salv-orlando has quit IRC | 06:59 | |
| *** annashen has joined #openstack-nova | 07:00 | |
| *** rdopiera has joined #openstack-nova | 07:00 | |
| *** rdopiera_ has joined #openstack-nova | 07:02 | |
| openstackgerrit | Shiina, Hironori proposed openstack/nova: Add a Nova API to inject an NMI https://review.openstack.org/202617 | 07:03 |
| *** scheuran has joined #openstack-nova | 07:05 | |
| *** annashen has quit IRC | 07:05 | |
| *** ramineni1 has left #openstack-nova | 07:06 | |
| *** markvoelker has joined #openstack-nova | 07:06 | |
| *** markvoelker has quit IRC | 07:11 | |
| *** achanda has quit IRC | 07:11 | |
| *** moshele has joined #openstack-nova | 07:11 | |
| *** apuimedo has joined #openstack-nova | 07:12 | |
| *** Nikolay_St has joined #openstack-nova | 07:14 | |
| *** otter768 has joined #openstack-nova | 07:16 | |
| *** achanda has joined #openstack-nova | 07:18 | |
| *** otter768 has quit IRC | 07:21 | |
| *** ishant__ has joined #openstack-nova | 07:21 | |
| *** bfic has joined #openstack-nova | 07:26 | |
| openstackgerrit | Maxim Nestratov proposed openstack/nova: libvirt: rename parallels driver to virtuozzo https://review.openstack.org/184311 | 07:27 |
| *** mpaolino has joined #openstack-nova | 07:29 | |
| *** atuvenie has joined #openstack-nova | 07:30 | |
| *** lpetrut has joined #openstack-nova | 07:34 | |
| *** sahid has quit IRC | 07:35 | |
| *** browne has quit IRC | 07:35 | |
| *** achanda has quit IRC | 07:35 | |
| *** sahid has joined #openstack-nova | 07:36 | |
| *** sahid has quit IRC | 07:36 | |
| *** alex_klimov has joined #openstack-nova | 07:38 | |
| *** sahid has joined #openstack-nova | 07:38 | |
| *** Jeffrey4l has quit IRC | 07:40 | |
| *** Kevin_Zheng has joined #openstack-nova | 07:40 | |
| *** ajayaa has joined #openstack-nova | 07:40 | |
| *** Administrator_ has joined #openstack-nova | 07:43 | |
| *** jlanoux has joined #openstack-nova | 07:44 | |
| *** yamahata has quit IRC | 07:44 | |
| *** achanda has joined #openstack-nova | 07:46 | |
| *** Kevin_Zheng has quit IRC | 07:46 | |
| *** Administrator_ has quit IRC | 07:48 | |
| *** Marga_ has quit IRC | 07:49 | |
| *** shuquan has quit IRC | 07:49 | |
| *** salv-orlando has joined #openstack-nova | 07:50 | |
| *** Marga_ has joined #openstack-nova | 07:50 | |
| *** Jeffrey4l has joined #openstack-nova | 07:53 | |
| *** shuquan has joined #openstack-nova | 07:55 | |
| *** Kevin_Zheng has joined #openstack-nova | 07:56 | |
| *** coolsvap is now known as coolsvap|away | 07:57 | |
| *** Administrator_ has joined #openstack-nova | 07:57 | |
| *** Marga_ has quit IRC | 07:57 | |
| *** Marga_ has joined #openstack-nova | 07:58 | |
| *** Administrator__ has joined #openstack-nova | 07:59 | |
| *** achanda has quit IRC | 08:00 | |
| *** annashen has joined #openstack-nova | 08:01 | |
| *** kmARC has joined #openstack-nova | 08:01 | |
| *** Kevin_Zheng has quit IRC | 08:01 | |
| *** achanda has joined #openstack-nova | 08:01 | |
| *** aloga has quit IRC | 08:02 | |
| *** jwcroppe has joined #openstack-nova | 08:02 | |
| *** aloga has joined #openstack-nova | 08:02 | |
| *** jistr has joined #openstack-nova | 08:03 | |
| *** Administrator_ has quit IRC | 08:03 | |
| *** Kevin_Zheng has joined #openstack-nova | 08:03 | |
| -openstackstatus- NOTICE: zuul has been restarted and queues restored. It may take some time to work through the backlog. | 08:03 | |
| *** ChanServ changes topic to "This channel is for Nova development. For support of Nova deployments, please use #openstack. -=- master is open for Lemming development" | 08:03 | |
| *** achanda has quit IRC | 08:04 | |
| *** Administrator__ has quit IRC | 08:04 | |
| *** Administrator_ has joined #openstack-nova | 08:04 | |
| *** annashen has quit IRC | 08:05 | |
| *** Administrator__ has joined #openstack-nova | 08:07 | |
| openstackgerrit | Marian Horban proposed openstack/nova: libvirt:Rsync remote FS driver was added https://review.openstack.org/146887 | 08:07 |
| *** Kevin_Zheng has quit IRC | 08:07 | |
| *** lucasagomes has joined #openstack-nova | 08:08 | |
| *** Kevin_Zheng has joined #openstack-nova | 08:09 | |
| *** Administrator_ has quit IRC | 08:10 | |
| *** Administrator_ has joined #openstack-nova | 08:11 | |
| *** karimb has joined #openstack-nova | 08:11 | |
| *** Administrator__ has quit IRC | 08:11 | |
| *** achanda has joined #openstack-nova | 08:11 | |
| *** Administrator__ has joined #openstack-nova | 08:12 | |
| *** ndipanov has quit IRC | 08:12 | |
| *** Kevin_Zheng has quit IRC | 08:14 | |
| *** Kevin_Zheng has joined #openstack-nova | 08:14 | |
| *** Administrator_ has quit IRC | 08:15 | |
| *** Administrator__ has quit IRC | 08:16 | |
| *** Administrator_ has joined #openstack-nova | 08:17 | |
| *** pixelbeat_ has joined #openstack-nova | 08:18 | |
| *** Administrator__ has joined #openstack-nova | 08:18 | |
| *** Kevin_Zheng has quit IRC | 08:19 | |
| *** Kevin_Zheng has joined #openstack-nova | 08:20 | |
| openstackgerrit | Adelina Tuvenie proposed openstack/nova: Moves OVS related code to a common location https://review.openstack.org/179727 | 08:21 |
| openstackgerrit | Adelina Tuvenie proposed openstack/nova: Adds Hyper-V OVS ViF driver https://review.openstack.org/140045 | 08:21 |
| *** Administrator_ has quit IRC | 08:21 | |
| *** shuquan has quit IRC | 08:21 | |
| *** shuquan has joined #openstack-nova | 08:22 | |
| *** Administrator__ has quit IRC | 08:22 | |
| *** Administrator_ has joined #openstack-nova | 08:24 | |
| *** Administrator__ has joined #openstack-nova | 08:25 | |
| *** Kevin_Zheng has quit IRC | 08:27 | |
| *** vishwanathj has quit IRC | 08:27 | |
| *** Kevin_Zheng has joined #openstack-nova | 08:27 | |
| *** Administrator_ has quit IRC | 08:28 | |
| *** vishwanathj has joined #openstack-nova | 08:28 | |
| *** Administrator_ has joined #openstack-nova | 08:29 | |
| *** Administrator__ has quit IRC | 08:30 | |
| *** Administrator__ has joined #openstack-nova | 08:32 | |
| *** achanda has quit IRC | 08:32 | |
| *** Kevin_Zheng has quit IRC | 08:32 | |
| *** alexpilotti has joined #openstack-nova | 08:33 | |
| *** Kevin_Zheng has joined #openstack-nova | 08:34 | |
| *** Administrator_ has quit IRC | 08:34 | |
| openstackgerrit | Adelina Tuvenie proposed openstack/nova: Adds RemoteFX support to the Hyper-V driver https://review.openstack.org/42529 | 08:34 |
| *** Administrator_ has joined #openstack-nova | 08:35 | |
| bauzas | morning nova | 08:35 |
| bauzas | lxsli: around ? | 08:35 |
| bauzas | lxsli: re: resource-objects | 08:35 |
| *** Administrator__ has quit IRC | 08:36 | |
| *** MentalRay has quit IRC | 08:36 | |
| lxsli | bauzas: morning | 08:36 |
| *** Administrator__ has joined #openstack-nova | 08:37 | |
| *** Kevin_Zheng has quit IRC | 08:38 | |
| *** Kevin_Zheng has joined #openstack-nova | 08:39 | |
| *** vishwanathj has quit IRC | 08:39 | |
| *** gszasz has joined #openstack-nova | 08:39 | |
| openstackgerrit | Alexis Lee proposed openstack/nova: Add non-negative integer and float fields https://review.openstack.org/206249 | 08:39 |
| openstackgerrit | Alexis Lee proposed openstack/nova: Add object models for comparing resource amounts https://review.openstack.org/128992 | 08:39 |
| bauzas | lxsli: mouuuurning | 08:39 |
| *** Administrator_ has quit IRC | 08:39 | |
| *** yassine__ has joined #openstack-nova | 08:40 | |
| bauzas | lxsli: that's exactly my thoughts, edleafe just pulled the trigger too fast | 08:40 |
| openstackgerrit | Alexis Lee proposed openstack/nova: Adds resource classes https://review.openstack.org/128999 | 08:40 |
| bauzas | lxsli: the whole series was needing a rebase | 08:40 |
| lxsli | bauzas: ah I just did that :) | 08:40 |
| bauzas | lxsli: coolness | 08:40 |
| *** Administrator_ has joined #openstack-nova | 08:40 | |
| lxsli | bauzas: how are the rest of my updates? | 08:41 |
| *** Administrator__ has quit IRC | 08:41 | |
| bauzas | lxsli: the commit msg is a bit weird for https://review.openstack.org/206249, I will provide some comments instead of directly fixing it - because I don't want to break the chain by providing a new PS | 08:41 |
| bauzas | lxsli: I'll do another pass by this morning | 08:41 |
| bauzas | lxsli: I left some comments on PS20 because you weren't exactly going in the same direction than the one I said in PS19 | 08:42 |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Register JSON-Home data for resources actions https://review.openstack.org/206360 | 08:42 |
| bauzas | lxsli: also, the __hash__ builtin implementation sounded good to me | 08:43 |
| *** Administrator__ has joined #openstack-nova | 08:43 | |
| lxsli | bauzas: unfortunately most NovaObjects are mutable so it's actively harmful to implement __hash__ based on values | 08:43 |
| bauzas | lxsli: maybe using the hash(field) was better than just returning the field, but that's fine since it's a positive integer | 08:43 |
| bauzas | lxsli: see my explanation, you can have mutable objects in a collection | 08:44 |
| *** Kevin_Zheng has quit IRC | 08:44 | |
| *** y_sawai has joined #openstack-nova | 08:44 | |
| bauzas | lxsli: I mean, you can key any mutable object | 08:44 |
| bauzas | the rules are pretty simple, the mutable object has to be hashable in order to be in an hashable collection | 08:45 |
| lxsli | but how can a mutable object's hash be stable? | 08:45 |
| *** aix has joined #openstack-nova | 08:45 | |
| *** Kevin_Zheng has joined #openstack-nova | 08:45 | |
| bauzas | lxsli: mutable just means that the reference is the same | 08:46 |
| bauzas | when updating the object | 08:46 |
| lxsli | a = [1]; b = []; c = set((a,b)); a.delete(0) | 08:46 |
| *** Administrator_ has quit IRC | 08:46 | |
| lxsli | now c has two objects with the same hash in it, which is illegal for a set | 08:46 |
| *** Administrator_ has joined #openstack-nova | 08:46 | |
| bauzas | lxsli: did you see my paste ? | 08:47 |
| *** Administrator__ has quit IRC | 08:48 | |
| *** Administrator__ has joined #openstack-nova | 08:48 | |
| lxsli | not yet, just looking at your comments on PS#20 | 08:48 |
| bauzas | lxsli: also, lists are unhashable, how can you have 'a' as a set item ? | 08:48 |
| bauzas | lxsli: I mean as a tuple item | 08:49 |
| kashyap | For your Tuesday Nova amusement, /me came across this yesterday :-) http://blog.oddbit.com/2015/07/26/in-which-we-are-amazed-it-doesnt-all-fall-apart/ | 08:49 |
| lxsli | bauzas: there's a reason lists are unhashable, that's because they're mutable :) | 08:49 |
| bauzas | lxsli: here, we're not talking about lists but objectds | 08:49 |
| lxsli | you can easily imagine breaking set's contract similarly with NovaObjects | 08:49 |
| bauzas | lxsli: and given your list can be an object field, it can become hashable | 08:50 |
| *** Kevin_Zheng has quit IRC | 08:50 | |
| lxsli | kashyap: morning | 08:50 |
| kashyap | Morning. | 08:50 |
| *** achanda has joined #openstack-nova | 08:51 | |
| *** Kevin_Zheng has joined #openstack-nova | 08:51 | |
| *** Administrator_ has quit IRC | 08:52 | |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: switch to oslo.cache lib https://review.openstack.org/203049 | 08:52 |
| lxsli | bauzas: https://docs.python.org/3.0/glossary.html#term-hashable | 08:52 |
| *** Administrator__ has quit IRC | 08:52 | |
| *** Administrator_ has joined #openstack-nova | 08:53 | |
| *** vishwanathj has joined #openstack-nova | 08:53 | |
| *** ZZelle has quit IRC | 08:53 | |
| *** ZZelle has joined #openstack-nova | 08:53 | |
| *** Administrator__ has joined #openstack-nova | 08:54 | |
| *** coolsvap|away is now known as coolsvap | 08:56 | |
| *** Kevin_Zheng has quit IRC | 08:56 | |
| bauzas | "An object is hashable if it has a hash value which never changes during its lifetime" | 08:57 |
| *** doude_ has joined #openstack-nova | 08:57 | |
| *** Administrator_ has quit IRC | 08:57 | |
| *** doude_ has quit IRC | 08:57 | |
| *** Kevin_Zheng has joined #openstack-nova | 08:57 | |
| bauzas | lxsli: I see your point, but I still don't see why we shouldn't accept having to implement __hash__() | 08:58 |
| openstackgerrit | Hans Lindgren proposed openstack/nova: Remove unused refresh_security_group_members() call https://review.openstack.org/182632 | 08:58 |
| openstackgerrit | Hans Lindgren proposed openstack/nova: Make security group refresh with refresh_instance_security_groups() https://review.openstack.org/182633 | 08:58 |
| lxsli | I'm replying on PS#20 | 08:59 |
| openstackgerrit | Adelina Tuvenie proposed openstack/nova: Adds Hyper-V vNic hot-plug implementation https://review.openstack.org/138255 | 08:59 |
| *** Administrator_ has joined #openstack-nova | 08:59 | |
| *** doude has quit IRC | 08:59 | |
| *** Administrator__ has quit IRC | 08:59 | |
| *** achanda has quit IRC | 09:00 | |
| bauzas | lxsli: http://paste.openstack.org/show/406028/ | 09:00 |
| bauzas | another example | 09:00 |
| *** Administrator__ has joined #openstack-nova | 09:01 | |
| *** matrohon has joined #openstack-nova | 09:01 | |
| *** annashen has joined #openstack-nova | 09:01 | |
| *** danpb has joined #openstack-nova | 09:02 | |
| bauzas | lxsli: anyway, sounds we're yakshaving | 09:02 |
| *** jwcroppe has quit IRC | 09:02 | |
| *** Kevin_Zheng has quit IRC | 09:03 | |
| bauzas | lxsli: my take is that if an instance of IntegerAmount is changing because the amount field changes, then even if the object keeps the same reference, it needs a different hash | 09:03 |
| *** Administrator_ has quit IRC | 09:04 | |
| *** Kevin_Zheng has joined #openstack-nova | 09:04 | |
| bauzas | lxsli: here, you consider deferring the hash value of a Facade to an internal value, which is immutable | 09:04 |
| *** annashen has quit IRC | 09:06 | |
| *** Administrator_ has joined #openstack-nova | 09:07 | |
| *** markvoelker has joined #openstack-nova | 09:07 | |
| *** Administrator__ has quit IRC | 09:07 | |
| *** Administrator__ has joined #openstack-nova | 09:08 | |
| openstackgerrit | Marc Fouché proposed openstack/nova-specs: Inheritable admin image properties https://review.openstack.org/206431 | 09:09 |
| *** vishwanathj has quit IRC | 09:10 | |
| *** Kevin_Zheng has quit IRC | 09:10 | |
| *** Kevin_Zheng has joined #openstack-nova | 09:10 | |
| *** boris-42 has quit IRC | 09:10 | |
| *** Administrator_ has quit IRC | 09:11 | |
| *** Administrator_ has joined #openstack-nova | 09:11 | |
| *** markvoelker has quit IRC | 09:11 | |
| *** vishwanathj has joined #openstack-nova | 09:12 | |
| *** ndipanov has joined #openstack-nova | 09:12 | |
| *** Administrator__ has quit IRC | 09:12 | |
| *** Administrator__ has joined #openstack-nova | 09:13 | |
| deepakcs | danpb: sahid: when u both have some time, https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/volume-snapshot-improvements,n,z | 09:15 |
| *** Kevin_Zheng has quit IRC | 09:15 | |
| *** Kevin_Zheng has joined #openstack-nova | 09:15 | |
| bauzas | lxsli: last point, that SO thread explains why it's allowed http://stackoverflow.com/questions/4418741/im-able-to-use-a-mutable-object-as-a-dictionary-key-in-python-is-this-not-disa | 09:15 |
| *** jwcroppe has joined #openstack-nova | 09:16 | |
| *** Administrator_ has quit IRC | 09:16 | |
| *** jwcroppe has quit IRC | 09:16 | |
| bauzas | lxsli: here, providing an hash method will make sure that your behaviour will be consistent with your internal representation | 09:16 |
| lxsli | which answer? | 09:16 |
| *** achanda has joined #openstack-nova | 09:17 | |
| *** Administrator_ has joined #openstack-nova | 09:17 | |
| bauzas | lxsli: the first one is good | 09:17 |
| *** otter768 has joined #openstack-nova | 09:17 | |
| bauzas | lxsli: if not providing __hash__, then the NovaObject as a vanilla object will still have an hash | 09:17 |
| *** Administrator__ has quit IRC | 09:17 | |
| lxsli | bauzas: that answer agrees with me completely | 09:17 |
| bauzas | lxsli: which will basically be its id | 09:17 |
| *** Administrator__ has joined #openstack-nova | 09:18 | |
| bauzas | lxsli: if you leave hash() unimplemented for an object, it means that you could have some unexpected surprises because an object is *by default* hashable | 09:18 |
| bauzas | even if it's mutable | 09:19 |
| lxsli | bauzas: put replies on PS#20 | 09:19 |
| lxsli | bauzas: yes but the default impl of __hash__ doesn't refer to those mutable attrs | 09:19 |
| lxsli | bauzas: so it never changes, maintaining the contract | 09:19 |
| *** Kevin_Zheng has quit IRC | 09:20 | |
| bauzas | but then A(1) and A(1) would be diffreent | 09:20 |
| lxsli | yes, this is the default behaviour | 09:20 |
| lxsli | (I presume you mean A(1) != A(1)) | 09:20 |
| *** atuvenie has quit IRC | 09:21 | |
| bauzas | lxsli: no, because it's valid to have set([A(1), A(1)]) if hash is on 'amount' | 09:21 |
| *** Administrator_ has quit IRC | 09:21 | |
| *** otter768 has quit IRC | 09:21 | |
| bauzas | lxsli: it will just make sure that we only have one set([A(1)]) because the amounts are identical | 09:21 |
| lxsli | bauzas: I don't seem to be getting across to you | 09:22 |
| *** Kevin_Zheng has joined #openstack-nova | 09:22 | |
| lxsli | bauzas: maybe read the set contract? it's based on __eq__ | 09:22 |
| *** Administrator__ has quit IRC | 09:22 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Show 'locked_by' information in server Show & List Detail https://review.openstack.org/202391 | 09:23 |
| bauzas | lxsli: it doesn't | 09:23 |
| *** Administrator_ has joined #openstack-nova | 09:24 | |
| *** abhishekk has joined #openstack-nova | 09:24 | |
| bauzas | lxsli: http://paste.openstack.org/show/406035/ | 09:24 |
| *** exploreshaifali has joined #openstack-nova | 09:24 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Show 'locked_by' information in server Show & List Detail https://review.openstack.org/202391 | 09:24 |
| *** Administrator__ has joined #openstack-nova | 09:25 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Show 'locked_by' information in server Show & List Detail https://review.openstack.org/202391 | 09:27 |
| *** Kevin_Zheng has quit IRC | 09:27 | |
| *** Kevin_Zheng has joined #openstack-nova | 09:28 | |
| *** Administrator_ has quit IRC | 09:28 | |
| *** y_sawai has quit IRC | 09:29 | |
| *** Administrator_ has joined #openstack-nova | 09:29 | |
| *** jasondotstar|afk has joined #openstack-nova | 09:29 | |
| *** Administrator__ has quit IRC | 09:30 | |
| *** Administrator__ has joined #openstack-nova | 09:31 | |
| *** Kevin_Zheng has quit IRC | 09:32 | |
| *** achanda has quit IRC | 09:33 | |
| *** Kevin_Zheng has joined #openstack-nova | 09:34 | |
| *** Administrator_ has quit IRC | 09:34 | |
| *** Administrator_ has joined #openstack-nova | 09:35 | |
| *** atuvenie has joined #openstack-nova | 09:35 | |
| *** Administrator__ has quit IRC | 09:36 | |
| *** romainh has quit IRC | 09:36 | |
| *** Administrator__ has joined #openstack-nova | 09:37 | |
| *** jwcroppe has joined #openstack-nova | 09:38 | |
| *** Kevin_Zheng has quit IRC | 09:38 | |
| *** Administrator_ has quit IRC | 09:40 | |
| *** Kevin_Zheng has joined #openstack-nova | 09:40 | |
| *** romainh has joined #openstack-nova | 09:41 | |
| *** jwcroppe has quit IRC | 09:42 | |
| *** Administrator__ has quit IRC | 09:42 | |
| *** Administrator_ has joined #openstack-nova | 09:43 | |
| *** Administrator__ has joined #openstack-nova | 09:45 | |
| *** gzhai has joined #openstack-nova | 09:45 | |
| *** Kevin_Zheng has quit IRC | 09:46 | |
| *** Kevin_Zheng has joined #openstack-nova | 09:47 | |
| lxsli | bauzas: http://paste.openstack.org/show/406039/ | 09:48 |
| *** Administrator_ has quit IRC | 09:48 | |
| lxsli | hmm hold on, typo | 09:48 |
| *** doude has joined #openstack-nova | 09:49 | |
| *** Administrator_ has joined #openstack-nova | 09:49 | |
| *** Administrator__ has quit IRC | 09:50 | |
| lxsli | http://paste.openstack.org/show/406040/ | 09:50 |
| lxsli | first paste was staggeringly wrong >.< | 09:51 |
| *** Administrator__ has joined #openstack-nova | 09:51 | |
| *** Kevin_Zheng has quit IRC | 09:52 | |
| *** atuvenie has quit IRC | 09:53 | |
| *** kjelly has quit IRC | 09:53 | |
| bauzas | lxsli: okay, I was just wondering what was the difference between Default and Mutable :p | 09:53 |
| bauzas | lxsli: okay, again, I don't say you're wrong | 09:53 |
| bauzas | lxsli: yes, there are some caveats | 09:53 |
| *** Kevin_Zheng has joined #openstack-nova | 09:54 | |
| bauzas | lxsli: yes, it doesn't prevent dups within sets if the object changes | 09:54 |
| lxsli | do you see now how providing hash based on mutable data leads to the hash changing and breaking the contract of set? | 09:54 |
| lxsli | well, that is entirely the problem | 09:54 |
| *** jasondotstar|afk has quit IRC | 09:54 | |
| *** Administrator_ has quit IRC | 09:54 | |
| lxsli | dicts are also affected if you use badly hashed objects as keys | 09:54 |
| *** Administrator_ has joined #openstack-nova | 09:55 | |
| *** aix has quit IRC | 09:55 | |
| *** Administrator__ has quit IRC | 09:55 | |
| bauzas | lxsli: okay, sounds like we discussed 30 mins for getting me convinced about something I already knew :/ | 09:56 |
| lxsli | bauzas: well as long as we now agree I'm right :p | 09:56 |
| bauzas | lxsli: again, I just feel that's better to keep an hash value | 09:56 |
| lxsli | why provide a hash function that breaks contract and hence is unusable? | 09:56 |
| *** Administrator__ has joined #openstack-nova | 09:57 | |
| *** Kevin_Zheng has quit IRC | 09:58 | |
| bauzas | lxsli: because jay introduced it ? :p | 09:58 |
| bauzas | lxsli: see my comment on PS22 | 09:58 |
| bauzas | I'm glad to nitpick for like 30 mins | 09:59 |
| lxsli | I think it was probably me that introduced it | 09:59 |
| lxsli | and I don't like appeal to authority | 09:59 |
| bauzas | because I think that could make sense if the proposer was wanting to have this behaviour | 09:59 |
| *** Administrator_ has quit IRC | 09:59 | |
| bauzas | it's all about the attended behaviour | 09:59 |
| lxsli | not if they're wrong | 10:00 |
| lxsli | a proposer might try to define 1 == 2, I'm not letting that in either, even if it was Jay proposing it | 10:00 |
| *** Kevin_Zheng has joined #openstack-nova | 10:00 | |
| bauzas | eh | 10:01 |
| bauzas | don't get me wrong, that's acceptable to have such behaviour - if we want to have it | 10:01 |
| *** Administrator_ has joined #openstack-nova | 10:02 | |
| *** Administrator__ has quit IRC | 10:02 | |
| *** shuquan has quit IRC | 10:02 | |
| *** annashen has joined #openstack-nova | 10:02 | |
| *** Administrator__ has joined #openstack-nova | 10:04 | |
| *** jwcroppe has joined #openstack-nova | 10:04 | |
| *** Kevin_Zheng has quit IRC | 10:04 | |
| lxsli | We'll have to agree to disagree | 10:05 |
| lxsli | and I guess that means your -1 remains | 10:05 |
| *** Kevin_Zheng has joined #openstack-nova | 10:06 | |
| *** Administrator_ has quit IRC | 10:06 | |
| *** Administrator_ has joined #openstack-nova | 10:07 | |
| *** annashen has quit IRC | 10:07 | |
| *** Administrator__ has quit IRC | 10:08 | |
| *** Administrator__ has joined #openstack-nova | 10:09 | |
| *** Kevin_Zheng has quit IRC | 10:11 | |
| lxsli | bauzas: about the __init__ args, do you have any reasons to back up your preference? | 10:11 |
| lxsli | what impact does this have? | 10:11 |
| *** Administrator_ has quit IRC | 10:12 | |
| *** jasondotstar|afk has joined #openstack-nova | 10:12 | |
| gzhai | sdague:hello | 10:12 |
| *** Kevin_Zheng has joined #openstack-nova | 10:12 | |
| ndipanov | bauzas, stupid question - what is the difference (in practice between) disabling a service and this http://specs.openstack.org/openstack/nova-specs/specs/liberty/approved/mark-host-down.html | 10:13 |
| gzhai | sdague: Do you have time to review one simple fix? https://review.openstack.org/#/c/203573/ | 10:13 |
| *** Administrator_ has joined #openstack-nova | 10:14 | |
| *** Administrator__ has quit IRC | 10:14 | |
| bauzas | lxsli: well, about __init__, it goes about leaving the possibility to have unset fields | 10:14 |
| bauzas | lxsli: but I don't have a strong -1 here, I at least wanted to get rid of fields validation within the init method | 10:15 |
| bauzas | ndipanov: oh that? | 10:15 |
| lxsli | bauzas: sure the field validation is gone | 10:16 |
| lxsli | bauzas: I don't see how what Jay's done can lead to unset fields? | 10:16 |
| ndipanov | bauzas, yes that :) | 10:16 |
| bauzas | ndipanov: it's because detecting that a service is down is pretty bad atm | 10:16 |
| lxsli | bauzas: if you don't supply total or used you'll get a TypeError | 10:16 |
| *** Administrator__ has joined #openstack-nova | 10:17 | |
| ndipanov | so bauzas the question is - how is taht different than disabling the host in practice | 10:17 |
| bauzas | ndipanov: so the poster wanted to leave the possibility to consider that the host is down by passing it from an external tool | 10:17 |
| bauzas | ndipanov: oh, because sometimes you want your service to be enabled but considered as down | 10:17 |
| *** Kevin_Zheng has quit IRC | 10:17 | |
| ndipanov | bauzas, what is the use case for that? | 10:18 |
| bauzas | ndipanov: https://github.com/openstack/nova/blob/master/nova/compute/api.py#L3281-L3284 | 10:19 |
| *** Kevin_Zheng has joined #openstack-nova | 10:19 | |
| *** jwcroppe has quit IRC | 10:19 | |
| openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: check min required qemu/libvirt versions on s390/s390x https://review.openstack.org/183486 | 10:20 |
| bauzas | ndipanov: service_is_up doesn't care about whether the service is enabled or not | 10:20 |
| ndipanov | bauzas, are you serious - we add all of that just because of one check in the code that could be changed to check for disabled | 10:20 |
| *** Administrator_ has quit IRC | 10:20 | |
| johnthetubaguy | ndipanov: thanks for all the reviews on those devref docs, looks good, I will try get back to you next week, once we have this freeze out the way, attempting to focus on code reviews and super time critical stuff this week | 10:20 |
| *** Administrator_ has joined #openstack-nova | 10:20 | |
| openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: mark NUMA huge page mappings as shared access https://review.openstack.org/176360 | 10:21 |
| ndipanov | johnthetubaguy, no probs - mostly it's about making it a bit more time-proof :) and removing some reliance on implicit knowledge of the reader | 10:21 |
| bauzas | ndipanov: that's not the only place where we verify that the service is up | 10:21 |
