| *** hfu_ has joined #openstack-nova | 00:00 | |
| *** hfu has quit IRC | 00:01 | |
| *** kaisers has joined #openstack-nova | 00:04 | |
| *** thorst_ has quit IRC | 00:04 | |
| *** kaisers has quit IRC | 00:09 | |
| *** hfu_ has quit IRC | 00:13 | |
| *** ducttape_ has quit IRC | 00:13 | |
| *** TTris has joined #openstack-nova | 00:24 | |
| *** CrashOverride has joined #openstack-nova | 00:37 | |
| *** hoangcx has joined #openstack-nova | 00:41 | |
| *** rcernin has quit IRC | 00:42 | |
| *** tuanluong has joined #openstack-nova | 00:46 | |
| *** huanxie has joined #openstack-nova | 00:47 | |
| *** tovin07 has joined #openstack-nova | 00:48 | |
| *** thorst_ has joined #openstack-nova | 00:50 | |
| *** thorst_ has quit IRC | 00:50 | |
| *** thorst_ has joined #openstack-nova | 00:54 | |
| *** CrashOverride has quit IRC | 00:59 | |
| *** thorst_ has quit IRC | 01:00 | |
| *** gongysh has joined #openstack-nova | 01:03 | |
| *** tovin07 has quit IRC | 01:04 | |
| *** phuongnh has joined #openstack-nova | 01:04 | |
| *** kaisers has joined #openstack-nova | 01:06 | |
| *** gongysh has quit IRC | 01:07 | |
| *** gus has quit IRC | 01:07 | |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: REST API for resource classes https://review.openstack.org/391959 | 01:10 |
|---|---|---|
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: adds ResourceClass.save() https://review.openstack.org/400448 | 01:10 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: adds ResourceClass.destroy() https://review.openstack.org/391918 | 01:10 |
| *** gus has joined #openstack-nova | 01:11 | |
| *** catinthe_ has quit IRC | 01:16 | |
| *** gus has quit IRC | 01:16 | |
| *** gus has joined #openstack-nova | 01:17 | |
| *** catintheroof has joined #openstack-nova | 01:17 | |
| *** tovin07 has joined #openstack-nova | 01:17 | |
| *** liusheng has quit IRC | 01:19 | |
| *** victorhe has joined #openstack-nova | 01:19 | |
| *** catintheroof has quit IRC | 01:21 | |
| *** huanxie has quit IRC | 01:24 | |
| *** liusheng has joined #openstack-nova | 01:26 | |
| *** jamielennox is now known as jamielennox|away | 01:26 | |
| openstackgerrit | alex proposed openstack/nova: there are some data left over after resize https://review.openstack.org/402361 | 01:26 |
| *** siva_krish has joined #openstack-nova | 01:33 | |
| *** hfu has joined #openstack-nova | 01:37 | |
| *** ljxiash has joined #openstack-nova | 01:38 | |
| *** ljxiash has quit IRC | 01:38 | |
| *** ljxiash has joined #openstack-nova | 01:38 | |
| *** hfu has quit IRC | 01:40 | |
| *** brad[] has quit IRC | 01:40 | |
| *** hfu has joined #openstack-nova | 01:41 | |
| *** brad[] has joined #openstack-nova | 01:42 | |
| *** hieulq has joined #openstack-nova | 01:44 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: Check destination_type when booting with bdm provided https://review.openstack.org/402372 | 01:49 |
| *** ducttape_ has joined #openstack-nova | 01:53 | |
| openstackgerrit | Jay Pipes proposed openstack/nova: rt: pass the nodename to public methods https://review.openstack.org/398470 | 01:53 |
| openstackgerrit | Jay Pipes proposed openstack/nova: rt: remove fluff from test_resource_tracker https://review.openstack.org/398471 | 01:53 |
| openstackgerrit | Jay Pipes proposed openstack/nova: rt: pass nodename to internal methods https://review.openstack.org/398469 | 01:53 |
| openstackgerrit | Jay Pipes proposed openstack/nova: rt: explicitly pass compute node to _update() https://review.openstack.org/398472 | 01:53 |
| openstackgerrit | Jay Pipes proposed openstack/nova: rt: use a single ResourceTracker object instance https://review.openstack.org/398473 | 01:53 |
| *** thorst_ has joined #openstack-nova | 01:57 | |
| *** baoli has joined #openstack-nova | 01:57 | |
| *** baoli has quit IRC | 02:02 | |
| *** thorst_ has quit IRC | 02:03 | |
| *** ducttape_ has quit IRC | 02:06 | |
| *** kaisers_ has joined #openstack-nova | 02:09 | |
| *** victorhe has quit IRC | 02:09 | |
| *** siva_krish has quit IRC | 02:11 | |
| *** baoli has joined #openstack-nova | 02:12 | |
| *** baoli has quit IRC | 02:12 | |
| *** baoli has joined #openstack-nova | 02:12 | |
| *** kaisers has quit IRC | 02:12 | |
| *** victorhe has joined #openstack-nova | 02:13 | |
| *** liuyulong has joined #openstack-nova | 02:13 | |
| *** kaisers_ has quit IRC | 02:14 | |
| *** jamielennox|away is now known as jamielennox | 02:15 | |
| *** kaisers has joined #openstack-nova | 02:15 | |
| *** baoli has quit IRC | 02:17 | |
| *** kaisers has quit IRC | 02:20 | |
| *** masayukig has joined #openstack-nova | 02:21 | |
| *** hfu has quit IRC | 02:22 | |
| *** gcb has joined #openstack-nova | 02:23 | |
| *** hfu has joined #openstack-nova | 02:24 | |
| *** claudiub has quit IRC | 02:25 | |
| *** Jeffrey4l has joined #openstack-nova | 02:28 | |
| *** shuquan has joined #openstack-nova | 02:30 | |
| *** markvoelker has joined #openstack-nova | 02:31 | |
| *** hfu has quit IRC | 02:41 | |
| openstackgerrit | Li Yingjun proposed openstack/nova: Optimize creating security_group https://review.openstack.org/315311 | 02:42 |
| *** hfu has joined #openstack-nova | 02:45 | |
| *** gongysh has joined #openstack-nova | 02:45 | |
| *** baoli has joined #openstack-nova | 02:46 | |
| *** baoli has quit IRC | 02:48 | |
| *** thorst_ has joined #openstack-nova | 02:52 | |
| *** thorst__ has joined #openstack-nova | 02:54 | |
| *** thorst_ has quit IRC | 02:57 | |
| *** thorst__ has quit IRC | 03:02 | |
| *** liuyulong_ has joined #openstack-nova | 03:03 | |
| *** liuyulong has quit IRC | 03:06 | |
| *** gongysh has quit IRC | 03:14 | |
| *** swamireddy has joined #openstack-nova | 03:15 | |
| *** kaisers has joined #openstack-nova | 03:17 | |
| openstackgerrit | Hieu LE proposed openstack/nova: Remove mox from unit/virt/libvirt/test_imagebackend.py (4) https://review.openstack.org/335820 | 03:17 |
| *** efried has quit IRC | 03:18 | |
| *** efried has joined #openstack-nova | 03:32 | |
| *** shihanzhang has joined #openstack-nova | 03:32 | |
| openstackgerrit | Eric Brown proposed openstack/nova: VMware: Update supported OS types for ESX 6.5 https://review.openstack.org/403492 | 03:33 |
| *** shuquan has quit IRC | 03:34 | |
| openstackgerrit | Eric Brown proposed openstack/nova: VMware: Update supported OS types for ESX 6.5 https://review.openstack.org/403492 | 03:36 |
| *** ayoung has joined #openstack-nova | 03:38 | |
| *** cfriesen has joined #openstack-nova | 03:39 | |
| *** shuquan has joined #openstack-nova | 03:42 | |
| *** kaisers has quit IRC | 03:46 | |
| *** liuyulong__ has joined #openstack-nova | 03:48 | |
| *** shuquan_ has joined #openstack-nova | 03:48 | |
| *** shuquan has quit IRC | 03:50 | |
| *** ljxiash has quit IRC | 03:50 | |
| *** ljxiash has joined #openstack-nova | 03:51 | |
| *** liuyulong_ has quit IRC | 03:52 | |
| *** victorhe has quit IRC | 03:53 | |
| *** hfu has quit IRC | 03:56 | |
| *** ljxiash has quit IRC | 03:59 | |
| *** thorst_ has joined #openstack-nova | 04:04 | |
| *** thorst_ has quit IRC | 04:05 | |
| *** victorhe has joined #openstack-nova | 04:05 | |
| *** ducttape_ has joined #openstack-nova | 04:06 | |
| *** code-R has quit IRC | 04:09 | |
| *** code-R has joined #openstack-nova | 04:09 | |
| *** psachin has joined #openstack-nova | 04:10 | |
| *** ducttape_ has quit IRC | 04:11 | |
| *** links has joined #openstack-nova | 04:13 | |
| *** liuyulong_ has joined #openstack-nova | 04:13 | |
| *** victorhe has quit IRC | 04:14 | |
| *** liuyulong__ has quit IRC | 04:16 | |
| *** sreeram_v has joined #openstack-nova | 04:16 | |
| *** victorhe has joined #openstack-nova | 04:16 | |
| *** code-R has quit IRC | 04:21 | |
| *** code-R has joined #openstack-nova | 04:21 | |
| *** victorhe has quit IRC | 04:24 | |
| *** victorhe has joined #openstack-nova | 04:27 | |
| *** mdnadeem has joined #openstack-nova | 04:28 | |
| *** mdnadeem has quit IRC | 04:28 | |
| *** mdnadeem has joined #openstack-nova | 04:31 | |
| *** mdnadeem_ has joined #openstack-nova | 04:32 | |
| *** mdnadeem has quit IRC | 04:32 | |
| *** shuquan_ has quit IRC | 04:37 | |
| *** code-R has quit IRC | 04:41 | |
| *** kaisers has joined #openstack-nova | 04:42 | |
| *** ayogi has joined #openstack-nova | 04:46 | |
| *** kaisers has quit IRC | 04:48 | |
| *** kaisers has joined #openstack-nova | 04:48 | |
| *** kaisers has quit IRC | 04:49 | |
| *** kaisers has joined #openstack-nova | 04:50 | |
| *** kaisers has quit IRC | 04:55 | |
| *** ratailor has joined #openstack-nova | 05:04 | |
| *** zhangjl has joined #openstack-nova | 05:09 | |
| *** thorst_ has joined #openstack-nova | 05:10 | |
| *** sudipto_ has joined #openstack-nova | 05:15 | |
| *** I has joined #openstack-nova | 05:15 | |
| *** I is now known as Guest35753 | 05:15 | |
| *** ecelik has joined #openstack-nova | 05:17 | |
| *** ecelik has left #openstack-nova | 05:17 | |
| *** Guest35753 has quit IRC | 05:18 | |
| *** sudipto_ has quit IRC | 05:18 | |
| *** thorst_ has quit IRC | 05:19 | |
| *** kaisers has joined #openstack-nova | 05:20 | |
| *** ducnc has quit IRC | 05:25 | |
| *** giri has quit IRC | 05:28 | |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: REST API for resource classes https://review.openstack.org/391959 | 05:33 |
| *** ljxiash has joined #openstack-nova | 05:33 | |
| *** shuquan has joined #openstack-nova | 05:38 | |
| openstackgerrit | Michael Still proposed openstack/nova-specs: Spec for continued vendordata work in Ocata. https://review.openstack.org/395959 | 05:46 |
| *** cfriesen has quit IRC | 05:46 | |
| *** cfriesen has joined #openstack-nova | 05:50 | |
| *** prateek has joined #openstack-nova | 05:51 | |
| *** shuquan has quit IRC | 05:52 | |
| *** shuquan has joined #openstack-nova | 05:52 | |
| *** victorhe has quit IRC | 05:54 | |
| *** victorhe has joined #openstack-nova | 05:55 | |
| *** jaosorior has joined #openstack-nova | 05:57 | |
| *** gjayavelu has joined #openstack-nova | 05:58 | |
| *** hfu has joined #openstack-nova | 05:58 | |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Port virt.disk.mount.test_nbd to Python 3 https://review.openstack.org/403508 | 05:59 |
| *** code-R has joined #openstack-nova | 06:00 | |
| *** hfu has quit IRC | 06:04 | |
| *** hfu has joined #openstack-nova | 06:06 | |
| *** moshele has joined #openstack-nova | 06:09 | |
| *** cfriesen has quit IRC | 06:09 | |
| *** hfu has quit IRC | 06:09 | |
| *** victorhe has quit IRC | 06:12 | |
| *** victorhe has joined #openstack-nova | 06:13 | |
| *** hfu has joined #openstack-nova | 06:14 | |
| *** nandal has joined #openstack-nova | 06:14 | |
| *** victorhe_ has joined #openstack-nova | 06:16 | |
| *** victorhe has quit IRC | 06:16 | |
| *** thorst_ has joined #openstack-nova | 06:17 | |
| *** yufei has joined #openstack-nova | 06:23 | |
| *** thorst_ has quit IRC | 06:24 | |
| *** lpetrut has joined #openstack-nova | 06:25 | |
| *** claudiub has joined #openstack-nova | 06:28 | |
| *** yufei has left #openstack-nova | 06:29 | |
| *** ircuser-1 has quit IRC | 06:30 | |
| *** hfu has quit IRC | 06:31 | |
| *** hfu has joined #openstack-nova | 06:32 | |
| *** gjayavelu has quit IRC | 06:34 | |
| *** mjura has joined #openstack-nova | 06:42 | |
| *** code-R_ has joined #openstack-nova | 06:49 | |
| *** mjura has quit IRC | 06:50 | |
| *** mjura has joined #openstack-nova | 06:50 | |
| *** code-R has quit IRC | 06:52 | |
| *** markus_z has joined #openstack-nova | 06:56 | |
| *** I has joined #openstack-nova | 06:57 | |
| *** sudipto has joined #openstack-nova | 06:57 | |
| *** I is now known as Guest81242 | 06:57 | |
| *** nkrinner_afk is now known as nkrinner | 06:59 | |
| *** oanson has joined #openstack-nova | 07:00 | |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: libvirt: Acquire TCP ports for console during live migration https://review.openstack.org/275801 | 07:03 |
| *** jaosorior has quit IRC | 07:04 | |
| *** jaosorior has joined #openstack-nova | 07:05 | |
| *** sridharg has joined #openstack-nova | 07:06 | |
| openstackgerrit | zhaolihui proposed openstack/nova: Fix keypair quota error https://review.openstack.org/403529 | 07:11 |
| *** shuquan has quit IRC | 07:11 | |
| *** victorhe has joined #openstack-nova | 07:13 | |
| *** victorhe_ has quit IRC | 07:13 | |
| *** shuquan has joined #openstack-nova | 07:15 | |
| *** victorhe has quit IRC | 07:15 | |
| *** victorhe has joined #openstack-nova | 07:16 | |
| *** Oku_OS-away is now known as Oku_OS | 07:16 | |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: live-migration test hook for serial console https://review.openstack.org/347471 | 07:17 |
| *** kaisers_ has joined #openstack-nova | 07:19 | |
| *** code-R_ has quit IRC | 07:20 | |
| *** shuquan has quit IRC | 07:21 | |
| *** kaisers has quit IRC | 07:22 | |
| *** thorst_ has joined #openstack-nova | 07:23 | |
| *** shuquan has joined #openstack-nova | 07:23 | |
| *** thorst_ has quit IRC | 07:29 | |
| *** lennyb has quit IRC | 07:29 | |
| openstackgerrit | Moshe Levi proposed openstack/nova: Enhance PCI passthrough whitelist to support regex https://review.openstack.org/199488 | 07:30 |
| *** ekuris has joined #openstack-nova | 07:31 | |
| *** paul-carlton2 has joined #openstack-nova | 07:32 | |
| *** victorhe has quit IRC | 07:32 | |
| *** lennyb has joined #openstack-nova | 07:32 | |
| *** lpetrut has quit IRC | 07:34 | |
| openstackgerrit | Alex Xu proposed openstack/nova-specs: Add whitelist for filter and sort query parameters for server list API https://review.openstack.org/393205 | 07:35 |
| *** Jeffrey4l has quit IRC | 07:35 | |
| *** code-R has joined #openstack-nova | 07:35 | |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/nova: Remove mox from nova/tests/unit/test_configdrive2.py https://review.openstack.org/322754 | 07:39 |
| *** andreas_s has joined #openstack-nova | 07:39 | |
| *** nherciu has joined #openstack-nova | 07:43 | |
| *** Jeffrey4l has joined #openstack-nova | 07:43 | |
| *** rcernin has joined #openstack-nova | 07:44 | |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Port api.openstack.compute.test_disk_config to Python 3 https://review.openstack.org/403536 | 07:44 |
| *** hfu has quit IRC | 07:44 | |
| *** gouthamr has joined #openstack-nova | 07:44 | |
| *** hfu has joined #openstack-nova | 07:45 | |
| *** pcaruana has joined #openstack-nova | 07:45 | |
| *** shuquan_ has joined #openstack-nova | 07:45 | |
| *** shuquan has quit IRC | 07:46 | |
| *** irenab has quit IRC | 07:46 | |
| *** irenab has joined #openstack-nova | 07:47 | |
| *** mnestratov has joined #openstack-nova | 07:48 | |
| openstackgerrit | Bin Zhou proposed openstack/nova: Avoid misremove of volume when live migrate fail https://review.openstack.org/387226 | 07:49 |
| *** yingjun has joined #openstack-nova | 07:50 | |
| yingjun | ping gibi | 07:50 |
| *** matrohon has joined #openstack-nova | 07:52 | |
| openstackgerrit | Hieu LE proposed openstack/nova: Remove mox from unit/virt/libvirt/test_imagebackend.py (5) https://review.openstack.org/335921 | 07:52 |
| *** ducnc has joined #openstack-nova | 07:53 | |
| *** ralonsoh has joined #openstack-nova | 07:53 | |
| openstackgerrit | Hieu LE proposed openstack/nova: Remove mox from unit/virt/libvirt/test_imagebackend.py (end) https://review.openstack.org/335925 | 07:54 |
| *** irenab has quit IRC | 07:55 | |
| *** irenab has joined #openstack-nova | 07:55 | |