| ndipanov | very useful work | 10:21 |
| *** Administrator__ has quit IRC | 10:21 | |
| *** aix has joined #openstack-nova | 10:22 | |
| ndipanov | bauzas, so what I am saying is: when is this distinction (between disabled and down) useful? | 10:22 |
| gmann | johnthetubaguy: sdague - please have a look (locked_by ) - https://review.openstack.org/#/c/202391/ | 10:22 |
| johnthetubaguy | ndipanov: yeah, totally good points, we touched on that a bit at the summit, its important to get this time-proof (and not miss leading vs what is not complete!) | 10:22 |
| bauzas | ndipanov: I wasn't part of the whole convo, so I only have a limited amount of understanding here, but we considered that it was better to have a separate field than just marking a service disabled | 10:22 |
| ndipanov | oookay | 10:22 |
| bauzas | ndipanov: lemme remember the whole stuff | 10:22 |
| bauzas | ndipanov: because it was a couple of months before | 10:23 |
| johnthetubaguy | ndipanov: bauzas: that marked down is really where an external system claims it has actually properly fenced a compute node, so evacuate is safe right? | 10:23 |
| ndipanov | johnthetubaguy, yes | 10:23 |
| bauzas | johnthetubaguy: yup, that is | 10:23 |
| johnthetubaguy | its really an external service integration point, in my head anyways | 10:23 |
| ndipanov | I mean I get the difference - it's just that the distinction is subtle enough that it may not be useful reall y | 10:23 |
| bauzas | ndipanov: I also remember that is_up is buggy | 10:23 |
| *** Administrator__ has joined #openstack-nova | 10:23 | |
| *** Kevin_Zheng has quit IRC | 10:24 | |
| bauzas | ndipanov: I mean, given you update a service, it will consider your service back up | 10:24 |
| johnthetubaguy | ndipanov: so I think its to do with the difference between nova-compute being dead, and the whole hypervisor being turned off, I guess | 10:24 |
| johnthetubaguy | ndipanov: so the cost of failure is high, its two writers to the same volume, and massive data corruption | 10:24 |
| bauzas | lemme remember that, long story, needs to read through the review | 10:24 |
| johnthetubaguy | if it wasn't for the data corruption risk, I would say fixing service group would be the better call | 10:25 |
| johnthetubaguy | does that make any more sense? | 10:25 |
| ndipanov | yeah I guess | 10:25 |
| bauzas | yeah, I remember it was pretty difficult to assume that we would mark the hosts just by putting the service disabled because it wasn't the same thing | 10:25 |
| johnthetubaguy | agreed, its getting boarder line, but its that horrid miss match between nova-compute dead, but volume still being accessed | 10:26 |
| ndipanov | the thing I am worried about is how it's being used in the code | 10:26 |
| *** Kevin_Zheng has joined #openstack-nova | 10:26 | |
| johnthetubaguy | ah, OK, you mean using it in too many places? | 10:26 |
| ndipanov | or interchange-ably which may cause subtle bugs | 10:27 |
| *** Administrator_ has quit IRC | 10:27 | |
| ndipanov | it should be something that makes sense for the business logic that we write - can_schedule_to vs host_down | 10:27 |
| ndipanov | which is I guess the same thing | 10:28 |
| *** Administrator_ has joined #openstack-nova | 10:28 | |
| bauzas | ndipanov: so the review is worthwhile https://review.openstack.org/#/c/169836/ | 10:28 |
| *** Administrator__ has quit IRC | 10:28 | |
| bauzas | ndipanov: in particular in PS15 | 10:29 |
| *** alex_klimov has quit IRC | 10:29 | |
| bauzas | ndipanov: so, yeah, the thing was | 10:29 |
| *** Administrator__ has joined #openstack-nova | 10:29 | |
| sdague | morning folks | 10:29 |
| bauzas | ndipanov: so, disabling a service is something orthogonal to checking whether the service is up because it's managing your controlplane | 10:30 |
| bauzas | ndipanov: on the other hand, we want to accelerate detecting whether a service is up, because we know that sometimes some tools know better than us | 10:30 |
| bauzas | whether a service is down, rather | 10:30 |
| *** dims has joined #openstack-nova | 10:31 | |
| *** Kevin_Zheng has quit IRC | 10:31 | |
| bauzas | ndipanov: you can see the consensus coming in by PS15 | 10:31 |
| ndipanov | yep reading that now very useful - too bad it's not really discoverable | 10:32 |
| *** Kevin_Zheng has joined #openstack-nova | 10:32 | |
| *** Administrator_ has quit IRC | 10:32 | |
| bauzas | ndipanov: I agree it's debatable | 10:32 |
| lucasagomes | hi, when you guys have a time please take a look at https://review.openstack.org/#/c/204995/ this is adding some better error handling in the destroy() method of the Ironic driver while preparing it to things we want to do in the future (allowing abort a deployment at any stage) | 10:32 |
| *** Administrator_ has joined #openstack-nova | 10:33 | |
| johnthetubaguy | so this will need some careful docs, which sucks, but I think its reality that sucks here, I think, and we are just modelling that here | 10:33 |
| bauzas | ndipanov: since the SG API needs also some work, one could argue that we should rather fix the service detection | 10:33 |
| *** salv-orl_ has joined #openstack-nova | 10:34 | |
| ndipanov | ok so it seems that the distinction is useful but poorly named | 10:34 |
| bauzas | ndipanov: but here, we consider that even if we strongly improve the mean time to detection, there could be some specific logic that we should consider | 10:34 |
| *** Marga_ has quit IRC | 10:34 | |
| bauzas | ndipanov: probably | 10:34 |
| *** Administrator__ has quit IRC | 10:34 | |
| *** shihanzhang has joined #openstack-nova | 10:34 | |
| johnthetubaguy | lucasagomes: can you please add this patch into the ironic subteam section in here: https://etherpad.openstack.org/p/liberty-nova-priorities-tracking | 10:35 |
| ndipanov | even if we fix it to be instanteanious with ZK we still want the added flag to say - these VMs are dead not doing IO | 10:35 |
| lucasagomes | johnthetubaguy, hey, sure! thanks | 10:35 |
| *** Administrator__ has joined #openstack-nova | 10:35 | |
| bauzas | ndipanov: zactly | 10:35 |
| johnthetubaguy | lucasagomes: I think thats a more scaleable way to get the right folks to see important bug fix patches, at least I would love that to work :) | 10:35 |
| *** shz has joined #openstack-nova | 10:35 | |
| ndipanov | thanks bauzas good tip | 10:35 |
| *** hanzhang has quit IRC | 10:35 | |
| johnthetubaguy | ndipanov: bauzas: yeah, I think thats it, even with perfect service group, we still need this | 10:36 |
| *** Kevin_Zheng has quit IRC | 10:36 | |
| ndipanov | well we could say that we don't and up to the mgmt tool to do things in the right order, but this way we provide nicer integration API for sure | 10:37 |
| lucasagomes | johnthetubaguy, ++ added! Thank you | 10:37 |
| *** salv-orlando has quit IRC | 10:37 | |
| sdague | gmann: reviewed, I'd really like to move that version compare one level deeper, I think it will make everything easier to understand | 10:37 |
| *** deepthi has quit IRC | 10:37 | |
| *** Administrator_ has quit IRC | 10:37 | |
| *** Kevin_Zheng has joined #openstack-nova | 10:38 | |
| johnthetubaguy | lucasagomes: no worries, I hope that helps get some attention on that, our next cycle (next week) is going to focus on merging bug fixes and priority features, I hope | 10:38 |
| *** Administrator_ has joined #openstack-nova | 10:39 | |
| *** Administrator__ has quit IRC | 10:39 | |
| *** shihanzhang has quit IRC | 10:40 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Enable python34 tests for nova/tests/unit/scheduler/test*.py https://review.openstack.org/187968 | 10:40 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Get py34 subunit.run test discovery to work https://review.openstack.org/204723 | 10:41 |
| *** Administrator__ has joined #openstack-nova | 10:41 | |
| *** bradjones has quit IRC | 10:41 | |
| lucasagomes | johnthetubaguy, ah that's great. I will see if I can push on https://review.openstack.org/#/c/204162/ at the Ironic side because that will affect the nova driver too | 10:42 |
| lucasagomes | and add to that etherpad | 10:42 |
| *** Kevin_Zheng has quit IRC | 10:42 | |
| *** claudiub has joined #openstack-nova | 10:42 | |
| johnthetubaguy | lucasagomes: ah, OK, not seen the etherpad used that way, but willing to try that out | 10:42 |
| *** Kevin_Zheng has joined #openstack-nova | 10:43 | |
| gmann | sdague: Thanks, actually i also though of doing that in _view_builder.show but that is being used by other methods also like update, rebuild etc | 10:44 |
| *** Administrator_ has quit IRC | 10:44 | |
| *** shyu has quit IRC | 10:45 | |
| openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: support management of downtime during migration https://review.openstack.org/162253 | 10:45 |
| *** Administrator_ has joined #openstack-nova | 10:45 | |
| *** Administrator__ has quit IRC | 10:45 | |
| *** Jeffrey4l has quit IRC | 10:45 | |
| sdague | gmann: ok, but if this is part of the representation for a server now, why would that be an issue? | 10:45 |
| *** doude has quit IRC | 10:46 | |
| dims | sdague: would the rootwrap daemon be appropriate for L2 or should it wait till L2 is out? (https://review.openstack.org/#/c/180695/) | 10:47 |
| *** Kevin_Zheng has quit IRC | 10:47 | |
| dims | good morning! :) | 10:47 |
| sdague | morning, so, the check queue is 450 deep, I don't think many new patches are really getting in this week :) | 10:47 |
| *** salv-orl_ has quit IRC | 10:47 | |
| gmann | sdague: humm, there are many extended attributes which only been shown in show /detail server response not in update/rebuild response. | 10:48 |
| dims | sdague: ack thanks | 10:48 |
| dims | whoa! just looked | 10:48 |
| *** Administrator__ has joined #openstack-nova | 10:48 | |
| gmann | sdague: but as we are adding new attribute (without any extended namespace etc) here, we can allow it for update/rebuild also. right? | 10:49 |
| dims | looks like something broke last night the charts at the bottom of the zuul queue page are flat last night | 10:49 |
| gmann | sdague: m just wondering about spec where I mentioned only for Show/Detail Response | 10:51 |
| *** Administrator_ has quit IRC | 10:52 | |
| *** eglynn has joined #openstack-nova | 10:52 | |
| sdague | gmann: so, it feels to me that if we have a detailed version of the server, and that's used a bunch of places, adding new detail should end up there as well. | 10:52 |
| *** Kevin_Zheng has joined #openstack-nova | 10:52 | |
| johnthetubaguy | markus_z: we should catch up on bugs, we have a triage day scheduled, I would love for you to help push that :) | 10:52 |
| markus_z | johnthetubaguy: sure. What day is that? I missed that. | 10:53 |
| sdague | gmann: it's also a little weird to me that locked_by is there even if it's 'null' | 10:53 |
| sdague | I don't think I'd have done it that way | 10:54 |
| *** Administrator__ has quit IRC | 10:54 | |
| gmann | sdague: but null convey server is not locked | 10:54 |
| *** Administrator_ has joined #openstack-nova | 10:54 | |
| sdague | gmann: sure, but you could also just not include the attribute | 10:54 |
| *** ankit_ag has joined #openstack-nova | 10:55 | |
| sdague | because for 99.999% of API calls, it won't have a value | 10:55 |
| *** scheuran has quit IRC | 10:55 | |
| gmann | sdague: yea, but that way users has to check that attributes presence first | 10:56 |
| markus_z | johnthetubaguy: 7th August is bug triage day? | 10:56 |
| johnthetubaguy | markus_z: all the dates should be on here: https://wiki.openstack.org/wiki/Nova/Liberty_Release_Schedule | 10:56 |
| johnthetubaguy | markus_z: yeah, I am thinking of moving it to the Tuesday though, if that works for you? | 10:57 |
| *** Administrator__ has joined #openstack-nova | 10:57 | |
| gmann | sdague: but that also make sense , if server is locked then only this attribute is present | 10:57 |
| *** Kevin_Zheng has quit IRC | 10:57 | |
| kashyap | markus_z: Utopic: almost one hour each morning should be a bug triage hour :-) | 10:57 |
| sdague | gmann: sure, but locked_by seems like a thing that you are only going to look for in very specific instances | 10:57 |
| markus_z | johnthetubaguy: Tuesday, August 4th is fine for me, sure | 10:58 |
| sdague | gmann: is there an example of what a real request looks like here where something is LOCKED? | 10:58 |
| johnthetubaguy | markus_z: ah, so thats really a monday in the UK, due to a holiday, maybe the wednesday is better? that still OK? | 10:58 |
| sdague | https://review.openstack.org/#/c/202391/12/doc/v3/api_samples/servers/v2.9/server-get-resp.json,cm doesn't seem to be one | 10:58 |
| markus_z | kashyap: Jepp, that's my morning routine :) Did miss it during the midcycle meetup | 10:59 |
| sdague | that's the null case | 10:59 |
| *** mpaolino has quit IRC | 10:59 | |
| openstackgerrit | Radomir Dopieralski proposed openstack/nova: Handle SSL termination proxies for version list https://review.openstack.org/206479 | 10:59 |
| gmann | sdague: let me check | 10:59 |
| sdague | like what status is expected? | 10:59 |
| markus_z | johnthetubaguy: Still fine, a few meetings but feasible. | 11:00 |
| *** mpaolino has joined #openstack-nova | 11:00 | |
| *** Administrator_ has quit IRC | 11:00 | |
| sdague | why did people know a think was locked in the first place | 11:00 |
| *** Kevin_Zheng has joined #openstack-nova | 11:00 | |
| *** salv-orlando has joined #openstack-nova | 11:00 | |
| *** salv-orl_ has joined #openstack-nova | 11:01 | |
| *** romainh has quit IRC | 11:01 | |
| *** vladikr has quit IRC | 11:02 | |
| markus_z | johnthetubaguy: Would be great if we could push down the number of new bugs at that day. | 11:02 |
| *** Administrator__ has quit IRC | 11:02 | |
| johnthetubaguy | markus_z: yeah, there is also a focus to try and stop release critical ones early | 11:02 |
| *** Administrator_ has joined #openstack-nova | 11:02 | |
| gmann | sdague: from tests actually i added for null only, which can be with locked server | 11:02 |
| *** mpaolino has quit IRC | 11:03 | |
| *** annashen has joined #openstack-nova | 11:03 | |
| gmann | sdague: its just here - https://review.openstack.org/#/c/202391/12/nova/api/openstack/rest_api_version_history.rst | 11:03 |
| markus_z | johnthetubaguy: Do you mean to determine faster if a bug is critical? | 11:03 |
| gmann | sdague: functional tests has only null case but unit tests has other combination of admin, owner etc | 11:04 |
| markus_z | johnthetubaguy: If yes, what criteria did we use in the past? | 11:04 |
| johnthetubaguy | markus_z: just make sure we spot the critical ones, yeah | 11:04 |
| johnthetubaguy | markus_z: really release blockers are usually "nasty" regressions, or any critical bug | 11:04 |
| *** salv-orlando has quit IRC | 11:05 | |
| *** ishant__ has quit IRC | 11:05 | |
| sdague | gmann: so, before this call, how did you know a thing was locked? | 11:05 |
| gmann | sdague: which are just inherited from DB https://github.com/openstack/nova/blob/master/nova/db/sqlalchemy/models.py#L293 | 11:05 |
| gmann | sdague: before this there was way to know | 11:05 |
| gmann | sdague: till user gets error msg saying server is locked | 11:05 |
| johnthetubaguy | markus_z: so I wonder if we want to have a bug triage day on 5th, then a review day on 12th, which leaves the Feature Proposal Freeze stuff on 18th, 1st being liberty-3? | 11:06 |
| gmann | * there was no | 11:06 |
| *** Kevin_Zheng has quit IRC | 11:06 | |
| gmann | sdague: from server info there was no way. | 11:06 |
| *** alex_klimov has joined #openstack-nova | 11:06 | |
| sdague | ok, I don't know if you have the tolerance to do this again, but it really seems like a lock is a kind of resource | 11:06 |
| sdague | so instead of an attribute on server | 11:06 |
| sdague | I'd have added servers/{ID}/lock | 11:07 |
| *** Administrator_ has quit IRC | 11:07 | |
| sdague | and fetching that tells you about that lock | 11:07 |
| *** annashen has quit IRC | 11:07 | |
| johnthetubaguy | markus_z: does that look good? https://wiki.openstack.org/wiki/Nova/Liberty_Release_Schedule#Special_review_days | 11:08 |
| markus_z | johnthetubaguy: Yes, sounds reasonable. I'll advertise the bug triage day via ML. | 11:08 |
| *** markvoelker has joined #openstack-nova | 11:08 | |
| johnthetubaguy | markus_z: I know if we do these things two often, people don't show up, but I think those three days should work? | 11:08 |
| johnthetubaguy | markus_z: cool, did you see our old guide here: https://wiki.openstack.org/wiki/BugTriage and here: https://wiki.openstack.org/wiki/Nova/BugTriage | 11:08 |
| johnthetubaguy | markus_z: the only thing Nova has extra is that list of bug tags, so the bug tag owners can help set the priority and get attention onto specific bugs, to help scale out the whole process, it would be good to get that happening again (I know some folks are still doing that) | 11:09 |
| markus_z | johnthetubaguy: Yes, I've seen and read these guides. I did put it together to the developer guide https://review.openstack.org/#/c/192232/ | 11:09 |
| gmann | sdague: yea that way also we can do. But i feel that makes little bit extra effort from user side to know lock information by separate call? | 11:09 |
| sdague | gmann: sure, but it's an exceptional event | 11:10 |
| *** humble__ has joined #openstack-nova | 11:10 | |
| sdague | 99.99999... a lot of nines, % of API calls don't need that info | 11:10 |
| johnthetubaguy | markus_z: awesome, that sounds good | 11:11 |
| markus_z | johnthetubaguy: I'll talk to the tag owners which are in the current wiki to see if they need anything for that. | 11:11 |
| *** shuquan has joined #openstack-nova | 11:11 | |
| gmann | sdague: hummm. yea that is exceptional event | 11:11 |
| *** dheeraj-gupta-4 has joined #openstack-nova | 11:12 | |
| *** markvoelker has quit IRC | 11:13 | |
| *** humble_ has quit IRC | 11:14 | |
| gmann | sdague: So shall i go ahead and update spec for that :)- https://github.com/openstack/nova-specs/blob/master/specs/liberty/approved/add-locking-information-in-server-get-response.rst | 11:15 |
| sdague | gmann: so that would throw your whole approach up in the air, but I don't really like the idea of adding a whole new param that is almost always going to be null | 11:15 |
| sdague | johnthetubaguy: do you have a couple of minutes to consider this API thing? | 11:16 |
| johnthetubaguy | sdague: API is important, so yes | 11:16 |
| sdague | gmann: let's at least get an ack from johnthetubaguy, I'd hate to send you down another path that ends up being a dead end | 11:16 |
| gmann | sdague: yea, i agree on that | 11:16 |
| johnthetubaguy | sdague: gmann: feels like that should only be in the detail response, is that the debate here? | 11:17 |
| sdague | johnthetubaguy: no, not really | 11:17 |
| sdague | well, sort of | 11:17 |
| *** otter768 has joined #openstack-nova | 11:18 | |
| sdague | the point is the detailed version of the server shows up a few places, like during rebuild. | 11:18 |
| johnthetubaguy | oh | 11:18 |
| sdague | however, more importantly, adding an attr to detailed response that is almost always set to 'null' | 11:18 |
| sdague | seems really weird | 11:18 |
| sdague | and it seems like the wrong pattern | 11:19 |
| gmann | sdague: so how about your first approach of not adding this if it is null | 11:19 |
| johnthetubaguy | so in my head I had detail showing everything, not that might be wrong | 11:19 |
| gmann | sdague: johnthetubaguy : like "progress" information in server | 11:19 |
| johnthetubaguy | rebuild returning a detail response seems dumb I guess, maybe we should change those ones, not sure | 11:20 |
| sdague | yeh, I don't know why that's the case, but honestly, it's probably fine | 11:20 |
| sdague | however, part of me thinks that a "lock" is really a sub resource | 11:20 |
| johnthetubaguy | sdague: yeah, that rebuild thing is probably not worth the cost of fixing it, but yeah | 11:21 |
| sdague | and that if you want the detail about a lock, you should ask for it, vs. it just being on the server representation | 11:21 |
| johnthetubaguy | sdague: hmm, maybe, so I have a skewed perspective here, I should share that | 11:21 |
| sdague | GET /servers/{ID}/lock | 11:21 |
| *** marcusvrn has joined #openstack-nova | 11:21 | |
| johnthetubaguy | sdague: so I keep hearing feedback from the folks that work on UI things, and they really hate multiple API calls, and hate proxies and caches to work around that | 11:22 |
| markus_z | kashyap: Would you consider this a bug? https://bugs.launchpad.net/nova/+bug/1478546 | 11:22 |
| openstack | Launchpad bug 1478546 in OpenStack Compute (nova) "nova GMR doesn't provide option to specify log_dir path" [Undecided,New] - Assigned to Arun Mani (arun-mani) | 11:22 |
| *** otter768 has quit IRC | 11:22 | |
| johnthetubaguy | sdague: now thats not a good reason on its own, just I think thats want makes be lean towards everything | 11:22 |
| kashyap | markus_z: Back in a few min, on phone. | 11:23 |
| sdague | johnthetubaguy: so instead we have a server representation which grows and has everything anyone could possibly want to know? | 11:23 |
| johnthetubaguy | sdague: I get your point though, if we include everything, the detail list becomes next to useless as it gets so big, which is also bad | 11:23 |
| sdague | right | 11:23 |
| sdague | plus, it becomes slow, because it's constantly joining everything | 11:24 |
| sdague | even when you don't need it | 11:24 |
| johnthetubaguy | sdague: agreed | 11:24 |
| *** shuquan has quit IRC | 11:24 | |
| * johnthetubaguy thinks about an include=lock,progress,foobar query string, then hates himself a little for thinking about that | 11:25 | |
| johnthetubaguy | gmann: do we have this info available in the API at all yet? | 11:26 |
| johnthetubaguy | gmann: I guess right now its just in the DB? | 11:26 |
| sdague | johnthetubaguy: so right now it's only exposed via exception | 11:26 |
| gmann | johnthetubaguy: humm, even i was thinking same adding query or something | 11:26 |
| sdague | in @check_instance_lock | 11:26 |
| sdague | decorator | 11:26 |
| gmann | johnthetubaguy: yea, only error msg | 11:26 |
| sdague | oh, except if you are admin. | 11:27 |
| *** humble__ has quit IRC | 11:27 | |
| sdague | then you ignore locks | 11:27 |
| *** Marga_ has joined #openstack-nova | 11:27 | |
| johnthetubaguy | sdague: unless its an admin lock, but yeah | 11:28 |
| johnthetubaguy | gmann: sdague: so I think we all agree its handy if the API tells you about the lock | 11:28 |
| sdague | yeh, I can see that | 11:29 |
| johnthetubaguy | sdague: I guess lock is /instance/uuid/action which kinda means going in detail fits the pattern, but that doesn't make it a good idea either | 11:29 |
| *** thorst has joined #openstack-nova | 11:29 | |
| sdague | so, how do we feel about locked_by being only included in the details if a lock exists | 11:31 |
| *** Murali_ has joined #openstack-nova | 11:31 | |
| *** Marga_ has quit IRC | 11:31 | |
| sdague | I can see from a UI perspective why you'd want to indicate all of this in a list view | 11:31 |
| sdague | and it's on instance | 11:31 |
| johnthetubaguy | sdague: so I have a feeling all API calls should care about the lock info, but they don't care today, because there is no information on that right now, is that fair? | 11:32 |
| *** Marga_ has joined #openstack-nova | 11:32 | |
| sdague | maybe, the entire lock workflow is weird | 11:32 |
| *** Murali has quit IRC | 11:32 | |
| johnthetubaguy | yeah, it is | 11:32 |
| *** Murali_ is now known as Murali | 11:32 | |
| johnthetubaguy | I mean it might get extended a little with trove instances | 11:32 |
| johnthetubaguy | when its locked by a service token on boot | 11:32 |
| johnthetubaguy | not sure that helps things much | 11:33 |
| johnthetubaguy | in the UI I could see people greying out options when the instance is locked, that seems quite important actually | 11:33 |
| sdague | when were server locks introduced? | 11:33 |
| sdague | oh, neat, locked was supposed to be removed in icehouse :) | 11:34 |
| gmann | sdague: is it? | 11:34 |
| sdague | https://review.openstack.org/#/c/38196/ | 11:35 |
| openstackgerrit | Roman Dobosz proposed openstack/nova: New nova API call to mark nova-compute down https://review.openstack.org/184086 | 11:35 |
| openstackgerrit | Roman Dobosz proposed openstack/nova: Service group drivers forced_down flag utilization https://review.openstack.org/197063 | 11:35 |
| sdague | gmann: in favor of locked_by | 11:35 |
| sdague | johnthetubaguy: ok, so the fact that it is locked, feels like that's important top level info for UI | 11:36 |
| gmann | sdague: ahh yea, I saw notes there earlier | 11:36 |
| sdague | however, the details of that lock feels like another API call | 11:36 |
| sdague | especially if this is going to extend over time | 11:36 |
| sdague | because, honestly, it feels like locks should have a timestamp and an optional reason | 11:37 |
| sdague | and I'm sure we'll want to add that over time | 11:37 |
| johnthetubaguy | sdague: ah, so thats a very sound argument, I like that | 11:37 |
| *** coolsvap is now known as coolsvap|away | 11:37 | |
| sdague | which means "locked" should should be exposed in show as a True / False | 11:37 |
| johnthetubaguy | sdague: locked: True/False at the top, and more details in /instance/uuid/lock, I like that distinction | 11:37 |
| sdague | yeh | 11:38 |
| gmann | humm, thats nice idea | 11:38 |
| *** karimb has quit IRC | 11:38 | |
| *** Murali has quit IRC | 11:38 | |
| sdague | because, like you said, trove is totally going to want to extend this later | 11:38 |
| *** lucasagomes is now known as lucas-hungry | 11:38 | |
| johnthetubaguy | gmann: so I wish we had spotted this in the spec review, its a failed spec review really, I quite like what sdague is proposing here | 11:38 |
| johnthetubaguy | sdague: totally | 11:38 |
| gmann | johnthetubaguy: yea, we should have thought about that during spec review | 11:40 |
| gmann | johnthetubaguy: sdague: so let me update the spec and make - locked: True/False in show | 11:40 |
| sdague | gmann: great, and I think it's ok to still have that show up in any detailed list | 11:41 |
| johnthetubaguy | gmann: sdague: how quick do we think we can turn around on the updated approach? I feel like we should give this an exception for a week, if that lets it happen in the face of the process screw up here. does that work? | 11:41 |
| gmann | johnthetubaguy: sdague : and further we can have /instance/uuid/lock may be in M (due to deadlines) :) | 11:41 |
| sdague | like the rebuild case, a detailed server definition is fine | 11:41 |
| *** romainh has joined #openstack-nova | 11:41 | |
| johnthetubaguy | gmann: maybe yeah, I guess, do we think its possible to get this merged next week ish? | 11:42 |
| sdague | gmann: yeh, that should go to M | 11:42 |
| *** karimb has joined #openstack-nova | 11:42 | |
| sdague | johnthetubaguy: honestly, making this patch be that interface is probably an hour of work | 11:42 |
| gmann | sdague: johnthetubaguy : this one i can update quickly | 11:42 |
| gmann | now only | 11:42 |
| gmann | and tomorrow i can put up info in spec | 11:43 |
| johnthetubaguy | sdague: gmann: so I am willing to *try* get a quick turn around on this | 11:43 |
| johnthetubaguy | sdague: gmann: I see we approved this in kilo in a similar form: http://specs.openstack.org/openstack/nova-specs/specs/kilo/approved/get-lock-status-of-instance.html | 11:43 |
| johnthetubaguy | sdague: gmann: for the sake of two days grace, it feels like we should attempt this for liberty? | 11:44 |
| sdague | johnthetubaguy: yes, I think this is a good add | 11:44 |
| gmann | johnthetubaguy: yea. | 11:44 |
| *** mwagner_lap has quit IRC | 11:44 | |
| sdague | I just wanted to get ahead of an API change we might regret and have to unwind. | 11:44 |
| *** dbite has joined #openstack-nova | 11:45 | |
| *** dguitarbite has quit IRC | 11:45 | |
| johnthetubaguy | sdague: totally | 11:45 |
| *** bkopilov has quit IRC | 11:45 | |
| johnthetubaguy | gmann: sdague: well lets try it, it might not make it, but it seems very, very close, and useful, so worth a try | 11:46 |
| openstackgerrit | Shiina, Hironori proposed openstack/nova: Add a Nova API to inject an NMI https://review.openstack.org/202617 | 11:46 |
| *** jasondotstar|afk has quit IRC | 11:46 | |
| kashyap | markus_z: Back, now. Sorry. Yeah, looking at that bug -- well, it looks more like an RFE. | 11:46 |
| gmann | johnthetubaguy: sdague : sure. Thanks | 11:46 |
| *** VW_ has joined #openstack-nova | 11:47 | |
| kashyap | markus_z: One can trivially redirect from stdrr w/ stream manipulation using something like "|&" | 11:47 |
| sdague | yeh, it feels like we should give this one a bit of a break, because we're holding it up a little for a better interface. And having to decide useful feature with worse interface vs. better interface but delayed by 6 months is going to make people pick bad interfaces in review. | 11:47 |
| johnthetubaguy | sdague: +1 | 11:48 |
| *** karimb has quit IRC | 11:48 | |
| *** dave-mcc_ has quit IRC | 11:49 | |
| johnthetubaguy | sdague: we have extra process here to stop mess ups, we spotted one, so lets cut this one some slack to compensate | 11:49 |
| * johnthetubaguy needs to go have some food :) | 11:49 | |
| johnthetubaguy | gmann: sdague: so one thought, as I run off, what if an "admin" does a detail on the server, its not locked from them, so what should they see for locked? True or False? | 11:51 |
| garyk | daily review request for https://review.openstack.org/#/c/188789 …. | 11:51 |
| *** jwcroppe has joined #openstack-nova | 11:51 | |
| gmann | johnthetubaguy: sdague : IMO True and further admin can override the lock which was locked by owner | 11:52 |
| sdague | gmann: yeh, all the locked semantics are an undocumented mess :) | 11:53 |
| sdague | but we can sort that out over time | 11:53 |
| *** rotbeard has joined #openstack-nova | 11:53 | |
| gmann | sdague: yea.. | 11:54 |
| *** jordanP has joined #openstack-nova | 11:55 | |
| *** Marga_ has quit IRC | 11:55 | |
| jordanP | sahid, hi. Could you (re)review https://review.openstack.org/#/c/192736/ (libvirt:when deleting snapshot use qemu-img to blockRebase if domain is stopped) please ? I have just rebased it. | 11:55 |
| *** Murali has joined #openstack-nova | 11:56 | |
| *** BobBall_AWOL is now known as BobBall | 11:56 | |
| sahid | danpb: perhaps you may want to look at it ^ | 11:57 |
| sahid | jordanP: yes i will do | 11:57 |
| jordanP | sahid, thanks | 11:57 |
| *** Marga_ has joined #openstack-nova | 11:58 | |
| *** hshiina has joined #openstack-nova | 12:02 | |
| *** annashen has joined #openstack-nova | 12:04 | |
| kashyap | jordanP: If you're respinning, please wrap the commit messages at 72 chars; if not, disregard me. | 12:06 |
| jordanP | kashyap, I'll keep that in mind. | 12:07 |
| kashyap | It's actually noted here (but who reads docs :-)) -- https://wiki.openstack.org/wiki/GitCommitMessages#Summary_of_Git_commit_message_structure | 12:08 |
| *** markvoelker has joined #openstack-nova | 12:09 | |
| *** annashen has quit IRC | 12:09 | |
| *** karimb has joined #openstack-nova | 12:09 | |
| *** kodoku has joined #openstack-nova | 12:10 | |
| *** dbite has quit IRC | 12:12 | |
| *** raildo has joined #openstack-nova | 12:13 | |
| *** markvoelker has quit IRC | 12:14 | |
| gabriel-bezerra | mriedem_away: Hi. Thanks for pointing that. | 12:16 |
| gabriel-bezerra | Is gate-nova-python27 still broken? | 12:16 |
| markus_z | kashyap: OK, cool, I was a bit unsure but I think you are right. | 12:18 |
| *** baoli has joined #openstack-nova | 12:18 | |
| *** jwcroppe has quit IRC | 12:19 | |
| *** doude has joined #openstack-nova | 12:19 | |
| *** jwcroppe has joined #openstack-nova | 12:20 | |
| kodoku | Hi, when I use nova image-list, nova use glance python client ? | 12:20 |
| *** kylek3h has joined #openstack-nova | 12:22 | |
| *** josecastroleon has joined #openstack-nova | 12:22 | |
| *** baoli has quit IRC | 12:23 | |
| *** baoli has joined #openstack-nova | 12:24 | |
| *** kylek3h has quit IRC | 12:26 | |
| dims | @pkholkin : for the oslo.cache patch, we'll need a way to check if memcached is configured and fail if it isn't (see servicegroup stuff), that's one use case. Another is if memcached is not configured then use in-memory. third is straight in-memory. | 12:26 |
| lxsli | kodoku: the python-glance-client library, yes I think so, why? | 12:26 |
| lxsli | oops, python-glanceclient | 12:27 |
| *** mestery has joined #openstack-nova | 12:27 | |
| *** aysyd has joined #openstack-nova | 12:27 | |
| *** scheuran has joined #openstack-nova | 12:27 | |
| kodoku | lxsli because I have this issue with nova image-list ==> https://bugs.launchpad.net/python-glanceclient/+bug/1442664 | 12:28 |
| openstack | Launchpad bug 1442664 in python-glanceclient "Since 0.16.1 client breaks nova when using https" [Critical,Fix released] - Assigned to Stuart McLaren (stuart-mclaren) | 12:28 |
| *** eharney has joined #openstack-nova | 12:28 | |
| *** markvoelker has joined #openstack-nova | 12:29 | |
| markus_z | kodoku: The project doesn't specify it in it's requirements: https://git.openstack.org/cgit/openstack/python-novaclient/tree/requirements.txt | 12:29 |
| *** mwagner_lap has joined #openstack-nova | 12:29 | |
| kodoku | markus_z ok but I think nova use glance python library | 12:31 |
| markus_z | kodoku: Yes, I would say so: https://git.openstack.org/cgit/openstack/nova/tree/requirements.txt#n31 | 12:32 |
| kodoku | hum 0.18.0 | 12:32 |
| kodoku | ok I test kilo with 0.17.2 | 12:32 |
| *** ip0353_ has joined #openstack-nova | 12:32 | |
| markus_z | "nova image-list" is made available by the "python-novaclient". | 12:32 |
| ip0353_ | Hi, has anyone seen this issue before with instances on the same compute being unable to connect to each other's ports? https://bugs.launchpad.net/nova/+bug/1478925 | 12:33 |
| openstack | Launchpad bug 1478925 in OpenStack Compute (nova) "Instances on the same compute node unable to connect to each other's ports" [Undecided,New] | 12:33 |
| kodoku | markus_z When I launch VM with nova I have this issue "Caught error: 'PoolManager' object has no attribute 'connection_kw'" Same issue with nova image-list | 12:33 |
| *** edmondsw has joined #openstack-nova | 12:33 | |
| *** slong_ has joined #openstack-nova | 12:34 | |
| *** dheeraj-gupta-4 has quit IRC | 12:34 | |
| *** slong__ has joined #openstack-nova | 12:34 | |
| *** lucas-hungry is now known as lucasagomes | 12:35 | |
| *** ijuwang has joined #openstack-nova | 12:35 | |
| *** slong has quit IRC | 12:36 | |
| *** rfolco has joined #openstack-nova | 12:37 | |
| markus_z | kodoku: Could you double-check it with version 0.18 please? | 12:37 |
| *** jasondotstar|afk has joined #openstack-nova | 12:37 | |
| *** slong_ has quit IRC | 12:38 | |
| pkholkin | @dims: Yes. Folks, what do you think about dims question in case of our patch https://review.openstack.org/#/c/203049/4 ? | 12:39 |
| *** karimb has quit IRC | 12:39 | |
| *** rlrossit has joined #openstack-nova | 12:39 | |
| dims | pkholkin: i believe the servicegroup stuff will not work as the design is based on shared memcached data | 12:39 |
| *** mdrabe has joined #openstack-nova | 12:40 | |
| markus_z | kodoku: The release letter of python-glanceclient says that version 0.18 contains the fix: http://docs.openstack.org/developer/python-glanceclient/#id7 | 12:40 |
| johnthetubaguy | ip0353_: honestly, its quite hard to get all the networking config correct, I think asking in #openstack might find someone who knows how to debug neutron config is likely to be the best bet, unless I am missing something | 12:40 |
| ZZelle | garyk, hi | 12:41 |
| *** ip0353__ has joined #openstack-nova | 12:41 | |
| *** VW_ has quit IRC | 12:41 | |
| *** VW_ has joined #openstack-nova | 12:41 | |
| *** zhenguo has quit IRC | 12:42 | |
| *** artom has joined #openstack-nova | 12:42 | |
| *** vladikr has joined #openstack-nova | 12:42 | |
| garyk | ZZelle: hi, whats up | 12:43 |
| ip0353_ | @johnthetubaguy Thanks, I'll try asking there. I am not completely sure whether this is a config issue. I seem to have exactly one compute node on which this works and after comparing all the configs, they all are the same | 12:43 |
| ZZelle | garyk, one question about https://review.openstack.org/#/c/58883 | 12:43 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: switch to oslo.cache lib https://review.openstack.org/203049 | 12:44 |
| garyk | ZZelle: sure. how can i help | 12:44 |
| *** dave-mccowan has joined #openstack-nova | 12:44 | |
| pkholkin | dims: but previously we wanted to remove common/memorycache.py and use our in-memory DictBackend | 12:45 |
| ZZelle | garyk, does nova handle correctly glance vsphere backend? (ie: avoid to copy glance image file from glance over http) | 12:45 |
| ZZelle | and copy it between datastores | 12:45 |
| *** coolsvap|away is now known as coolsvap | 12:45 | |
| *** ip0353 has joined #openstack-nova | 12:45 | |
| garyk | ZZelle: it should unless those patches are still in review in nova (which they very well may be). give me a few minutes to check | 12:46 |
| *** hshiina has quit IRC | 12:46 | |
| ZZelle | garyk, ok | 12:46 |
| *** VW_ has quit IRC | 12:46 | |
| *** ip0353__ has quit IRC | 12:47 | |
| *** ip0353_ has quit IRC | 12:47 | |
| johnthetubaguy | garyk: are multiple swap disks possible in BDMs? thinking about this one: https://review.openstack.org/#/c/142174 | 12:47 |
| *** eglynn has quit IRC | 12:48 | |
| kodoku | markus_z I upgrade but look my trace http://paste.openstack.org/show/406059/ | 12:48 |
| danpb | johnthetubaguy: yep, pretty sure we allow multiple swaps | 12:48 |
| kodoku | markus_z Why nova use http.py library and not https.py ????? My endpoint is full https :/ | 12:49 |
| danpb | there's no conceptua reason why we'd limit it | 12:49 |
| garyk | johnthetubaguy: the code for the parsing of the bdms implies that it is. | 12:49 |
| *** jasondotstar|afk is now known as jasondotstar | 12:49 | |
| garyk | ZZelle: the idea is that the glance images are kept on a datastore and when importing the image there is a direct copy from one datastore to another. | 12:50 |
| ip0353 | @johnthetubaguy it's either a very hidden config file that I don't know of, or maybe a bug? I don't know, I've been stuck on trying to figure out what's wrong with this for a couple of days | 12:50 |
| johnthetubaguy | garyk: danpb: so I wondered about that, but I am seeing the opposite here: https://github.com/openstack/nova/blob/master/nova/virt/driver.py#L115 | 12:50 |
| *** karimb has joined #openstack-nova | 12:50 | |
| *** changbl has joined #openstack-nova | 12:50 | |
| *** y_sawai has joined #openstack-nova | 12:50 | |
| ZZelle | garyk, that's what i understood but i assume nova icehouse would support such behavior because it's implemented on glance side | 12:51 |
| garyk | ZZelle: i think that code may have only landed in nova in k. i am still checking | 12:51 |
| danpb | johnthetubaguy: hmm, wouldn't be surprised if we allow it in the BDM and then don't support it properly in drivers :-) | 12:51 |
| ZZelle | garyk, but it's not working and according to nova code i am not sure i was merged in k. | 12:51 |
| johnthetubaguy | ip0353: so this is really for nova development topics, where as this seems like a neutron usage issue, so I think #openstack is a better place to get an answer quickly, I don't know the details of neutron configuration, but folks on those other channels should be able to help you better | 12:51 |
| garyk | johnthetubaguy: i think that this is my bad. it can only support one and not multiple | 12:51 |
| danpb | ndipanov: is probably person who knows for sure | 12:52 |
| *** karmatronic has joined #openstack-nova | 12:52 | |
| johnthetubaguy | danpb: yeah, I was just thinking the same thing, ndipanov swaps, do we allow multiples or maybe just not implement multiples? | 12:52 |
| garyk | ZZelle: the different projects work at their own pace…. | 12:52 |
| garyk | ZZelle: for example the neutron plugin - has nsxv support in K and the nova code has yet to land in l :( | 12:53 |
| *** mpaolino has joined #openstack-nova | 12:53 | |
| markus_z | kodoku: Do you use the Juno release? | 12:53 |
| ZZelle | garyk, :/ | 12:54 |
| garyk | so whoever wants to use this stuff in k is between a rock and a nhard place | 12:54 |
| ip0353 | johnthetubaguy thanks a lot! | 12:54 |
| *** y_sawai has quit IRC | 12:54 | |
| *** karimb has quit IRC | 12:54 | |
| *** moshele has quit IRC | 12:55 | |
| garyk | ZZelle: give me a few minutes to check with the latest nova code. i just need to configure my devstack env (by default i use the filesystem for glance) | 12:55 |
| johnthetubaguy | ip0353: ah, looks like the folks on #openstack have done their support magic, all good, security groups could well be the answer | 12:55 |
| *** changbl has quit IRC | 12:55 | |
| kodoku | markus_z no kilo | 12:56 |
| ndipanov | johnthetubaguy, we don't and it would raise an error in the aPI | 12:56 |
| kodoku | markus_z I try to upgrade 0.17.2 to 0.18 | 12:56 |
| *** clduser has joined #openstack-nova | 12:56 | |
| *** rushiagr is now known as rushiagr_away | 12:56 | |
| johnthetubaguy | ndipanov: ah, cool, thats consistent then, sweet. | 12:57 |
| garyk | johnthetubaguy: i need to change my score. only issue is an indentation - but that is not worth blocking something as it passes pep8 | 12:57 |
| ndipanov | johnthetubaguy, https://github.com/openstack/nova/blob/master/nova/compute/api.py#L1280-L1282 | 12:57 |
| danpb | hmm, mriedem_away's massive patch series is going to conflict nicely with os-brick | 12:58 |
| *** moshele has joined #openstack-nova | 12:58 | |
| * danpb wonders whether to just approve them both and see which one wins | 12:58 | |
| garyk | danpb: gate is down so it will be a race in anycase … | 12:58 |
| johnthetubaguy | ndipanov: cool, so I was being lazy, thats actually quite obvious in the end, I only got this far: https://github.com/openstack/nova/blob/master/nova/virt/driver.py#L115 | 12:59 |
| kodoku | markus_z Ok new error but maybe nova 2.23.0 no works with 0.18.0 | 12:59 |
| garyk | i think that the brick stuff should land first as moving all of that stuff around may break things with the merge, but that is my two cents | 12:59 |
| ndipanov | johnthetubaguy, np that's what IRC is for :) | 12:59 |
| johnthetubaguy | danpb: I would be +1 for brick at this point | 12:59 |
| johnthetubaguy | ndipanov: very true | 12:59 |
| danpb | nb, mriedem_away's stuff is already pending with +A | 12:59 |
| johnthetubaguy | danpb: ah, gotcha | 13:00 |
| danpb | so its really whether to also approve brick at the same time | 13:00 |
| johnthetubaguy | looking foward to the next version of gerrit that tells us about those clashes, at least I think it did | 13:00 |
| johnthetubaguy | danpb: I am tempted to say approve it, so the trivial re-approve process can save it, but maybe thats just dumb? | 13:01 |
| *** gszasz has quit IRC | 13:02 | |
| *** afazekas has quit IRC | 13:02 | |
| johnthetubaguy | I guess rebasing on top is the best way? | 13:02 |
| markus_z | kodoku: The version number 2.23.0 is used in the python-novaclient. The python-novaclient *doesn't* use the python-glanceclient. This should be listed in the requirements.txt | 13:03 |
| kodoku | kragniz ok I found my issue. It's nova client. Wrong usage with endpoint | 13:03 |
| *** moshele has quit IRC | 13:03 | |
| garyk | johnthetubaguy: danpb: can you guys please look at https://review.openstack.org/187942 it was rebased. needless to say it is based on https://review.openstack.org/#/c/188789 (which i hope may land in the next few weeks) | 13:04 |
| *** Kevin_Zheng has joined #openstack-nova | 13:04 | |
| kodoku | my endpoint is https://proxy/compute for computeV2 | 13:04 |
| kodoku | kragniz with nova --debug I see "GET /compute/v2/d6f612ae52a9411f86ab5e2a79b4fd7e/images/detail HTTP/1.1" 500 128 | 13:04 |
| kragniz | kodoku: cool | 13:04 |
| johnthetubaguy | garyk: did I change much since I did +A? I guess it was all rebase stuff? | 13:05 |
| *** annashen has joined #openstack-nova | 13:05 | |
| kodoku | But in log nova-api.log I see http://proxy/v2/d6f612ae52a9411f86ab5e2a79b4fd7e/images/detail | 13:05 |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Show 'locked_by' information in server details https://review.openstack.org/202391 | 13:05 |
| *** Administrator_ has joined #openstack-nova | 13:05 | |
| kodoku | But it's wrong | 13:05 |
| kodoku | I need to see https://proxy/compute/v2 | 13:05 |
| *** gszasz has joined #openstack-nova | 13:06 | |
| kodoku | you see markus_z ? sorry i quote kragniz and not you ^^ | 13:07 |
| garyk | johnthetubaguy: the code did not change - there was just a conflict with the unit tests - that is the console patch landed and that added a new test and git dit not like that - kind of adding new tests at the end of files is a recipe for conflicts | 13:07 |
| johnthetubaguy | garyk: I figured, looked like the adds and removes were the same number | 13:07 |
| gmann | sdague: johnthetubaguy : updated locked info one. please have look- https://review.openstack.org/#/c/202391/13 | 13:07 |
| gmann | If any comments, I will check tomorrow. see ya | 13:08 |
| markus_z | kodoku: Good to hear that this is solved for you. Would be great if you could add your observation to the bug report for later references and other contributors. | 13:08 |
| *** amotoki has joined #openstack-nova | 13:08 | |
| kodoku | markus_z I don't know HOW nova make his URL for contact API | 13:08 |
| *** Kevin_Zheng has quit IRC | 13:08 | |
| *** moshele has joined #openstack-nova | 13:08 | |
| sdague | gmann: yeh, I like that much better | 13:09 |
| johnthetubaguy | garyk: follow question on the neutron stuff I was reviewing the other day | 13:09 |
| sdague | johnthetubaguy do you need a respin of the spec based on that? | 13:09 |
| garyk | johnthetubaguy: all ears | 13:09 |
| johnthetubaguy | garyk: what happens to the instance networking if a call up update the vnic index failes when its required | 13:09 |
| *** Kevin_Zheng has joined #openstack-nova | 13:09 | |
| *** annashen has quit IRC | 13:09 | |
| garyk | johnthetubaguy: to the instance nothing. | 13:10 |
| johnthetubaguy | sdague: we probably should do that, give its the API, but be sure we review that | 13:10 |
| johnthetubaguy | garyk: I guess if it fails you get no networking setup though? | 13:10 |
| sdague | zomg, getting graphviz flow charts to work, is a bit interesting | 13:10 |
| garyk | if for some unlikely reason the vnic index was not updated in the VC for the security groups the instance will not be able to get a DHCP address if there is no vmware tools installed | 13:10 |
| garyk | johnthetubaguy: in addition to this port security may not work | 13:11 |
| *** bfic has quit IRC | 13:11 | |
| johnthetubaguy | garyk: no port security sounds bad... | 13:11 |
| kodoku | markus_z Can you tell me why that ?? http://paste.openstack.org/show/406065/ | 13:11 |
| *** Administrator_ has quit IRC | 13:11 | |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Added server tags support in nova-api https://review.openstack.org/128940 | 13:11 |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Fixed incorrect behaviour of method _check_instance_exists https://review.openstack.org/203037 | 13:11 |
| johnthetubaguy | garyk: I guess whenever we see the extension, we really need to call it? or is that not true either? | 13:11 |
| *** xyang1 has joined #openstack-nova | 13:12 | |
| *** Administrator_ has joined #openstack-nova | 13:12 | |
| garyk | johnthetubaguy: i am not sure that i understand? | 13:12 |
| garyk | basically this is only something that is supported by the NSXv plugin. All other NSX plugins do not require this | 13:13 |
| markus_z | kodoku: I'm sorry, but I know next to nothing about the API. | 13:13 |
| johnthetubaguy | garyk: so I am thinking about this bit: https://review.openstack.org/#/c/147126/20/nova/network/neutronv2/api.py,cm | 13:13 |
| *** nelsnelson has joined #openstack-nova | 13:13 | |
| *** kodoku_ has joined #openstack-nova | 13:13 | |
| *** nelsnelson has quit IRC | 13:13 | |
| *** tpeoples has joined #openstack-nova | 13:13 | |
| johnthetubaguy | garyk: what should we do when the exception is raised, right now we silently ignore it, which feels bad if that means we might fail to add security groups | 13:13 |
| *** nelsnelson has joined #openstack-nova | 13:14 | |
| garyk | johnthetubaguy: no, the security groups are configured | 13:14 |
| *** VW_ has joined #openstack-nova | 13:14 | |
| garyk | the problem is that the spoofguard policy for the VM will not be updated. | 13:14 |
| garyk | that means that the VM will not get a DHCP address | 13:14 |
| *** edmondsw has quit IRC | 13:14 | |
| garyk | there are 2 options here: | 13:14 |
| garyk | 1. fail the instance boot | 13:14 |
| *** Kevin_Zheng has quit IRC | 13:14 | |
| garyk | 2. let the instance boot and troubleshoot | 13:15 |
| garyk | i prefer the latter | 13:15 |
| *** Kevin_Zheng has joined #openstack-nova | 13:15 | |
| johnthetubaguy | garyk: so its not quite as bad as 1, as I think it should be able to retry the build (although that might not work right now, due to other bugs) | 13:15 |
| *** kodoku has quit IRC | 13:15 | |
| *** gzhai has quit IRC | 13:15 | |
| *** rdo has quit IRC | 13:15 | |
| *** xyang has joined #openstack-nova | 13:16 | |
| johnthetubaguy | garyk: I think we generally go for (1) though, report the error, rather than hide it in the logs silently | 13:16 |
| johnthetubaguy | garyk: although I would be the first to admit we are far from consistent on these things | 13:16 |
| garyk | why do you think that? | 13:16 |
| *** zzzeek has joined #openstack-nova | 13:17 | |
| garyk | in the VC case there may be only one host - that is a cluster | 13:17 |
| garyk | so the instance will not be booted and there will be a failed boot. this could even be due to a problem with the DB that is spurious | 13:17 |
| johnthetubaguy | garyk: ah, so I totally forgot that doesn't work for you folks :( | 13:17 |
| *** cpschult has joined #openstack-nova | 13:17 | |
| garyk | an admin that has proper tools to troubleshoot should easily be able to see what has happened here and rectify. | 13:17 |
| *** xyang1 has quit IRC | 13:17 | |
| *** Administrator_ has quit IRC | 13:18 | |
| gabriel-bezerra | Good morning. Is gate-nova-python27 still broken? | 13:18 |
| *** Administrator_ has joined #openstack-nova | 13:18 | |
| edleafe | lxsli: saw your comment on https://review.openstack.org/#/c/206249. Sorry for not giving you co-author, but the patchset I took it from only had jaypipes' name on it | 13:18 |
| *** otter768 has joined #openstack-nova | 13:19 | |
| openstackgerrit | Jorge Niedbalski R. proposed openstack/nova: Detach iSCSI latest path for latest disk https://review.openstack.org/135382 | 13:19 |
| *** browne has joined #openstack-nova | 13:19 | |
| *** Kevin_Zheng has quit IRC | 13:20 | |
| *** READ10 has joined #openstack-nova | 13:20 | |
| rlrossit | dansmith: https://review.openstack.org/#/c/190145/14/nova/tests/unit/objects/test_objects.py,cm I'm having trouble keeping up... I need my unit tests to catch everyone on this instead of hoping I stumble upon them. | 13:20 |
| *** Kevin_Zheng has joined #openstack-nova | 13:20 | |
| johnthetubaguy | garyk: but failing and making the user retry (or fixing nova's regular build retry mechanism for vmware could avoid forcing the user do that) | 13:20 |
| *** rushiagr_away is now known as rushiagr | 13:21 | |
| lxsli | morning edleafe | 13:21 |
| johnthetubaguy | garyk: so I am looking in the neutron file, and we are far from consistent, so I guess its wrong to block this approach | 13:21 |
| lxsli | yeah I should possibly be co-author on that one by now but I'm not too bothered | 13:21 |
| edleafe | lxsli: just sayin' that if I had seen your name, I would have given you credit. | 13:22 |
| lxsli | edleafe: cool cool :) | 13:22 |
| *** Administrator__ has joined #openstack-nova | 13:23 | |
| *** Administrator_ has quit IRC | 13:23 | |
| *** otter768 has quit IRC | 13:23 | |
| garyk | johnthetubaguy: correct. there are different flows and different ways in handling it. with our stress testing and validations (which have been done since kilo) this has not been an issue | 13:23 |
| *** nelsnelson has quit IRC | 13:23 | |
| garyk | then again i am not sure that will help in favor or against | 13:24 |
| *** Administrator_ has joined #openstack-nova | 13:24 | |
| *** nelsnelson has joined #openstack-nova | 13:24 | |
| johnthetubaguy | garyk: true, if its a very rare failure, I prefer failing early more, that argues against bothering with a retry, in some ways | 13:25 |
| johnthetubaguy | garyk: anyways, appreciate describing the impact of the exception being hidden, that was really my question here | 13:26 |