| *** WillUK2010 has joined #openstack-nova | 07:56 | |
| *** code-R_ has joined #openstack-nova | 07:57 | |
| tdurakov | markus_z: hi | 07:57 |
| markus_z | tdurakov: hey ho | 07:57 |
| tdurakov | markus_z: about test hook, I'm investigating that issues locally, it looks like tempest deletes instance even before live migration, trying to figure out what causes tempest to do so | 07:58 |
| tdurakov | markus_z: will submit new version once figure out what goes wrong | 07:59 |
| tdurakov | jfyi^ | 07:59 |
| *** code-R has quit IRC | 07:59 | |
| markus_z | tdurakov: hm, weird. ok, thanks for that. I pushed a new ps in tempest too, which solves the "resource not found" thing. | 08:00 |
| tdurakov | markus_z: which one? | 08:00 |
| tdurakov | when you are trying to get serial console? | 08:00 |
| markus_z | https://review.openstack.org/346815 | 08:01 |
| markus_z | Yes | 08:01 |
| openstackgerrit | Xuanzhou Perry Dong proposed openstack/nova: Add service status enforcement for consoleauth https://review.openstack.org/336364 | 08:01 |
| tdurakov | markus_z: what was the root cause? | 08:02 |
| tdurakov | I saw that nova already deletes instance before tempest attempts to get serial console | 08:02 |
| markus_z | tdurakov: I used an REST API which was dropped in an microversion. Microversion 2.6 introduced the unified console, but I used the deprecated "get-serial-console" REST API. | 08:04 |
| markus_z | tdurakov: The doc change is here: https://review.openstack.org/#/c/402138/2/api-ref/source/servers-action-remote-consoles.inc | 08:04 |
| markus_z | (for the sake of completeness) | 08:04 |
| tdurakov | markus_z: hmmm, ok, let's wait for the tests are finished, my understanding was a bit different) | 08:05 |
| gibi | yingjun: ping | 08:05 |
| *** gmann has quit IRC | 08:06 | |
| *** yingjun has quit IRC | 08:07 | |
| *** ducttape_ has joined #openstack-nova | 08:08 | |
| openstackgerrit | Timofey Durakov proposed openstack/nova: live-migration test hook for serial console https://review.openstack.org/347471 | 08:10 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: live-migration test hook for serial console https://review.openstack.org/347471 | 08:10 |
| tdurakov | markus_z: you forgot about the block-migration flag:-p | 08:11 |
| *** sreeram_v has quit IRC | 08:11 | |
| *** sreeram_v has joined #openstack-nova | 08:12 | |
| *** irenab has quit IRC | 08:12 | |
| *** ducttape_ has quit IRC | 08:13 | |
| *** irenab has joined #openstack-nova | 08:13 | |
| openstackgerrit | Balazs Gibizer proposed openstack/nova: Transform aggregate.create notification https://review.openstack.org/394511 | 08:15 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova: Transform aggregate.add_host notification https://review.openstack.org/396210 | 08:15 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova: Transform aggregate.remove_host notification https://review.openstack.org/396211 | 08:15 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova: Add aggregate notification related enum values https://review.openstack.org/396209 | 08:15 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova: Transform aggregate.delete notification https://review.openstack.org/394512 | 08:15 |
| *** sam_nowitzki has joined #openstack-nova | 08:15 | |
| *** sreeram_v has quit IRC | 08:16 | |
| *** ihrachys has joined #openstack-nova | 08:22 | |
| *** nherciu_ has joined #openstack-nova | 08:22 | |
| sudipto | hi can someone help me with a few queries on availability zones? | 08:23 |
| *** nherciu has quit IRC | 08:26 | |
| *** shuquan_ has quit IRC | 08:26 | |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Port api.openstack.compute.test_extended* to Python 3 https://review.openstack.org/403548 | 08:28 |
| *** yingjun has joined #openstack-nova | 08:28 | |
| markus_z | tdurakov: damn, thanks, I did forget that you made a change. | 08:30 |
| *** code-R_ has quit IRC | 08:30 | |
| *** bkopilov has quit IRC | 08:30 | |
| tdurakov | markus_z: np) | 08:30 |
| *** sreeram_v has joined #openstack-nova | 08:31 | |
| *** bkopilov has joined #openstack-nova | 08:32 | |
| *** shuquan has joined #openstack-nova | 08:34 | |
| *** code-R has joined #openstack-nova | 08:34 | |
| openstackgerrit | Bin Zhou proposed openstack/nova: Avoid misremove of volume when live migrate fail https://review.openstack.org/387226 | 08:39 |
| sudipto | johnthetubaguy, is it a good idea to have a availability zone per host? | 08:39 |
| openstackgerrit | Bin Zhou proposed openstack/nova: Avoid misremove of volume when live migrate fail https://review.openstack.org/387226 | 08:40 |
| sudipto | that is if someone wants to do targeted deployment of instances... | 08:40 |
| openstackgerrit | Maxim Nestratov proposed openstack/nova: don't store vnc/spice listen addrs in migration data if disabled https://review.openstack.org/396218 | 08:41 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/nova: DNM PoC Enable VNC console access to Ironic nodes https://review.openstack.org/398270 | 08:44 |
| *** gjayavelu has joined #openstack-nova | 08:46 | |
| *** bauwser is now known as bauzas | 08:47 | |
| bauzas | sudipto: what do you wanna know about AZs ? | 08:47 |
| sudipto | bauzas, hi, thanks. Basically I have a few questions - 1. Is it a good idea to group individual hosts into one availability zone each to do targeted deployments to hosts? 2. If at a later point in time, the hosts change the AZs, will there be an impact on the instances on them? | 08:48 |
| bauzas | sudipto: you know what's a Nova AZ, right? | 08:49 |
| sudipto | bauzas, yup. | 08:49 |
| sudipto | think i know :) | 08:49 |
| bauzas | sudipto: so, it's just an aggregate metadata | 08:49 |
| sudipto | bauzas, yeah - if you do a flavor based extra_specs - then we can do it via aggregates is what you are saying right? | 08:49 |
| bauzas | while aggregates are hidden from end-users, AZs are known | 08:50 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/nova: Remove mox from nova/tests/unit/virt/test_virt.py https://review.openstack.org/323288 | 08:50 |
| bauzas | sudipto: not sure I understand your concern | 08:50 |
| bauzas | sudipto: if you want to provide some extra specs for flavors, they're also hiddent | 08:51 |
| *** gjayavelu has quit IRC | 08:51 | |
| bauzas | hidden | 08:51 |
| sudipto | bauzas, ok sorry. I want to know - 1. What's the best way to do targeted deployments of instances to a given host? | 08:51 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/nova: Remove mox from nova/tests/unit/compute/test_virtapi.py https://review.openstack.org/323213 | 08:52 |
| bauzas | sudipto: heh, there are a LOOOOOOT of possibilities | 08:52 |
| bauzas | sudipto: first, you mean for a end-user ? | 08:52 |
| *** amoralej|off is now known as amoralej | 08:52 | |
| bauzas | sudipto: because end-users don't know hosts, right? | 08:53 |
| sudipto | bauzas, yeah | 08:53 |
| bauzas | sudipto: end-users don't know the deployment, they just want to make sure that 2 instances are not within the same host | 08:53 |
| bauzas | or they want to trust the operator for creating 2 zones | 08:54 |
| sudipto | bauzas, ok | 08:54 |
| openstackgerrit | Michael Still proposed openstack/nova: List support for force-completing a live migration in Feature support matrix https://review.openstack.org/397871 | 08:54 |
| tdurakov | markus_z: so, it works https://review.openstack.org/#/c/347471/38 need to figure out what's happened with grenade job | 08:54 |
| sudipto | bauzas, and if it's an operator who's deploying some pets? | 08:55 |
| *** unicell1 has joined #openstack-nova | 08:55 | |
| bauzas | sudipto: that's doesn't mean they know how the 2 different zones will be separated, they just "trust" that if two instances are in two different zones, they expect to have a failure domain different | 08:55 |
| bauzas | sudipto: well, an end-user doesn't know if that's a pet or not | 08:56 |
| bauzas | sudipto: they just ask for something specific that the operator would get to them | 08:56 |
| *** pilgrimstack has joined #openstack-nova | 08:56 | |
| sudipto | bauzas, got it. | 08:56 |
| *** unicell has quit IRC | 08:57 | |
| bauzas | sudipto: they can say to the operator 'I want a pet' and then the operator will tell them "okay, you have to do this or that in order to get a pet" | 08:57 |
| *** sreeram_v has quit IRC | 08:57 | |
| *** sreeram_v has joined #openstack-nova | 08:58 | |
| bauzas | like for example using a specific flavor, or an image, or a specific AZ etc. | 08:58 |
| sudipto | bauzas, got it.. | 08:58 |
| *** aloga_ has joined #openstack-nova | 08:58 | |
| bauzas | or all the instances in some cloud could be treated equally and be all pets | 08:58 |
| sudipto | from a operator standpoint, if the AZ is changed for a given host, will it have any implication? | 08:58 |
| *** swamireddy has quit IRC | 08:59 | |
| bauzas | you can change an AZ metadata on an aggregate, the corollar being that all the hosts belonging to that aggregate will change from AZ | 08:59 |
| bauzas | but there are verifications | 08:59 |
| *** zzzeek has quit IRC | 09:00 | |
| bauzas | like, we verify first that as an operator, if you modify the agg AZ info, you don't turn hosts belonging to two AZs | 09:00 |
| sudipto | yeah because one host is unique for one AX | 09:00 |
| sudipto | AZ | 09:00 |
| *** code-R has quit IRC | 09:00 | |
| *** kevinz has joined #openstack-nova | 09:00 | |
| *** lpetrut has joined #openstack-nova | 09:01 | |
| *** shuquan has quit IRC | 09:01 | |
| openstackgerrit | Tovin Seven proposed openstack/nova: Integrate OSProfiler and Nova https://review.openstack.org/254703 | 09:01 |
| *** zzzeek has joined #openstack-nova | 09:01 | |
| *** sreeram_v has quit IRC | 09:02 | |
| sudipto | bauzas, but since there's a AZ change, will the scheduler to be alright with any operations performed on an instance which was a part of a different (older) AZ? | 09:02 |
| *** paul-carlton2 has quit IRC | 09:02 | |
| *** shuquan has joined #openstack-nova | 09:03 | |
| sudipto | basically host 1 --> AZ_1 -> instance 1 changed to host 1 --> AZ_2 --> instance 1 --> now let's say a resize on the same host is performed for instance 1. | 09:03 |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: Don't delete neutron port when attach failed https://review.openstack.org/403568 | 09:03 |
| sudipto | or let's say host 1 --> AZ_1 --> instance 1, host2,host1 --> AZ_2 --> instance 1 - now a scheduler activity is performed on instance 1. | 09:04 |
| bauzas | sudipto: AFAIR, if the user didn't specify an AZ, that's fine | 09:05 |
| *** hshiina has joined #openstack-nova | 09:05 | |
| *** CristinaPauna has joined #openstack-nova | 09:05 | |
| markus_z | tdurakov: Looks like it is this one: https://bugs.launchpad.net/zuul/+bug/1635111 | 09:05 |
| openstack | Launchpad bug 1635111 in Zuul "All grenade jobs are failing with error" [High,New] | 09:05 |
| bauzas | sudipto: but if the user asked for a specific AZ, we deny to change the AZ | 09:05 |
| *** andrey-mp has joined #openstack-nova | 09:05 | |
| bauzas | sudipto: there is the notion of "default AZ" for an instance standpoint | 09:05 |
| sudipto | bauzas, I see that i deployed instance 1 with AZ_1 and then re-added the host to another AZ_2 -- now i perform a resize operation on the instance and all of the instance Availability zones are flipped to AZ_2 | 09:06 |
| sudipto | or maybe there's a periodic task that modifies it? | 09:06 |
| tdurakov | markus_z: origin issue is hidden by ansible bug that causes post-failure, see http://logs.openstack.org/71/347471/38/check/gate-grenade-dsvm-neutron-multinode-live-migration-nv/55c5533/console.html#_2016-11-28_08_47_42_196474 | 09:06 |
| bauzas | sudipto: tl;dr: if you *explicitely* specified an AZ at boot time, your instance will be stuck within that AZ | 09:07 |
| sudipto | bauzas, ok... | 09:07 |
| *** ayogi has quit IRC | 09:07 | |
| *** ayogi has joined #openstack-nova | 09:07 | |
| andrey-mp | Hi. What happens with reviews for nova? How it should be proceeded? https://review.openstack.org/#/c/352612/4 Why core reviewers didn't check this and review was abandoned by them with reason 'no activity'? | 09:07 |
| *** pilgrimstack has left #openstack-nova | 09:08 | |
| kashyap | andrey-mp: mikal should've first asked a question before abandoning with that blanet reason | 09:09 |
| kashyap | s/blanet/blanket/ | 09:09 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/nova: Remove mox from nova/tests/unit/db/test_db_api.py https://review.openstack.org/323165 | 09:09 |
| andrey-mp | kashyap: but I didn't see this question... | 09:10 |
| kashyap | andrey-mp: Which question? | 09:10 |
| openstackgerrit | Maxim Nestratov proposed openstack/nova: libvirt: enable console log for virtuozzo https://review.openstack.org/397166 | 09:11 |
| andrey-mp | kashyap: You've wrote that he should asked a question | 09:11 |
| andrey-mp | or I misunderstood the process? | 09:12 |
| markus_z | andrey-mp: A script abandons patches with certain criteria automatically | 09:12 |
| kashyap | andrey-mp: Well, I don't know what rationale he used to just close it | 09:12 |
| kashyap | (Apart from that 12-week thing) | 09:12 |
| markus_z | andrey-mp: That script needs a credential to to the auto-abandon | 09:13 |
| kashyap | markus_z: Ah, right. | 09:13 |
| markus_z | andrey-mp: mikal gives his credentials for this script. | 09:13 |
| markus_z | andrey-mp: There was *never* such a scenario like "hm, andrey's patch sucks, I'll close it" :) | 09:13 |
| markus_z | andrey-mp: The script is relatively simple, it checks "has it a -1 for more than 12 weeks?", if that's the case, an abandon happens | 09:14 |
| markus_z | andrey-mp: Reading the comment history of your patch, it looks like the VMware CI voted with -1 some time ago. | 09:15 |
| andrey-mp | markus_z: ok, got it. So, how to proceed such reviews? It's has +1 from jenkins and awaiting for review. | 09:15 |
| markus_z | andrey-mp: You should be able to revive it. Maybe under "reply". | 09:15 |
| *** ralonsoh_ has joined #openstack-nova | 09:15 | |
| bauzas | sudipto: https://github.com/openstack/nova/blob/master/nova/compute/api.py#L4277 | 09:16 |
| andrey-mp | markus_z: I can't because it's not my review - it fixes my bug ) | 09:16 |
| markus_z | andrey-mp: I think your change was an exception, as the VMware CI isn't voting anymore AFAICT. | 09:16 |
| markus_z | andrey-mp: Ah, right. Anyone here with the name "Anusha Unnam"? | 09:17 |
| *** portdirect_away is now known as portdirect | 09:17 | |
| markus_z | andrey-mp: A core member can revive patches from other people too. | 09:17 |
| kashyap | markus_z: The script didn't take it into account that the -1 _was_ addressed? | 09:18 |
| markus_z | bauzas: Can you revive https://review.openstack.org/#/c/352612/ please? It got auto-closed by mistake. | 09:18 |
| kashyap | The "-1" was for adding tests | 09:18 |
| kashyap | markus_z: Okay, you note why it's an exception above | 09:18 |
| markus_z | kashyap: I think it checks for *any* -1 by either human or machine | 09:18 |
| *** ralonsoh has quit IRC | 09:19 | |
| sudipto | bauzas, this is case, where you aren't allowed to change the AZ of a host unless that host is not present in any other AZ right? | 09:19 |
| kashyap | Uh, that's dumb as hell | 09:19 |
| *** ralonsoh_ is now known as ralonsoh | 09:19 | |
| markus_z | kashyap: I simply assume that, I don't know where this script is to double-check my assumption. | 09:19 |
| andrey-mp | ok, thanks all for help! | 09:19 |
| markus_z | andrey-mp: I know it can be frustrating sometimes, but sometimes it's just an honest mistake :) | 09:20 |
| kashyap | markus_z: Thanks for clarification, either way | 09:21 |
| markus_z | andrey-mp: FWIW, the review backlog is far too huge to check everything manually | 09:21 |
| markus_z | *before abandoning it | 09:21 |
| sudipto | s/unless, if | 09:21 |
| sudipto | s/unless/if | 09:22 |
| andrey-mp | yeah, there are many reviews for nova... ) | 09:22 |
| *** WillUK2010 has quit IRC | 09:22 | |
| markus_z | To "Anusha Unnam", please consider adding your IRC handle to your Launchpad profile, so that it comes up when "stalking" you here: http://www.openstack.org/community/members/profile/52206 | 09:23 |
| bauzas | markus_z: well, I can, sure | 09:23 |
| markus_z | bauzas: thanks a lot! | 09:24 |
| bauzas | markus_z: but the patch probably would love some rebase | 09:24 |
| *** mgoddard has joined #openstack-nova | 09:24 | |
| markus_z | andrey-mp: ^ can you do that? | 09:24 |
| bauzas | do we have someone stepping up for possible follow-ups ? | 09:24 |
| bauzas | andrey-mp: you seem interested in that change, volunteer for handing over the patch ? | 09:25 |
| andrey-mp | markus_z: sure | 09:25 |
| markus_z | andrey-mp: coolio, thanks | 09:26 |
| andrey-mp | bauzas: I'll try to track it | 09:26 |
| *** lucas-afk is now known as lucasagomes | 09:27 | |
| bauzas | andrey-mp: cool, thanks I appreciate | 09:27 |
| openstackgerrit | Andrey Pavlov proposed openstack/nova: Fix floating-ip-list to show associated instance info https://review.openstack.org/352612 | 09:27 |
| bauzas | andrey-mp: here, the fact is that we want to make sure that we have resources for ensuring that this change can get merged | 09:27 |
| bauzas | andrey-mp: thanks for the rebase | 09:28 |
| markus_z | andrey-mp: For the sake of completeness, the freeze is earlier this time https://releases.openstack.org/ocata/schedule.html It should aim to get it merged until middle of January | 09:28 |
| markus_z | s/It/You | 09:28 |
| bauzas | that's a bug ? | 09:28 |
| *** shuquan has quit IRC | 09:28 | |
| markus_z | yeah | 09:29 |
| markus_z | hm, but nova-network... | 09:29 |
| bauzas | sec, we stopped showing information from the API ? | 09:30 |
| *** yassine has joined #openstack-nova | 09:30 | |
| *** yassine is now known as Guest89691 | 09:31 | |
| *** thorst_ has joined #openstack-nova | 09:31 | |
| *** swamireddy has joined #openstack-nova | 09:31 | |
| andrey-mp | bauzas: yes, from my point of view. | 09:31 |
| bauzas | andrey-mp: if you deploy with n-net ? | 09:32 |
| andrey-mp | yeah | 09:32 |
| bauzas | :/ | 09:32 |
| andrey-mp | it's still supported... | 09:33 |
| openstackgerrit | Roman Podoliaka proposed openstack/nova: Integrate OSProfiler and Nova https://review.openstack.org/254703 | 09:34 |
| bauzas | andrey-mp: my point is more that we broke the API contract in the past | 09:34 |
| bauzas | is s/more/rather/ | 09:35 |
| *** rmart04 has joined #openstack-nova | 09:36 | |
| openstackgerrit | Stephen Finucane proposed openstack/nova: Add PCIWeigher https://review.openstack.org/379524 | 09:37 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: Prefer non-PCI host nodes for non-PCI instances https://review.openstack.org/379625 | 09:37 |
| *** thorst_ has quit IRC | 09:39 | |
| *** nandal has quit IRC | 09:39 | |
| *** ihrachys has quit IRC | 09:42 | |
| *** yingjun has quit IRC | 09:43 | |
| *** yingjun has joined #openstack-nova | 09:43 | |
| *** yingjun has quit IRC | 09:46 | |
| *** yingjun has joined #openstack-nova | 09:46 | |
| *** erlon-airlong has joined #openstack-nova | 09:47 | |
| *** shuquan has joined #openstack-nova | 09:50 | |
| *** mdnadeem_ has quit IRC | 09:50 | |
| *** yingjun has quit IRC | 09:51 | |
| *** mdnadeem_ has joined #openstack-nova | 09:53 | |
| *** code-R has joined #openstack-nova | 09:55 | |
| *** Yingxin has quit IRC | 09:56 | |
| *** code-R_ has joined #openstack-nova | 09:57 | |
| *** derekh has joined #openstack-nova | 09:59 | |
| *** huanxie has joined #openstack-nova | 09:59 | |
| *** shuquan has quit IRC | 09:59 | |
| *** hfu has quit IRC | 09:59 | |
| *** code-R has quit IRC | 10:00 | |
| *** cdent has joined #openstack-nova | 10:00 | |
| *** gabor_antal has joined #openstack-nova | 10:02 | |
| *** aloga_ has quit IRC | 10:02 | |
| *** mgoddard_ has joined #openstack-nova | 10:03 | |
| *** gabor_antal_ has quit IRC | 10:05 | |
| *** mgoddard has quit IRC | 10:06 | |
| *** gouthamr has quit IRC | 10:07 | |
| *** gouthamr has joined #openstack-nova | 10:09 | |
| *** gouthamr has quit IRC | 10:09 | |
| *** sahid has joined #openstack-nova | 10:09 | |
| *** gouthamr has joined #openstack-nova | 10:09 | |
| *** ducttape_ has joined #openstack-nova | 10:09 | |
| *** gouthamr has quit IRC | 10:12 | |
| *** gouthamr has joined #openstack-nova | 10:13 | |
| *** ducttape_ has quit IRC | 10:14 | |
| alex_xu | johnthetubaguy: good morning, I updated https://review.openstack.org/#/c/393205/8, but still have problem with sort which I'm not sure what is best choice, and I commented in the spec. hope you can take a look at | 10:15 |
| alex_xu | I also will catch Jay when he online | 10:16 |
| johnthetubaguy | alex_xu: yeah, its a good question, I don't know the answer to that | 10:16 |
| tdurakov | markus_z: see http://logs.openstack.org/71/347471/38/check/gate-grenade-dsvm-neutron-multinode-live-migration-nv/9b31271/logs/n-cpu.txt 2016-11-28T09:31:35.488874Z | 10:17 |
| johnthetubaguy | alex_xu: I assume the index would still help keep sort efficient on the smaller set, but thats random guessing | 10:17 |
| *** andrey-mp has quit IRC | 10:18 | |
| alex_xu | johnthetubaguy: i only test on huge table, the index doesn't help at here. i can do another test. | 10:19 |
| *** stevemar has quit IRC | 10:19 | |
| alex_xu | let me catch jay :) | 10:19 |
| johnthetubaguy | alex_xu: what was the example you were looking at? | 10:20 |
| alex_xu | johnthetubaguy: I just did 'select * from instances order by project_id', we have index 'instances_project_id_deleted_idx', but it doesn't help. It only help for 'select id from instances order by project_id' or 'select project_id, deleted_at from instances order by project_id' | 10:22 |
| alex_xu | for nova, we basically equal to 'select * from instances' | 10:22 |
| johnthetubaguy | right, if you look at the query we make, we almost always have deleted=0 in a where clause | 10:24 |
| johnthetubaguy | because we don't list delete instances for most users | 10:24 |
| johnthetubaguy | try add the where clause, I think thats the bit that matters | 10:24 |
| *** andrey-mp has joined #openstack-nova | 10:25 | |
| *** xinliang has joined #openstack-nova | 10:25 | |
| *** kaisers_ has quit IRC | 10:27 | |
| *** kaisers has joined #openstack-nova | 10:27 | |
| johnthetubaguy | alex_xu: also, I would try using EXPLAIN EXTENDED in your query, it should tell you what is going on: http://dev.mysql.com/doc/refman/5.7/en/explain-extended.html | 10:29 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: hardware: Flatten functions https://review.openstack.org/367470 | 10:30 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: Standardize '_get_XXX_constraints' functions https://review.openstack.org/385071 | 10:30 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: Rename '_numa_get_constraints_XXX' functions https://review.openstack.org/385072 | 10:30 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: De-duplicate _numa_get_flavor_XXX_map_list https://review.openstack.org/385074 | 10:30 |
| * sfinucan thought johnthetubaguy had finally cracked and was yelling at people | 10:30 | |
| johnthetubaguy | sfinucan: heh | 10:30 |
| alex_xu | johnthetubaguy: yea, EXPLAIN told me the index doesn't help at here | 10:30 |
| johnthetubaguy | alex_xu: gotcha | 10:31 |
| openstackgerrit | Stephen Finucane proposed openstack/nova-specs: Fix additional mistakes in the NUMA spec document https://review.openstack.org/323252 | 10:31 |
| openstackgerrit | Stephen Finucane proposed openstack/nova-specs: priorities: Populate Ocata priorities https://review.openstack.org/400326 | 10:31 |
| alex_xu | johnthetubaguy: if there is huge size of tenant, the thing should be ok i guess. for the non-admin user, the result always filter by project_id and user_id first, then the db only sort on a small set. | 10:32 |
| johnthetubaguy | alex_xu: you find totally brings up a good point though, show_delete=True is probably not very efficient, more so than the fact its going to return lots of things | 10:32 |
| alex_xu | s/there is/there isn't/ | 10:32 |
| johnthetubaguy | alex_xu: so don't forget the tenant with thousands of VMs, but yeah, there is damage limited with the quota | 10:32 |
| markus_z | tdurakov: That issue is solved with https://review.openstack.org/#/c/275801/ I think the rebase removed it from the dependency list. | 10:32 |
| tdurakov | markus_z: resubmit it then | 10:33 |
| tdurakov | otherwise looks good:) | 10:33 |
| markus_z | tdurakov: review 347471 (nova hook) needs review 346815 (tempest test) needs review 275801 (nova ports fix). | 10:34 |
| tdurakov | so, hook should depends on both patches, right? | 10:35 |
| markus_z | tdurakov: yes | 10:35 |
| markus_z | one in nova and one in tempest | 10:35 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: live-migration test hook for serial console https://review.openstack.org/347471 | 10:36 |
| tdurakov | markus_z: done^ | 10:36 |
| *** thorst_ has joined #openstack-nova | 10:37 | |
| markus_z | tdurakov: yeah, that should work too, thanks. | 10:37 |
| *** hoangcx has quit IRC | 10:37 | |
| tdurakov | markus_z: btw, https://review.openstack.org/#/c/275801/ - need to be backported, right? | 10:37 |
| markus_z | tdurakov: would be awesome but I'm not sure if/how it can be done | 10:38 |
| *** CristinaPauna has quit IRC | 10:39 | |
| *** CristinaPauna has joined #openstack-nova | 10:39 | |
| alex_xu | johnthetubaguy: yea | 10:40 |
| markus_z | I will be happy as a clam if it merges for Ocata :) | 10:40 |
| alex_xu | johnthetubaguy: and after filter by project_id, the index 'instances_project_id_deleted_idx' already used. and only one index can be used in a query. so nothing can help if there is sort | 10:41 |
| *** phuongnh has quit IRC | 10:42 | |
| *** ihrachys has joined #openstack-nova | 10:43 | |
| *** zhangjl has quit IRC | 10:44 | |
| *** thorst_ has quit IRC | 10:44 | |
| alex_xu | johnthetubaguy: btw, I will go to the bug smash this week. there is event at Wednesday, i'm not sure i can back in time. If i'm not, would you mind running the meeting? | 10:44 |
| johnthetubaguy | alex_xu: no problem, do you the agenda covering what you think we need to cover? | 10:45 |
| *** rfolco has joined #openstack-nova | 10:45 | |
| alex_xu | johnthetubaguy: the only thing I have in mind is query parameters | 10:46 |
| johnthetubaguy | alex_xu: trimming it down to a small number of results quickly is probably the important part I guess, so maybe thats the best we can hope for | 10:46 |
| johnthetubaguy | alex_xu: yeah, thats the hot topic I guess | 10:46 |
| alex_xu | johnthetubaguy: yea, agree with you | 10:46 |
| * alex_xu go home now, back later | 10:48 | |
| johnthetubaguy | so. has anyone noticed the ceph job has got very angry recently? | 10:50 |
| *** ducnc has quit IRC | 10:50 | |
| bauzas | johnthetubaguy: any chance you could look at https://review.openstack.org/#/c/403035/ soon ? | 10:52 |
| lyarwood | johnthetubaguy: nope, can't say I have, it has been green in my libvirt block volume changes for the last week AFAIK | 10:52 |
| johnthetubaguy | lyarwood: maybe its just bad luck then, see a few red ones some up for stuff I have +W ed | 10:53 |
| johnthetubaguy | bauzas: whats the alternative one from roman again? | 10:53 |
| johnthetubaguy | bauzas: I see tdurakov is prefering that approach | 10:53 |
| bauzas | johnthetubaguy: yup, but I don't agree with tdurakov | 10:53 |
| bauzas | johnthetubaguy: there are 2 reasons for that | 10:53 |
| *** paul-carlton2 has joined #openstack-nova | 10:54 | |
| *** kragniz has quit IRC | 10:54 | |
| bauzas | #1 the object is having a Boolean field type, and we understood on Friday why | 10:54 |
| bauzas | the change is saying that we should change it to an integer, which means roman doesn't understand why it's different from SQL2A | 10:54 |
| bauzas | SQLA | 10:54 |
| *** kragniz has joined #openstack-nova | 10:54 | |
| *** mgoddard_ has quit IRC | 10:54 | |
| johnthetubaguy | bauzas: do you have the link to his change? sounds different to what I saw before | 10:55 |
| bauzas | #2 because roman's change is changing all the classes by using a different Mixin which is exactly the same than oslo.db but just for something about postgre | 10:55 |
| johnthetubaguy | bauzas: regardless, I think we should do your change as its a direct fix of the regression, which totally makes sense to me (like a partial revert) | 10:55 |
| bauzas | johnthetubaguy: https://review.openstack.org/#/c/402090/ | 10:56 |
| *** mgoddard has joined #openstack-nova | 10:56 | |
| *** gouthamr has quit IRC | 10:57 | |
| bauzas | johnthetubaguy: we can wait for US folks coming around tho | 10:58 |
| *** sambetts|afk is now known as sambetts | 10:58 | |
| *** ljxiash has quit IRC | 11:00 | |
| *** paul-carlton has joined #openstack-nova | 11:02 | |
| *** paul-carlton has left #openstack-nova | 11:03 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Add rudimentary CORS support to placement API https://review.openstack.org/392891 | 11:06 |
| *** irenab has quit IRC | 11:11 | |
| *** irenab has joined #openstack-nova | 11:12 | |
| johnthetubaguy | bauzas: I added my comments around what I think we should do | 11:12 |
| openstackgerrit | Lee Yarwood proposed openstack/nova: libvirt: Ensure to_xml always returns an ASCII string https://review.openstack.org/402382 | 11:13 |
| openstackgerrit | Lee Yarwood proposed openstack/nova: libvirt: utils.get_disk return the target dev of the root device https://review.openstack.org/402716 | 11:13 |
| openstackgerrit | Lee Yarwood proposed openstack/nova: libvirt: Use the mirror element to detect when to pivot https://review.openstack.org/399735 | 11:13 |
| openstackgerrit | Lee Yarwood proposed openstack/nova: libvirt: Make guest.BlockDevice only use the target dev of a disk https://review.openstack.org/402163 | 11:13 |
| openstackgerrit | Lee Yarwood proposed openstack/nova: libvirt: Wait for abort_job to complete when not resizing https://review.openstack.org/382449 | 11:13 |
| bauzas | johnthetubaguy: I agree with your opinion : #1 fix the regression, #2 discuss on the direction we want to get for the deleted field | 11:13 |
| bauzas | johnthetubaguy: tbh, I thought about that on Friday night, and I think we can just modify https://github.com/openstack/nova/blob/master/nova/objects/base.py#L219 to be read_only=True | 11:15 |
| johnthetubaguy | bauzas: +1 | 11:15 |
| openstackgerrit | Bob Ball proposed openstack/nova: Use physical utilisation for cached images https://review.openstack.org/353862 | 11:15 |
| johnthetubaguy | bauzas: that seems a good starting point to me | 11:15 |
| bauzas | johnthetubaguy: I haven't found any call modifying the object field directly | 11:16 |
| johnthetubaguy | bauzas: do you fancy throwing up at patch, and seeing what it does to the tests? | 11:16 |