| *** Kevin_Zheng has quit IRC | 13:26 | |
| johnthetubaguy | garyk: I know as an operator I prefer the obvious failures, but I don't know if thats true for all operators | 13:27 |
| *** lucas____ has joined #openstack-nova | 13:27 | |
| *** Administrator__ has quit IRC | 13:28 | |
| *** Kevin_Zheng has joined #openstack-nova | 13:28 | |
| *** dguitarbite has joined #openstack-nova | 13:28 | |
| garyk | johnthetubaguy: i am fine to update the code to raise the exception. | 13:28 |
| *** xyang1 has joined #openstack-nova | 13:28 | |
| *** dguitarbite is now known as Guest11859 | 13:28 | |
| garyk | the fact that it did not happen does not mean that it will not happen. | 13:28 |
| garyk | i'll update | 13:28 |
| dansmith | rlrossit: yeah, but thanks for catching things manually in the meantime | 13:29 |
| dansmith | rlrossit: you said something to me yesterday and then went offline before I could answer.. but I don't remember what it was | 13:29 |
| rlrossit | dansmith: it gives me a good spike in reviews :) | 13:29 |
| dansmith | yeah | 13:29 |
| rlrossit | to the chat logs! | 13:29 |
| rlrossit | copy-pasta: dansmith: possibly totally unreasonable test, but if there is any change in an object adding a field, there has to be something in obj_make_compatible to take out that field if it goes to the previous version right? So we should be testing that | 13:30 |
| *** Administrator_ has quit IRC | 13:30 | |
| johnthetubaguy | garyk: OK, thanks | 13:31 |
| dansmith | rlrossit: oh right.. so that would be an awesome test, except that I don't think we know what version an attribute was added in, do we? | 13:31 |
| johnthetubaguy | garyk: sorry to rat hole on that, just trying to understand the code, I think I am happy with the rest of that now, thanks for the updates around the base_api and stuff, that reads better to me now | 13:31 |
| bauzas | dansmith: do you have a few mins for discussing about NovaObjects ? | 13:31 |
| bauzas | dansmith: see you discussing with rlrossit so taking the next item :) | 13:32 |
| garyk | johnthetubaguy: np. i am just happy that someone is looking at this | 13:32 |
| dansmith | bauzas: yeah, in a sec | 13:32 |
| *** dboik has joined #openstack-nova | 13:32 | |
| bauzas | dansmith: np | 13:32 |
| *** cfriesen has joined #openstack-nova | 13:32 | |
| openstackgerrit | Zhai, Edwin proposed openstack/nova: Fix notify_decorator errors https://review.openstack.org/203573 | 13:32 |
| dansmith | rlrossit: so what's the status of your patch now? I saw you talking with mriedem about getting stable properly running the other stuff | 13:32 |
| rlrossit | dansmith: so I was thinking we could do it by taking current version, generating a primitive, backport that to current-1, and then assert it doesn't have keys that it doesn't have fields for | 13:32 |
| rlrossit | dansmith: none of it got solved, still infinite recursion :( | 13:33 |
| johnthetubaguy | garyk: did we get anyone from the security team to review the metadata service proxy at all, I guess I am wondering if the looked at this when it merged in neutron? | 13:34 |
| *** Kevin_Zheng has quit IRC | 13:34 | |
| johnthetubaguy | garyk: its just crypto is hard, and I really don't feel qualified when I am looking at that change, and thats a crap answer, but its the truth | 13:34 |
| dansmith | rlrossit: but the thing you're testing is the thing that would remove the keys/fields | 13:35 |
| dansmith | rlrossit: we don't actually store programatic history of the objects.. i.e. we can't randomly get version 1.23 if our code to generate 1.23 is wrong | 13:35 |
| johnthetubaguy | garyk: so I have a feeling I was miss-reading this, there is not as much new here as I was first thinking, let me get back to you on that | 13:35 |
| *** gzhai has joined #openstack-nova | 13:36 | |
| garyk | johnthetubaguy: not sure that i understand? why do we need someone from the security team to review traffic being redirected by a load balancer? | 13:36 |
| garyk | johnthetubaguy: i am not sure what crypto part you are referring to. the neutron secret? that is what exists in the neutron code today (and it is pretty easy t break) | 13:36 |
| garyk | we are just supporting the same interfaces | 13:37 |
| rlrossit | dansmith: but if we knew 1.23 was good, and we made 1.24 where we added the field 'foo', couldn't we try going from 1.24 to 1.23 and assert there isn't a 'foo' in the primitive after the backport? | 13:37 |
| johnthetubaguy | garyk: it felt like a totally new system, but its more that I didn't see the shared code, I am looking more closely now | 13:37 |
| openstackgerrit | dane-fichter proposed openstack/nova: libvirt: Raise error when booting from encrypted volumes https://review.openstack.org/203784 | 13:37 |
| johnthetubaguy | garyk: yeah, I just noticed that, its basically the same, looking closer now | 13:37 |
| dansmith | rlrossit: yes, but where is the source of truth that says 1.23 should be missing? or are you talking about only fields that cause us to have a relationship with another object (and thus be in the relationships map)? | 13:38 |
| garyk | johnthetubaguy: for the metadata the issue here is that we do not have a neutron proxy that injects the headers to determine which instance sent the request. instead we have a load balancer running as part of the dhcp/routing service inject a hhtp header that helps us locate the instance id | 13:38 |
| dansmith | rlrossit: mriedem_away: did the patch to stable to fix the version used in the backport call get landed? | 13:38 |
| *** shuquan has joined #openstack-nova | 13:38 | |
| *** gzhai has quit IRC | 13:39 | |
| dansmith | looks like not | 13:39 |
| johnthetubaguy | garyk: yeah, thats why I wasn't expecting the code path to be so different, just trying to work out whats different with the regular case | 13:39 |
| *** chuckcarmack has joined #openstack-nova | 13:39 | |
| rlrossit | dansmith: this might be because I'm misunderstanding it, but basically in obj_make_compatible() if we add a field, we should always be doing if version < (major, minor): del primitive['thing_i_added']? | 13:39 |
| garyk | johnthetubaguy: in the ref arch guest <-> dhcp agent <-> proxy <-> metadata service api | 13:40 |
| dansmith | rlrossit: so you need your patch to depends-on that patch from mriedem against stable, right? | 13:40 |
| garyk | the proxy interfaces with neutron to get the instance id | 13:40 |
| rlrossit | dansmith: it didn't land but I did a depends-on the stable fix, but that still didn't solve the infinite recursion | 13:40 |
| dansmith | rlrossit: yep | 13:40 |
| dansmith | rlrossit: you did it and then removed the depends-on? | 13:40 |
| garyk | johnthetubaguy: nsx support guest <-> nsx dhcp with load balancer <-> metadata service | 13:40 |
| dansmith | rlrossit: https://review.openstack.org/#/c/198730/ | 13:40 |
| garyk | here the metadata service queries neutron | 13:40 |
| dansmith | rlrossit: that just depends-on your logging change | 13:41 |
| rlrossit | dansmith: our 2 conversations are merging... I have no idea what you're talking about | 13:41 |
| johnthetubaguy | garyk: yeah, thats cool | 13:41 |
| dansmith | rlrossit: sorry, lets table the additional test chat for a sec | 13:41 |
| garyk | johnthetubaguy: advantages - client caching for tokens, on less point of failure on the way | 13:41 |
| rlrossit | dansmith: but I rebased the logging change on mriedem_away's backport | 13:41 |
| *** dane-fichter has joined #openstack-nova | 13:41 | |
| rlrossit | cool | 13:41 |
| johnthetubaguy | garyk: I think its really just that I think the code needs to be a bit more DRY | 13:41 |
| garyk | it is very dry, a desert out there... | 13:41 |
| rlrossit | bauzas is going to start becoming impatient :) | 13:41 |
| bauzas | rlrossit: ah ah no worries | 13:41 |
| dansmith | rlrossit: this: https://review.openstack.org/#/c/198730/ needs to depends-on this: https://review.openstack.org/#/c/205686/ | 13:41 |
| dansmith | rlrossit: and it doesn't right? | 13:42 |
| garyk | johnthetubaguy: can you please be more specifc regarding how we can improve the code? | 13:42 |
| *** lucas____ has quit IRC | 13:42 | |
| *** ctrath has joined #openstack-nova | 13:42 | |
| johnthetubaguy | garyk: I am adding comments now, just digging into it still | 13:42 |
| garyk | ok, thanks | 13:42 |
| johnthetubaguy | garyk: I will try be specific about what I wanting to see, really just a heads up, then I realised its not all that different after all | 13:43 |
| rlrossit | dansmith: so here's the chain of what's going on: https://review.openstack.org/#/c/198730/ depends on https://review.openstack.org/#/c/205611/ which is rebased (depends on) https://review.openstack.org/#/c/205686/ | 13:43 |
| dansmith | oh right, I missed that you had a non-zuul dependency on mriedem_away's | 13:44 |
| *** bkopilov has joined #openstack-nova | 13:45 | |
| dansmith | rlrossit: okay, but this clearly made an improvement, right? | 13:46 |
| *** lucas____ has joined #openstack-nova | 13:46 | |
| dansmith | we're no longer ping-ponging back and forth between 1.16 and 1.19 | 13:46 |
| rlrossit | dansmith: nope, it's still a problem | 13:46 |
| rlrossit | lemme find the logs | 13:46 |
| *** deepakcs has quit IRC | 13:47 | |
| garyk | johnthetubaguy: ok, thanks | 13:47 |
| lxsli | bauzas: so if I push a new version of 128992 removing total, used and overcommit_ratio from __init__, you could +1? | 13:47 |
| lxsli | bauzas: I don't see a problem with that | 13:47 |
| bauzas | lxsli: I just added a comment now | 13:48 |
| rlrossit | dansmith: uhhhhh... this is weird | 13:48 |
| lxsli | bauzas: yeah I read it, I'm just doublechecking we're on the same page before I push a new version | 13:48 |
| rlrossit | dansmith: it is just retrying 1.16 -> 1.16 -> 1.16 over and over again | 13:48 |
| *** nelsnelson has quit IRC | 13:48 | |
| dansmith | right | 13:48 |
| dansmith | which makes me think something else silly got broken | 13:48 |
| dansmith | like some logic moved in the process of adding debug or something | 13:49 |
| lxsli | bauzas: in fact I can delete the whole __init__ method | 13:49 |
| bauzas | lxsli: tbh, we don't probably need an init method, but I left the use of obj_set_defaults() to keep the previous behaviour, ie. making sure that the ratio is set to 1.0 anyway | 13:49 |
| rlrossit | dansmith: so this is a ryan error, not a nova error now? | 13:49 |
| bauzas | lxsli: why ? | 13:49 |
| bauzas | lxsli: oh, because that's a mixin, right? | 13:49 |
| lxsli | bauzas: I can put the default in the field declaration | 13:49 |
| bauzas | lxsli: yeah, that's just I added | 13:50 |
| dansmith | rlrossit: just sayin' that's what it kinda feels like, but still looking | 13:50 |
| dansmith | bauzas: using obj_set_defaults() in __init__ is dangerous | 13:50 |
| rlrossit | dansmith: I am more than willing to be proven the one with a bug if it means I get to finally solve this | 13:50 |
| *** thangp has joined #openstack-nova | 13:50 | |
| bauzas | dansmith: why so ? pretty interested in knowing why :) | 13:50 |
| dansmith | bauzas: let me code up an example, just a sec | 13:51 |
| bauzas | dansmith: sure thing | 13:51 |
| *** belmoreira has quit IRC | 13:51 | |
| bauzas | (if so, I have to modify my own patch series= | 13:51 |
| *** edmondsw has joined #openstack-nova | 13:51 | |
| bauzas | which is understandable | 13:51 |
| bauzas | dansmith: just for your records, lxsli and I are debating since this morning about https://review.openstack.org/#/c/128992/22 | 13:52 |
| lxsli | can't see where obj_set_defaults is normally called from | 13:52 |
| bauzas | lxsli: it's never called, you have to call it explicitely | 13:52 |
| lxsli | bauzas: but no-one else calls it explicitly yet? that seems odd | 13:52 |
| *** mriedem_away is now known as mriedem | 13:53 | |
| dansmith | bauzas: http://paste.openstack.org/show/406072/ | 13:53 |
| * bauzas looking | 13:53 | |
| *** andrewbogott has joined #openstack-nova | 13:53 | |
| bauzas | dansmith: oooooh, right, tbh I also saw that | 13:53 |
| openstackgerrit | Sean Dague proposed openstack/nova: Document when we should have a microversion https://review.openstack.org/191188 | 13:54 |
| *** andrewbogott has left #openstack-nova | 13:54 | |
| bauzas | lxsli: dansmith is right, I made a pattern that way https://review.openstack.org/#/c/145528/35/nova/objects/request_spec.py,cm | 13:54 |
| lxsli | dansmith: is obj_set_defaults called from o.vo? | 13:54 |
| dansmith | ? | 13:54 |
| *** karmatronic has quit IRC | 13:54 | |
| dansmith | it's provided in o.vo, but it's never called implicitly, no | 13:54 |
| claudiub | danpb: hello! If you have a bit of time, can you take a look at this? You've already did some reviews on it, it has a +2 already: https://review.openstack.org/#/c/190997/ | 13:54 |
| *** fawadkhaliq has joined #openstack-nova | 13:55 | |
| johnthetubaguy | garyk: I added some suggestions on https://review.openstack.org/#/c/147539, on the upside, I have a feeling the unit tests are good, so they should keep working across my suggested refactor, which is handy. | 13:55 |
| lxsli | I'm quite confused as no-one ever seems to call it | 13:55 |
| danpb | claudiub: aren't we too late to be dealing with specs now (johnthetubaguy ^^^ ?) | 13:56 |
| bauzas | dansmith: are you considering then https://review.openstack.org/#/c/145528/35/nova/objects/request_spec.py,cm L63 to be valid ? | 13:56 |
| *** ctrath has quit IRC | 13:56 | |
| *** tpeoples has quit IRC | 13:56 | |
| dansmith | bauzas: I really need to continue focusing on rlrossit's thing for a bit, I just wanted to end the obj_set_defaults() speculation | 13:56 |
| claudiub | danpb: it can still be approved, and it will good for M? | 13:57 |
| bauzas | dansmith: okay, take your time, we can discuss about that later on | 13:57 |
| garyk | johnthetubaguy: ok, thanks | 13:57 |
| bauzas | I'm just wondering the correct pattern for objects instantiation with defaults | 13:57 |
| bauzas | dansmith: ^ | 13:57 |
| *** ajayaa has quit IRC | 13:58 | |
| johnthetubaguy | claudiub: danpb: the spec deadline was the end of last month, but mitaka is now open for specs (I will send out an email once this review push is over) | 13:58 |
| johnthetubaguy | claudiub: danpb: the exceptions are bug fixes that needs a spec (API changes), priority stuff (although that should be done too really), and updates to existing specs | 13:59 |
| dansmith | rlrossit: so this first line here: http://logs.openstack.org/30/198730/14/check/gate-grenade-dsvm-partial-ncpu/a29229b/logs/old/screen-n-cpu.txt.gz?level=INFO#_2015-07-27_15_35_20_300 | 14:00 |
| danpb | ok, so claudiub's spec needs to be reposted against M tree then | 14:01 |
| dansmith | rlrossit: is confusing because the old side can take instance list 1.16, which means it shouldn't fault on it, which means we're not logging the actual thing it's choking on right? | 14:01 |
| *** shuquan has quit IRC | 14:01 | |
| *** Murali has quit IRC | 14:01 | |
| johnthetubaguy | yes, I think thats the comment mriedem added, I just added the same comment | 14:01 |
| *** tpeoples has joined #openstack-nova | 14:01 | |
| dansmith | rlrossit: ah, I see, yeah you're printing the objprim version/name | 14:02 |
| claudiub | danpb: johnthetubaguy: sure, I can do that, but there is no mitaka subfolder in nova-specs yet. | 14:02 |
| *** gszasz has quit IRC | 14:02 | |
| dansmith | rlrossit: but what must be happening at this point, I think, is that we're legitimately failing to backport some sub-object and we just need to know what that object is so we can fix our obj_relationships map, right? | 14:02 |
| *** burt has joined #openstack-nova | 14:02 | |
| *** neiljerram has joined #openstack-nova | 14:02 | |
| rlrossit | dansmith: but the backport of the subobject should still be logging right? | 14:03 |
| rlrossit | (totally agree, just wondering where the logs are) | 14:03 |
| dansmith | rlrossit: we need to log the thing that caused the error on the old side (so your DNM patch) | 14:03 |
| dansmith | rlrossit: the new side doesn't know what the problem is, because we're (properly) asking for InstanceList 1.16, which is the top, and we expect everything inside to be good | 14:04 |
| lxsli | claudiub: just make it | 14:04 |
| *** Piet has joined #openstack-nova | 14:04 | |
| dansmith | rlrossit: but since that's not happening, we need a bit of logging on the old side to see what it keeps complaining about | 14:04 |
| danpb | claudiub: yes it does exist | 14:04 |
| danpb | claudiub: you must not have up2date gitcheckout | 14:04 |
| rlrossit | dansmith: ok, so basically put logging in the exact same spot except in the kilo change? | 14:04 |
| dansmith | rlrossit: lemme push it up and I'll show you | 14:05 |
| *** abhishekk has quit IRC | 14:05 | |
| *** MentalRay has joined #openstack-nova | 14:05 | |
| *** ctrath has joined #openstack-nova | 14:05 | |
| rlrossit | dansmith: you just going to hijack my review and put good logging in? | 14:05 |
| dansmith | rlrossit: yes | 14:05 |
| dansmith | and by "put good logging in" I mean "change one line slightly" | 14:05 |
| lxsli | bauzas: check out nova/objects/base.py:610 | 14:05 |
| *** annashen has joined #openstack-nova | 14:06 | |
| mnestratov | dansmith: hi, I answered to your questions in https://review.openstack.org/#/c/184311/ . Can you take a look? | 14:06 |
| openstackgerrit | Daniel Berrange proposed openstack/nova-specs: Virtual guest device role tagging https://review.openstack.org/195662 | 14:06 |
| bauzas | lxsli: mmm, you mean https://github.com/openstack/nova/blob/master/nova/objects/base.py#L610 ? | 14:06 |
| dansmith | mnestratov: you're #4 in line, but yes :) | 14:06 |
| mnestratov | dansmith: :) | 14:07 |
| lxsli | bauzas: yep. I suspect that's all specifying a default in a field does for you | 14:07 |
| bauzas | lxsli: correct | 14:07 |
| *** ivasev has joined #openstack-nova | 14:07 | |
| *** ankit_ag has quit IRC | 14:08 | |
| bauzas | lxsli: well, I looked at that a couple of months before, so I'm mostly rusty here | 14:08 |
| openstackgerrit | Claudiu Belu proposed openstack/nova-specs: Adds Hyper-V UEFI Secure Boot spec https://review.openstack.org/190997 | 14:08 |
| lxsli | bauzas: tempted to cut the knot by not providing a default for now :p | 14:08 |
| claudiub | danpb: yeah, you're right, I didn't rebase it on origin/master, sorry. :) | 14:08 |
| rlrossit | dansmith: I rebased that chain of two, so when you get the logging up I can push those back up to see what happens | 14:08 |
| bauzas | lxsli: but IIRC, defaulting a field doesn't mean it will be instanciated to the default value when the object is created | 14:08 |
| lxsli | bauzas: that appears to be correct yes | 14:08 |
| ndipanov | bauzas, lxsli that's a bug imho but it's not so easy to fix | 14:09 |
| *** yamahata has joined #openstack-nova | 14:09 | |
| bauzas | lxsli: unless you explicitely call obj_set_defaults(field) for the reason dansmith provided | 14:09 |
| lxsli | ndipanov: +1 | 14:09 |
| bauzas | lxsli: ie. that really depends on the object and how it's hydrated | 14:09 |
| dansmith | rlrossit: okay pushed | 14:10 |
| *** boris-42 has joined #openstack-nova | 14:10 | |
| lxsli | bauzas: so if the object has ever been saved to the DB, it should already have the default in it | 14:10 |
| bauzas | lxsli: here, we need to agree on a pattern | 14:10 |
| *** mpaolino has quit IRC | 14:10 | |
| *** annashen has quit IRC | 14:11 | |
| bauzas | lxsli: if the object is given by a DB call like a obj.get_by_id(), it's populated from what it gets | 14:11 |
| bauzas | lxsli: in that case, an obj_set_defaults() sounds risky | 14:11 |
| lxsli | bauzas: I think we need to set the default a) when the attr is requested, once we realise the object isn't in the DB; b) before we save to the DB | 14:11 |
| lxsli | b cont) assuming there is no value set (which is different to None being set) | 14:12 |
| bauzas | lxsli: I'd rather say depending on whether the object is orphaned or not | 14:12 |
| openstackgerrit | Ryan Rossiter proposed openstack/nova: Fix test_relationships() for subobject versions https://review.openstack.org/198904 | 14:13 |
| openstackgerrit | Ryan Rossiter proposed openstack/nova: Change List objects to use obj_relationships https://review.openstack.org/198730 | 14:13 |
| bauzas | lxsli: if that's an orphaned object, who cares about what could give the DB? | 14:13 |
| dansmith | rlrossit: so, I'm hoping that the next log run we get will tell us that "foo object at 1.700 failed" and we can go check our relationships and see that you missed one or something | 14:13 |
| dansmith | rlrossit: entirely possible something shifted underneath you since your first version and you just didn't catch it | 14:13 |
| rlrossit | dansmith: time to sit and stare at zuul for a good hour | 14:13 |
| dansmith | rlrossit: yeah, and 800+ things in the queue it'll be a while :/ | 14:13 |
| dansmith | rlrossit: you could, of course, spend the time getting grenade to run locally :) | 14:14 |
| rlrossit | dansmith: yeah there's a good chance one of the 13 rebases involved a mistake | 14:14 |
| rlrossit | but that involves working.... | 14:14 |
| mnestratov | taking into account the fact that non priority feature freeze is approaching could cores pay some attention to virtuozzo related patches? They are absolutely trivial: | 14:14 |
| mnestratov | snapshot feature support: | 14:14 |
| mnestratov | - https://review.openstack.org/#/c/184153/ | 14:14 |
| mnestratov | resize support: | 14:14 |
| mnestratov | - https://review.openstack.org/#/c/182390/ | 14:14 |
| mnestratov | - https://review.openstack.org/#/c/182257/ | 14:14 |
| dansmith | rlrossit: yeah, but if so, it'll be an easy fix and we'll be on our way | 14:14 |
| rlrossit | yeah I might do that in the meantime if it's pretty simple | 14:14 |
| bauzas | man, I really need to get my yellow belt for objects :) | 14:14 |
| lxsli | bauzas: I don't know what orphaned means in o.vo context well, can you help me? Do new objects start out orphaned? Can orphaned objects be saved? Can an object become un-orphaned? | 14:14 |
| *** gszasz has joined #openstack-nova | 14:14 | |
| dansmith | lxsli: orphaned means no context stored in the object | 14:14 |
| lxsli | bauzas: the o.vo doc could be considerably better | 14:14 |
| *** jecarey_ has joined #openstack-nova | 14:14 | |
| bauzas | lxsli: no context | 14:15 |
| bauzas | lxsli: argh, jinxed by dansmith | 14:15 |
| lxsli | right, which means what? :) | 14:15 |
| dansmith | bauzas: okay, point me at what you want me to look at? | 14:15 |
| lxsli | ah well we have the expert now | 14:15 |
| dansmith | lxsli: no nova.context.RequestContext stored in the object | 14:15 |
| lxsli | right and what do we use the RequestContext for... | 14:15 |
| bauzas | dansmith: so, we're yakshaving since this morning about https://review.openstack.org/#/c/128999/ | 14:15 |
| dansmith | lxsli: any call we make back to conductor | 14:16 |
| dansmith | lxsli: so any remotable call, any lazy-load, etc | 14:16 |
| lxsli | ok thanks | 14:16 |
| bauzas | dansmith: so, the convo is not exactly about that specific change, but rather more on what should be a proper pattern for initing default values | 14:16 |
| lxsli | I have many more questions but lets focus on the issue for now | 14:16 |
| bauzas | dansmith: and that hits lxsli's quest for knowledge :) | 14:17 |
| lxsli | dansmith: in fact the link is https://review.openstack.org/#/c/128992/22/nova/objects/resources.py | 14:17 |
| dansmith | bauzas: okay, I don't see that that patch sets any defaults, so interesting linkage but... | 14:17 |
| lxsli | line 164 | 14:17 |
| bauzas | dansmith: because PEBKAC maybe ? | 14:17 |
| bauzas | https://review.openstack.org/#/c/128992/22 | 14:18 |
| bauzas | oh, no, correct change | 14:18 |
| *** tongli has joined #openstack-nova | 14:18 | |
| bauzas | heh :) | 14:18 |
| bauzas | no offense | 14:18 |
| bauzas | (was jk on me) | 14:18 |
| lxsli | eventually bauzas will see I'm right, this has been a pattern today ;) | 14:18 |
| dansmith | okay, so the question is about jaypipes' use of the __init__ method on IntegerInventory/ | 14:18 |
| bauzas | dansmith: zactly | 14:19 |
| dansmith | right, so, | 14:19 |
| dansmith | I hassled him about this in person last week | 14:19 |
| bauzas | dansmith: originally, the init was coercing a few fields | 14:19 |
| dansmith | doing obj_set_defaults() in that method breaks convention and could be a sign to others to do it that way | 14:19 |
| bauzas | dansmith: then, it has been moved to new field types but we're still leaving the overcommit_ratio to 1.0 by default | 14:19 |
| dansmith | bauzas: yeah, I know, I reviewed it | 14:19 |
| *** garyk1 has joined #openstack-nova | 14:20 | |
| bauzas | dansmith: so, given your point, what could be a good way for ensuring that an object would be created with default values ? | 14:20 |
| dansmith | bauzas: so, my preference is that we not have an __init__ at all, and just be explicit whenever we use the object to initialize the things we need there | 14:20 |
| bauzas | I see | 14:20 |
| dansmith | but, of course, for very specialized objects like this, I certainly understand the desire to have shortcuts like this | 14:21 |
| bauzas | dansmith: because of the possibility to wedge the existing DB values ? | 14:21 |
| *** tsekiyama has joined #openstack-nova | 14:21 | |
| *** dane-fichter has quit IRC | 14:21 | |
| dansmith | if this is an object that is only ever created and never updated, then it's not a problem really, other than setting a precedent for others to do it when they don't understand the implications | 14:21 |
| dansmith | bauzas: yeah | 14:21 |
| openstackgerrit | Andrea Rosa proposed openstack/nova: Raise nova exception for missing/empty machine-id https://review.openstack.org/202640 | 14:21 |
| bauzas | dansmith: ok I see, so we should be very clear about that - no defaults() in a persisted object | 14:21 |
| dansmith | well, | 14:22 |
| bauzas | dansmith: I need to amend my own change to mark that very explictely | 14:22 |
| *** garyk has quit IRC | 14:22 | |
| dansmith | a defaults() method is okay if it's used carefully | 14:22 |
| bauzas | dansmith: so the save() method has to care about it | 14:22 |
| dansmith | but calling it implicitly from __init__ is bad, and this is like half of that | 14:22 |
| *** alex_klimov has quit IRC | 14:22 | |
| *** alex_klimov has joined #openstack-nova | 14:23 | |