| bauzas | sure, I can do that | 11:16 |
| bauzas | the only time we modify the deleted field in SQLA is by using the oslo.db method | 11:16 |
| johnthetubaguy | bauzas: honestly, all of those fields should be read_only | 11:16 |
| bauzas | perhaps | 11:16 |
| bauzas | I haven't checked the others | 11:16 |
| johnthetubaguy | I suspect we do strange things with updated_at, but the other two should be read_only | 11:17 |
| *** GnomeKing has joined #openstack-nova | 11:18 | |
| bauzas | johnthetubaguy: yeah, I remember some discussion about modifying updated_at | 11:19 |
| bauzas | at least the DB servicegroup | 11:20 |
| bauzas | which is touching that field | 11:20 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: Check Config Options Consistency for xenserver.py https://review.openstack.org/350284 | 11:20 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: conf: Covert StrOpt -> PortOpt https://review.openstack.org/394392 | 11:20 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: conf: Deprecate 'remap_vbd_dev' option https://review.openstack.org/394393 | 11:20 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: conf: Deprecate 'torrent_' options https://review.openstack.org/403629 | 11:20 |
| GnomeKing | johnthetubaguy: Any chance you could take a gander at https://review.openstack.org/#/c/353862/ ? I'm curious if the DiskFilter is used at rax? | 11:20 |
| openstackgerrit | Andrey Volkov proposed openstack/nova: PoC: add hierarchy quota driver https://review.openstack.org/391072 | 11:20 |
| johnthetubaguy | GnomeKing: I don't believe we use that at all, in public cloud at least | 11:21 |
| johnthetubaguy | GnomeKing: everything is sized around RAM slots | 11:21 |
| *** TTris has quit IRC | 11:22 | |
| GnomeKing | Right. Would explain why you don't hit this bug. | 11:22 |
| *** portdirect is now known as portdirect_away | 11:22 | |
| *** Bobba has quit IRC | 11:23 | |
| johnthetubaguy | GnomeKing: so physical_utilisation isn't really a thing | 11:23 |
| GnomeKing | Why? | 11:23 |
| GnomeKing | It is for base VDIs - clearly not for the leaf ones used by a VM | 11:24 |
| gcb | johnthetubaguy, thanks for your review commits of Python 3 porting :-) | 11:26 |
| johnthetubaguy | gcb: no worries, I got chance to do some reviews while it was quiet :) | 11:27 |
| *** Guest81242 has quit IRC | 11:27 | |
| *** sudipto has quit IRC | 11:27 | |
| johnthetubaguy | GnomeKing: I was looking at the SR total one in that patch I guess | 11:28 |
| *** gszasz has joined #openstack-nova | 11:28 | |
| GnomeKing | It isn't getting the SR total physical used; it's getting all VDIs on the SR and counting virtual allocation for any VDIs attached to a VM, and physical for any VDIs not attached (since they cannot expand) | 11:29 |
| *** tuanluong has quit IRC | 11:29 | |
| *** jaosorior has quit IRC | 11:30 | |
| *** oanson has quit IRC | 11:30 | |
| gcb | I have more commits of python 3 porting in https://etherpad.openstack.org/p/nova_py35, Nova cores, please help review when you are free :-) | 11:30 |
| *** mpaolino has joined #openstack-nova | 11:30 | |
| *** mpaolino has quit IRC | 11:31 | |
| *** oanson has joined #openstack-nova | 11:31 | |
| *** zenoway has joined #openstack-nova | 11:32 | |
| *** sam_nowitzki has quit IRC | 11:32 | |
| paul-carlton2 | kaisers, I'm struggling to get https://review.openstack.org/#/c/268053 to pass nova-quobyteci-dsvm-volume ci, seems to keep failing | 11:33 |
| *** irenab has quit IRC | 11:34 | |
| pkholkin | nova-specs cores, please review the spec! https://review.openstack.org/#/c/381912 | 11:34 |
| *** TTris has joined #openstack-nova | 11:34 | |
| johnthetubaguy | pkholkin: just for context, we are well past spec freeze for ocata now | 11:34 |
| *** kaisers_ has joined #openstack-nova | 11:36 | |
| pkholkin | johnthetubaguy: yes, John, I know, but I think we can make an exception for this, the spec is already in one 'special' list https://etherpad.openstack.org/p/nova-ocata-spec-review-sprint | 11:36 |
| *** nandal has joined #openstack-nova | 11:36 | |
| johnthetubaguy | pkholkin: its worth asking the question | 11:36 |
| pkholkin | I had many positive feedback there | 11:36 |
| *** irenab has joined #openstack-nova | 11:37 | |
| *** kaisers has quit IRC | 11:37 | |
| pkholkin | and I tried to be very active on this | 11:38 |
| paul-carlton2 | salv-orlando still struggling with Vmware NSX CI for https://review.openstack.org/#/c/268053? | 11:39 |
| *** irenab has quit IRC | 11:42 | |
| *** thorst_ has joined #openstack-nova | 11:42 | |
| *** irenab has joined #openstack-nova | 11:42 | |
| openstackgerrit | Gergely Csatari proposed openstack/nova: Optional name in Update Server description in api-ref https://review.openstack.org/403638 | 11:44 |
| *** nicolasbock has joined #openstack-nova | 11:45 | |
| *** ayogi has quit IRC | 11:46 | |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: WIP: [scheduler] fixes aggregate_image_properties_isolation filter https://review.openstack.org/403640 | 11:47 |
| *** irenab has quit IRC | 11:47 | |
| *** slagle has joined #openstack-nova | 11:49 | |
| *** thorst_ has quit IRC | 11:49 | |
| *** irenab has joined #openstack-nova | 11:54 | |
| openstackgerrit | Matthew Booth proposed openstack/nova: libvirt: Resize an existing disk https://review.openstack.org/403647 | 11:55 |
| *** rtheis has joined #openstack-nova | 11:56 | |
| *** code-R_ has quit IRC | 11:59 | |
| *** irenab has quit IRC | 12:02 | |
| *** code-R has joined #openstack-nova | 12:04 | |
| GnomeKing | Good call on the potential race with VDIs being deleted - thanks johnthetubaguy | 12:04 |
| *** irenab has joined #openstack-nova | 12:08 | |
| *** GnomeKing has quit IRC | 12:10 | |
| *** yingjun has joined #openstack-nova | 12:10 | |
| *** ducttape_ has joined #openstack-nova | 12:10 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Optional separate database for placement API https://review.openstack.org/362766 | 12:14 |
| *** ducttape_ has quit IRC | 12:15 | |
| *** bvanhav has joined #openstack-nova | 12:17 | |
| *** Serlex has joined #openstack-nova | 12:17 | |
| *** rodrigods has quit IRC | 12:17 | |
| *** rodrigods has joined #openstack-nova | 12:17 | |
| *** dave-mccowan has joined #openstack-nova | 12:19 | |
| openstackgerrit | Alex Szarka proposed openstack/nova: Transform instance-exists notification https://review.openstack.org/403660 | 12:20 |
| *** ljxiash has joined #openstack-nova | 12:22 | |
| *** ratailor has quit IRC | 12:24 | |
| *** ljxiash has quit IRC | 12:26 | |
| openstackgerrit | Alex Szarka proposed openstack/nova: Transform instance.rebuild notification https://review.openstack.org/384922 | 12:27 |
| *** yuntongjin has joined #openstack-nova | 12:28 | |
| *** oanson has quit IRC | 12:29 | |
| *** jaosorior has joined #openstack-nova | 12:29 | |
| *** alexpilo_ has quit IRC | 12:29 | |
| tdurakov | johnthetubaguy: hi, seen your comments on postgre gate fixes, my understanding, that rpodolyaka's patch will be transparent for code users. I'm also not a big fan of hot-fixing for special cases, like if field.name == 'special_case': continue:-p | 12:30 |
| *** ducttape_ has joined #openstack-nova | 12:32 | |
| *** alexpilotti has joined #openstack-nova | 12:32 | |
| *** yingjun_ has joined #openstack-nova | 12:32 | |
| *** oanson has joined #openstack-nova | 12:33 | |
| *** bauzas has quit IRC | 12:34 | |
| *** mdnadeem_ has quit IRC | 12:35 | |
| *** yingjun has quit IRC | 12:36 | |
| *** ketann has joined #openstack-nova | 12:36 | |
| ketann | can someone help review this spec: https://review.openstack.org/#/c/390413/ for Veritas HyperScale driver for Nova | 12:37 |
| *** nandal has quit IRC | 12:38 | |
| *** bauzas has joined #openstack-nova | 12:39 | |
| *** mdnadeem_ has joined #openstack-nova | 12:39 | |
| *** prateek has quit IRC | 12:39 | |
| *** yuntongjin has quit IRC | 12:40 | |
| bauzas | tdurakov: have you read our discussion ? | 12:41 |
| bauzas | tdurakov: and why I don't agree with you ? | 12:41 |
| tdurakov | bauzas: yep | 12:42 |
| tdurakov | bauzas: have you read my comment? | 12:42 |
| bauzas | tdurakov: as well | 12:42 |
| *** lucasagomes is now known as lucas-hungry | 12:44 | |
| *** gouthamr has joined #openstack-nova | 12:44 | |
| *** sudipto has joined #openstack-nova | 12:44 | |
| bauzas | tdurakov: so, you prefer to modify *all* the SQLA classes just for a postgre problem, and you think it's not something a tech debt ? | 12:44 |
| *** Guest81242 has joined #openstack-nova | 12:44 | |
| tdurakov | bauzas: yes, I am | 12:45 |
| bauzas | for example, you think that someone looking in 2 years at the module wouldn't have problems understanding why we added a special layer for just postgre ? | 12:45 |
| bauzas | and you don't think that someone looking at that would say "meh, why nova isn't just using oslo.db ?" | 12:46 |
| *** amoralej is now known as amoralej|off | 12:46 | |
| *** amoralej|off is now known as amoralej|lunch | 12:46 | |
| tdurakov | bauzas: the same is valid for your idea, tbh | 12:46 |
| tdurakov | bauzas: let's just wait for feedback from other cores | 12:46 |
| bauzas | tdurakov: no, because for 2 reasons | 12:47 |
| bauzas | tdurakov: one, I explained in my comment why we should NEVER try to modify the deleted field | 12:47 |
| bauzas | tdurakov: secondly, because it's just in one place about something really small => only the instance object | 12:48 |
| *** thorst_ has joined #openstack-nova | 12:48 | |
| *** mnestratov|3 has joined #openstack-nova | 12:48 | |
| bauzas | either way, I'm adding a new followup change for really fixing the real problem | 12:48 |
| bauzas | the real problem is not that postgre should have the right type, rather that we should never try to modify the deleted field | 12:48 |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Add volume status check when deleting instances https://review.openstack.org/260339 | 12:48 |
| tdurakov | > rather that we should never try to modify the deleted field - won't it affect *all* classes? | 12:49 |
| tdurakov | bauzas: ^ | 12:49 |
| bauzas | tdurakov: see my changer going in 30 mins | 12:50 |
| tdurakov | bauzas: ok | 12:50 |
| *** mnestratov has quit IRC | 12:51 | |
| bauzas | tdurakov: have you seen my comment on why I also dislike saying that we should modify the field type for deleted ? | 12:51 |
| tdurakov | bauzas: yes, I've read comments on that 2 patches | 12:51 |
| *** andrey-mp has quit IRC | 12:52 | |
| *** tbachman has joined #openstack-nova | 12:53 | |
| *** Guest81242 has quit IRC | 12:55 | |
| *** sudipto has quit IRC | 12:55 | |
| *** jaypipes has joined #openstack-nova | 12:55 | |
| *** ababich has joined #openstack-nova | 12:56 | |
| *** gcb has quit IRC | 12:56 | |
| *** lamt has quit IRC | 12:58 | |
| *** andrey-mp has joined #openstack-nova | 13:00 | |
| *** pradk has joined #openstack-nova | 13:00 | |
| *** markus_z has quit IRC | 13:01 | |
| openstackgerrit | Roman Podoliaka proposed openstack/nova: db: coerce deleted param to be an integer value https://review.openstack.org/402090 | 13:03 |
| alex_xu | jaypipes: good morning, I updated the https://review.openstack.org/#/c/393205/, but I have one problem, found the sort acutally can't get any benefit from the index, so not sure what i should do , I commented in the spec. looking for your help | 13:06 |
| *** kylek3h has quit IRC | 13:07 | |
| *** ducttape_ has quit IRC | 13:10 | |
| *** ketann has quit IRC | 13:10 | |
| *** baoli has joined #openstack-nova | 13:16 | |
| *** markus_z has joined #openstack-nova | 13:16 | |
| *** jaosorior is now known as jaosorior_brb | 13:16 | |
| jaypipes | alex_xu: will take a look shortly. just getting through holiday email backlog ;) | 13:18 |
| alex_xu | jaypipes: thanks :) | 13:18 |
| *** code-R_ has joined #openstack-nova | 13:20 | |
| johnthetubaguy | tdurakov: my take is, lets merge the direct fix of the regression, while we work out the best way to fix things in general | 13:21 |
| rpodolyaka | johnthetubaguy: thanks for the context on the deleted field in objects! I'm fine with either solution | 13:22 |
| rpodolyaka | it's just that we probably still want to see something like this on oslo.db | 13:22 |
| johnthetubaguy | rpodolyaka: yeah, not against making oslo.db do the right thing on the conversion too | 13:23 |
| rpodolyaka | so... you may end up using this transparently | 13:23 |
| johnthetubaguy | rpodolyaka: yeah, I suspect we might want both longer term | 13:23 |
| *** code-R has quit IRC | 13:23 | |
| *** jwcroppe has quit IRC | 13:24 | |
| *** tbachman_ has joined #openstack-nova | 13:24 | |
| *** jwcroppe has joined #openstack-nova | 13:25 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Block updating o.vo fields only DB related https://review.openstack.org/403694 | 13:25 |
| bauzas | rpodolyaka: johnthetubaguy: tdurakov: ^ | 13:25 |
| *** tbachman has quit IRC | 13:26 | |
| *** tbachman_ is now known as tbachman | 13:26 | |
| bauzas | that's I think the best approach for making sure those fields should only be returned from the DB, and never tried to be updated | 13:26 |
| rpodolyaka | bauzas: as I replied to John's comment on review - I'm all for making the field read-only | 13:26 |
| rpodolyaka | it's just I'm not sure what OVO behaviour is when the object is created | 13:27 |
| rpodolyaka | whether it will use the default of the field for initialization or not | 13:27 |
| *** ljxiash has joined #openstack-nova | 13:27 | |
| jaypipes | alex_xu: done. | 13:28 |
| johnthetubaguy | rpodolyaka: I assume read_only means you can't set the object attribute directly? which would have fixed the case we ran into I believe. | 13:28 |
| johnthetubaguy | rpodolyaka: I am not certain tough | 13:29 |
| johnthetubaguy | though | 13:29 |
| alex_xu | jaypipes: thanks, checking now | 13:29 |
| rpodolyaka | johnthetubaguy: yeah, I get that, but what happens when you do Instance(...).create() without passing the value for deleted? | 13:29 |
| rpodolyaka | I'd expect OVO to use the field default value | 13:29 |
| rpodolyaka | which is False | 13:30 |
| rpodolyaka | not 0 | 13:30 |
| *** jwcroppe has quit IRC | 13:30 | |
| alex_xu | jaypipes: do you means it is ok keep those sort keys even they didn't get benefit on the indexes, just focus on the filters? | 13:31 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: WIP: [scheduler] fixes aggregate_image_properties_isolation filter https://review.openstack.org/403640 | 13:31 |
| johnthetubaguy | rpodolyaka: so it wasn't set, so I don't think its in the changes list, so its not passed, but again, I am not 100% sure | 13:32 |
| *** paul-carlton has joined #openstack-nova | 13:33 | |
| *** dims has quit IRC | 13:33 | |
| *** kylek3h has joined #openstack-nova | 13:34 | |
| *** shuquan has joined #openstack-nova | 13:34 | |
| *** paul-carlton has quit IRC | 13:34 | |
| rpodolyaka | if it's not passed, then we use the model column's default which is 0 and everything's fine | 13:34 |
| rpodolyaka | so what's the purpose of an object field default then? is it only used when you create an object instance in the code? (i.e. not load it from the DB) | 13:36 |
| bauzas | rpodolyaka: so, the purpose of readonly is to only make sure we set a field once | 13:38 |
| rpodolyaka | bauzas: I'm more interested in field default value and whether it's going to be used in create() when the object is persisted or not | 13:38 |
| bauzas | rpodolyaka: in the case of most of our objects that reflect the DB (ie. those that are using the PersistentObject Mixin), we do create the object in SQLA and then we're hydrating a fresh new object from what we read in DB | 13:39 |
| *** sreeram_v has joined #openstack-nova | 13:39 | |
| *** shuquan has quit IRC | 13:39 | |
| edleafe | Scheduler subteam meeting in 20 minutes in #openstack-meeting-alt | 13:40 |
| *** shuquan has joined #openstack-nova | 13:40 | |
| *** mdnadeem_ has quit IRC | 13:41 | |
| *** mrsoul has quit IRC | 13:41 | |
| *** mdrabe has joined #openstack-nova | 13:42 | |
| *** dims has joined #openstack-nova | 13:44 | |
| *** stevemar__ has joined #openstack-nova | 13:47 | |
| *** mrsoul has joined #openstack-nova | 13:47 | |
| *** paul-carlton2 has quit IRC | 13:47 | |
| *** amoralej|lunch is now known as amoralej | 13:50 | |
| *** jwcroppe has joined #openstack-nova | 13:50 | |
| *** ducttape_ has joined #openstack-nova | 13:52 | |
| *** paul-carlton2 has joined #openstack-nova | 13:53 | |
| *** jheroux has joined #openstack-nova | 13:53 | |
| *** lucas-hungry is now known as lucasagomes | 13:55 | |
| *** ducttape_ has quit IRC | 13:56 | |
| *** sreeram_v has quit IRC | 13:57 | |
| sfinucan | macsz: RE conf options - there's a reason for not indenting the bullet points under 'Related Options' etc, right? | 13:58 |
| *** sreeram_v has joined #openstack-nova | 13:58 | |
| macsz | sfinucan: yes, if you do indent, the doc generates the list in <blockquote> | 13:59 |
| sfinucan | macsz: Thought so. Cheers :) | 13:59 |
| *** yingjun has joined #openstack-nova | 13:59 | |
| macsz | sfinucan: example: http://docs.openstack.org/newton/config-reference/compute/config-options.html | 14:00 |
| macsz | and grep for consoleauth_topic = consoleauth | 14:00 |
| *** lamt has joined #openstack-nova | 14:00 | |
| macsz | sfinucan: cheers, buddy :) | 14:00 |
| *** prateek has joined #openstack-nova | 14:00 | |
| *** ralonsoh_ has joined #openstack-nova | 14:01 | |
| johnthetubaguy | macsz: sfinucan: my horrendous hack to make that all work is very whitespace dependent right now (that could be tweaked... but its tricky to do a good general job). | 14:02 |
| *** yingjun_ has quit IRC | 14:02 | |
| *** sreeram_v has quit IRC | 14:03 | |
| *** ralonsoh has quit IRC | 14:03 | |
| *** sam_nowitzki has joined #openstack-nova | 14:04 | |
| *** oanson has quit IRC | 14:04 | |
| *** irenab_ has joined #openstack-nova | 14:04 | |
| *** links has quit IRC | 14:04 | |
| *** thorst_ is now known as thorst_coffee_ru | 14:05 | |
| *** irenab has quit IRC | 14:06 | |
| macsz | johnthetubaguy: i found to force consistency 'non-space' approach in reviews easier than mess with the doc scripts ;) | 14:06 |
| *** bvanhav has quit IRC | 14:07 | |
| macsz | johnthetubaguy kudos to you for even doing this doc parsing stuff | 14:07 |
| *** irenab_ has quit IRC | 14:08 | |
| *** sam_nowitzki has quit IRC | 14:11 | |
| *** irenab has joined #openstack-nova | 14:12 | |
| *** tblakes has joined #openstack-nova | 14:12 | |
| johnthetubaguy | macsz: meh, I just crawled in a hole with a sledgehammer for a few hours, then the docs folks merged it | 14:13 |
| *** lpetrut has quit IRC | 14:13 | |
| *** bvanhav has joined #openstack-nova | 14:14 | |
| *** markus_z has quit IRC | 14:15 | |
| *** thorst_coffee_ru is now known as thorst_ | 14:16 | |
| *** bvanhav_ has joined #openstack-nova | 14:16 | |
| *** tblakes has quit IRC | 14:16 | |
| *** lpetrut has joined #openstack-nova | 14:17 | |
| *** tblakes has joined #openstack-nova | 14:17 | |
| *** clenimar has joined #openstack-nova | 14:18 | |
| *** ducttape_ has joined #openstack-nova | 14:18 | |
| *** tblakes_ has joined #openstack-nova | 14:19 | |
| *** bvanhav has quit IRC | 14:19 | |
| *** shuquan has quit IRC | 14:21 | |
| *** yingjun has quit IRC | 14:22 | |
| *** yingjun has joined #openstack-nova | 14:22 | |
| *** tblakes has quit IRC | 14:22 | |
| *** tblakes_ is now known as tblakes | 14:22 | |
| *** Lord_balancer is now known as umbSublime | 14:26 | |
| *** yingjun has quit IRC | 14:26 | |
| *** mriedem has joined #openstack-nova | 14:27 | |
| *** rcernin has quit IRC | 14:27 | |
| *** yingjun has joined #openstack-nova | 14:27 | |
| *** openstackstatus has quit IRC | 14:28 | |
| *** openstackstatus has joined #openstack-nova | 14:29 | |
| *** ChanServ sets mode: +v openstackstatus | 14:29 | |
| *** yingjun has quit IRC | 14:31 | |
| *** baoli has quit IRC | 14:35 | |
| *** esberglu has joined #openstack-nova | 14:36 | |
| *** rcernin has joined #openstack-nova | 14:36 | |
| *** eharney has joined #openstack-nova | 14:37 | |
| *** victorhe has joined #openstack-nova | 14:42 | |
| mriedem | bauzas: i'm a bit confused looking at https://review.openstack.org/#/c/403035/ - and why we blow up when creating the instance from the buildrequest when instance.deleted is set in this case but not when creating a 'normal' instance which also has a boolean 'deleted' field | 14:42 |
| mriedem | bauzas: the only difference i can see would be if the instance.deleted field was never set before this | 14:42 |
| *** sridharg has quit IRC | 14:43 | |
| *** eharney has quit IRC | 14:43 | |
| *** vishwana_ has quit IRC | 14:44 | |
| *** moshele has quit IRC | 14:44 | |
| *** vishwanathj has joined #openstack-nova | 14:44 | |
| *** baoli has joined #openstack-nova | 14:46 | |
| *** jamesdenton has joined #openstack-nova | 14:46 | |
| mriedem | dansmith: have you taken a look at ^ yet? | 14:46 |
| *** kevinz has quit IRC | 14:46 | |
| dansmith | no | 14:46 |
| dansmith | oh boy | 14:47 |
| *** moshele has joined #openstack-nova | 14:47 | |
| *** victorhe has quit IRC | 14:47 | |
| *** victorhe has joined #openstack-nova | 14:48 | |
| dansmith | mriedem: yeah, I don't get that at all | 14:48 |
| *** markus_z has joined #openstack-nova | 14:48 | |
| mriedem | o | 14:49 |
| mriedem | *i'm glad i'm not alone | 14:49 |
| dansmith | makes no sense to me actually | 14:50 |
| mriedem | >>> from nova.objects import instance as instance_obj | 14:50 |
| mriedem | >>> instance = instance_obj.Instance() | 14:50 |
| mriedem | >>> instance.obj_attr_is_set('deleted') | 14:50 |
| mriedem | False | 14:50 |
| dansmith | even still, deleted=False is fine, and we're still using instance.create(), which can only ever see the boolean of the field anyway | 14:51 |
| bauzas | dansmith: mriedem: sorry was afk for a bit | 14:52 |
| *** victorhe has quit IRC | 14:52 | |
| bauzas | so, lemme tell you a looooong story | 14:52 |
| bauzas | while you folks were eating turkey | 14:52 |
| *** gcb has joined #openstack-nova | 14:53 | |
| *** tblakes has quit IRC | 14:54 | |
| bauzas | dansmith: mriedem: all of that is blowing up because https://github.com/openstack/nova/blob/master/nova/objects/build_request.py#L85 | 14:54 |
| dansmith | that should be cast to a boolean though | 14:55 |
| bauzas | previously, we were setting some fields in the compute.api module, then calling .create() which was hydrating all the Instance object fields by getting the DB | 14:55 |
| bauzas | dansmith: agreed, that's another problem, but we're casting integers in the coercing method | 14:55 |
| bauzas | dansmith: anyway, the point is that we have a different type between SQLA and the object | 14:56 |
| bauzas | that's fine by me | 14:56 |
| bauzas | because we only want to return False or True for the user | 14:56 |
| dansmith | yeah, but that's intentional | 14:56 |
| bauzas | exactly | 14:56 |
| *** nherciu_ has quit IRC | 14:56 | |
| *** eharney has joined #openstack-nova | 14:57 | |
| dansmith | explain how we're not casting 0 to False? | 14:57 |
| dansmith | https://github.com/openstack/oslo.versionedobjects/blob/master/oslo_versionedobjects/fields.py#L396-L397 | 14:57 |
| mriedem | i think we are | 14:58 |
| mriedem | from the sqla dump | 14:58 |
| mriedem | 'deleted': False, | 14:58 |
| *** oanson has joined #openstack-nova | 14:58 | |
| dansmith | oh, I see, | 14:58 |
| dansmith | the problem is actually that we're not going back to an integer on create() | 14:58 |
| openstackgerrit | Hironori Shiina proposed openstack/nova: Add soft reboot support to ironic driver https://review.openstack.org/403745 | 14:58 |
| mriedem | right | 14:58 |
| dansmith | I thought the opposite was being asserted | 14:58 |
| dansmith | gotcha | 14:58 |
| dansmith | that needs to be fixed, so why skip the deleted field? | 14:59 |
| mriedem | and was never a problem before because instance.deleted was never set before we called create on it | 14:59 |
| dansmith | yea | 14:59 |
| *** cfriesen has joined #openstack-nova | 14:59 | |
| bauzas | the problem is that we want to set again that field | 14:59 |
| *** paw has quit IRC | 14:59 | |
| dansmith | the problem is that we're not properly changing a bool into an int in create() if it's set :) | 15:00 |
| * dansmith cooks up a patch for that | 15:00 | |
| bauzas | dansmith: why should we modify that ? | 15:01 |
| dansmith | well, actually, I guess that's hard because we don't yet know the id, so we need to do it in the db layer | 15:01 |
| bauzas | dansmith: the only reason for modifying the deleted field should be when we destroy the instance, right? | 15:01 |
| dansmith | bauzas: yes, we should, IMHO | 15:01 |
| dansmith | the create() method is trying to create something that doesn't match the model | 15:01 |
| bauzas | dansmith: so, when we're calling instance.destroy(), it goes into oslo.db which sets the deleted biy | 15:01 |
| bauzas | bit | 15:01 |
| bauzas | dansmith: to the objec tid | 15:02 |
| *** hshiina has quit IRC | 15:02 | |
| dansmith | bauzas: yeah | 15:02 |
| bauzas | dansmith: that's why I provided a follow-up | 15:02 |
| mriedem | i guess you could convert False to 0 in https://github.com/openstack/nova/blob/520ee65a55bf9894d5915d3cebc892439d703312/nova/db/sqlalchemy/api.py#L1664 | 15:02 |
| bauzas | dansmith: putting read-only to the PersistentObject mixin fields | 15:02 |
| bauzas | mriedem: honestly, I want to go the way where noone updates that field :) | 15:03 |
| mriedem | https://review.openstack.org/#/c/403035/2/nova/objects/build_request.py as a hack works because we only ever use that method during the create flow | 15:03 |
| *** victorhe has joined #openstack-nova | 15:03 | |
| dansmith | mriedem: but not True | 15:03 |
| mriedem | dansmith: i know | 15:03 |
| *** cfriesen_ has joined #openstack-nova | 15:03 | |
| mriedem | i don't know what that would be, error? | 15:03 |
| dansmith | http://pastebin.com/EtjGqDfa | 15:04 |
| bauzas | mriedem: dansmith: https://review.openstack.org/#/c/403694/1 is the real fix IMHO | 15:04 |
| dansmith | not very hard | 15:04 |
| *** cfriesen has quit IRC | 15:04 | |
| dansmith | bauzas: I disagree | 15:04 |
| bauzas | dansmith: why so ? | 15:04 |
| *** ralonsoh__ has joined #openstack-nova | 15:04 | |
| dansmith | bauzas: that those fields are backed by oslo.db right now may not be the case in the future | 15:04 |
| dansmith | it's a detail | 15:04 |
| *** psachin has quit IRC | 15:05 | |
| *** jaosorior_brb has quit IRC | 15:05 | |
| *** clenimar has quit IRC | 15:05 | |
| mdbooth | jaypipes: Are you up and about, yet? Responded to https://review.openstack.org/#/c/337790/ | 15:05 |
| dansmith | bauzas: we set some of those fields in a few places right? | 15:05 |
| *** jaosorior_brb has joined #openstack-nova | 15:05 | |
| mriedem | dansmith: that patch seems ok, instance_ref.delete() is a bit weird though, not sure we'd ever expect to create a deleted instance | 15:05 |
| mriedem | and does calling instance_ref.delete() work before context.session.add(instance_ref)? | 15:06 |
| *** ralonsoh_ has quit IRC | 15:06 | |
| bauzas | dansmith: I looked at deleted, and I actually don't find any of those | 15:06 |
| dansmith | mriedem: we've got id there at that point so it should be fine, I'll write tests | 15:06 |
| bauzas | I mean, no places where we actually set deleted | 15:06 |
| dansmith | bauzas: I see several | 15:06 |
| bauzas | dansmith: where ? :) | 15:06 |
| bauzas | I'm curious | 15:06 |
| *** Guest81242 has joined #openstack-nova | 15:06 | |
| *** sudipto has joined #openstack-nova | 15:06 | |
| dansmith | bauzas: https://github.com/openstack/nova/blob/master/nova/objects/network.py#L165-L169 | 15:07 |
| openstackgerrit | Corentin Ardeois proposed openstack/nova: Add CORS filter to versions pipeline https://review.openstack.org/400785 | 15:07 |
| dansmith | bauzas: we do some delete operations in-place where we don't re-load the object from the db layer | 15:07 |
| openstackgerrit | Corentin Ardeois proposed openstack/nova: Add CORS filter to versions pipeline https://review.openstack.org/400785 | 15:07 |
| bauzas | dansmith: ah, that's for another object | 15:07 |
| *** clenimar has joined #openstack-nova | 15:07 | |
| bauzas | I see | 15:07 |
| dansmith | bauzas: even still, anything that does _from_db_object() won't ever be able to update the deleted field | 15:07 |
| bauzas | and given people don't re-load it from the DB, they prefer to set it straight and mark it non-persistent | 15:08 |
| dansmith | like instance.refresh() after something else has deleted | 15:08 |
| bauzas | _from_db_object() sets the right value for deleted | 15:08 |
| bauzas | that's how I understood why it's intentional to have a boolean | 15:08 |
| bauzas | either way, folks, it's just 2 days I'm passing around that problem | 15:09 |
| dansmith | bauzas: but it can't if we make it read only | 15:09 |
| bauzas | feel free to cook whatever you want, I just provided https://review.openstack.org/#/c/403035/2 as a possible fixup for short term | 15:09 |
| dansmith | heh, we have no test_instance_create() in test_db_api ? | 15:09 |
| *** daemontool has joined #openstack-nova | 15:10 | |
| mriedem | while we're fixing things, shouldn't https://github.com/openstack/nova/blob/master/nova/objects/build_request.py#L85 match the logic in https://github.com/openstack/nova/blob/master/nova/objects/instance.py#L347 ? | 15:10 |
| *** ljxiash_ has joined #openstack-nova | 15:10 | |
| bauzas | it should | 15:10 |
| dansmith | mriedem: definitely | 15:10 |
| mriedem | i guess https://github.com/openstack/nova/blob/master/nova/objects/build_request.py#L85 is assuming that because the build request is short lived until the instance is created, the build request won't ever hold a deleted instance | 15:10 |
| dansmith | but it's doing the right thing at the middle layer anyway | 15:11 |
| mriedem | dansmith: ok separate patch or also in yours? | 15:11 |
| bauzas | yup, because we delete the BuildRequest when we delete the instance | 15:11 |
| mriedem | bauzas: well we delete the build request after we create the instance | 15:11 |
| mriedem | in conductor i thought | 15:11 |
| *** vladikr has joined #openstack-nova | 15:11 | |
| mriedem | build request is only around to provide a response to the api until the instance lives in a cell | 15:11 |
| jaypipes | mdbooth: just got back from breakfast... looking now. | 15:12 |
| dansmith | mriedem: correct | 15:13 |
| *** ljxiash has quit IRC | 15:13 | |
| bauzas | either way, there are millions of ways to correct the problem, the real thing being that we set something in https://github.com/openstack/nova/blob/master/nova/objects/build_request.py#L85 that we want to persist | 15:14 |