| bauzas | dansmith: so, given we know that we're working with non-persisted objects, is this really a problem but the precedent it raises ? | 14:23 |
| *** Guest11859 has quit IRC | 14:23 | |
| dansmith | bauzas: well, that's what I said to jaypipes, and then he said "well, it's not persisted now, but will be later" so .. :D | 14:23 |
| bauzas | eheh | 14:24 |
| dansmith | it's actually okay in persisted objects sometimes, right? | 14:24 |
| nagyz_ | hmm, what would be the best way to remove a symlink that I need to create so the encryptor function properly inside a libvirt driver?... | 14:24 |
| dansmith | like, | 14:24 |
| dansmith | if we only ever create/destroy but never update an object, then it won't be a problem | 14:24 |
| dansmith | it's just a complicated story, no way to make it otherwise I think | 14:24 |
| nagyz_ | os.remove() won't work and I guess I cannot call the utils with rm | 14:24 |
| bauzas | right, and if we take from DB first, it's not also a problem | 14:24 |
| bauzas | okay, I need to consider that | 14:24 |
| kodoku_ | Hi/Re When I create cache for my image, I have error in nova ==> error: [Errno 9] Bad file descriptor and error in glance ==> error: [Errno 104] Connection reset by peer | 14:24 |
| nagyz_ | or I can just add it to the patch at the moment and seek guidance in the form or reviews ;-) | 14:25 |
| dansmith | bauzas: well, you can't really ensure that it's always taken from the db first | 14:25 |
| dansmith | bauzas: but anyway, you get the idea | 14:25 |
| lxsli | dansmith: bauzas: so in terms of getting this merged, removing the default seems the sanest thing to do? | 14:25 |
| bauzas | lxsli: on my take, I'm just going to remove https://review.openstack.org/#/c/145528/35/nova/objects/request_spec.py,cm L61 | 14:26 |
| bauzas | lxsli: so, yeah I would suggest you to do so as well | 14:26 |
| *** annegentle has joined #openstack-nova | 14:26 | |
| openstackgerrit | sajeesh cs proposed openstack/nova: Nested Quota Driver: Adding v2 and v3 support https://review.openstack.org/149828 | 14:26 |
| bauzas | lxsli: that doesn't mean we can't leave the field having a default attr | 14:26 |
| dansmith | rlrossit: wanna sniff test these version changes; https://review.openstack.org/#/c/184311/26/nova/tests/unit/objects/test_objects.py,cm ? | 14:27 |
| bauzas | lxsli: on my own, I'll keep the ReqSpec.num_instances field to have a default param, but I won't set it as init time | 14:27 |
| lxsli | bauzas: ok works for me | 14:27 |
| rlrossit | dansmith: sure | 14:27 |
| *** jaypipes has joined #openstack-nova | 14:29 | |
| openstackgerrit | Tiago Mello proposed openstack/nova: Handle unsupported Spice feature for PowerKVM https://review.openstack.org/203809 | 14:29 |
| openstackgerrit | Alexis Lee proposed openstack/nova: Adds resource classes https://review.openstack.org/128999 | 14:29 |
| openstackgerrit | Alexis Lee proposed openstack/nova: Add object models for comparing resource amounts https://review.openstack.org/128992 | 14:30 |
| *** fawadk has joined #openstack-nova | 14:30 | |
| *** fawadkhaliq has quit IRC | 14:30 | |
| *** gszasz has quit IRC | 14:31 | |
| *** ctrath has quit IRC | 14:32 | |
| *** VW__ has joined #openstack-nova | 14:32 | |
| rlrossit | dansmith: I feel bad for just pulling down these changes, rebasing mine on top of theirs, and then just reviewing based on the results from my tests..... | 14:33 |
| *** fawadkhaliq has joined #openstack-nova | 14:33 | |
| *** yamahata has quit IRC | 14:33 | |
| *** fawadk has quit IRC | 14:33 | |
| dansmith | rlrossit: why? you should s/bad/proud/ IMHO :) | 14:34 |
| rlrossit | it totally feels like I'm cheating | 14:35 |
| *** VW_ has quit IRC | 14:35 | |
| lxsli | cheating is for winners | 14:35 |
| lxsli | in a no-holds-barred sport like opensource development anyway | 14:36 |
| dansmith | rlrossit: if writing tests to catch people making mistakes is cheating, then I'm going to hell | 14:36 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: Don't check flavor disk attributes when booting from volume https://review.openstack.org/204635 | 14:36 |
| dansmith | I don't even believe in hell, but I know I have a first class ticket there anyway | 14:36 |
| *** VW__ has quit IRC | 14:37 | |
| lxsli | https://www.google.com/maps/place/Hell,+MI+48169/@42.4347194,-83.985,16z/data=!3m1!4b1!4m2!3m1!1s0x882334578757ad71:0xd45079770956618a | 14:37 |
| lxsli | You can get icecream | 14:37 |
| rlrossit | of course hell would be in michigan | 14:37 |
| lxsli | I was expecting Finland, but ^^ | 14:39 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: compute: improve exceptions related to disk size checks https://review.openstack.org/206565 | 14:39 |
| *** gszasz has joined #openstack-nova | 14:43 | |
| *** angdraug has joined #openstack-nova | 14:44 | |
| openstackgerrit | Roman Podoliaka proposed openstack/nova: rebuild: make sure server is shut down before volumes are detached https://review.openstack.org/176892 | 14:44 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: objects: add hw_vif_multiqueue_enabled image property https://review.openstack.org/202168 | 14:45 |
| *** ajayaa has joined #openstack-nova | 14:45 | |
| *** meghal has joined #openstack-nova | 14:46 | |
| kodoku_ | Who can tell me why I have issue with image-list (0.19.0 glance version) "InvalidSchema: No connection adapters were found for 'https://controlle:9292:/v1/images/detail?sort_key=name&sort_dir=asc&limit=20' | 14:47 |
| kodoku_ | If I request witch curl and good token, it's works :/ | 14:47 |
| kodoku_ | with* | 14:47 |
| *** ddaskal has joined #openstack-nova | 14:48 | |
| *** erlon has joined #openstack-nova | 14:49 | |
| rlrossit | dansmith: that change looked good to me. They passed my tests at least and I didn't see anything wrong with them (in case I need to improve my tests, turns out they're perfect) | 14:50 |
| dansmith | rlrossit: cool | 14:51 |
| *** jwcroppe has quit IRC | 14:52 | |
| openstackgerrit | Stephen Finucane proposed openstack/nova: Store correct VirtCPUTopology https://review.openstack.org/197125 | 14:52 |
| *** jwcroppe has joined #openstack-nova | 14:52 | |
| *** sudipto has joined #openstack-nova | 14:52 | |
| *** Nikolay_St has quit IRC | 14:55 | |
| *** ctrath has joined #openstack-nova | 14:56 | |
| *** jwcroppe has quit IRC | 14:56 | |
| *** achanda has joined #openstack-nova | 14:56 | |
| *** sudipto has quit IRC | 14:57 | |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: objects: add hw_vif_multiqueue_enabled image property https://review.openstack.org/202168 | 14:58 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: tests: change assertEqual args to expected, observed for VIF https://review.openstack.org/204190 | 14:58 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: enable virtio-net multiqueue https://review.openstack.org/128829 | 14:58 |
| *** ctrath has quit IRC | 14:59 | |
| *** sudipto has joined #openstack-nova | 14:59 | |
| *** pixelbeat_ has quit IRC | 14:59 | |
| gszasz | bauzas: ping | 15:02 |
| bauzas | gszasz: yup ? | 15:02 |
| *** tdurakov has joined #openstack-nova | 15:03 | |
| bauzas | naked ping? | 15:03 |
| openstackgerrit | Zoltan Arnold Nagy proposed openstack/nova: libvirt: add encrypted rbd volume support https://review.openstack.org/206576 | 15:03 |
| bauzas | https://blogs.gnome.org/markmc/2014/02/20/naked-pings/ | 15:03 |
| gszasz | bauzas: I have a question according https://bugs.launchpad.net/nova/+bug/1214943 | 15:03 |
| 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) | 15:03 |
| bauzas | yep? | 15:04 |
| gszasz | bauzas: sorry for naked ping | 15:04 |
| bauzas | np | 15:04 |
| bauzas | just for others too | 15:04 |
| bauzas | gszasz: what do you want to know ? | 15:04 |
| gszasz | bauzas: we are hitting this issue in customer scenario, what is the actual status of this bug | 15:05 |
| bauzas | gszasz: so, there are 3 points here | 15:06 |
| *** agireud has joined #openstack-nova | 15:06 | |
| *** annashen has joined #openstack-nova | 15:06 | |
| bauzas | gszasz: 1/ a workaround https://review.openstack.org/#/c/180151/ | 15:07 |
| *** tonytan4ever has joined #openstack-nova | 15:07 | |
| bauzas | I need to discuss with johnthetubaguy about this one but not had time yet | 15:07 |
| bauzas | gszasz: 2/ allocation ratios should be per compute so we could access them | 15:07 |
| bauzas | that's something I'm currently working on, based on https://blueprints.launchpad.net/nova/+spec/allocation-ratio-to-resource-tracker | 15:08 |
| bauzas | gszasz: 3/ we should call the scheduler to verify the live migration even if a dest host is provided, that's something approved by https://blueprints.launchpad.net/nova/+spec/check-destination-on-migrations | 15:09 |
| bauzas | gszasz: all of them are currently either on review or approved but waiting for implementation | 15:09 |
| gszasz | aix: ^^ | 15:09 |
| *** achanda has quit IRC | 15:09 | |
| *** VW_ has joined #openstack-nova | 15:10 | |
| gszasz | bauzas: splendid | 15:10 |
| gszasz | bauzas: thanks for your time | 15:10 |
| *** nkrinner has quit IRC | 15:10 | |
| bauzas | gszasz: np | 15:10 |
| *** annashen has quit IRC | 15:11 | |
| *** pixelbeat_ has joined #openstack-nova | 15:13 | |
| johnthetubaguy | bauzas: am I blocking you on something? | 15:15 |
| *** gszasz has quit IRC | 15:15 | |
| bauzas | johnthetubaguy: not really | 15:16 |
| bauzas | johnthetubaguy: I have a bugfix change that you reviewed but I had no time to work on it yet | 15:16 |
| bauzas | johnthetubaguy: since you're quite busy with your PTL job, I prefer to discuss that once FF is done | 15:16 |
| johnthetubaguy | bauzas: sounds good, thank you | 15:17 |
| danpb | sorry mriedem, you loose today, os-brick merged first | 15:17 |
| *** karimb has joined #openstack-nova | 15:18 | |
| *** mtanino has joined #openstack-nova | 15:18 | |
| mriedem | danpb: np | 15:18 |
| * johnthetubaguy giggles, something about throwing a brick through a window | 15:19 | |
| danpb | i even +2d your stuff first, but random bogus jenkins failure lost your place in the queue | 15:19 |
| *** otter768 has joined #openstack-nova | 15:20 | |
| mriedem | yeah, it's not an issue, i figured this would happen anyway | 15:20 |
| *** vishwanathj has quit IRC | 15:20 | |
| *** mestery has quit IRC | 15:20 | |
| *** zsais has joined #openstack-nova | 15:21 | |
| *** exploreshaifali has quit IRC | 15:21 | |
| *** dannywilson has joined #openstack-nova | 15:22 | |
| *** markus_z has quit IRC | 15:23 | |
| *** agireud has quit IRC | 15:24 | |
| *** mpaolino has joined #openstack-nova | 15:24 | |
| *** otter768 has quit IRC | 15:24 | |
| mriedem | however, the scaleio and hgst patches should rebase on top of mine | 15:24 |
| mriedem | hemna: congrats https://review.openstack.org/#/c/175569/ | 15:24 |
| hemna | mriedem, thanks man. thanks for all the help with that guy. | 15:25 |
| *** cfriesen has quit IRC | 15:25 | |
| *** agireud has joined #openstack-nova | 15:25 | |
| sudipto | dansmith, jaypipes : one final review today? https://review.openstack.org/#/c/198972/ (hopefully) :) | 15:25 |
| *** meghal has quit IRC | 15:25 | |
| *** jasondotstar has quit IRC | 15:25 | |
| dansmith | sudipto: yep, will do those after I finish my current thing | 15:26 |
| sudipto | dansmith, thanks :) | 15:27 |
| *** tongli_ has joined #openstack-nova | 15:28 | |
| *** garyk1 has quit IRC | 15:29 | |
| *** cfriesen has joined #openstack-nova | 15:30 | |
| *** Kevin_Zheng has joined #openstack-nova | 15:31 | |
| *** tongli has quit IRC | 15:31 | |
| openstackgerrit | Dmitry Guryanov proposed openstack/nova: libvirt: do relative block rebase only with non-null base https://review.openstack.org/202442 | 15:32 |
| *** browne has quit IRC | 15:32 | |
| *** Administrator_ has joined #openstack-nova | 15:33 | |
| *** rdopiera_ has quit IRC | 15:33 | |
| *** mestery has joined #openstack-nova | 15:34 | |
| *** Administrator__ has joined #openstack-nova | 15:34 | |
| mriedem | rlrossit: chuckcarmack: easy code review https://review.openstack.org/#/c/206236/ | 15:34 |
| *** Kevin_Zheng has quit IRC | 15:36 | |
| *** garyk has joined #openstack-nova | 15:37 | |
| *** Kevin_Zheng has joined #openstack-nova | 15:37 | |
| *** Administrator_ has quit IRC | 15:38 | |
| *** aix has quit IRC | 15:38 | |
| *** Administrator__ has quit IRC | 15:39 | |
| *** VW_ has quit IRC | 15:40 | |
| openstackgerrit | garyk proposed openstack/nova: Neutron: add in API method for updating VNIC index https://review.openstack.org/147126 | 15:40 |
| *** Administrator_ has joined #openstack-nova | 15:40 | |
| dansmith | jaypipes: around? | 15:40 |
| *** VW_ has joined #openstack-nova | 15:40 | |
| jaypipes | dansmith: I have an all-day meeting :( | 15:40 |
| dansmith | jaypipes: veto | 15:40 |
| jaypipes | heh | 15:40 |
| dansmith | jaypipes: sudo around? | 15:41 |
| dansmith | jaypipes: okay, well, tomorrow then | 15:41 |
| *** josecastroleon has quit IRC | 15:41 | |
| jaypipes | dansmith: tomorrow afternoon, yes :) | 15:41 |
| dansmith | sheesh | 15:41 |
| *** ijw has joined #openstack-nova | 15:42 | |
| *** Administrator__ has joined #openstack-nova | 15:42 | |
| *** VW__ has joined #openstack-nova | 15:43 | |
| *** zsais1 has joined #openstack-nova | 15:43 | |
| *** zsais1 has quit IRC | 15:43 | |
| *** Kevin_Zheng has quit IRC | 15:43 | |
| *** VW_ has quit IRC | 15:43 | |
| *** Administrator_ has quit IRC | 15:45 | |
| *** Kevin_Zheng has joined #openstack-nova | 15:45 | |
| *** zsais has quit IRC | 15:46 | |
| dansmith | sudipto: we don't need to merge this until you have a patch on top to add this functionality to a virt driver right? https://review.openstack.org/#/c/202623/3 | 15:47 |
| *** alexpilotti has quit IRC | 15:47 | |
| *** claudiub has quit IRC | 15:47 | |
| *** Administrator__ has quit IRC | 15:47 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify Scheduler RPC API select_dests() https://review.openstack.org/199205 | 15:48 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Create RequestSpec object https://review.openstack.org/145528 | 15:48 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Consider that all scheduler calls are IO Ops https://review.openstack.org/188486 | 15:48 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Use RequestSpec object in the FilterScheduler https://review.openstack.org/188487 | 15:48 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify left filters for RequestSpec https://review.openstack.org/202678 | 15:48 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Prepare filters for using RequestSpec object https://review.openstack.org/197912 | 15:48 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify conductor to use RequestSpec object https://review.openstack.org/203078 | 15:48 |
| *** annashen has joined #openstack-nova | 15:48 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify metric-related filters for RequestSpec https://review.openstack.org/202676 | 15:48 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Use RequestSpec in the ChanceScheduler https://review.openstack.org/199051 | 15:48 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify NUMA, PCI and num_instances filters for RequestSpec https://review.openstack.org/202677 | 15:48 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify Aggregate filters for RequestSpec https://review.openstack.org/202674 | 15:48 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify default filters for RequestSpec https://review.openstack.org/202675 | 15:48 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Add RequestSpec methods for primitiving into dicts https://review.openstack.org/187197 | 15:48 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify Affinity filter for RequestSpec https://review.openstack.org/202673 | 15:48 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Use RequestSpec object in HostManager https://review.openstack.org/191251 | 15:48 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Use the same pci_requests field for all filters and HostManager https://review.openstack.org/191250 | 15:48 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Add DictOfListOfStrings type of field https://review.openstack.org/203596 | 15:48 |
| sudipto | dansmith, yeah agreed. | 15:48 |
| *** Administrator_ has joined #openstack-nova | 15:48 | |
| dansmith | sudipto: okay cool | 15:48 |
| *** meghal has joined #openstack-nova | 15:49 | |
| dansmith | sudipto: I just put a -2 on there so we don't merge it accidentally since it's such a tiny patch | 15:50 |
| *** jlanoux has quit IRC | 15:50 | |
| dansmith | sudipto: just let me know when you have something on top and I'll drop it | 15:50 |
| sudipto | dansmith, sure i will. thx. Also read your comments just now on the other patch. Once you and jaypipes have an agreement, please let me know :) | 15:51 |
| dansmith | sudipto: yeah, sorry we keep crossing each other | 15:51 |
| *** annashen has quit IRC | 15:51 | |
| sudipto | dansmith, np :) | 15:51 |
| *** Kevin_Zheng has quit IRC | 15:51 | |
| *** Administrator__ has joined #openstack-nova | 15:51 | |
| dansmith | sudipto: now you should concentrate on doing PTO properly :) | 15:51 |
| *** scheuran has quit IRC | 15:52 | |
| sudipto | dansmith, all is coz of that deadline thing on the 31st for non-prio specs :( | 15:52 |
| dansmith | sdague: can you tell me what the magic browser thing is to show in-progress test data? | 15:52 |
| dansmith | sudipto: sounds like someone scheduled their PTO wrong :D | 15:52 |
| *** changbl has joined #openstack-nova | 15:52 | |
| sudipto | dansmith, :( :( | 15:53 |
| dansmith | heh | 15:53 |
| *** nic has joined #openstack-nova | 15:53 | |
| *** Kevin_Zheng has joined #openstack-nova | 15:54 | |
| openstackgerrit | Petrut Lucian proposed openstack/nova: Hyper-V: Implement nova rescue https://review.openstack.org/127159 | 15:54 |
| mriedem | ndipanov: replied in https://review.openstack.org/#/c/199297/ - let me know if my response doesn't make sense | 15:54 |
| *** whenry has quit IRC | 15:54 | |
| *** Administrator_ has quit IRC | 15:55 | |
| *** tjones has joined #openstack-nova | 15:55 | |
| *** logan2 has quit IRC | 15:55 | |
| *** sahid has quit IRC | 15:55 | |
| ndipanov | mriedem oh man | 15:56 |
| *** Administrator_ has joined #openstack-nova | 15:56 | |
| *** Administrator__ has quit IRC | 15:56 | |
| *** unicell has joined #openstack-nova | 15:57 | |
| *** Nikolay_St has joined #openstack-nova | 15:57 | |
| dansmith | mriedem: I think I addressed all your feedback on this: https://review.openstack.org/#/c/197770/ | 15:57 |
| dansmith | mriedem: just needs a +W to unblock two patches | 15:57 |
| *** pixelbeat_ has quit IRC | 15:57 | |
| sdague | dansmith: zuul_inline = true; | 15:57 |
| mriedem | am i actually expected to do code reviews today? o-) | 15:57 |
| sdague | in the javascript console on a gerrit page | 15:58 |
| ndipanov | mriedem, but from what I could tell - pbr just calls sphinx | 15:58 |
| dansmith | mriedem: no, just +W without looking | 15:58 |
| ndipanov | I really could not find any refernces to any of these settings | 15:58 |
| mriedem | ndipanov: not necessarily | 15:58 |
| *** unicell1 has quit IRC | 15:58 | |
| mriedem | ndipanov: finding the code | 15:58 |
| ndipanov | I mean I believe you it's just super confusing | 15:58 |
| mriedem | i dug in the guts of pbr for about a day to figure this out | 15:58 |
| mriedem | yes | 15:58 |
| mriedem | that's why i updated the pbr docs on building docs via pbr :) | 15:58 |
| dansmith | sdague: I should have just tried autocomplete of zuul | 15:58 |
| sdague | yeh, it's important that the value is true, 1 doesn't work | 15:59 |
| mriedem | ndipanov: this is where pbr generates the docs for you, not via sphinx http://git.openstack.org/cgit/openstack-dev/pbr/tree/pbr/builddoc.py#n80 | 15:59 |
| dansmith | sdague: so is that supposed to be active immediately, or do I do something to trigger it to start updating? | 15:59 |
| ndipanov | right so that reads the conf.py | 15:59 |
| sdague | it goes with a timer pop, it's like every 60s or something | 15:59 |
| *** Kevin_Zheng has quit IRC | 15:59 | |
| *** Administrator__ has joined #openstack-nova | 16:00 | |
| dansmith | hmm | 16:00 |
| *** cfriesen_ has joined #openstack-nova | 16:00 | |
| sdague | http://dl.dropbox.com/u/6514884/screenshot_232.png | 16:00 |
| *** lucas____ has quit IRC | 16:00 | |
| mriedem | ndipanov: well, this reads the config http://git.openstack.org/cgit/openstack-dev/pbr/tree/pbr/builddoc.py#n120 | 16:00 |
| sdague | it's not the whole zuul thing, just an indicator that it's being run | 16:00 |
| *** rotbeard has quit IRC | 16:00 | |
| mriedem | ndipanov: http://git.openstack.org/cgit/openstack-dev/pbr/tree/pbr/builddoc.py#n170 is where the autodoc is generated depending on flags passed into pbr | 16:01 |
| ndipanov | mriedem, I see what you mean | 16:01 |
| mriedem | in nova's case we use the auto_index path | 16:01 |
| dansmith | sdague: ooh, I thought this was the thing that imported the bar graph view | 16:01 |
| mriedem | ndipanov: then depending on if you use warnerrors or not, it hits here http://git.openstack.org/cgit/openstack-dev/pbr/tree/pbr/builddoc.py#n187 | 16:01 |
| mriedem | and ether delegates up to sphinx or calls it out separately | 16:01 |
| openstackgerrit | Alex Xu proposed openstack/python-novaclient: Adds support for x509 certificates as keypairs https://review.openstack.org/136458 | 16:02 |
| openstackgerrit | Alex Xu proposed openstack/python-novaclient: Add version discover and check in CLI https://review.openstack.org/184281 | 16:02 |
| openstackgerrit | Alex Xu proposed openstack/python-novaclient: Set "latest" as default compute api version https://review.openstack.org/184658 | 16:02 |
| ndipanov | mriedem, sigh | 16:02 |
| mriedem | ndipanov: so yeah, i guess here it reads conf.py http://git.openstack.org/cgit/openstack-dev/pbr/tree/pbr/builddoc.py#n132 | 16:02 |
| mriedem | and expects that name | 16:02 |
| ndipanov | otp now - will +2 it with a snarky comment after | 16:02 |
| kashyap | mriedem: Urgh, I just looked up the definition of "gouge", wow, sounds pretty grotesque. | 16:03 |
| mriedem | ndipanov: ok - i guess you could maybe change the confoverrides that get passed into the sphinx app builder thing, for exclude_modules based on what we have for autoindex_exclude_modules, i think i actually tried that when i was playing with this | 16:03 |
| *** Administrator_ has quit IRC | 16:03 | |
| mriedem | ndipanov: i'm happy to open a bug against pbr on that also | 16:03 |
| *** Kevin_Zheng has joined #openstack-nova | 16:03 | |
| *** cfriesen has quit IRC | 16:03 | |
| sdague | dansmith: no, I didn't think the whole bar graph was useful here | 16:03 |
| ndipanov | in the long run probably not a good use of our time tbh | 16:04 |
| dansmith | sdague: okay | 16:04 |
| *** Administrator_ has joined #openstack-nova | 16:04 | |
| mriedem | dansmith: so on https://review.openstack.org/#/c/197770/ - i had a few comments in ps5 | 16:05 |
| mriedem | are you saying you addressed those w/o reply that you'd do it and i have to diff to make sure, or other? | 16:05 |
| *** Administrator__ has quit IRC | 16:06 | |
| *** coolsvap is now known as coolsvap|dinner | 16:06 | |
| *** kmARC has quit IRC | 16:06 | |
| *** penick has joined #openstack-nova | 16:06 | |
| mriedem | dansmith: i guess overall i was confused about why some of these were converted, but not others: self.stubs.Set(db, 'instance_get_by_uuid', | 16:06 |
| mriedem | and i still see some, so the pattern is alluding me a bit | 16:06 |
| dansmith | mriedem: I think the only things you had that weren't covered by melwitt's comments was a question on a thing that I answered.. oh and the instance delete case | 16:06 |
| openstackgerrit | Nicolas Simonds proposed openstack/nova: libvirt: Fix/implement revert-resize for RBD-backed images https://review.openstack.org/187395 | 16:06 |
| openstackgerrit | Nicolas Simonds proposed openstack/nova: libvirt: Split out resize_image logic from create_image https://review.openstack.org/187857 | 16:06 |
| *** baoli has quit IRC | 16:07 | |
| *** annashen has joined #openstack-nova | 16:07 | |
| dansmith | mriedem: some of them I converted to use mock because I needed mock's call records to figure out why it was failing, and didn't think it was worth converting back | 16:07 |
| *** Kevin_Zheng has quit IRC | 16:07 | |
| *** Administrator__ has joined #openstack-nova | 16:08 | |
| dansmith | mriedem: I can convert them back if you want, but I'm going to be annoyed because I'm just trying to fix these tests functionally so I can drop the effing flavor compat bits :) | 16:08 |
| *** unicell has quit IRC | 16:09 | |
| *** jyuso1 has quit IRC | 16:10 | |
| mriedem | dansmith: my confusion is that the commit message makes it sound like, in these tests, if anything stubs out the database api to get instances, convert those, but that's not the case throughout | 16:10 |
| *** abhijeetm has quit IRC | 16:11 | |
| mriedem | so it's not so much mock vs mox | 16:11 |
| *** moshele has quit IRC | 16:11 | |
| mriedem | it's about deciding which things were stubbing the db api | 16:11 |
| *** Administrator_ has quit IRC | 16:11 | |
| *** annegentle has quit IRC | 16:11 | |
| *** Kevin_Zheng has joined #openstack-nova | 16:11 | |
| dansmith | okay, I thought you meant mock vs. stub | 16:12 |
| *** annashen has quit IRC | 16:12 | |
| dansmith | mriedem: so, can I limit scope in the commit message to address that? | 16:12 |
| lxsli | bauzas: if an instance is resized, will the requestspec be updated? | 16:12 |
| *** ip0353 has quit IRC | 16:12 | |
| dansmith | I'm frustrated that I spent half the last cycle getting flavors migrated and have spent half this cycle trying to remove it, you know :) | 16:13 |
| *** Administrator_ has joined #openstack-nova | 16:13 | |
| *** pixelbeat_ has joined #openstack-nova | 16:13 | |
| mriedem | dansmith: just explain it to me here, i don't need the commit message updated | 16:13 |
| *** lpetrut has quit IRC | 16:14 | |
| dansmith | this just hits the cases where we're pulling out an instance that might have fake flavor information that the test depends on | 16:14 |
| bauzas | lxsli: should be, but that's not part of the original series | 16:14 |
| dansmith | so, any test that the later patch fails because it seeds the db call with an instance that has flavor stuff in sysmeta | 16:14 |