| mriedem | alright i've got scrum in 15 minutes so i'll assume https://github.com/openstack/nova/blob/master/nova/objects/build_request.py#L85 is being handled by someone else, and i'll review after that | 15:14 |
| mriedem | i also saw no takers in the ML thread on getting pg coverage back in the integrated gate :) | 15:14 |
| bauzas | so, either we consider just removing that L85 setting, or we drop deleted from the items we wanna set in the SQLA layer, or we make sure we're not passing it before calling .save()... | 15:14 |
| mriedem | bauzas: i think we're doing the first two | 15:15 |
| mriedem | well, not removing L85 | 15:15 |
| mriedem | but fixing it to be accurate | 15:15 |
| bauzas | I chose the 3rd option :) | 15:15 |
| bauzas | but okay | 15:15 |
| mriedem | we don't know how get_new_instance is going to be used in the furture | 15:15 |
| mriedem | *future | 15:15 |
| mriedem | so i'm less inclined to go with that option | 15:15 |
| markus_z | I need one last change to finish bp/libvirt-virtlogd but cannot reach danpb since 2 weeks: https://review.openstack.org/#/c/323765/ Whom can I ask for a libvirt driver review? | 15:15 |
| bauzas | mriedem: I'm fine with that, I just wanted to provide a short-term fix | 15:16 |
| jaypipes | markus_z: I can do one. | 15:16 |
| bauzas | mriedem: you need to remember the special timing that was Friday afternoon my time | 15:16 |
| markus_z | jaypipes: That would be really great, thank you! | 15:16 |
| jaypipes | np | 15:16 |
| bauzas | mriedem: where litterally nobody except john and me around in the channel | 15:16 |
| mriedem | markus_z: i'd try to get sahid to take a look | 15:16 |
| bauzas | dansmith: you okay with me touching https://github.com/openstack/nova/blob/master/nova/objects/build_request.py#L85 ? | 15:17 |
| mriedem | bauzas: yes i know you were trying to put out a fire, i appreciate that | 15:17 |
| mriedem | and rpodolyaka for also helping out | 15:17 |
| dansmith | bauzas: yep | 15:17 |
| markus_z | sahid: Could you maybe also have a look at https://review.openstack.org/#/c/323765/ please? | 15:17 |
| bauzas | okay, that should be short | 15:17 |
| *** jlvill-travel is now known as jlvillal | 15:17 | |
| bauzas | s/short/quick | 15:17 |
| rpodolyaka | mriedem: np! | 15:17 |
| sahid | markus_z, markus_z sure will look at it shortly | 15:19 |
| sahid | mriedem: ^ | 15:19 |
| markus_z | sahid: Thanks! Let me know if you want a review on one of your patches. | 15:19 |
| mriedem | markus_z: danpb has a series up for review related to secure console proxy, might be in your wheelhouse | 15:20 |
| jaypipes | mdbooth: done. | 15:20 |
| mdbooth | jaypipes: Thank you! | 15:20 |
| markus_z | mriedem: I read the spec. Yep, will look for them and review. | 15:20 |
| mriedem | https://review.openstack.org/#/q/topic:bp/websocket-proxy-to-host-security | 15:20 |
| *** tbachman has quit IRC | 15:23 | |
| *** tblakes_ has joined #openstack-nova | 15:24 | |
| *** ekuris has quit IRC | 15:27 | |
| openstackgerrit | Moshe Levi proposed openstack/nova: Enhance PCI passthrough whitelist to support regex https://review.openstack.org/199488 | 15:27 |
| dansmith | hmm, no openstack bot | 15:27 |
| dansmith | for me | 15:27 |
| dansmith | mriedem: bauzas: https://review.openstack.org/403756 | 15:28 |
| *** tblakes_ is now known as tblakes | 15:28 | |
| *** READ10 has joined #openstack-nova | 15:29 | |
| dansmith | mriedem: your comment about session.add() was right, I thought you meant "when the context manager returns", but I had just put it there because I wanted it at the end and saw the closure as the next method in the file incorrectly | 15:29 |
| bauzas | dansmith: mmm, so if someone updates instance.deleted = True, then we would call soft_delete() ? | 15:29 |
| dansmith | mriedem: so yeah, after session.add(), and seems to work fine | 15:29 |
| mnestratov|3 | mriedem: thanks for updating https://review.openstack.org/#/c/282398/ for us :) | 15:29 |
| diana_clarke | I've run into a couple of snafus for the simple tenant usage pagination work. I wonder if I could pick your brain for a minute, dansmith. | 15:29 |
| dansmith | bauzas: yes | 15:29 |
| diana_clarke | dansmith: The 'nova usage' CLI only shows the usage totals (not details), making it impossible to determine the marker (an instance uuid) to pass to the CLI to get subsequent pages. | 15:29 |
| diana_clarke | dansmith: For example: http://paste.openstack.org/show/589771/ | 15:29 |
| diana_clarke | dansmith: Do you think I should dog food the next links for the python-novaclient CLI and continue to show the complete totals? Other ideas? Thanks! | 15:30 |
| bauzas | dansmith: that adds a new behaviour, but okay | 15:30 |
| bauzas | anyway, I need to disappear for the next 30 mins :( | 15:30 |
| dansmith | bauzas: well, corrects a behavior IMHO but yeah :) | 15:30 |
| bauzas | dansmith: agreed | 15:30 |
| *** moshele has quit IRC | 15:30 | |
| *** mnestratov|3 is now known as mnestratov | 15:30 | |
| bauzas | oh man, I need to change all the mocks for just changing .deleted = 0 to something like db['deleted'] == db['id'] | 15:31 |
| bauzas | graah | 15:31 |
| dansmith | diana_clarke: the client should handle the paging internally right? | 15:31 |
| mriedem | mnestratov: np | 15:32 |
| diana_clarke | dansmith: It wasn't, but now I'm thinking it should. 'nova usage' wouldn't accept limit & marker, but rather always show the full totals (page under the hood). | 15:33 |
| *** andrey-mp has quit IRC | 15:33 | |
| *** paw has joined #openstack-nova | 15:33 | |
| diana_clarke | dansmith: And then I wonder how the plain old python object client interface should work. It should accept limit & marker I think for maximum flexibility. | 15:34 |
| diana_clarke | dansmith: Or not. That way horizon wouldn | 15:34 |
| diana_clarke | oops | 15:34 |
| mriedem | does http://status.openstack.org/zuul/?openstack/nova work for others? | 15:34 |
| mriedem | does that populate the filters box? | 15:35 |
| diana_clarke | dansmith: Or not. That way horizon wouldn't need to change at all. The json payload would still be large, but the memory usage wouldn't balloon b/c it would be constructed with multiple smaller DB result sets. | 15:35 |
| dansmith | diana_clarke: yeah, I would expect novaclient to just do the thing internally, especially since we have to fix up the totals | 15:35 |
| jaypipes | markus_z: done. | 15:36 |
| diana_clarke | dansmith: For both cases? python-novaclient CLI and POPOs? | 15:36 |
| *** dims has quit IRC | 15:36 | |
| *** bnemec has joined #openstack-nova | 15:37 | |
| *** jaosorior_brb is now known as jaosorior | 15:37 | |
| mriedem | nope i guess it doesn't | 15:37 |
| openstackgerrit | Francesco Santoro proposed openstack/os-vif: os-vif: add vif_name to VIFVHostUser class https://review.openstack.org/390225 | 15:38 |
| openstackgerrit | Francesco Santoro proposed openstack/os-vif: os-vif: add new classes to enable fast path vhostuser https://review.openstack.org/375533 | 15:38 |
| dansmith | diana_clarke: um, who you calling a poopoo? | 15:38 |
| *** slaweq_ has quit IRC | 15:39 | |
| mriedem | i assume she meant police officers | 15:40 |
| mriedem | writing REST API clients for nova | 15:40 |
| mriedem | to catch criminals | 15:40 |
| *** mlavalle has joined #openstack-nova | 15:40 | |
| dansmith | mriedem: heh | 15:41 |
| *** dims has joined #openstack-nova | 15:41 | |
| dansmith | I assume she means "horizon" | 15:41 |
| diana_clarke | Plain Old Python Objects (the Usage API in python-novaclient as opposed to the CLI) | 15:42 |
| *** rfolco_ has joined #openstack-nova | 15:42 | |
| *** mtanino has joined #openstack-nova | 15:42 | |
| markus_z | "popo" means ass in German :D | 15:42 |
| diana_clarke | lol | 15:43 |
| *** rfolco has quit IRC | 15:44 | |
| *** mtanino has quit IRC | 15:46 | |
| dansmith | diana_clarke: the python api probably shouldn't hide that from the layer above, I would think | 15:47 |
| dansmith | or in german, | 15:47 |
| dansmith | "the ASS probably shouldn't..." | 15:47 |
| diana_clarke | dansmith: Okay, thanks for the sanity check. I'll give that a shot (hide paging in the CLI, but expose it in the python API). | 15:48 |
| diana_clarke | My next snafu has to do with the change in sorting that came out of spec review. Sorting my uuid makes functional tests hard... | 15:48 |
| diana_clarke | s/my/by | 15:48 |
| *** ralonsoh_ has joined #openstack-nova | 15:49 | |
| diana_clarke | When it was sorted by an auto-incrementing id, I could create two instances and deterministicly predict the order for the next links etc. With uuids I can't do that... | 15:50 |
| *** mtanino has joined #openstack-nova | 15:50 | |
| *** ralonsoh__ has quit IRC | 15:50 | |
| *** prateek_ has joined #openstack-nova | 15:50 | |
| *** code-R_ has quit IRC | 15:51 | |
| *** harlowja_at_home has joined #openstack-nova | 15:51 | |
| *** slaweq_ has joined #openstack-nova | 15:52 | |
| *** xyang1 has joined #openstack-nova | 15:52 | |
| *** jdurgin1 has joined #openstack-nova | 15:53 | |
| *** prateek has quit IRC | 15:53 | |
| *** Sukhdev has joined #openstack-nova | 15:54 | |
| *** hongbin has joined #openstack-nova | 15:54 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova-specs: Spec for continued vendordata work in Ocata. https://review.openstack.org/395959 | 15:56 |
| *** yamahata has joined #openstack-nova | 15:56 | |
| openstackgerrit | Stephen Finucane proposed openstack/nova: conf: Deprecate 'torrent_' options https://review.openstack.org/403629 | 15:57 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: Check Config Options Consistency for xenserver.py https://review.openstack.org/350284 | 15:57 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: conf: Covert StrOpt -> PortOpt https://review.openstack.org/394392 | 15:57 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: conf: Deprecate 'remap_vbd_dev' option https://review.openstack.org/394393 | 15:57 |
| mriedem | dansmith: johnthetubaguy: i'd still like jpenick to comment on the user roles part of this https://review.openstack.org/#/c/395959/ but i'm +2 as there are other things in that spec we agreed to get done in ocata | 15:57 |
| *** avolkov has quit IRC | 15:59 | |
| johnthetubaguy | mriedem: ah, I will try take a look at that ASAP | 15:59 |
| *** ducttape_ has quit IRC | 15:59 | |
| *** ducttape_ has joined #openstack-nova | 16:00 | |
| *** avolkov has joined #openstack-nova | 16:00 | |
| *** ducttape_ has quit IRC | 16:01 | |
| *** mlavalle has quit IRC | 16:01 | |
| *** erlon-airlong has quit IRC | 16:01 | |
| *** swamireddy has quit IRC | 16:01 | |
| *** hieulq has quit IRC | 16:01 | |
| *** tovin07 has quit IRC | 16:01 | |
| *** alineb has quit IRC | 16:01 | |
| *** shaohe_feng has quit IRC | 16:01 | |
| *** mirrorbo1 has quit IRC | 16:01 | |
| *** mvk has quit IRC | 16:01 | |
| *** jklare has quit IRC | 16:01 | |
| *** david-lyle has quit IRC | 16:01 | |
| *** HenryG has quit IRC | 16:01 | |
| *** purplerbot has quit IRC | 16:01 | |
| *** andreaf has quit IRC | 16:01 | |
| *** ameade has quit IRC | 16:01 | |
| *** karlamrhein has quit IRC | 16:01 | |
| *** Kimmo__ has quit IRC | 16:01 | |
| *** zigo has quit IRC | 16:01 | |
| *** ccard_ has quit IRC | 16:01 | |
| *** scottynomad has quit IRC | 16:01 | |
| *** keekz has quit IRC | 16:01 | |
| *** haukebruno has quit IRC | 16:01 | |
| *** zigo has joined #openstack-nova | 16:01 | |
| *** haukebruno has joined #openstack-nova | 16:01 | |
| *** Kimmo__ has joined #openstack-nova | 16:02 | |
| *** mirrorbo1 has joined #openstack-nova | 16:02 | |
| *** keekz has joined #openstack-nova | 16:02 | |
| *** mvk has joined #openstack-nova | 16:02 | |
| *** ccard_ has joined #openstack-nova | 16:02 | |
| *** tovin07_ has joined #openstack-nova | 16:02 | |
| *** jklare has joined #openstack-nova | 16:02 | |
| *** david-lyle has joined #openstack-nova | 16:02 | |
| *** hieulq has joined #openstack-nova | 16:02 | |
| *** scottynomad has joined #openstack-nova | 16:02 | |
| *** ducttape_ has joined #openstack-nova | 16:02 | |
| *** shaohe_feng has joined #openstack-nova | 16:02 | |
| *** mlavalle has joined #openstack-nova | 16:02 | |
| *** swamireddy has joined #openstack-nova | 16:02 | |
| *** HenryG has joined #openstack-nova | 16:02 | |
| *** eharney has quit IRC | 16:03 | |
| *** mjura has quit IRC | 16:03 | |
| *** purplerbot has joined #openstack-nova | 16:03 | |
| *** gouthamr_ has joined #openstack-nova | 16:04 | |
| *** mlavalle has quit IRC | 16:04 | |
| *** erlon-airlong has joined #openstack-nova | 16:04 | |
| *** ameade has joined #openstack-nova | 16:05 | |
| *** gouthamr has quit IRC | 16:05 | |
| *** gouthamr_ is now known as gouthamr | 16:06 | |
| *** karlamrhein has joined #openstack-nova | 16:07 | |
| *** Sukhdev has quit IRC | 16:08 | |
| sahid | markus_z: i have a question https://review.openstack.org/#/c/275801 ? | 16:11 |
| *** andreaf has joined #openstack-nova | 16:12 | |
| markus_z | sahid: I'll answer in the review. Thanks for reviewing. | 16:12 |
| *** danpawlik has quit IRC | 16:12 | |
| *** zenoway has quit IRC | 16:14 | |
| *** zenoway has joined #openstack-nova | 16:14 | |
| openstackgerrit | Corentin Ardeois proposed openstack/nova: Add CORS filter to versions pipeline https://review.openstack.org/400785 | 16:15 |
| markus_z | sahid: Hm, maybe here too: My only reason was to have one method argument less. That's all. | 16:17 |
| markus_z | sahid: Do you prefer a different way? | 16:17 |
| *** mvk has quit IRC | 16:17 | |
| *** eharney has joined #openstack-nova | 16:18 | |
| openstackgerrit | Danil Akhmetov proposed openstack/nova: Use proper user and tenant in the owner section of libvirt.xml. https://review.openstack.org/399679 | 16:18 |
| *** Sukhdev_ has joined #openstack-nova | 16:18 | |
| *** zenoway has quit IRC | 16:19 | |
| *** nkrinner is now known as nkrinner_afk | 16:19 | |
| sahid | markus_z: no, no it's that "guess" parameter passed to the method I thought it was a guest object but it's actually a config object | 16:21 |
| *** prometheanfire has joined #openstack-nova | 16:22 | |
| markus_z | sahid: Yeah, that causes a lot of confusion. There's a TODO in line 4346 for that. I was thinking to do the renaming with this change, but it felt like an additional thing out of scope for this change. | 16:23 |
| prometheanfire | mriedem: I'd like your advice on how I handle this since it's been marked invalid | 16:23 |
| prometheanfire | mriedem: https://bugs.launchpad.net/nova/+bug/1629555 | 16:23 |
| openstack | Launchpad bug 1629555 in OpenStack Compute (nova) "Allow the usage of mkisofs instead of genisoimage" [Undecided,Invalid] | 16:23 |
| sahid | markus_z: yes sure, don't do that | 16:23 |
| markus_z | sahid: After the ports thing is fixed and virtlogd is merged, I'll push a proposal to refactor this method. The rename will be part of it. | 16:23 |
| sahid | markus_z: that'd be nice, we really have to dom something in that part of code | 16:24 |
| prometheanfire | mriedem: nova currently uses genisoimage from cdrkit, which is unmaintained and dead code. I could patch the relevent code, but really think the default should change upstream, given the deadness of cdrkit | 16:24 |
| markus_z | sahid: yep, it's a mess | 16:24 |
| *** code-R has joined #openstack-nova | 16:25 | |
| prometheanfire | mriedem: I'm aware I can use a config option to override it, but I can't force my users to do so, which means false bug reports and the like | 16:26 |
| *** tbachman has joined #openstack-nova | 16:26 | |
| prometheanfire | might be able to specify an 'admin' config file that includes that info though | 16:27 |
| *** rcernin has quit IRC | 16:27 | |
| mriedem | prometheanfire: so i'm assuming you don't have genisoimage in gentoo? | 16:27 |
| mriedem | and that's the issue? | 16:27 |
| prometheanfire | we are removing it | 16:28 |
| prometheanfire | https://bugs.gentoo.org/show_bug.cgi?id=591778 | 16:28 |
| openstack | bugs.gentoo.org bug 591778 in Current packages "app-cdr/cdrkit removal request" [Normal,Confirmed] - Assigned to maintainer-needed | 16:28 |
| mriedem | hyper-v has to workaround this by setting mkisofs_cmd to the full path of mkisofs.exe | 16:28 |
| mriedem | prometheanfire: do your packages not ship a nova.conf.sample? | 16:29 |
| mriedem | or just straight up default nova.conf? | 16:29 |
| prometheanfire | ya, we ship the sample | 16:29 |
| mriedem | so why not just set mkisofs_cmd in the nova.conf that you ship? | 16:29 |
| prometheanfire | I think I'll update our init scripts and use a second config file with that option | 16:29 |
| mriedem | if you're worried about users | 16:29 |
| prometheanfire | didn't think of that just now | 16:29 |
| mriedem | i seem to remember gcb or someone trying to change to using mkisofs by default | 16:30 |
| mriedem | but it got hung up on something | 16:30 |
| mriedem | oh no i'm thinking of this https://review.openstack.org/#/c/262744/ | 16:30 |
| sahid | markus_z: something seems to be wrong in post_live_migration_at_destination, i commented on the review | 16:31 |
| *** code-R_ has joined #openstack-nova | 16:31 | |
| *** paul-carlton2 has quit IRC | 16:32 | |
| mriedem | prometheanfire: if you wanted to work on making mkisofs the default for the mkisofs_cmd option, that'd be fine, but probably requires some changes to devstack to make sure that package is installed | 16:32 |
| prometheanfire | ya, it would | 16:34 |
| *** code-R has quit IRC | 16:34 | |
| *** liuyulong__ has joined #openstack-nova | 16:34 | |
| prometheanfire | I'm considering it, you think it'd be well recieved? | 16:34 |
| mriedem | prometheanfire: if genisoimage is going the way of the dinosaur, then sure | 16:35 |
| mriedem | i can't speak for ubuntu or fedora or centos though | 16:35 |
| *** liuyulong_ has quit IRC | 16:38 | |
| prometheanfire | mriedem: I'd expect them to follow, upstream is dead, so it should be pruned imo | 16:38 |
| markus_z | sahid: yeah, true. I think that would even introduce a bug. | 16:38 |
| *** rama_y has joined #openstack-nova | 16:38 | |
| markus_z | extremely unlikely, but possible | 16:38 |
| *** claudiub has quit IRC | 16:38 | |
| mriedem | prometheanfire: if you want to lead that charge be my guest | 16:39 |
| mriedem | would be good to checkpoint with some other packaging people first | 16:39 |
| mriedem | i'm assuming you know those people involved in packaging upstream better tha nme | 16:39 |
| *** claudiub has joined #openstack-nova | 16:39 | |
| *** josecastroleon has quit IRC | 16:40 | |
| prometheanfire | yarp | 16:41 |
| *** nherciu_ has joined #openstack-nova | 16:41 | |
| *** kristian__ has joined #openstack-nova | 16:41 | |
| kristian__ | Hi. May someone help me with this gpu passthrough bug? | 16:41 |
| kristian__ | https://bugs.launchpad.net/nova/+bug/1642419 | 16:41 |
| openstack | Launchpad bug 1642419 in kolla-ansible "GPU Passthrough isn't working" [Low,Confirmed] | 16:41 |
| mriedem | sfinucan: cfriesen_: sounds like something you might know about ^ | 16:42 |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: libvirt: Acquire TCP ports for console during live migration https://review.openstack.org/275801 | 16:42 |
| *** harlowja_at_home has quit IRC | 16:42 | |
| sfinucan | mriedem, kristian__: I'll take a look | 16:43 |
| kristian__ | sfinucan: If you will need ssh for local debugging I may give you access | 16:44 |
| *** rama_y has quit IRC | 16:44 | |
| openstackgerrit | Pavel Gluschak proposed openstack/nova: don't store VNC/SPICE listen addrs in migration data if disabled https://review.openstack.org/396218 | 16:44 |
| diana_clarke | Here's an easy api-ref review [1]. This fix is the result of multiple ops being unable to figure out how to see server groups when then have more than osapi-max-limit. | 16:45 |
| diana_clarke | [1] https://review.openstack.org/#/c/400851/ | 16:45 |
| * markus_z leaves | 16:45 | |
| *** markus_z has quit IRC | 16:45 | |
| diana_clarke | Turns out paging exists for server groups, but that it's just not documented. | 16:45 |
| *** irenab has quit IRC | 16:49 | |
| cdent | macsz: is https://bugs.launchpad.net/nova/+bug/1623483 still relevant now that the aggregates stuff has merged, or are you planning to be nice and tidy and do some kind of microversion-based cleanup? | 16:49 |
| openstack | Launchpad bug 1623483 in OpenStack Compute (nova) "In placement api links in resource provider representation links to aggregates but we never merged aggregates support" [Low,In progress] - Assigned to Maciej Szankin (mszankin) | 16:49 |
| *** claudiub has quit IRC | 16:50 | |
| *** jschwarz is now known as jschwarz|mtg | 16:50 | |
| macsz | cdent: i was wandering wether this bug is really a bug right now | 16:51 |
| cdent | I think it may not be anymore | 16:51 |
| *** gjayavelu has joined #openstack-nova | 16:51 | |
| *** kaisers_ has quit IRC | 16:51 | |
| macsz | cdent: and yeah, it seems like there is none right now, so we can close this one and maybe, as you suggested, just do some refactors / micro | 16:51 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: docs: cleanup wording for 'SOFT_DELETED' in api-guide https://review.openstack.org/403802 | 16:52 |
| pkholkin | mriedem: dansmith: hey folks, could you please review https://review.openstack.org/#/c/381912/ | 16:52 |
| lbeliveau | kristian__: it's definitively not a nova issue, something with libvirt, looking ... | 16:52 |
| pkholkin | I tried to be very active on this | 16:52 |
| mriedem | diana_clarke: that would probably be an easy change to the api-ref docs | 16:52 |
| cdent | macsz: yeah, I'll go ahead and comment and mark it dead | 16:52 |
| pkholkin | and got many positive feedback | 16:52 |
| mriedem | diana_clarke: you could report a bug against the nova api-ref to keep track of that | 16:52 |
| mriedem | pkholkin: spec freeze was 11/17 | 16:52 |
| pkholkin | the implementation is ready, also I found some things to discuss | 16:52 |
| diana_clarke | mriedem: I did the fix here (https://review.openstack.org/#/c/400851/) unless I've missunderstood. | 16:52 |
| kristian__ | lbeliveau: so what are you suggesting? | 16:52 |
| pkholkin | mriedem: yeah, I know, I think we can make an exception for this | 16:53 |
| *** tbachman has quit IRC | 16:53 | |
| pkholkin | I found this list e.g. https://etherpad.openstack.org/p/nova-ocata-spec-review-sprint | 16:53 |
| lbeliveau | kristian__: which platform/linux distro are you using ? | 16:53 |
| pkholkin | this cycle is short, but I am sure I can make it | 16:54 |
| kristian__ | Ubuntu 16.04.1 x64 | 16:54 |
| *** irenab has joined #openstack-nova | 16:54 | |
| kristian__ | with kolla | 16:54 |
| kristian__ | and they said that I should talk with you | 16:54 |
| *** claudiub has joined #openstack-nova | 16:54 | |
| *** rama_y has joined #openstack-nova | 16:54 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Honor deleted flag in instance_create() https://review.openstack.org/403756 | 16:55 |
| bauzas | mriedem: dansmith: interesting things I discovered when looking at how to fix https://github.com/openstack/nova/blob/master/nova/objects/build_request.py#L85 | 16:55 |
| dansmith | mriedem: bauzas: ^ one test definitely thought it could create with deleted=True :) | 16:55 |
| *** claudiub|2 has joined #openstack-nova | 16:55 | |
| bauzas | mriedem: dansmith: we basically assume that the primitived instance object is correct, so when doing https://github.com/openstack/nova/blob/master/nova/objects/build_request.py#L64-L65 we set the right value to deleted | 16:56 |
| dansmith | bauzas: if it's set in the blob, sure | 16:56 |
| mriedem | diana_clarke: ok, so offset is basically equivalent to marker? | 16:56 |
| *** andreas_s has quit IRC | 16:56 | |
| pkholkin | mriedem: I got +1 from two cores in nova (but not cores in specs), +1 from the original author of the spec, and some others | 16:56 |
| bauzas | dansmith: I can't find situations where we're primitiving an instance not having that field set | 16:57 |
| mriedem | pkholkin: it's not a matter of *you* making it, it's a matter of review bandwidth | 16:57 |
| *** anush has joined #openstack-nova | 16:57 | |
| *** rmart04 has quit IRC | 16:57 | |
| pkholkin | mriedem: and that's the problem | 16:58 |
| bauzas | dansmith: the problem is that if that blob doesn't have the deleted field set, I'm unable to know whether the instance was deleted or not | 16:58 |
| diana_clarke | mriedem: No, offset is old-school. It's more like: "start at item 50, and give me the next 25 items" Just simple integers. | 16:58 |
| pkholkin | I think we should not lose another cycle waiting | 16:58 |
| bauzas | dansmith: as I don't have the DB value, just the object boolean | 16:58 |
| *** claudiub has quit IRC | 16:59 | |
| dansmith | bauzas: um, what? if it's unset it's not deleted | 16:59 |
| diana_clarke | mriedem: I manually tested that it;s working: http://paste.openstack.org/show/589962/ | 17:00 |
| pkholkin | mriedem: bauzas promised to support this, too | 17:00 |
| lbeliveau | kristian__: same guest without GPU passthrough works right ? | 17:00 |
| openstackgerrit | Pavel Gluschak proposed openstack/nova: libvirt: make live migration possible with Virtuozzo https://review.openstack.org/355805 | 17:00 |
| bauzas | pkholkin: you mean adding the pike structure ? | 17:01 |
| kristian__ | yes | 17:01 |
| *** sahid has quit IRC | 17:01 | |
| mriedem | bauzas: no he means this https://review.openstack.org/#/c/381912/ | 17:01 |
| *** gjayavelu has quit IRC | 17:01 | |
| pkholkin | yes | 17:01 |
| *** BobBall is now known as BobBall_AWOL | 17:01 | |
| lbeliveau | kristian__: it doesn't makes sense that libvirt tries to do something with ethtool on this kind of device | 17:01 |
| lbeliveau | looking ... | 17:01 |
| kristian__ | yes | 17:02 |
| kristian__ | weird | 17:02 |
| bauzas | well, those two filters are disabled by default | 17:02 |
| bauzas | so, the problem is really with operators deploying them | 17:02 |
| mriedem | diana_clarke: do you know if we have any functional tests for that? | 17:02 |
| bauzas | dansmith: okay, I can turn that way, and consider that if the field is unset, then it's for a non-deleted instance | 17:03 |
| bauzas | we actually default "deleted" to False | 17:03 |
| *** matrohon has quit IRC | 17:04 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Small improvements to placement.rst https://review.openstack.org/403811 | 17:04 |
| mriedem | diana_clarke: not that you'd need to add one in that docs patch | 17:04 |
| *** lpetrut has quit IRC | 17:04 | |
| *** daemontool has quit IRC | 17:04 | |
| *** armax has joined #openstack-nova | 17:04 | |
| diana_clarke | mriedem: It looks like only the utility method is tested, but not the 5 or so endpoints that employ simple paging with limit & offset. | 17:05 |
| mriedem | diana_clarke: also, looks like we could get an IndexError out of that offset being wrong? like in your example, if i specify offset=2? | 17:05 |
| *** gjayavelu has joined #openstack-nova | 17:05 | |
| mriedem | i'm assuming you'd get a 500 for passing a bad offset | 17:05 |
| mriedem | it's pretty nasty that we limit in python rather than the db | 17:06 |
| pkholkin | do we have some exceptions to specs? | 17:06 |
| diana_clarke | mriedem: it's using python slices under the hood, so I don't think so. When foo = [1, 2, 3] foo[100] is fine. | 17:06 |
| diana_clarke | I mean, foo[:100] | 17:06 |
| lbeliveau | kristian__: I tried something similar not so long ago with an older version of libvirt (1.2.17), and it used to work | 17:06 |
| *** tbachman has joined #openstack-nova | 17:07 | |
| *** lpetrut has joined #openstack-nova | 17:07 | |
| lbeliveau | kristian__: Maybe you should try with an older release of libvirt | 17:07 |
| mriedem | diana_clarke: yeah i guess we just get an empty list then | 17:07 |
| diana_clarke | mriedem: I'll look at adding a functional test. I also have looking at the other 4 endpoints on my TODO list. But this one is stumping actual users in the wild. | 17:08 |
| kristian__ | ok so 1.2.17 | 17:08 |
| *** jschwarz|mtg is now known as jschwarz | 17:08 | |
| pkholkin | bauzas: I also think that we have a bug in current implementation of aggregate_image_properties_isolation filter | 17:08 |
| kristian__ | I will get in touch with kolla team to help me out | 17:08 |
| lbeliveau | kristian__: The libvirt XML for the PCI passthrough looks good ... so that's why I'm encline to think it is a libvirt issue (googled it a bit, but no luck) | 17:08 |
| lbeliveau | kristian__: ok let me know, you can ping me | 17:08 |
| pkholkin | but in general it is not related to strict case in my spec | 17:09 |
| penick | mriedem: i’ll take a look at https://review.openstack.org/#/c/395959/ later today (vendordata) | 17:09 |
| *** andrey-mp has joined #openstack-nova | 17:09 | |
| *** penick has quit IRC | 17:09 | |
| kristian__ | ok we will see | 17:09 |
| *** gjayavelu has quit IRC | 17:09 | |
| kristian__ | lbeliveau: how can I get specific libvirt version? | 17:11 |
| mriedem | diana_clarke: yup, +2 | 17:11 |
| mriedem | nice find | 17:11 |
| *** tbachman has quit IRC | 17:12 | |
| mriedem | pkholkin: we don't have an exception process for ocata specs | 17:12 |
| mriedem | pkholkin: there are some that will be marged after the spec freeze but that's because they are tied to priorities | 17:12 |
| pkholkin | that's too bad to hear | 17:13 |
| pkholkin | I pinged folks many times, I think I got too few attention to this ( | 17:13 |
| mriedem | pkholkin: yes, but also, take a look at this list https://blueprints.launchpad.net/nova/ocata | 17:14 |
| *** archies has joined #openstack-nova | 17:14 | |
| lbeliveau | kristian__: look at this http://askubuntu.com/questions/138284/how-to-downgrade-a-package-via-apt-get (worst case get the source and building it from scratch) | 17:15 |
| mriedem | pkholkin: if you want to make a case for an exception in the mailing list and try to get some support there, that's fine with me, but i don't really want to be opening the door for a bunch of other spec exception requests for ocata given how much we already have on our plate and how little time we have left in the release | 17:15 |
| *** gjayavelu has joined #openstack-nova | 17:15 | |
| *** mdrabe has quit IRC | 17:15 | |
| *** mdrabe_ has joined #openstack-nova | 17:15 | |
| pkholkin | mriedem: ok, I think I will give a try, thanks! | 17:16 |
| *** mlavalle has joined #openstack-nova | 17:16 | |
| kristian__ | ok | 17:17 |
| pkholkin | bauzas: I think you should look at this, what's your opinion? https://review.openstack.org/#/c/403640/ | 17:17 |
| pkholkin | this changes the behaviour of the old filter, but I think it is a bug | 17:18 |
| kristian__ | lbeliveau: no such package found | 17:18 |
| *** pcaruana has quit IRC | 17:19 | |