| *** Administrator__ has quit IRC | 16:14 | |
| *** davideagnello has joined #openstack-nova | 16:14 | |
| *** MentalRay_ has joined #openstack-nova | 16:14 | |
| lxsli | bauzas: OK thanks, it'll be important for check-migr-dest | 16:14 |
| dansmith | which is almost anything that calls show or detail on an instance that came from the DB | 16:14 |
| bauzas | lxsli: ie. any conductor method should check the original ReqSpec, put it to the scheduler for verifying the migration, then should update the ReqSpec | 16:14 |
| bauzas | lxsli: that's part of the persist-reqspec BP IMHO | 16:15 |
| *** Administrator__ has joined #openstack-nova | 16:15 | |
| lxsli | bauzas: I'll look at that next, thanks | 16:15 |
| dansmith | mriedem: and since these tests are supposed to be testing the REST API and not the db layer, I (in the previous patches) got the mocks set up for doing this at the compute/api layer, which is what they call | 16:15 |
| mriedem | dansmith: ok, so like test_start_with_bogus_id doesn't get converted b/c it's stubbed to raise an exception before we ever use anything off the instance | 16:15 |
| dansmith | yeah | 16:15 |
| *** Kevin_Zheng has quit IRC | 16:16 | |
| dansmith | in the goal of cleaning up tests, that should be converted too | 16:16 |
| dansmith | but my goal here is to get the flavor compat bits out before we snap L :) | 16:16 |
| mriedem | and anything using _get_with_state isn't a problem b/c those only show the status value in the response | 16:16 |
| mriedem | doens't touch flavor | 16:16 |
| dansmith | mriedem: I'm really happy to add a sentence to the last paragraph of the commit message explaining the scope further | 16:17 |
| dansmith | if you think it'd be clearer | 16:17 |
| *** Kevin_Zheng has joined #openstack-nova | 16:17 | |
| mriedem | no | 16:17 |
| bauzas | lxsli: (et al.) FYI, I'll be on PTO from this Friday to Aug 13th incl. (back on Friday) | 16:17 |
| mriedem | i'm just adding inline review comments before +2 | 16:17 |
| *** meghal has quit IRC | 16:17 | |
| dansmith | okay | 16:17 |
| *** Administrator_ has quit IRC | 16:17 | |
| lxsli | bauzas: PTO = vacation? | 16:17 |
| dansmith | note to all: I'm not doing anything flavor-related in M, by protest :D | 16:17 |
| lxsli | or holiday as we call it in the UK :) | 16:17 |
| bauzas | lxsli: yup, paid time off | 16:17 |
| *** MentalRay has quit IRC | 16:17 | |
| *** ctrath has joined #openstack-nova | 16:18 | |
| *** kodoku_ has quit IRC | 16:18 | |
| lxsli | bauzas: ah yes I remember you saying, enjoy it! | 16:18 |
| bauzas | CP we call that in France :) | 16:18 |
| *** tonytan4ever has quit IRC | 16:18 | |
| bauzas | aka. the French shutdown | 16:19 |
| *** Administrator_ has joined #openstack-nova | 16:20 | |
| *** Administrator__ has quit IRC | 16:20 | |
| *** alex_klimov has quit IRC | 16:20 | |
| *** tpeoples has quit IRC | 16:21 | |
| thangp | dansmith: I think I hit a oslo_versionedobject bug - https://bugs.launchpad.net/oslo.versionedobjects/+bug/1479041 ... I am not sure | 16:22 |
| openstack | Launchpad bug 1479041 in oslo.versionedobjects "fields.from_primitive does not support subclasses of VersionedObject" [Undecided,New] | 16:22 |
| dansmith | thangp: not possible | 16:22 |
| *** Administrator__ has joined #openstack-nova | 16:22 | |
| *** jistr has quit IRC | 16:22 | |
| thangp | dansmith: ok...so what am I doing wrong? | 16:22 |
| *** Kevin_Zheng has quit IRC | 16:23 | |
| *** logan2 has joined #openstack-nova | 16:24 | |
| dansmith | thangp: that was sarcasm :D | 16:24 |
| dansmith | thangp: looking | 16:24 |
| thangp | dansmith: irc doesnt convey that well :) | 16:24 |
| dansmith | thangp: sorry :) | 16:24 |
| johnthetubaguy | has anyone taken a look at cinder-multi-attach | 16:25 |
| jroll | if nick == 'dansmith': message_type = 'sarcasm' | 16:25 |
| *** browne has joined #openstack-nova | 16:25 | |
| dansmith | thangp: that ^ | 16:25 |
| thangp | haha | 16:25 |
| *** Kevin_Zheng has joined #openstack-nova | 16:25 | |
| *** Administrator_ has quit IRC | 16:25 | |
| thangp | so I should take "superdan" on friday as no-sarcasm? | 16:25 |
| thangp | lol | 16:25 |
| *** ftersin has left #openstack-nova | 16:26 | |
| lxsli | I see, so your disavowal of flavors was sarcastic and you actually want all the flavor work for yourself? noted. | 16:26 |
| dansmith | lxsli: touche | 16:26 |
| dansmith | thangp: I think your fix is good | 16:26 |
| thangp | dansmith: ok, I can quickly put up a patch for it | 16:26 |
| thangp | thx! | 16:27 |
| dansmith | thangp: thanks! | 16:27 |
| mriedem | johnthetubaguy: in what context? | 16:27 |
| thangp | np :) | 16:27 |
| *** Administrator__ has quit IRC | 16:27 | |
| johnthetubaguy | mriedem: good question, was just looking at the patches that are up, the API is WIP, and the spec says no API impact, which on reflect can't have ever been true | 16:27 |
| mriedem | johnthetubaguy: link to the spec? | 16:28 |
| johnthetubaguy | mriedem: I think it was a fast approve from last release, so that got missed | 16:28 |
| *** karimb has quit IRC | 16:28 | |
| johnthetubaguy | mriedem: http://specs.openstack.org/openstack/nova-specs/specs/liberty/approved/multi-attach-volume.html#rest-api-impact | 16:28 |
| *** Administrator_ has joined #openstack-nova | 16:28 | |
| *** matrohon has quit IRC | 16:28 | |
| johnthetubaguy | is it only me that gerrit is being really slow for, by the way? | 16:29 |
| *** rlrossit has left #openstack-nova | 16:29 | |
| mriedem | it's a bit slow | 16:29 |
| dansmith | it's slow because it's tuesday | 16:29 |
| dansmith | or rather, it's slow because the day name ends with "-day" | 16:29 |
| melwitt | slow gerrit tuesdays? | 16:29 |
| dansmith | melwitt: sounds like a wing place promotion | 16:29 |
| johnthetubaguy | yeah, I should do more reviews in the morning, its quite performant at that time of day | 16:30 |
| melwitt | dansmith: that's a good idea | 16:30 |
| mriedem | johnthetubaguy: yeah, so https://review.openstack.org/#/c/153038/ is definitely an API impact, | 16:31 |
| mriedem | and is adding the support to the v2 API | 16:31 |
| *** Administrator__ has joined #openstack-nova | 16:31 | |
| *** Kevin_Zheng has quit IRC | 16:31 | |
| *** meghal has joined #openstack-nova | 16:32 | |
| openstackgerrit | Sean Dague proposed openstack/nova: Document when we should have a microversion https://review.openstack.org/191188 | 16:33 |
| *** Kevin_Zheng has joined #openstack-nova | 16:34 | |
| *** jasondotstar has joined #openstack-nova | 16:34 | |
| openstackgerrit | Marc Fouché proposed openstack/nova-specs: Inheritable admin image properties https://review.openstack.org/206431 | 16:34 |
| *** Administrator_ has quit IRC | 16:35 | |
| *** tpeoples has joined #openstack-nova | 16:35 | |
| *** Administrator__ has quit IRC | 16:36 | |
| *** Administrator_ has joined #openstack-nova | 16:36 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/nova: Ironic: Do not wait for the node to be unprovisioned https://review.openstack.org/206614 | 16:36 |
| bauzas | johnthetubaguy: Zuul is also liking the EU mornings :) | 16:37 |
| *** annashen has joined #openstack-nova | 16:38 | |
| mriedem | johnthetubaguy: so -1 on https://review.openstack.org/#/c/153038/ with a few comments | 16:38 |
| *** Kevin_Zheng has quit IRC | 16:39 | |
| *** Administrator__ has joined #openstack-nova | 16:40 | |
| *** thangp has quit IRC | 16:40 | |
| melwitt | been seeing activity on this lately, whether or not we allow detach root device volume in shutoff state https://bugs.launchpad.net/nova/+bug/1396965 apparently we used to allow it, then removed it | 16:40 |
| openstack | Launchpad bug 1396965 in OpenStack Compute (nova) "Add capability to detach root device volume of an instance, when in shutoff state" [Wishlist,Opinion] - Assigned to swapnil (swap-kamble) | 16:40 |
| *** lucas__ has joined #openstack-nova | 16:40 | |
| *** thangp has joined #openstack-nova | 16:41 | |
| *** jordanP has quit IRC | 16:41 | |
| mriedem | ndipanov: ^ | 16:41 |
| dansmith | melwitt: yeah, we talked about that at the midcycle | 16:41 |
| johnthetubaguy | mriedem: yeah, cool, I was just wondering if this was hitting peoples radar yet, I am just looking at the low blueprints that have been up for a long time again, this is sure one of those | 16:41 |
| *** lucas__ has quit IRC | 16:41 | |
| dansmith | melwitt: and the summit | 16:41 |
| melwitt | dansmith: oh. I must not have understood it at all because I don't remember that | 16:41 |
| *** lucas__ has joined #openstack-nova | 16:41 | |
| dansmith | melwitt: we "allowed" it before in that it slipped through with minimal checking, but had some major caveats, and now we have checks in place to make it fail clearly | 16:41 |
| *** lpetrut has joined #openstack-nova | 16:42 | |
| *** unicell has joined #openstack-nova | 16:43 | |
| *** Administrator_ has quit IRC | 16:43 | |
| melwitt | dansmith: interesting, I'll check out the patch that added the checks and update the lp bug | 16:43 |
| *** Kevin_Zheng has joined #openstack-nova | 16:43 | |
| dansmith | melwitt: you should talk to PaulMurray | 16:45 |
| *** Administrator__ has quit IRC | 16:45 | |
| *** Administrator_ has joined #openstack-nova | 16:45 | |
| *** lucas__ has quit IRC | 16:46 | |
| *** Administrator__ has joined #openstack-nova | 16:48 | |
| *** Kevin_Zheng has quit IRC | 16:48 | |
| melwitt | dansmith: is he working on something related to it? | 16:50 |
| dansmith | melwitt: yeah | 16:50 |
| dansmith | he wants to work on re-enabling it with proper safeguards | 16:50 |
| *** Administrator_ has quit IRC | 16:51 | |
| melwitt | dansmith: oh, cool. I'll ask him about it | 16:51 |
| dansmith | bauzas: want to look at my feedback on your patch just now and see if it makes sense while I still have it in my head? | 16:51 |
| *** dims has quit IRC | 16:52 | |
| *** rlrossit has joined #openstack-nova | 16:52 | |
| *** Kevin_Zheng has joined #openstack-nova | 16:52 | |
| *** dims has joined #openstack-nova | 16:52 | |
| *** dims has quit IRC | 16:52 | |
| *** lucas__ has joined #openstack-nova | 16:53 | |
| *** dims has joined #openstack-nova | 16:53 | |
| bauzas | dansmith: oh sure | 16:53 |
| bauzas | dansmith: lemme check your thoughts | 16:53 |
| dansmith | don't check my thoughts, just the review | 16:54 |
| * dansmith puts on his tinfoil hat | 16:54 | |
| *** jaypipes has quit IRC | 16:54 | |
| bauzas | call me mandrake | 16:54 |
| *** Administrator_ has joined #openstack-nova | 16:54 | |
| *** jaypipes has joined #openstack-nova | 16:54 | |
| *** Administrator__ has quit IRC | 16:54 | |
| *** leeantho has joined #openstack-nova | 16:54 | |
| openstackgerrit | Tracy Jones proposed openstack/nova: Cleanup of compute api reboot method https://review.openstack.org/127427 | 16:56 |
| *** Administrator__ has joined #openstack-nova | 16:56 | |
| bauzas | dansmith: well, nothing special to complain :) | 16:56 |
| bauzas | dansmith: your comments are good :) | 16:56 |
| bauzas | dansmith: the obj_set_defaults() was limited to having a new object hydrated | 16:57 |
| openstackgerrit | Marian Horban proposed openstack/nova: libvirt:Rsync remote FS driver was added https://review.openstack.org/146887 | 16:57 |
| *** annashen has quit IRC | 16:57 | |
| dansmith | bauzas: yeah, I think that's kosher | 16:57 |
| *** whenry has joined #openstack-nova | 16:57 | |
| *** Kevin_Zheng has quit IRC | 16:58 | |
| bauzas | dansmith: for the moment, we don't persist the limits | 16:58 |
| *** Kevin_Zheng has joined #openstack-nova | 16:58 | |
| dansmith | right | 16:58 |
| dansmith | hence the (eventually) | 16:58 |
| *** toure is now known as toure|biab | 16:58 | |
| bauzas | dansmith: but even if we would persist those limits, having those set with defaults is okay to me because we really hydrate from a primitive | 16:58 |
| dansmith | yes | 16:59 |
| bauzas | dansmith: that's a from_dict() stuff so we assume to create a new object | 16:59 |
| *** toure|biab is now known as toure | 16:59 | |
| dansmith | and they're a blob, which means we're always writing the whole thing, sparse or not | 16:59 |
| bauzas | from a primitive I mean | 16:59 |
| *** armax has joined #openstack-nova | 16:59 | |
| *** Administrator_ has quit IRC | 16:59 | |
| bauzas | dansmith: well, the primitive is really weak | 16:59 |
| *** annashen has joined #openstack-nova | 16:59 | |
| bauzas | dansmith: we don't really know what we'll get from it | 17:00 |
| *** BobBall is now known as BobBall_AWOL | 17:00 | |
| *** Administrator_ has joined #openstack-nova | 17:00 | |
| bauzas | dansmith: but we assume that the object will assume to know all the limits and persist them correctly if we want | 17:00 |
| dansmith | right | 17:00 |
| dansmith | but my point is, you never have the case where you would need to write just obj.foo and obj.baz to the database without touching foo.bar, | 17:00 |
| *** Administrator__ has quit IRC | 17:00 | |
| bauzas | dansmith: I'm just thinking of the next step where alaski or other could just want to persist the Limits object | 17:01 |
| dansmith | because you're going to store these as a blob in the DB, which means you always write the whole thing, yes? | 17:01 |
| bauzas | dansmith: mmm, I need to consider that before answering | 17:01 |
| *** coolsvap|dinner is now known as coolsvap|mt | 17:01 | |
| *** coolsvap|mt is now known as coolsvap|mtg | 17:01 | |
| *** dboik has quit IRC | 17:01 | |
| bauzas | dansmith: so, the limits are there for being given by the Scheduler | 17:01 |
| bauzas | dansmith: each filter can set a limit | 17:02 |
| *** Administrator__ has joined #openstack-nova | 17:02 | |
| dansmith | hmm | 17:02 |
| bauzas | dansmith: but this is not persisted by the scheduler | 17:03 |
| dansmith | I don't understand "each filter can set a limit" | 17:03 |
| bauzas | dansmith: sec, showing you | 17:03 |
| *** Kevin_Zheng has quit IRC | 17:03 | |
| *** exploreshaifali has joined #openstack-nova | 17:04 | |
| ndipanov | bauzas, limits are always the same across restarts right - they mostly come from config options | 17:04 |
| bauzas | ndipanov: right | 17:05 |
| ndipanov | note that limits are super racy in the current implementation | 17:05 |
| ndipanov | they can be clobbered on the host_state at any point | 17:05 |
| ndipanov | and the last one will be sent to all | 17:05 |
| bauzas | ndipanov: the thing is, we use limits on 2 sides, in the conductor (using the filter_props) and in the scheduler using the HostState | 17:05 |
| *** Kevin_Zheng has joined #openstack-nova | 17:05 | |
| *** Administrator_ has quit IRC | 17:05 | |
| *** lpetrut has quit IRC | 17:06 | |
| *** vilobhmm has joined #openstack-nova | 17:06 | |
| bauzas | ndipanov: but the scheduler doesn't give the limits back to the conductor, right? | 17:06 |
| *** danpb has quit IRC | 17:06 | |
| ndipanov | basically scheduler is like mr. burns in an old simpsons episode where he has all diseases in the world that together keep him alive | 17:06 |
| bauzas | ndipanov: I mean, the scheduler fences off the limits using the HostManager, | 17:06 |
| bauzas | ndipanov: but it doesn't pass back the limits to the conductor | 17:06 |
| ndipanov | bauzas, limits are needed for claims | 17:06 |
| bauzas | ndipanov: I like the analogy | 17:06 |
| ndipanov | they need to be sent to computes | 17:07 |
| bauzas | ndipanov: right, I know that but they are not given by the scheduler | 17:07 |
| ndipanov | they are | 17:07 |
| ndipanov | set by the filters | 17:07 |
| *** Administrator__ has quit IRC | 17:07 | |
| bauzas | ndipanov: from the returned tuple of select_dests() ? | 17:07 |
| bauzas | ndipanov: lemme check quickyl | 17:07 |
| ndipanov | they are set on the host_state | 17:08 |
| vladikr | dansmith, Hi, just wondering, was there already a conversation about libosinfo | 17:08 |
| vladikr | ? | 17:08 |
| *** Administrator_ has joined #openstack-nova | 17:08 | |
| *** shaohe_robot has quit IRC | 17:08 | |
| dansmith | vladikr: I think the plan was to schedule something this week, but I wasn't in the group of potential attendees, so I'm not sure | 17:08 |
| bauzas | ndipanov: oh man, you're correct, they're part of the return payload | 17:09 |
| dansmith | vladikr: I would check with johnthetubaguy danpb sdague jaypipes | 17:09 |
| bauzas | https://github.com/openstack/nova/blob/master/nova/scheduler/filter_scheduler.py#L91-L92 | 17:09 |
| vladikr | dansmith, I see.. | 17:09 |
| *** eliqiao has joined #openstack-nova | 17:09 | |
| ndipanov | bauzas, and racily set on the host state | 17:09 |
| ndipanov | it's a mess | 17:09 |
| bauzas | thinking about that quickly | 17:10 |
| vladikr | dansmith, do you think it will still get merged, even if it will pass the deadline on the 30th? | 17:10 |
| *** shaohe_robot has joined #openstack-nova | 17:10 | |
| bauzas | so, we have something obviously empty when hydrating the Spec | 17:10 |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Replace timeutils.isotime() with utcnow() and isoformat() https://review.openstack.org/206624 | 17:10 |
| *** mpaolino has quit IRC | 17:10 | |
| vladikr | dansmith, given that people will agree on a solution.. | 17:10 |
| mriedem | vladikr: link? | 17:10 |
| bauzas | it goes to the scheduler, then the scheduler racily updates the HostState.limits and then passes that back to the conductor | 17:10 |
| vladikr | mriedem, https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/libvirt-hardware-policy-from-libosinfo,n,z | 17:11 |
| dansmith | vladikr: I dunno, depends on the outcome I guess.. maybe the outcome will be "a feature freeze exception" | 17:11 |
| *** lucas__ has quit IRC | 17:11 | |
| bauzas | by any time, it doesn't update the Spec obj | 17:11 |
| vladikr | dansmith, ok, got it | 17:11 |
| *** Kevin_Zheng has quit IRC | 17:11 | |
| *** Administrator__ has joined #openstack-nova | 17:11 | |
| bauzas | that's only when we would pass the spec_obj in the conductor that we would update the Spec with those values | 17:12 |
| bauzas | dansmith: ndipanov: ^ | 17:12 |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Replace timeutils.isotime() with datetime.isoformat() https://review.openstack.org/206624 | 17:12 |
| bauzas | I'm not sure you followed me | 17:12 |
| mriedem | vladikr: did the -2 here get resolved? https://review.openstack.org/#/c/149625/ | 17:12 |
| dansmith | I didn't | 17:12 |
| dansmith | mriedem: that's what he's asking about | 17:12 |
| dansmith | mriedem: the convo that was supposed to happen to resolve it | 17:12 |
| bauzas | dansmith: okay, rewinding | 17:12 |
| mriedem | yeah, i don't see that happening before thursday | 17:13 |
| dansmith | especially with jaypipes being tied up until tomorrow afternoon | 17:13 |
| bauzas | dansmith: so, in the last patch of the series, I'm hydrating the Spec object using the from_prims() method in the conductor, right before calling select_dests() | 17:13 |
| *** Kevin_Zheng has joined #openstack-nova | 17:14 | |
| openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: set caps on maximum live migration time https://review.openstack.org/162254 | 17:14 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: support management of downtime during migration https://review.openstack.org/162253 | 17:14 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: ensure LibvirtConfigGuestDisk parses readonly/shareable flags https://review.openstack.org/206630 | 17:14 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: add helper methods for getting guest devices/disks https://review.openstack.org/206631 | 17:14 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: take account of disks in migration data size https://review.openstack.org/206632 | 17:14 |
| dansmith | burt: yeah | 17:14 |
| dansmith | oops | 17:14 |
| dansmith | bauzas: yeah | 17:14 |
| *** doude has quit IRC | 17:14 | |
| *** Administrator_ has quit IRC | 17:14 | |
| bauzas | dansmith: here, the thing is, even for migrations, we are passing the limits from the scheduler to the conductor and then to the RT for claiming | 17:15 |
| bauzas | dansmith: at the moment, there is no persistence | 17:15 |
| dansmith | yeah, and RT will use the limits to write the claims? | 17:15 |
| bauzas | dansmith: but Claims are not persisted too :) | 17:16 |
| *** davideagnello has quit IRC | 17:16 | |
| dansmith | ...yet | 17:16 |
| bauzas | dansmith: zactly | 17:16 |
| *** davideagnello has joined #openstack-nova | 17:16 | |
| dansmith | but this all goes to the "not a problem here" point right? | 17:16 |
| *** Administrator__ has quit IRC | 17:16 | |
| *** Administrator_ has joined #openstack-nova | 17:16 | |
| bauzas | dansmith: zactly (bis) | 17:16 |
| dansmith | okay | 17:16 |
| bauzas | dansmith: because either we store the limits or not, but by each time, we should update those as a whole | 17:17 |
| *** Administrator__ has joined #openstack-nova | 17:17 | |
| bauzas | dansmith: ie. given the scheduler should return back the limits, we should update the Spec.SchedulerLimits at once | 17:17 |
| bauzas | dansmith: and then, persist that | 17:18 |
| dansmith | yep | 17:18 |
| bauzas | so, yeah, I don't think it's a problem | 17:18 |
| openstackgerrit | Anthony Lee proposed openstack/nova: Fix live-migrations usage of the wrong connector information https://review.openstack.org/202770 | 17:18 |
| dansmith | so I'm confused.. I said this didn't look like a problem in the beginning, are we arguing or in violent agreement? | 17:18 |
| dansmith | okay | 17:18 |
| *** artom_ has joined #openstack-nova | 17:19 | |
| *** artom has quit IRC | 17:19 | |
| bauzas | dansmith: well, I had to think a bit more because it wasn't really clear on whether we were passing the limits or not | 17:19 |
| dansmith | okay | 17:19 |
| bauzas | but ndipanov helped me on that :) | 17:19 |
| *** Kevin_Zheng has quit IRC | 17:19 | |
| bauzas | dansmith: thanks for having reviewed that | 17:20 |
| dansmith | yep | 17:20 |
| ndipanov | imho limits should never ever have been in the scheduler | 17:20 |
| *** boris-42 has quit IRC | 17:20 | |
| ndipanov | jaypipes, had a bp to remove them | 17:20 |
| *** moshele has joined #openstack-nova | 17:20 | |
| *** Administrator_ has quit IRC | 17:20 | |
| *** neiljerr` has joined #openstack-nova | 17:20 | |
| *** sudipto has quit IRC | 17:20 | |
| *** mestery has quit IRC | 17:20 | |
| *** otter768 has joined #openstack-nova | 17:20 | |
| *** ssurana has joined #openstack-nova | 17:20 | |
| ndipanov | but not so easy because of our couple everything to everything and support live upgrades policy | 17:20 |
| *** Kevin_Zheng has joined #openstack-nova | 17:21 | |
| *** alex_xu has quit IRC | 17:21 | |
| *** neiljerram has quit IRC | 17:21 | |
| bauzas | ndipanov: yeah | 17:22 |
| *** Administrator_ has joined #openstack-nova | 17:22 | |
| *** boris-42 has joined #openstack-nova | 17:22 | |
| *** Administrator__ has quit IRC | 17:23 | |
| *** alex_xu has joined #openstack-nova | 17:23 | |
| *** Administrator__ has joined #openstack-nova | 17:24 | |
| *** otter768 has quit IRC | 17:25 | |
| *** Kevin_Zheng has quit IRC | 17:25 | |
| *** lucas__ has joined #openstack-nova | 17:26 | |
| *** neiljerr` has quit IRC | 17:26 | |
| *** Administrator_ has quit IRC | 17:27 | |
| *** meghal has quit IRC | 17:27 | |
| *** Administrator_ has joined #openstack-nova | 17:28 | |
| ndipanov | melwitt, mriedem dansmith there are cases when we may want to do that, but they need to be either carefully considered, or not supported at all | 17:29 |
| ndipanov | imho it's a solution looking for a problem | 17:29 |
| *** patrickeast has joined #openstack-nova | 17:29 | |
| ndipanov | but that's just me | 17:29 |
| *** annegentle has joined #openstack-nova | 17:30 | |
| mriedem | how did i get roped into this conversation? | 17:30 |
| mriedem | :P | 17:30 |
| ndipanov | hehe | 17:30 |
| *** romainh has quit IRC | 17:31 | |
| dansmith | ndipanov: you mean detach the root volume? | 17:31 |
| ndipanov | yes | 17:31 |
| mriedem | oh, i thought you were talking about limits + scheduler | 17:31 |
| dansmith | uses for detaching the root volume all seem to come from legit problems | 17:31 |
| dansmith | so I think it's a problem in need of a solution, myself | 17:31 |
| ndipanov | mriedem, that's a problem looking for more problems and they always show up | 17:31 |
| *** Administrator__ has quit IRC | 17:31 | |
| dansmith | but I also wish people wouldn't need it | 17:31 |
| *** Kevin_Zheng has joined #openstack-nova | 17:32 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Move to using ovo's remotable decorators https://review.openstack.org/202209 | 17:32 |
| openstackgerrit | Dan Smith proposed openstack/nova: Base NovaObject on VersionedObject https://review.openstack.org/201655 | 17:32 |
| sdague | mriedem: 193223,13 is in merge conflict with the gate | 17:32 |
| mriedem | sdague: that whole series' sh is f'ed | 17:33 |
| mriedem | i'm rebasing it locally now | 17:33 |
| sdague | mriedem: gotcha | 17:33 |
| *** romainh has joined #openstack-nova | 17:33 | |
| *** rdopiera has quit IRC | 17:33 | |
| *** Administrator__ has joined #openstack-nova | 17:34 | |
| *** Administrator_ has quit IRC | 17:34 | |
| *** annegentle has quit IRC | 17:35 | |
| *** yassine__ has quit IRC | 17:36 | |
| *** Administrator_ has joined #openstack-nova | 17:36 | |
| *** Kevin_Zheng has quit IRC | 17:37 | |
| *** Kevin_Zheng has joined #openstack-nova | 17:38 | |
| *** andrewbogott has joined #openstack-nova | 17:38 | |
| *** Administrator__ has quit IRC | 17:39 | |
| *** Administrator__ has joined #openstack-nova | 17:40 | |
| *** romainh has left #openstack-nova | 17:40 | |
| *** Administrator_ has quit IRC | 17:40 | |
| *** baoli has joined #openstack-nova | 17:41 | |
| *** achanda has joined #openstack-nova | 17:42 | |
| *** Administrator_ has joined #openstack-nova | 17:42 | |
| *** annashen has quit IRC | 17:42 | |
| *** Kevin_Zheng has quit IRC | 17:43 | |
| *** Kevin_Zheng has joined #openstack-nova | 17:45 | |
| *** Administrator__ has quit IRC | 17:45 | |
| *** garyk has quit IRC | 17:45 | |
| *** eliqiao has quit IRC | 17:45 | |
| dims | nova cores...i've updated "Currently reviewable" list at https://etherpad.openstack.org/p/liberty-nova-priorities-tracking | 17:46 |
| *** lucasagomes is now known as lucas-dinner | 17:47 | |
| *** eliqiao has joined #openstack-nova | 17:47 | |
| *** moshele has quit IRC | 17:47 | |
| *** Administrator_ has quit IRC | 17:47 | |
| *** markvoelker has quit IRC | 17:47 | |
| *** Administrator_ has joined #openstack-nova | 17:48 | |
| *** amotoki has quit IRC | 17:48 | |
| *** ssurana has left #openstack-nova | 17:49 | |
| *** meghal has joined #openstack-nova | 17:50 | |
| *** Administrator__ has joined #openstack-nova | 17:51 | |
| *** haomaiwang has quit IRC | 17:51 | |
| *** Kevin_Zheng has quit IRC | 17:51 | |
| *** annashen has joined #openstack-nova | 17:52 | |
| *** haomaiwang has joined #openstack-nova | 17:52 | |
| *** baoli has quit IRC | 17:52 | |
| *** baoli has joined #openstack-nova | 17:53 | |
| *** mestery has joined #openstack-nova | 17:54 | |
| *** Kevin_Zheng has joined #openstack-nova | 17:54 | |
| *** armax has quit IRC | 17:54 | |
| *** Administrator_ has quit IRC | 17:54 | |
| *** fawadkhaliq has quit IRC | 17:55 | |
| *** agireud has quit IRC | 17:55 | |
| *** mestery has quit IRC | 17:55 | |
| *** annashen has quit IRC | 17:56 | |
| *** Administrator__ has quit IRC | 17:56 | |
| *** Administrator_ has joined #openstack-nova | 17:56 | |
| *** changbl has quit IRC | 17:57 | |
| *** Administrator__ has joined #openstack-nova | 17:57 | |
| *** dboik has joined #openstack-nova | 17:57 | |
| mriedem | chuckcarmack: here is a low hanging fruit cleanup bug https://bugs.launchpad.net/nova/+bug/1479066 | 17:58 |
| openstack | Launchpad bug 1479066 in OpenStack Compute (nova) "DeprecationWarning: BaseException.message has been deprecated as of Python 2.6" [Undecided,New] | 17:58 |
| *** jasondotstar has quit IRC | 17:58 | |
| *** Piet has quit IRC | 17:58 | |
| mriedem | assign it to yourself if you plan on working on it | 17:58 |
| chuckcarmack | ok | 17:59 |
| *** Kevin_Zheng has quit IRC | 17:59 | |
| *** lpetrut has joined #openstack-nova | 18:00 | |
| *** kylek3h has joined #openstack-nova | 18:00 | |
| *** mriedem is now known as mriedem_meeting | 18:00 | |
| *** Administrator_ has quit IRC | 18:01 | |
| *** Kevin_Zheng has joined #openstack-nova | 18:01 | |
| *** neiljerr` has joined #openstack-nova | 18:01 | |
| *** achanda has quit IRC | 18:02 | |
| *** achanda has joined #openstack-nova | 18:02 | |
| *** Administrator_ has joined #openstack-nova | 18:02 | |
| *** Administrator__ has quit IRC | 18:03 | |
| *** fawadkhaliq has joined #openstack-nova | 18:04 | |
| *** shaohe_robot has quit IRC | 18:04 | |
| *** Administrator__ has joined #openstack-nova | 18:04 | |
| *** shaohe_robot has joined #openstack-nova | 18:04 | |
| *** alex_xu has quit IRC | 18:04 | |
| *** penick has quit IRC | 18:05 | |
| *** fawadkhaliq has quit IRC | 18:05 | |
| *** Kevin_Zheng has quit IRC | 18:05 | |
| *** xyang1 has quit IRC | 18:05 | |
| *** Administrator_ has quit IRC | 18:07 | |
| *** abhijeetm has joined #openstack-nova | 18:07 | |
| *** Administrator_ has joined #openstack-nova | 18:08 | |
| *** romainh has joined #openstack-nova | 18:08 | |
| *** lucas__ has quit IRC | 18:09 | |
| *** alex_xu has joined #openstack-nova | 18:09 | |
| *** lucas__ has joined #openstack-nova | 18:09 | |
| *** changbl has joined #openstack-nova | 18:10 | |
| *** Administrator__ has quit IRC | 18:11 | |
| *** davideag_ has joined #openstack-nova | 18:11 | |
| *** abhijeetm has quit IRC | 18:12 | |
| *** Kevin_Zheng has joined #openstack-nova | 18:12 | |
| *** davideagnello has quit IRC | 18:12 | |
| *** meghal has quit IRC | 18:12 | |
| *** armax has joined #openstack-nova | 18:13 | |
| *** davideagnello has joined #openstack-nova | 18:13 | |
| *** rdopiera has joined #openstack-nova | 18:14 | |
| *** Administrator__ has joined #openstack-nova | 18:14 | |
| *** Administrator_ has quit IRC | 18:15 | |
| *** davideag_ has quit IRC | 18:16 | |
| *** Administrator_ has joined #openstack-nova | 18:16 | |
| *** rfolco has quit IRC | 18:17 | |
| *** davideagnello has quit IRC | 18:17 | |
| *** davideagnello has joined #openstack-nova | 18:17 | |
| *** meghal has joined #openstack-nova | 18:17 | |
| *** Kevin_Zheng has quit IRC | 18:17 | |
| *** Kevin_Zheng has joined #openstack-nova | 18:18 | |
| *** renatoarmani has joined #openstack-nova | 18:19 | |
| *** lucas__ has quit IRC | 18:19 | |
| *** Administrator__ has quit IRC | 18:19 | |
| *** penick has joined #openstack-nova | 18:19 | |
| *** lucas__ has joined #openstack-nova | 18:19 | |
| *** Administrator__ has joined #openstack-nova | 18:20 | |
| *** Administrator_ has quit IRC | 18:20 | |
| *** lpetrut has quit IRC | 18:20 | |
| *** vishwanathj has joined #openstack-nova | 18:21 | |
| *** davideagnello has quit IRC | 18:22 | |
| *** Administrator_ has joined #openstack-nova | 18:22 | |
| *** Kevin_Zheng has quit IRC | 18:23 | |
| *** baoli has quit IRC | 18:23 | |
| *** davideagnello has joined #openstack-nova | 18:23 | |
| *** vishwana_ has joined #openstack-nova | 18:25 | |
| *** vishwanathj has quit IRC | 18:25 | |
| *** Kevin_Zheng has joined #openstack-nova | 18:25 | |
| *** Administrator__ has quit IRC | 18:25 | |
| *** annashen has joined #openstack-nova | 18:27 | |
| *** vishwana_ has quit IRC | 18:27 | |
| *** xyang1 has joined #openstack-nova | 18:27 | |
| *** Administrator_ has quit IRC | 18:27 | |
| *** Administrator_ has joined #openstack-nova | 18:28 | |
| *** haomaiwang has quit IRC | 18:29 | |
| *** haomaiwang has joined #openstack-nova | 18:30 | |
| *** moshele has joined #openstack-nova | 18:31 | |
| *** Kevin_Zheng has quit IRC | 18:31 | |
| *** Administrator__ has joined #openstack-nova | 18:31 | |
| *** Marga_ has quit IRC | 18:33 | |
| *** Kevin_Zheng has joined #openstack-nova | 18:34 | |
| *** Administrator_ has quit IRC | 18:34 | |
| *** artom_ has quit IRC | 18:35 | |
| *** artom has joined #openstack-nova | 18:36 | |
| *** Administrator__ has quit IRC | 18:36 | |
| *** Administrator_ has joined #openstack-nova | 18:36 | |
| *** rdo has joined #openstack-nova | 18:37 | |
| *** Marga_ has joined #openstack-nova | 18:37 | |
| *** Administrator__ has joined #openstack-nova | 18:37 | |
| moshele | can nova core review https://review.openstack.org/#/c/187052/ and https://review.openstack.org/#/c/199488/ ? | 18:37 |
| *** josecastroleon has joined #openstack-nova | 18:38 | |
| *** Kevin_Zheng has quit IRC | 18:39 | |
| *** fawadkhaliq has joined #openstack-nova | 18:39 | |
| openstackgerrit | Nicolas Simonds proposed openstack/nova: libvirt: make snapshot use RBD snapshot/clone when available https://review.openstack.org/205282 | 18:40 |
| *** Kevin_Zheng has joined #openstack-nova | 18:40 | |
| *** Administrator_ has quit IRC | 18:40 | |
| *** Piet has joined #openstack-nova | 18:41 | |
| *** markvoelker has joined #openstack-nova | 18:42 | |
| *** Administrator_ has joined #openstack-nova | 18:42 | |
| *** Administrator__ has quit IRC | 18:43 | |
| *** Administrator__ has joined #openstack-nova | 18:44 | |
| *** ddaskal has quit IRC | 18:44 | |
| *** Kevin_Zheng has quit IRC | 18:45 | |
| *** markvoelker_ has joined #openstack-nova | 18:46 | |
| *** burt has quit IRC | 18:47 | |
| *** Administrator_ has quit IRC | 18:48 | |
| *** markvoelker has quit IRC | 18:48 | |
| *** Administrator_ has joined #openstack-nova | 18:48 | |
| *** mragupat has joined #openstack-nova | 18:49 | |
| *** Administrator__ has quit IRC | 18:52 | |
| *** baoli has joined #openstack-nova | 18:52 | |
| *** Kevin_Zheng has joined #openstack-nova | 18:52 | |
| *** Administrator__ has joined #openstack-nova | 18:54 | |
| *** Administrator_ has quit IRC | 18:54 | |
| *** baoli has quit IRC | 18:55 | |
| *** baoli has joined #openstack-nova | 18:55 | |
| *** rfolco has joined #openstack-nova | 18:56 | |
| *** Administrator_ has joined #openstack-nova | 18:56 | |
| *** Kevin_Zheng has quit IRC | 18:57 | |
| *** Kevin_Zheng has joined #openstack-nova | 18:58 | |
| *** Administrator__ has quit IRC | 18:59 | |
| *** Administrator__ has joined #openstack-nova | 19:00 | |
| *** Administrator_ has quit IRC | 19:01 | |
| *** Administrator_ has joined #openstack-nova | 19:02 | |
| dansmith | rlrossit: looks like I made pep8 fail for you | 19:02 |
| dansmith | rlrossit: you're welcome. | 19:02 |
| rlrossit | dansmith: thanks a lot, bud | 19:02 |
| dansmith | np | 19:02 |
| *** Kevin_Zheng has quit IRC | 19:03 | |
| rlrossit | looks like grenade finished up though | 19:03 |
| dansmith | not the partial job | 19:03 |
| dansmith | which is the one that fails | 19:03 |
| rlrossit | It did on the second patch which is basically the same thing | 19:03 |
| dansmith | ah | 19:04 |
| dansmith | but you can't see those logs until everything else finished | 19:04 |
| dansmith | I have an 11h wait on the patches I just submitted | 19:04 |
| dansmith | so I guess it's beer o clock already | 19:04 |
| *** ijw has quit IRC | 19:05 | |
| rlrossit | yeah I've been super lazy today seeing as these things are taking ages to make it through | 19:05 |
| *** Administrator__ has quit IRC | 19:05 | |
| rlrossit | what's up with zuul today? | 19:05 |
| dansmith | Administrator_: yo | 19:05 |
| dansmith | rlrossit: it's busy | 19:05 |
| *** Kevin_Zheng has joined #openstack-nova | 19:05 | |
| *** Piet has quit IRC | 19:05 | |
| dansmith | Kevin_Zheng: ping | 19:05 |
| *** chuckcarmack has left #openstack-nova | 19:05 | |
| *** chuckcarmack has joined #openstack-nova | 19:06 | |
| *** zzzeek has quit IRC | 19:06 | |
| *** Administrator_ has quit IRC | 19:07 | |
| *** hemna is now known as hemnalunch | 19:07 | |
| *** ijw has joined #openstack-nova | 19:08 | |
| *** zzzeek has joined #openstack-nova | 19:08 | |
| *** josecastroleon has quit IRC | 19:08 | |
| rlrossit | dansmith: go nuts, let's see who finds what went wrong first | 19:08 |
| rlrossit | http://logs.openstack.org/04/198904/5/check/gate-grenade-dsvm-partial-ncpu/594d5a3/logs/ | 19:08 |
| *** Administrator_ has joined #openstack-nova | 19:08 | |
| dansmith | fsck | 19:09 |
| *** armax has quit IRC | 19:09 | |
| rlrossit | dansmith: wut.... | 19:09 |
| dansmith | I suck\ | 19:09 |
| rlrossit | things didn't blow up? | 19:09 |
| dansmith | 4.5 hours wasted | 19:10 |
| rlrossit | *headdesk* | 19:10 |
| dansmith | e.kwargs['objname'] is what I meant to do I think | 19:10 |
| *** fawadk has joined #openstack-nova | 19:11 | |
| *** Kevin_Zheng has quit IRC | 19:11 | |
| *** Administrator__ has joined #openstack-nova | 19:11 | |
| dansmith | Administrator_: yo | 19:12 |
| *** chhavi has quit IRC | 19:12 | |
| *** fawadkhaliq has quit IRC | 19:12 | |
| *** ijw has quit IRC | 19:12 | |
| *** Kevin_Zheng has joined #openstack-nova | 19:13 | |
| *** moshele has quit IRC | 19:13 | |
| dansmith | okay, so Kevin_Zheng appears to be signed in three times from the same machine, which has been flapping on and off irc all morning | 19:14 |
| dansmith | all in favor of a ban with a note? | 19:14 |
| *** moshele has joined #openstack-nova | 19:14 | |
| *** Administrator_ has quit IRC | 19:14 | |
| ctrath | is that Administrator_ ? | 19:15 |
| dansmith | as Kevin_Zhang, Administrator_ and Administrator__ | 19:15 |
| dansmith | yeah | 19:15 |
| melwitt | nice | 19:15 |
| *** Administrator_ has joined #openstack-nova | 19:15 | |
| ctrath | seems suspect | 19:15 |
| *** Administrator__ has quit IRC | 19:16 | |
| *** Administrator__ has joined #openstack-nova | 19:16 | |
| *** ijw has joined #openstack-nova | 19:17 | |
| *** Piet has joined #openstack-nova | 19:17 | |
| *** Kevin_Zheng has quit IRC | 19:18 | |
| *** Administrator_ has quit IRC | 19:19 | |
| *** Kevin_Zheng has joined #openstack-nova | 19:19 | |
| *** Administrator_ has joined #openstack-nova | 19:20 | |
| *** Administrator__ has quit IRC | 19:20 | |
| dansmith | sdague: you're usually good for +1 on a ban | 19:20 |
| *** baoli has quit IRC | 19:21 | |
| *** ijw has quit IRC | 19:21 | |
| *** otter768 has joined #openstack-nova | 19:21 | |
| *** exploreshaifali has quit IRC | 19:21 | |
| *** Administrator__ has joined #openstack-nova | 19:22 | |
| *** Kevin_Zheng has quit IRC | 19:24 | |
| *** matrohon has joined #openstack-nova | 19:25 | |
| *** Kevin_Zheng has joined #openstack-nova | 19:25 | |
| *** Administrator_ has quit IRC | 19:26 | |
| *** otter768 has quit IRC | 19:26 | |
| mriedem_meeting | ban em all and let god sort em out | 19:27 |
| *** mriedem_meeting is now known as mriedem | 19:27 | |
| *** Administrator__ has quit IRC | 19:27 | |
| *** ChanServ sets mode: +o dansmith | 19:27 | |
| *** ddaskal has joined #openstack-nova | 19:28 | |
| *** Administrator_ has joined #openstack-nova | 19:28 | |
| *** markvoelker_ has quit IRC | 19:29 | |
| *** armax has joined #openstack-nova | 19:29 | |
| *** mestery has joined #openstack-nova | 19:30 | |
| *** annashen has quit IRC | 19:30 | |
| *** Administrator__ has joined #openstack-nova | 19:31 | |
| *** Kevin_Zheng has quit IRC | 19:31 | |
| *** dansmith sets mode: +b nickmask!Kevin_Zhen@205.177.226.216 | 19:32 | |
| *** dansmith sets mode: +b nickmask!Administr@205.177.226.216 | 19:33 | |
| *** baoli has joined #openstack-nova | 19:33 | |
| *** armax has quit IRC | 19:33 | |
| *** armax has joined #openstack-nova | 19:34 | |
| *** Kevin_Zheng has joined #openstack-nova | 19:34 | |
| *** dansmith sets mode: -b +b!*@* | 19:34 | |
| *** Administrator_ has quit IRC | 19:34 | |
| *** dansmith sets mode: +b nickmask!Administ*@205.177.226.216 | 19:34 | |
| *** suresh12 has joined #openstack-nova | 19:34 | |
| *** dansmith sets mode: -b nickmask!Kevin_Zhen@205.177.226.216 | 19:34 | |
| *** dansmith sets mode: -b nickmask!Administr@205.177.226.216 | 19:35 | |
| *** Piet has quit IRC | 19:36 | |
| *** lucas__ has quit IRC | 19:36 | |
| *** Administrator__ has quit IRC | 19:36 | |
| *** Administrator_ has joined #openstack-nova | 19:36 | |
| *** armax has quit IRC | 19:37 | |
| *** lucas__ has joined #openstack-nova | 19:37 | |
| *** Administrator__ has joined #openstack-nova | 19:37 | |
| *** Kevin_Zheng has quit IRC | 19:39 | |
| *** ivasev has quit IRC | 19:39 | |
| *** shaohe_robot1 has joined #openstack-nova | 19:40 | |
| *** annegentle has joined #openstack-nova | 19:40 | |
| *** Kevin_Zheng has joined #openstack-nova | 19:41 | |
| *** Administrator_ has quit IRC | 19:41 | |
| *** shaohe_robot has quit IRC | 19:42 | |
| *** Administrator_ has joined #openstack-nova | 19:42 | |
| *** Administrator__ has quit IRC | 19:43 | |
| *** jasondotstar has joined #openstack-nova | 19:44 | |
| *** Administrator__ has joined #openstack-nova | 19:44 | |
| *** Kevin_Zheng has quit IRC | 19:45 | |
| *** Administrator_ has quit IRC | 19:47 | |
| *** mpaolino has joined #openstack-nova | 19:47 | |
| *** moshele has quit IRC | 19:47 | |
| *** VW__ has quit IRC | 19:48 | |
| *** READ10 has quit IRC | 19:48 | |
| *** shaohe_robot has joined #openstack-nova | 19:48 | |
| *** Administrator_ has joined #openstack-nova | 19:48 | |
| *** VW_ has joined #openstack-nova | 19:48 | |
| *** shaohe_robot1 has quit IRC | 19:49 | |
| *** suresh12 has quit IRC | 19:50 | |
| *** casusbelli has joined #openstack-nova | 19:50 | |
| *** ijw has joined #openstack-nova | 19:51 | |
| *** suresh12 has joined #openstack-nova | 19:51 | |
| *** Administrator__ has quit IRC | 19:51 | |
| *** Kevin_Zheng has joined #openstack-nova | 19:52 | |
| *** annashen has joined #openstack-nova | 19:53 | |
| *** VW_ has quit IRC | 19:53 | |
| *** mpaolino has quit IRC | 19:54 | |
| *** Administrator__ has joined #openstack-nova | 19:54 | |
| *** Administrator_ has quit IRC | 19:54 | |
| *** ijw has quit IRC | 19:56 | |
| *** tonytan4ever has joined #openstack-nova | 19:56 | |
| *** Administrator_ has joined #openstack-nova | 19:56 | |
| *** annashen has quit IRC | 19:57 | |
| *** Kevin_Zheng has quit IRC | 19:57 | |
| *** tjones has quit IRC | 19:57 | |
| *** Kevin_Zheng has joined #openstack-nova | 19:58 | |
| dansmith | gah | 19:59 |
| *** Administrator__ has quit IRC | 19:59 | |
| jroll | dansmith: *!*@205.177.226.216 | 19:59 |
| *** renatoarmani has quit IRC | 19:59 | |
| dansmith | yeah, I was trying to be more selective in case that's the NAT gateway for someone | 19:59 |
| jroll | mmm | 20:00 |
| *** Administrator__ has joined #openstack-nova | 20:00 | |
| jroll | *!~Administr@205... | 20:00 |
| jroll | I guess | 20:00 |
| *** dansmith sets mode: +b *!*@205.177.226.216 | 20:00 | |
| *** Administrator_ has quit IRC | 20:01 | |
| *** VW_ has joined #openstack-nova | 20:02 | |
| *** Kevin_Zheng has quit IRC | 20:03 | |
| rlrossit | dansmith: do you want to reupload a different version of that logging stuff and we can just let zuul handle the crap overnight? | 20:05 |
| dansmith | rlrossit: oh I thought you were goin gto | 20:05 |
| *** Administrator__ has quit IRC | 20:05 | |
| *** ChanServ sets mode: -o dansmith | 20:05 | |
| rlrossit | dansmith: woops I thought you knew what to do so you were going to do it... | 20:06 |
| dansmith | I can | 20:06 |
| rlrossit | either way it won't be done before we leave :P | 20:06 |
| *** alexpilotti has joined #openstack-nova | 20:06 | |
| *** rushiagr is now known as rushiagr_away | 20:06 | |
| dansmith | there | 20:07 |
| *** mestery has quit IRC | 20:08 | |
| *** mdrabe has quit IRC | 20:10 | |
| *** tjones has joined #openstack-nova | 20:10 | |
| *** mdrabe has joined #openstack-nova | 20:10 | |
| *** pixelbeat_ has quit IRC | 20:12 | |
| *** alexpilotti has quit IRC | 20:12 | |
| openstackgerrit | Ryan Rossiter proposed openstack/nova: Fix test_relationships() for subobject versions https://review.openstack.org/198904 | 20:12 |
| openstackgerrit | Ryan Rossiter proposed openstack/nova: Change List objects to use obj_relationships https://review.openstack.org/198730 | 20:12 |
| rlrossit | dansmith: welp.... here we go again | 20:12 |
| *** baoli has quit IRC | 20:14 | |
| *** ijw has joined #openstack-nova | 20:14 | |
| *** baoli has joined #openstack-nova | 20:14 | |
| *** igordcard has joined #openstack-nova | 20:16 | |
| *** agireud has joined #openstack-nova | 20:17 | |
| *** alex_klimov has joined #openstack-nova | 20:18 | |
| *** suresh12 has quit IRC | 20:18 | |
| *** rdopiera has quit IRC | 20:19 | |
| *** annashen has joined #openstack-nova | 20:20 | |
| *** jaypipes has quit IRC | 20:20 | |
| *** y_sawai has joined #openstack-nova | 20:21 | |
| *** mtanino has quit IRC | 20:23 | |
| *** lucas__ has quit IRC | 20:23 | |
| *** suresh12 has joined #openstack-nova | 20:23 | |
| *** suresh12 has quit IRC | 20:25 | |
| *** baoli has quit IRC | 20:26 | |
| *** moshele has joined #openstack-nova | 20:31 | |
| *** jaypipes has joined #openstack-nova | 20:31 | |
| *** salv-orlando has joined #openstack-nova | 20:31 | |
| *** armax has joined #openstack-nova | 20:32 | |
| *** vladikr has quit IRC | 20:32 | |
| nagyz_ | mriedem, I've pushed in the ceph change and added you for review, hope you don't mind | 20:33 |
| mriedem | ok, i'm in rebase hell today so won't get to it for awhile | 20:33 |
| nagyz_ | just wanted to give you a heads up so it's not out of the blue :) | 20:34 |
| nagyz_ | also what's up with the flavor disk too small issue? finally I have the time to debug that if someone didn't do it further | 20:34 |
| *** salv-orl_ has quit IRC | 20:34 | |
| mriedem | i have no idea | 20:37 |
| mriedem | nagyz_: i'd ask mdbooth since you guys were talking about it | 20:37 |
| *** penick has quit IRC | 20:38 | |
| nagyz_ | oh, sorry, the second question was more generic not pointed towards you :-) I indeed wanted to pick mdbooth's brain about it | 20:38 |
| mriedem | i take everything personally | 20:39 |
| mriedem | like the fact we merged the os-brick change before my series, i'm pretty sure that was out of spite | 20:39 |
| mriedem | regardless of me having the other +2 on it | 20:39 |
| *** lucas__ has joined #openstack-nova | 20:39 | |
| nagyz_ | does this mean that now nova is going to use the same code that cinder and others use instead of having it's own drives for everything? | 20:40 |
| nagyz_ | AFAIK that's the goal os os-brick? | 20:40 |
| mriedem | https://review.openstack.org/#/c/175569/ | 20:40 |
| mriedem | yes | 20:40 |
| mriedem | there are still some shims in nova that connect to brick | 20:40 |
| nagyz_ | also can I push in a patch that depends-on my current ceph patch that re-enables the ceph volume encryption scenario in tempest? | 20:41 |
| *** mtanino has joined #openstack-nova | 20:41 | |
| *** alexpilotti has joined #openstack-nova | 20:41 | |
| nagyz_ | I guess disabling it was just a stop-gap measure until the support is there | 20:41 |
| mriedem | yes | 20:42 |
| nagyz_ | great | 20:42 |
| *** logan2 has quit IRC | 20:42 | |
| nagyz_ | then I only have one missing feature for which I haven't heard a blip in ages: https://review.openstack.org/#/c/125963/ | 20:42 |
| *** baoli has joined #openstack-nova | 20:42 | |
| *** zzz_pczesno has joined #openstack-nova | 20:43 | |
| *** mestery has joined #openstack-nova | 20:43 | |
| *** lucas__ has quit IRC | 20:44 | |
| *** pczesno has quit IRC | 20:44 | |
| mriedem | nagyz_: so, your bp link doesn't work in the commit message | 20:45 |
| mriedem | for one | 20:45 |
| *** logan2 has joined #openstack-nova | 20:45 | |
| nagyz_ | well I actually haven't worked on this for a while just wrote the initial patch | 20:45 |
| mriedem | is there a blueprint? | 20:45 |
| mriedem | because jdurgin had a spec for rbd snapshots | 20:45 |
| nagyz_ | I would have assumed that all dependencies have been taken care of :) | 20:46 |
| nagyz_ | jdurgin, are you around? | 20:46 |
| mriedem | nagyz_: https://review.openstack.org/#/c/188244/ | 20:47 |
| mriedem | i suggest you guys collaborate for the bottom line web 3.0 great success | 20:47 |
| nagyz_ | lol | 20:48 |
| mriedem | your change is called out in his spec | 20:49 |
| mriedem | under references | 20:49 |
| mriedem | and there is this https://review.openstack.org/205282 | 20:49 |
| mriedem | https://review.openstack.org/#/c/205282/ | 20:49 |
| nagyz_ | that's a new one I haven't seen yet | 20:50 |
| mriedem | nic: ^ | 20:50 |
| nagyz_ | seems like ther ewere a couple people working towards the same goal :) | 20:50 |
| mriedem | yes | 20:50 |
| mriedem | poorly | 20:50 |
| nagyz_ | agreed | 20:50 |
| *** lucas__ has joined #openstack-nova | 20:50 | |
| *** lucas__ has quit IRC | 20:51 | |
| *** alexpilotti has quit IRC | 20:52 | |
| *** lucas__ has joined #openstack-nova | 20:52 | |
| *** lucas__ has quit IRC | 20:56 | |
| *** fawadk has quit IRC | 20:57 | |
| bauzas | banzaï | 20:57 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify Scheduler RPC API select_dests() https://review.openstack.org/199205 | 20:57 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Create RequestSpec object https://review.openstack.org/145528 | 20:57 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Consider that all scheduler calls are IO Ops https://review.openstack.org/188486 | 20:58 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Use RequestSpec object in the FilterScheduler https://review.openstack.org/188487 | 20:58 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify left filters for RequestSpec https://review.openstack.org/202678 | 20:58 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Prepare filters for using RequestSpec object https://review.openstack.org/197912 | 20:58 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify conductor to use RequestSpec object https://review.openstack.org/203078 | 20:58 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify metric-related filters for RequestSpec https://review.openstack.org/202676 | 20:58 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Use RequestSpec in the ChanceScheduler https://review.openstack.org/199051 | 20:58 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify NUMA, PCI and num_instances filters for RequestSpec https://review.openstack.org/202677 | 20:58 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify Aggregate filters for RequestSpec https://review.openstack.org/202674 | 20:58 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify default filters for RequestSpec https://review.openstack.org/202675 | 20:58 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Add RequestSpec methods for primitiving into dicts https://review.openstack.org/187197 | 20:58 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify Affinity filter for RequestSpec https://review.openstack.org/202673 | 20:58 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Use RequestSpec object in HostManager https://review.openstack.org/191251 | 20:58 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Use the same pci_requests field for all filters and HostManager https://review.openstack.org/191250 | 20:58 |
| bauzas | combo | 20:58 |
| *** moshele has quit IRC | 20:59 | |
| *** raildo has quit IRC | 21:01 | |
| *** jaypipes has quit IRC | 21:04 | |
| nic | @mriedem, nagyz_: we looked at the spec and associated patch set when trying to implement it for our product and found it… a touch unfinished | 21:04 |
| *** pixelbeat_ has joined #openstack-nova | 21:05 | |
| nic | Being cautious people who like simple implementations, we figured that the clone-then-flatten approach had better long-term viability from a maintenance perspective | 21:06 |
| *** rfolco has quit IRC | 21:06 | |
| nic | i.e., dependency hell is an inevitability if you chain a bunch of RBD layers together | 21:06 |
| *** aysyd has quit IRC | 21:06 | |
| *** aix has joined #openstack-nova | 21:06 | |
| *** thorst has quit IRC | 21:07 | |
| nic | Fortunately, that approach was already described in the spec as an "alternative", so we implemented the alternative spec | 21:07 |
| nic | And it's scarily fast compared to the download-and-upload approach in current master | 21:08 |
| nic | 200%-1300% performance increase | 21:08 |
| mriedem | if it's not an increase of 1500% it's not worth doing imo :) | 21:09 |
| mriedem | i mean, go big or go home right? | 21:09 |
| nic | True that | 21:09 |
| *** burt has joined #openstack-nova | 21:09 | |
| nic | It also elegantly dodges the "how do you clean this thing up?" problem the prior implementation had | 21:10 |
| mriedem | if it's all rainbows and sunshine what were the holdups on the spec? | 21:11 |
| mriedem | i never read into it | 21:11 |
| nic | You did see the priorities list at the mid-cycle, yes? This was nowhere near being on it | 21:11 |
| *** jasondotstar has quit IRC | 21:11 | |
| *** armax has quit IRC | 21:11 | |
| nagyz_ | nic, great work. I'm running currently with a pre-patched version and indeed it's really fast. | 21:12 |
| nagyz_ | but we really need it to land in liberty | 21:12 |
| nagyz_ | without it ceph ephemeral is unusable | 21:12 |
| nic | Yes. | 21:12 |
| nagyz_ | a simple user just creating snapshot can do a DOS by filling up /tmp | 21:13 |
| nic | I can vouch for my implementation being ready to merge | 21:13 |
| nic | But I am a tad non-objective | 21:13 |
| nagyz_ | and the flatten really gets around the glance complexity, so I'm +1 on that | 21:13 |
| mriedem | nic: well, it wasn't a topic for discussion either from what i remember | 21:13 |
| mriedem | hell, smatzek is local so you guys could have both talked about it | 21:13 |
| nic | @mriedem I didn't quite have it finished, and it was my very first spec, so I was trying to be as nice as possible | 21:14 |
| mriedem | alright, well, for future reference, the meetups are a great place to just dump something on the table that needs talking about, in any form | 21:15 |
| nic | (I mean, we had our internal implementation against stable/icehouse finished, but I hadn't adapted it to master yet @ the mid-cycle) | 21:15 |
| mriedem | especially given jay and you were both there | 21:15 |
| *** thangp has quit IRC | 21:15 | |
| mriedem | not as easy at summits for this kind of open-ended stuff since the summit is so time-boxed | 21:15 |
| nic | I did, as much as I was able. Various attendees might recall me asking really, REALLY dumb questions about how specs work | 21:15 |
| nagyz_ | nic, I probably have even less of a clue of how they work than you do ;-) | 21:16 |
| nic | I also bugged melwitt about tests, it was all in service of this thing we're now talking about | 21:16 |
| mriedem | well, i mean less about the spec process and more about this particular rbd snapshots spec | 21:16 |
| mriedem | irc is good for dumb questions about specs | 21:16 |
| nagyz_ | I did discuss this a while back with smatzek in e-mails and he didn't mention any movement in this area tho | 21:17 |
| nagyz_ | I mean the missing rbd ephemeral snapshotting piece | 21:17 |
| nic | There was hand-waving involved. Can't wave my hands on Freenode very effectively | 21:17 |
| mriedem | nagyz_: because smatzek was doing those emails privately | 21:17 |
| mriedem | from what i can tell | 21:17 |
| *** rlrossit has left #openstack-nova | 21:17 | |
| *** moshele has joined #openstack-nova | 21:17 | |
| nagyz_ | well I see a mail from him to you with no reply ;) maybe I was missed on CCs:P | 21:17 |
| *** penick has joined #openstack-nova | 21:17 | |
| mriedem | yeah, he said, i care about this, so you should review it | 21:18 |
| mriedem | i typically ignore that kind of email | 21:18 |
| nagyz_ | ;d | 21:18 |
| *** y_sawai has quit IRC | 21:18 | |
| nagyz_ | so nic you think the current patch is good? is it missing anything for the flatten-based implementation? | 21:18 |
| *** smathews has quit IRC | 21:19 | |
| *** chuckcarmack has left #openstack-nova | 21:19 | |
| mriedem | sounds like there are also some test limitations with devstack | 21:19 |
| mriedem | since we have a ceph job, it'd be good to sort that out so we could actually test this new way of doing snapshots | 21:19 |
| nic | Yes, I think it's fit to merge; seb has helped get devstack whipped into shape and things seem to work over there | 21:19 |
| *** lucas-dinner has quit IRC | 21:19 | |
| nagyz_ | mriedem, I see a change for devstack there that enabled the direct url access so that's good. re the image that could be still pending | 21:20 |
| mriedem | link? | 21:20 |
| *** moshele has quit IRC | 21:20 | |
| nic | https://review.openstack.org/206653 | 21:20 |
| *** pixelbeat_ has quit IRC | 21:21 | |
| nagyz_ | I swear I did see the devstack change just half an hour ago but can't find it | 21:22 |
| nic | There were three of them | 21:22 |
| nic | "git log" helps :-P | 21:22 |
| *** otter768 has joined #openstack-nova | 21:22 | |
| *** tonytan4ever has quit IRC | 21:23 | |
| nic | But while we're talking about Ceph, there's this bug that is also good to go: https://review.openstack.org/187395 | 21:23 |
| *** baoli has quit IRC | 21:24 | |
| *** matrohon has quit IRC | 21:25 | |
| *** mestery has quit IRC | 21:25 | |
| nagyz_ | does that actually fixes the instance rebuild scenario? | 21:25 |
| mriedem | nic: so if resize/revert doesn't work with ceph today, why does the resize/revert test pass in tempest with the ceph job? | 21:26 |
| *** meghal has quit IRC | 21:27 | |
| *** otter768 has quit IRC | 21:27 | |
| *** meghal has joined #openstack-nova | 21:27 | |
| nic | That's an excellent question for the Tempest guys. But if you stand up a devstack and try it, it does precisely nothing | 21:27 |
| mriedem | nic: well i assume it's a question of verification done, | 21:28 |
| *** meghal has quit IRC | 21:28 | |
| mriedem | i believe the test just does the resize, waits for confirm, then does the revert and waits for active | 21:28 |
| *** meghal has joined #openstack-nova | 21:29 | |
| nagyz_ | mriedem, lol another thing where tempest fails us?! :-) | 21:29 |
| nagyz_ | like with the rbd encryption | 21:29 |
| mriedem | well, | 21:29 |
| mriedem | tempest is only as good as what people are putting into it for stuff like this, | 21:30 |
| mriedem | so if the nova API says, 'yup, i did what you asked' | 21:30 |
| mriedem | then it's kind of doing 90% of it's job | 21:30 |
| openstackgerrit | Tiago Mello proposed openstack/nova: Handle unsupported Spice feature for PowerKVM https://review.openstack.org/203809 | 21:30 |
| mriedem | if resize with ceph is busted in nova, trying to do that should fail | 21:30 |
| nagyz_ | some of our users reported resizing issues but I haven't looked at it in the past - but I guess if the error bubbles up to do a horizon error it should be visible to tempest as well | 21:31 |
| *** vishwanathj has joined #openstack-nova | 21:31 | |
| nagyz_ | but if nic has a patch fixing it that's excellent ;- | 21:32 |
| nagyz_ | ;-) | 21:32 |
| mriedem | yeah but, | 21:32 |
| mriedem | if we know it's busted today, we shouldn't be saying it's ok | 21:32 |
| *** meghal has quit IRC | 21:33 | |
| *** lucas-dinner has joined #openstack-nova | 21:33 | |
| *** meghal has joined #openstack-nova | 21:33 | |
| mriedem | like this https://review.openstack.org/#/c/193830/ | 21:33 |
| nagyz_ | right | 21:35 |
| nagyz_ | nic, do you want me to take a look tomorrow at why the test is passing and not actually erroring out or will you investigate and fix it as well as part of the instance rebuild cleanup patch series? :-) | 21:35 |
| *** tonytan4ever has joined #openstack-nova | 21:35 | |
| *** vishwanathj has quit IRC | 21:36 | |
| mriedem | i think that means you do it | 21:37 |
| *** lucas__ has joined #openstack-nova | 21:39 | |
| nagyz_ | in my book silence is not agreement :p | 21:39 |
| nagyz_ | mriedem, seems we got you off track from rebase hell? :) | 21:41 |
| mriedem | nagyz_: i'm still doing it | 21:41 |
| *** dims has quit IRC | 21:42 | |
| *** lucas__ has quit IRC | 21:43 | |
| *** eharney has quit IRC | 21:45 | |
| nic | naygz_ I don't even know what that means. :-) | 21:46 |
| nagyz_ | ok then I'll check the test case | 21:47 |
| mriedem | nic: we have a ceph job and a resize/revert test that passes, which means if something is busted with resizing (or reverting a resize of a ceph-backed instance), the test is a false positive, | 21:47 |
| mriedem | nic: so what i was saying is, shouldn't nova be raising an error in the case that we know it doesn't work? | 21:47 |
| mriedem | so the test fails | 21:48 |
| nagyz_ | just to bring up a stupid question: is the instance rebuild the same as the resize/revert case? | 21:48 |
| mriedem | no | 21:48 |
| nic | mriedem: it "works", in that it will migrate an instance. but the code is bugged such that it never resizes anything | 21:49 |
| vilobhmm | melwitt : regarding review for https://review.openstack.org/#/c/151677/31/nova/quota.py , in first place IMHO we don't need to create a seperate class for handling sub-projects..the way sajeesh has done for class NestedQuotaDriver….just by modifyin the API's of DbQuotaDriver should suffice the purpose as I have done for Cinder here https://review.openstack.org/#/c/205369/….inside the API although code can be modified whether parent_id is set | 21:49 |
| nic | And if Tempest isn't looking at the size of the backing file, it would assume everything worked | 21:49 |
| nic | (backing file, backing RBD, etc.) | 21:49 |
| mriedem | yeah, tempest doesn't dig that far | 21:50 |
| mriedem | just checks the api responses | 21:50 |
| *** cpschult has quit IRC | 21:50 | |
| nagyz_ | so then we have the instance rebuild broken as well besides this? swell. | 21:50 |
| nic | Yeah, if you look at the bugfix, "create" and "resize" are conjoined for no good reason | 21:50 |
| nic | So since the RBD is there, it assumes there's nothing to do | 21:50 |
| ctrath | mriedem: can I push the "Get quota and usage from neutron for limit APIs" spec to mitaka? | 21:51 |
| mriedem | ctrath: s/can/should/ yes | 21:51 |
| nagyz_ | nic, you understand this piece of code much better than I do so you looking at fixing the test case would be faster | 21:51 |
| *** jwcroppe has joined #openstack-nova | 21:52 | |
| mriedem | nagyz_: i don't think tempest is going to probably have a change for checking the backing file/rbd before and after the resize | 21:52 |
| nic | Once I split them into two discrete calls, nova was magically able to say "hey, the image is here, but it's the wrong size" and fix it | 21:52 |
| mriedem | i was tihnking more of a fail fast change in nova in this scenario | 21:52 |
| nic | So, yeah. Total mystery. | 21:52 |
| *** vishwanathj has joined #openstack-nova | 21:52 | |
| ctrath | no substitution needed. There aren't any other specs in that directory yet, so I wanted to make sure it was ok to do so | 21:52 |
| mriedem | ctrath: none merged | 21:52 |
| mriedem | there are some proposed | 21:52 |
| melwitt | vilobhmm: could you add that comment to the review? it would be good to capture that feedback, also I want to look at the spec and see if any feedback there caused the split out into a different class | 21:52 |
| *** changbl has quit IRC | 21:53 | |
| mriedem | nothing will be approved for mitaka until after summit | 21:53 |
| *** annashen_ has joined #openstack-nova | 21:53 | |
| mriedem | well this has been a thrilling hour+ or so, but it's time for me to leave for awhile | 21:53 |
| mriedem | hopefully ceph doesn't break me on the way home somehow | 21:53 |
| *** agireud has quit IRC | 21:54 | |
| nagyz_ | gl getting home:) I'll find my bed as it's midnight in 6 mins. | 21:54 |
| *** mriedem has quit IRC | 21:54 | |
| vilobhmm | melwitt : sure..alrite | 21:55 |
| melwitt | vilobhmm: I think your suggestion would be ideal. I'm curious if something during the spec review caused the direction of subclassing and duplicating most of the code | 21:56 |
| *** annashen_ has quit IRC | 21:58 | |
| *** mdrabe has quit IRC | 21:59 | |
| *** mdrabe has joined #openstack-nova | 21:59 | |
| *** vishwanathj has quit IRC | 22:00 | |
| *** coolsvap|mtg is now known as coolsvap|away | 22:00 | |
| openstackgerrit | Cale Rath proposed openstack/nova-specs: Get quota and usage from neutron for limit APIs spec https://review.openstack.org/206735 | 22:02 |
| cfriesen_ | ndipanov: are you around? | 22:02 |
| *** tonytan4ever has quit IRC | 22:02 | |
| *** dims has joined #openstack-nova | 22:02 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/nova: Scheduler: Respect tracks_instance_changes in periodic task https://review.openstack.org/206736 | 22:02 |
| jroll | edleafe: ^^ this one is for you | 22:02 |
| openstackgerrit | Mike Dorman proposed openstack/nova: Add console allowed origins setting https://review.openstack.org/201677 | 22:02 |
| jroll | edleafe: that code brings absolute panic and craziness to an ironic environment, fwiw | 22:02 |
| * jroll looks for tests to possibly add there | 22:04 | |
| *** ddaskal has quit IRC | 22:04 | |
| vilobhmm | melwitt : I think intitally not everyone was sure that a seperate class would be needed or not….but as I procedded ahead with the implementation I pointed out to sajeesh that NestedQuotaDriver as well as DBQuotaDriver share all the methods just that NestedQuotaDriver will behave slighty different as it will handle sub-projects. I hoep that is taken into consideration and we minimize code duplication. Similar change in cinder has already been mer | 22:04 |
| *** vishwanathj has joined #openstack-nova | 22:04 | |
| melwitt | vilobhmm: okay, cool. I checked the spec https://review.openstack.org/#/c/129420/ and didn't find any requirement of a separate class either. I commented on the review | 22:05 |
| vilobhmm | melwitt : ok cool. sounds good! | 22:06 |
| *** jecarey_ has quit IRC | 22:08 | |
| *** burt has quit IRC | 22:09 | |
| *** hemnalunch is now known as hemna | 22:09 | |
| *** vishwanathj has quit IRC | 22:09 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/nova: Scheduler: Respect tracks_instance_changes in periodic task https://review.openstack.org/206736 | 22:11 |
| *** eliqiao1 has joined #openstack-nova | 22:11 | |
| edleafe | jroll: looking... | 22:11 |
| *** eliqiao has quit IRC | 22:12 | |
| *** haomaiwang has quit IRC | 22:15 | |
| *** VW_ has quit IRC | 22:16 | |
| *** haomaiwang has joined #openstack-nova | 22:17 | |
| *** VW_ has joined #openstack-nova | 22:17 | |
| *** mragupat has quit IRC | 22:17 | |
| *** annegentle has quit IRC | 22:18 | |
| *** dboik has quit IRC | 22:19 | |
| edleafe | hmmm... did gerrit die? | 22:19 |
| *** VW_ has quit IRC | 22:21 | |
| *** tpeoples has quit IRC | 22:22 | |
| *** tpeoples has joined #openstack-nova | 22:22 | |
| *** kylek3h has quit IRC | 22:22 | |
| *** mwagner_lap has quit IRC | 22:23 | |
| clarkb | no... | 22:25 |
| edleafe | annnd... it's back | 22:26 |
| clarkb | its been around since july 18th according to ps | 22:26 |
| clarkb | and I have been actively using it. What are you seeing to make it look like it has gone away? | 22:26 |
| clarkb | general slowness or apache proxy errors or ? | 22:26 |
| jroll | yeah, been fine for me | 22:26 |
| edleafe | clarkb: git review -d would hang | 22:27 |
| edleafe | clarkb: didn't matter what repo | 22:27 |
| edleafe | also git pull would hang | 22:27 |
| clarkb | edleafe: is git pull talking to gerrit or git.openstack.org or github? | 22:27 |
| clarkb | edleafe: check with git remote -v | 22:28 |
| openstackgerrit | Jim Rollenhagen proposed openstack/nova: Scheduler: Respect tracks_instance_changes in periodic task https://review.openstack.org/206736 | 22:28 |
| jroll | edleafe: ^ now with passing tests \o/ | 22:28 |
| openstackgerrit | Maxim Nestratov proposed openstack/nova: libvirt: rename parallels driver to virtuozzo https://review.openstack.org/184311 | 22:28 |
| edleafe | clarkb: looks like it was github | 22:29 |
| jroll | github loves to go down | 22:29 |
| *** tpeoples has quit IRC | 22:29 | |
| clarkb | though if git review -d and git pull were having trouble that sound smore like local network issues | 22:29 |
| edleafe | clarkb: I assumed gerrit because it said "Downloading refs/changes/36/206736/3 from gerrit" | 22:30 |
| edleafe | network seems fine for everything else | 22:30 |
| clarkb | edleafe: yes the git review -d was likely gerrit, but git pull rarely is if you use our documentation | 22:30 |
| clarkb | typically origin is github or git.openstack.org depending | 22:30 |
| edleafe | clarkb: git pull is working, but git review -d is still hanging | 22:31 |
| clarkb | edleafe: can you try git review -v -d ? | 22:31 |
| edleafe | maybe because it's jroll's patch :) | 22:31 |
| *** jnc8651 has quit IRC | 22:31 | |
| clarkb | that should narrow it down to a specific operation | 22:32 |
| jroll | edleafe: hey, I'm just fixing your code man ;D | 22:32 |
| * jroll shots fired | 22:32 | |
| *** mdrabe has quit IRC | 22:32 | |
| clarkb | one commonish thing we see is ipv6 woes | 22:32 |
| clarkb | especially with comcast rolling it out more and more and still having major problems with it | 22:32 |
| edleafe | clarkb: it hangs on the line: | 22:32 |
| edleafe | 2015-07-28 22:31:59.282486 Running: git fetch gerrit refs/changes/36/206736/3 | 22:32 |
| *** tonytan4ever has joined #openstack-nova | 22:32 | |
| jroll | anyway, this is totally reviewable on gerrit regardless of review -d problems :) | 22:32 |
| jroll | edleafe: I mostly just want to talk about if this is expected behavior or if I actually fixed a thing | 22:33 |
| edleafe | jroll: yeah, but I wanted to play with the tests that used to be failing | 22:33 |
| *** meghal has quit IRC | 22:33 | |
| jroll | edleafe: oh, they're passing now :P | 22:33 |
| *** Piet has joined #openstack-nova | 22:34 | |
| *** meghal has joined #openstack-nova | 22:35 | |
| edleafe | jroll: ah, I see you fixed the 'self.scheduler_tracks_instance_changes' reference | 22:36 |
| *** slong__ has quit IRC | 22:36 | |
| *** tonytan4ever has quit IRC | 22:37 | |
| *** ctrath has quit IRC | 22:37 | |
| edleafe | clarkb: git review -v -d finally finished. Took 3:15 to run | 22:37 |
| clarkb | edleafe: I am not able to reproduce from here, that is very odd | 22:37 |
| clarkb | edleafe: does your gerrit remote look like ssh://edleafe@review.openstdack.org:29418/openstack/nova ? | 22:38 |
| edleafe | clarkb: gerrit ssh://ed-leafe@review.openstack.org:29418/openstack/nova.git | 22:38 |
| jroll | edleafe: yeah, basically Instance.get_by_host completely kills an ironic environment, given that a single host can have thousands of instances | 22:39 |
| *** igordcard has quit IRC | 22:39 | |
| *** alex_klimov has quit IRC | 22:43 | |
| *** jnc8651 has joined #openstack-nova | 22:43 | |
| *** marcusvrn has quit IRC | 22:45 | |
| edleafe | jroll: so is it true that schedulers in ironic environments can never filter on instance information? | 22:46 |
| jroll | edleafe: well, besides the fact that this query is insane in an ironic environment, the ironci drive only supports on compute host | 22:46 |
| jroll | one compute host* | 22:47 |
| openstackgerrit | Tracy Jones proposed openstack/nova: Cleanup of compute api reboot method https://review.openstack.org/127427 | 22:47 |
| jroll | and many compute nodes | 22:47 |
| jroll | and a compute node can only have one instance | 22:47 |
| *** mriedem has joined #openstack-nova | 22:47 | |
| edleafe | jroll: because before the 'tracks_instance_info' change, some filters would make that call | 22:47 |
| jroll | so scheduling based on instance info doesn't really make sense, an ironic node either has all resources or zero resources | 22:47 |
| jroll | we didn't use those filters, apparently, idk | 22:47 |
| edleafe | jroll: wouldn't they have been insane then? | 22:47 |
| jroll | they would | 22:48 |
| jroll | but, there's recommended bare metal filters | 22:48 |
| jroll | edleafe: https://github.com/openstack/nova/blob/master/nova/scheduler/ironic_host_manager.py#L33-49 | 22:48 |
| edleafe | well, your change may help ironic, but it will break anyone using the instance-based filters running older computes | 22:49 |
| edleafe | actually, running those filters with any version of compute | 22:49 |
| jroll | edleafe: it only breaks them if they turn off instance tracking, right? | 22:49 |
| jroll | edleafe: this call was added with the instance tracking thing... | 22:49 |
| edleafe | jroll: yes | 22:49 |
| jroll | edleafe: so you expect instance tracking to work when instance tracking is disabled? | 22:50 |
| edleafe | no - that's what the _add_instance_info method is for | 22:50 |
| edleafe | if they don't track, then the call is made with each request | 22:50 |
| *** VW_ has joined #openstack-nova | 22:51 | |
| edleafe | the difference is that the host manager is making the call, not the filters | 22:51 |
| jroll | ok, so here's how I understand this: | 22:51 |
| edleafe | i.e., one call per request instead of one call per host | 22:51 |
| jroll | CONF.scheduler_tracks_instance_changes is there in case the performance impact of tracking instance changes is too large | 22:51 |
| edleafe | yes | 22:51 |
| jroll | but turning that off just makes that performance impact happen on every request instead of at startup? | 22:52 |
| edleafe | not exactly | 22:52 |
| jroll | or at first request, I should say | 22:52 |
| edleafe | it turns off the gathering of info at startup | 22:52 |
| edleafe | but it forces that info to be gathered once per request | 22:52 |
| edleafe | previously, the db was hit once per host per request | 22:53 |
| edleafe | now, even turned off, it's only once per request | 22:53 |
| edleafe | turned on, there is no db hit per request | 22:53 |
| jroll | right | 22:54 |
| jroll | so in an environment with hundreds of instances per host, this still is a massive perf impact | 22:54 |
| *** gzhai has joined #openstack-nova | 22:54 | |
| jroll | whether that's realistic for anything non-ironic, I'm not sure. containers probably. | 22:54 |
| openstackgerrit | melanie witt proposed openstack/nova: Add missing policy for limits extension https://review.openstack.org/173178 | 22:55 |
| edleafe | jroll: the performance impact was there in the filters | 22:55 |
| edleafe | jroll: it also tied the scheduler to the nova db | 22:55 |
| jroll | right, so it's now non-configurable, and before it was configurable | 22:56 |
| edleafe | jroll: well, not directly configurable | 22:56 |
| *** Guest47142 is now known as dan | 22:56 | |
| edleafe | jroll: it was part of the various instance filters | 22:56 |
| jroll | sure, by choosing filters | 22:56 |
| *** zzzeek has quit IRC | 22:56 | |
| jroll | edleafe: also, I'm having a hard time digging through all of this, but I assume this ignores deleted instances, right? | 22:57 |
| edleafe | jroll: I think a better fix would be to add an option that says "no instance info is ever going to be needed" | 22:57 |
| edleafe | jroll: yes - deleted instances are ignored | 22:58 |
| jroll | ok | 22:59 |
| jroll | so do we agree this doesn't scale at all to hundreds of instances on a single host? | 22:59 |
| jroll | and is that a case that nova in general, as a project, cares about? | 22:59 |
| *** dannywilson has quit IRC | 23:00 | |
| jroll | for a data point: this OOM'd nova-scheduler running by itself on a 4GB machine, both in the "at startup" and "on the fly" case (since they're the same for a single host | 23:00 |
| edleafe | yes, it sure doesn't sound optimal :) | 23:00 |
| jroll | ) | 23:00 |
| *** penick has quit IRC | 23:01 | |
| jroll | and in the startup case, ate enough CPU that the rabbit handshake happening in another thread timed out because packets weren't leaving fast enough | 23:01 |
| jroll | :| | 23:01 |
| edleafe | and yes, of course nova as a project cares about stuff that doesn't work | 23:01 |
| *** haomaiwang has quit IRC | 23:01 | |
| *** lucas__ has joined #openstack-nova | 23:02 | |
| *** haomaiwang has joined #openstack-nova | 23:02 | |
| edleafe | jroll: I have to head out in a little while, but I'll definitely take a look at this more in-depth tomorrow | 23:02 |
| jroll | edleafe: cool, thanks | 23:02 |
| * jroll wonders if new config options are allowed during feature freeze | 23:03 | |
| * jroll is thinking he remembers not | 23:03 | |
| edleafe | jroll: I fought to get that one CONF option added. They are strongly discouraging new CONF options | 23:03 |
| jroll | great! | 23:03 |
| edleafe | jroll: FF doesn't affect a bug. I would seriously consider this a bug | 23:03 |
| *** mariojv has joined #openstack-nova | 23:04 | |
| jroll | sure, but config option tends to imply feature | 23:04 |
| jroll | I'll talk to johnthetubaguy in the morning | 23:04 |
| edleafe | jroll: not if it is something that should have been there all along | 23:05 |
| jroll | I'm just ranting at this point :) | 23:05 |
| edleafe | jroll: only if you're making it do something new and shiny | 23:05 |
| *** dims has quit IRC | 23:05 | |
| edleafe | jroll: ok, gotta go. I'll read the scrollback in the morning to see what you and johnthetubaguy discuss | 23:06 |
| *** lucas__ has quit IRC | 23:06 | |
| jroll | cool, see ya, thanks for talk | 23:06 |
| *** edmondsw has quit IRC | 23:07 | |
| *** VW_ has quit IRC | 23:09 | |
| *** VW_ has joined #openstack-nova | 23:10 | |
| openstackgerrit | Giridhar Jayavelu proposed openstack/nova: Create a new dictionary for type_data in VMwareAPIVMTestCase class https://review.openstack.org/202768 | 23:11 |
| *** dave-mccowan has quit IRC | 23:14 | |
| *** eliqiao has joined #openstack-nova | 23:14 | |
| *** veena has joined #openstack-nova | 23:15 | |
| *** romainh has left #openstack-nova | 23:15 | |
| *** eliqiao1 has quit IRC | 23:16 | |
| openstackgerrit | Nicolas Simonds proposed openstack/nova: libvirt: make snapshot use RBD snapshot/clone when available https://review.openstack.org/205282 | 23:16 |
| *** casusbelli has quit IRC | 23:17 | |
| *** xek has quit IRC | 23:18 | |
| *** xek has joined #openstack-nova | 23:21 | |
| *** MentalRay_ has quit IRC | 23:21 | |
| *** otter768 has joined #openstack-nova | 23:23 | |
| *** r-daneel has joined #openstack-nova | 23:24 | |
| *** r-daneel has quit IRC | 23:24 | |
| *** Piet has quit IRC | 23:25 | |
| *** r-daneel has joined #openstack-nova | 23:25 | |
| *** Piet has joined #openstack-nova | 23:26 | |
| *** otter768 has quit IRC | 23:28 | |
| *** ijw has quit IRC | 23:29 | |
| *** neiljerr` has quit IRC | 23:34 | |
| *** dims has joined #openstack-nova | 23:34 | |
| *** ijw has joined #openstack-nova | 23:34 | |
| mriedem | jroll: new config options are possible for bug fixes too, depends on context | 23:40 |
| mriedem | also, new config options make backporting to stable harder, so something to keep in mind | 23:40 |
| *** jasondotstar has joined #openstack-nova | 23:41 | |
| *** VW_ has quit IRC | 23:41 | |
| *** kaisers has quit IRC | 23:42 | |
| *** VW_ has joined #openstack-nova | 23:42 | |
| *** davideagnello has quit IRC | 23:42 | |
| *** erlon has quit IRC | 23:43 | |
| *** artom has quit IRC | 23:43 | |
| *** kaisers has joined #openstack-nova | 23:44 | |
| *** davideagnello has joined #openstack-nova | 23:45 | |
| *** pczesno has joined #openstack-nova | 23:46 | |
| *** VW_ has quit IRC | 23:46 | |
| *** zzz_pczesno has quit IRC | 23:47 | |
| *** lucas-dinner has quit IRC | 23:51 | |
| *** salv-orlando has quit IRC | 23:52 | |
| *** annashen_ has joined #openstack-nova | 23:54 | |
| *** _gryf has quit IRC | 23:55 | |
| *** annashen has quit IRC | 23:55 | |
| *** _gryf has joined #openstack-nova | 23:56 | |
| *** annashen_ has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