| mriedem | jaypipes: is https://review.openstack.org/#/c/386710/1/specs/ocata/approved/nested-resource-providers.rst@100 something you can flesh out a bit in the spec? that's all that's holding me up there, documenting the plan from the whiteboard session at the summit so we don't lose that. | 17:19 |
| jaypipes | mriedem: yeah, gimme about an hour to push a rev. | 17:19 |
| mriedem | thanks | 17:19 |
| jaypipes | np, thx for your patience matt. | 17:19 |
| jaypipes | mriedem: too bad about your vikings, btw. | 17:20 |
| jaypipes | mriedem: if it makes you feel any better, Browns are 0-12. | 17:20 |
| *** gjayavelu has quit IRC | 17:20 | |
| mriedem | jaypipes: as i said in the ML, i knew they'd lose | 17:21 |
| jaypipes | heh | 17:21 |
| *** kaisers has joined #openstack-nova | 17:21 | |
| mriedem | 0-12 means high draft picks to squander | 17:22 |
| kristian__ | lbliveau: downloading 1.2.16 from ubuntu | 17:22 |
| *** ducttape_ has quit IRC | 17:22 | |
| *** ducttape_ has joined #openstack-nova | 17:22 | |
| *** stewie925 has joined #openstack-nova | 17:26 | |
| kristian__ | dberrange: Please may you help me with installing custom libvirt? | 17:29 |
| *** unicell1 has quit IRC | 17:29 | |
| *** nic has joined #openstack-nova | 17:33 | |
| archies | @mriedem: hi about the https://bugs.launchpad.net/nova/+bug/1636489. As discussed would like to share the script with you to get your views on it. | 17:34 |
| openstack | Launchpad bug 1636489 in OpenStack Compute (nova) "Volume attachment fails for all the available instances after running different volume operations for 1-2 hours or more." [Undecided,New] | 17:34 |
| *** code-R_ has quit IRC | 17:35 | |
| *** code-R has joined #openstack-nova | 17:35 | |
| mriedem | archies: just attach the script in the bug report | 17:35 |
| archies | ok | 17:35 |
| openstackgerrit | Gábor Antal proposed openstack/nova: Transform instance.volume_attach notification https://review.openstack.org/401992 | 17:36 |
| *** Apoorva has joined #openstack-nova | 17:37 | |
| *** mlavalle has quit IRC | 17:39 | |
| *** lpetrut has quit IRC | 17:39 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Default deleted if the instance from BuildRequest is not having it https://review.openstack.org/403821 | 17:40 |
| bauzas | dansmith: mriedem: first approach for fixing the bad behaviour of instance.deleted = 0 | 17:40 |
| * bauzas afk for 2 hours | 17:40 | |
| mriedem | gibi: i'm looking at https://blueprints.launchpad.net/nova/+spec/additional-notification-fields-for-searchlight and noting the fields to add to the instance payload, and i see those are based on what's in the first bug report, but i'm also wondering about some other fields that we return for server show in microversion >= 2.3: http://docs.openstack.org/developer/nova/api_microversion_history.html#maximum-in-kilo | 17:41 |
| mriedem | gibi: maybe we only care to make incremental changes as searchlight needs them? | 17:41 |
| *** prometheanfire has left #openstack-nova | 17:41 | |
| *** smatzek has joined #openstack-nova | 17:42 | |
| *** mlavalle has joined #openstack-nova | 17:42 | |
| *** gjayavelu has joined #openstack-nova | 17:43 | |
| openstackgerrit | Gábor Antal proposed openstack/nova: Transform instance.volume_attach notification https://review.openstack.org/401992 | 17:43 |
| *** claudiub has joined #openstack-nova | 17:44 | |
| *** gjayavelu has quit IRC | 17:44 | |
| openstackgerrit | Gábor Antal proposed openstack/nova: Transform instance.volume_attach notification https://review.openstack.org/401992 | 17:45 |
| mriedem | dansmith: you're against this right? https://review.openstack.org/#/c/403694/ | 17:46 |
| mriedem | might want to drop a hammer on it | 17:46 |
| *** claudiub|2 has quit IRC | 17:46 | |
| *** paw has quit IRC | 17:47 | |
| openstackgerrit | STEW TY proposed openstack/nova: Complete verification of servers-action-fixed-ip.inc https://review.openstack.org/338606 | 17:47 |
| mriedem | bauzas: i'm not sure about the fix here https://review.openstack.org/#/c/403821/1/nova/objects/build_request.py | 17:50 |
| mriedem | and why you're not doing what we talked about earlier | 17:50 |
| openstackgerrit | Jay Pipes proposed openstack/nova-specs: Add concept of nested resource providers https://review.openstack.org/386710 | 17:50 |
| jaypipes | mriedem: ^ | 17:50 |
| mriedem | thanks | 17:50 |
| *** prateek_ has quit IRC | 17:50 | |
| *** Sukhdev_ has quit IRC | 17:51 | |
| *** weshay is now known as weshay_pto | 17:52 | |
| *** bvanhav_ has quit IRC | 17:55 | |
| *** gcb has quit IRC | 17:55 | |
| *** bvanhav_ has joined #openstack-nova | 17:56 | |
| openstackgerrit | Jay Pipes proposed openstack/nova-specs: Add concept of nested resource providers https://review.openstack.org/386710 | 17:57 |
| openstackgerrit | Matt Riedemann proposed openstack/nova-specs: Add concept of nested resource providers https://review.openstack.org/386710 | 17:57 |
| jaypipes | mriedem: sorry, missed your reviiew typo fixes. now done | 17:57 |
| jaypipes | err... | 17:57 |
| mriedem | heh | 17:57 |
| mriedem | teamwork | 17:58 |
| *** sreeram_v has joined #openstack-nova | 18:00 | |
| openstackgerrit | STEW TY proposed openstack/nova: Complete verification of servers-action-fixed-ip.inc https://review.openstack.org/338606 | 18:00 |
| dansmith | mriedem: yeah, I'll hit it | 18:01 |
| *** Serlex has quit IRC | 18:01 | |
| *** ihrachys has quit IRC | 18:02 | |
| *** sudipto has quit IRC | 18:02 | |
| *** Guest81242 has quit IRC | 18:02 | |
| *** ralonsoh_ has quit IRC | 18:02 | |
| *** catintheroof has joined #openstack-nova | 18:04 | |
| *** portdirect_away is now known as portdirect | 18:04 | |
| *** gjayavelu has joined #openstack-nova | 18:04 | |
| *** sreeram_v has quit IRC | 18:04 | |
| *** lucasagomes is now known as lucas-afk | 18:05 | |
| *** gjayavelu has quit IRC | 18:06 | |
| *** derekh has quit IRC | 18:06 | |
| *** jaosorior has quit IRC | 18:08 | |
| *** TTris has quit IRC | 18:10 | |
| *** eharney has quit IRC | 18:11 | |
| openstackgerrit | Gábor Antal proposed openstack/nova: Transform instance.volume_attach notification https://review.openstack.org/401992 | 18:11 |
| dansmith | mriedem: replied to your comments on my patch | 18:11 |
| dansmith | if you're cool, I'll do another on top for the instance create change | 18:11 |
| mriedem | looking | 18:12 |
| mriedem | https://www.youtube.com/watch?v=IJJzlXy2YWo | 18:12 |
| dansmith | heh | 18:12 |
| *** slaweq_ has quit IRC | 18:13 | |
| *** jdurgin1 has quit IRC | 18:13 | |
| *** archies has quit IRC | 18:17 | |
| *** sambetts is now known as sambetts|afk | 18:17 | |
| mriedem | dansmith: replied https://review.openstack.org/#/c/403756/1/nova/db/sqlalchemy/api.py | 18:18 |
| mriedem | i'm not sure why we don't just error out if you're trying to create a new soft-deleted instance, it's not something we support | 18:19 |
| dansmith | mriedem: we don't support it at the api controller layer, but this is model, right? | 18:19 |
| dansmith | I mean, we can throw an error too, but I don't see why we should | 18:19 |
| dansmith | s/too/instead | 18:20 |
| dansmith | like, what happens when we start wanting to move instances to cell0 when they're deleted (making this up)? | 18:20 |
| *** nherciu_ has quit IRC | 18:20 | |
| mriedem | i don't feel great either way :) we could also handle this for any other soft-deletable resource in the model | 18:20 |
| dansmith | user doesn't care | 18:20 |
| *** lpetrut has joined #openstack-nova | 18:21 | |
| mriedem | so maybe my issue is precedent | 18:21 |
| dansmith | I dunno why I'm arguing, I don't care | 18:21 |
| dansmith | I'll just change it to what you want | 18:21 |
| *** anush has quit IRC | 18:23 | |
| *** nherciu has joined #openstack-nova | 18:25 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Check deleted flag in Instance.create() https://review.openstack.org/403756 | 18:27 |
| *** browne has joined #openstack-nova | 18:28 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Check deleted flag in Instance.create() https://review.openstack.org/403756 | 18:29 |
| *** Swami has joined #openstack-nova | 18:30 | |
| *** penick has joined #openstack-nova | 18:31 | |
| dansmith | cripes | 18:31 |
| *** kristian__ has quit IRC | 18:31 | |
| penick | egads | 18:32 |
| * dansmith learns to indent properly | 18:32 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Check deleted flag in Instance.create() https://review.openstack.org/403756 | 18:32 |
| *** eharney has joined #openstack-nova | 18:35 | |
| *** code-R has quit IRC | 18:38 | |
| *** code-R has joined #openstack-nova | 18:38 | |
| openstackgerrit | STEW TY proposed openstack/nova: Transform instance.rescue notifications https://review.openstack.org/385644 | 18:39 |
| *** kristian__ has joined #openstack-nova | 18:39 | |
| *** gszasz has quit IRC | 18:39 | |
| *** bvanhav__ has joined #openstack-nova | 18:40 | |
| *** bvanhav__ is now known as bvanahv | 18:40 | |
| *** bvanahv is now known as bvanhav | 18:40 | |
| *** bvanhav_ has quit IRC | 18:41 | |
| *** unicell has joined #openstack-nova | 18:41 | |
| *** lpetrut has quit IRC | 18:42 | |
| *** rmart04 has joined #openstack-nova | 18:42 | |
| *** mgoddard_ has joined #openstack-nova | 18:44 | |
| mriedem | dansmith: ok i've updated https://review.openstack.org/#/c/403101/ | 18:44 |
| mriedem | to depend on your change | 18:44 |
| openstackgerrit | Dan Smith proposed openstack/nova: Make RPCFixture support multiple connections https://review.openstack.org/396417 | 18:44 |
| dansmith | okay | 18:44 |
| mriedem | i have a feeling it won't fix the bug though b/c we'll be trading one failure for another | 18:44 |
| dansmith | you need bauzas too | 18:44 |
| mriedem | yeah | 18:44 |
| dansmith | my previous one would have fixed it | 18:45 |
| dansmith | now it doesn't | 18:45 |
| mriedem | so maybe i'll roll his on top of yours and address my issue in his | 18:45 |
| dansmith | yeah | 18:45 |
| *** bnemec has quit IRC | 18:45 | |
| dansmith | eff, I might have another pep8 fail in that one | 18:46 |
| *** rmart04 has quit IRC | 18:47 | |
| *** amrith has joined #openstack-nova | 18:47 | |
| mriedem | ok update yours and i'll cherry pick it on top of bauzas' | 18:47 |
| amrith | mriedem, ping. | 18:47 |
| *** mgoddard has quit IRC | 18:47 | |
| amrith | mriedem, we've reverted the nova-networking change in trove (see: https://review.openstack.org/#/c/397371/) | 18:47 |
| *** mgoddard_ has quit IRC | 18:48 | |
| *** lpetrut has joined #openstack-nova | 18:52 | |
| mriedem | amrith: and things are passing on master with the gate-trove-functional-dsvm-* job change? | 18:52 |
| amrith | dansmith, mriedem I am proposing a change to undo the earlier reverts ... stay tuned | 18:52 |
| dansmith | ahead of schedule, I like it | 18:52 |
| amrith | mriedem, no they are not all passing but we are far enough along that we have reverted the changes in our ci jobs to force nova networking on. | 18:53 |
| amrith | so now, we are running the CI with neutron and the changes for that are almost ready to go | 18:53 |
| amrith | we will be disabling specific tests which we have to work on | 18:53 |
| amrith | we are doing it this way because another significant set of changes (ubuntu-trusty to xenial) is also queued | 18:53 |
| amrith | and we want to get them both as soon as possible. | 18:54 |
| amrith | and this (the jump off the deep end) approach seems to be the best one at this point. | 18:54 |
| jaypipes | cdent: you mind if you or I fix that "containing application will catch a not found here" thing in a separate follow-up patch? | 18:55 |
| cdent | jaypipes: Yeah, sure, or even not at all. I only mentioned it to make reference to the other changes. the -1 was solely for the microversion methods | 18:56 |
| jaypipes | cdent: k, cool. I like the suggestion, just don't want to hold this up any further. | 18:56 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: REST API for resource classes https://review.openstack.org/391959 | 18:56 |
| cdent | +1 | 18:56 |
| jaypipes | cdent: ^^ | 18:56 |
| *** tlian has joined #openstack-nova | 18:58 | |
| openstackgerrit | amrith proposed openstack/nova: Revert "Log a warning when starting nova-net in non-cellsv1 deployments" https://review.openstack.org/403853 | 19:00 |
| openstackgerrit | Dan Smith proposed openstack/nova: Check deleted flag in Instance.create() https://review.openstack.org/403756 | 19:01 |
| dansmith | mriedem: another test depended on it just working ^ | 19:01 |
| *** jgriffith is now known as jgriffith_away | 19:02 | |
| openstackgerrit | amrith proposed openstack/nova: Revert "Revert "Make n-net refuse to start unless using CellsV1"" https://review.openstack.org/403855 | 19:02 |
| amrith | dansmith, mriedem please see commits ^^ | 19:03 |
| amrith | thanks for the help with these | 19:03 |
| *** irenab has quit IRC | 19:05 | |
| *** amoralej is now known as amoralej|off | 19:11 | |
| *** kristian__ has quit IRC | 19:13 | |
| mriedem | mnestratov: questions about https://github.com/openstack-infra/devstack-gate/blob/master/functions.sh#L666 in #openstack-infra | 19:13 |
| mriedem | gah this test_build_request is messed up | 19:19 |
| *** kristian__ has joined #openstack-nova | 19:19 | |
| *** kristian__ has quit IRC | 19:24 | |
| *** lpetrut has quit IRC | 19:28 | |
| openstackgerrit | Ed Leafe proposed openstack/nova: Add a retry loop to ResourceClass creation https://review.openstack.org/399170 | 19:29 |
| *** kristian__ has joined #openstack-nova | 19:30 | |
| *** rajathagasthya has joined #openstack-nova | 19:30 | |
| dansmith | melwitt: I meant to ask you about this last week, but on that rpc fixture patch, I'm getting this in devstack setup: http://logs.openstack.org/17/396417/13/check/gate-tempest-dsvm-neutron-full-ubuntu-xenial/db9c7a3/logs/devstacklog.txt.gz#_2016-11-28_19_05_24_610 | 19:33 |
| dansmith | melwitt: I'm fairly confused about the presence (or not) of transport_url, and whether we own it, o.msg owns it, etc | 19:33 |
| dansmith | it's clearly there in our config: http://logs.openstack.org/17/396417/13/check/gate-tempest-dsvm-neutron-full-ubuntu-xenial/db9c7a3/logs/etc/nova/nova.conf.txt.gz | 19:35 |
| dansmith | melwitt: any ideas? | 19:35 |
| *** rajathagasthya has quit IRC | 19:35 | |
| *** victorhe has quit IRC | 19:38 | |
| *** victorhe has joined #openstack-nova | 19:39 | |
| *** swann has joined #openstack-nova | 19:40 | |
| *** josecastroleon has joined #openstack-nova | 19:41 | |
| *** swann has left #openstack-nova | 19:42 | |
| *** sam_nowitzki has joined #openstack-nova | 19:42 | |
| *** matrohon has joined #openstack-nova | 19:44 | |
| *** swann has joined #openstack-nova | 19:44 | |
| *** READ10 has quit IRC | 19:45 | |
| *** josecastroleon has quit IRC | 19:46 | |
| *** slaweq_ has joined #openstack-nova | 19:51 | |
| *** Sukhdev_ has joined #openstack-nova | 19:53 | |
| *** ijw has joined #openstack-nova | 19:53 | |
| *** kfarr has joined #openstack-nova | 19:59 | |
| *** rfolco_ has quit IRC | 20:01 | |
| *** bnemec has joined #openstack-nova | 20:03 | |
| *** paw has joined #openstack-nova | 20:04 | |
| *** slaweq_ has quit IRC | 20:04 | |
| *** whenry has joined #openstack-nova | 20:04 | |
| *** mnasiadka has joined #openstack-nova | 20:05 | |
| *** woodster_ has joined #openstack-nova | 20:05 | |
| *** whenry has quit IRC | 20:06 | |
| *** browne has quit IRC | 20:08 | |
| *** tbachman has joined #openstack-nova | 20:10 | |
| *** browne has joined #openstack-nova | 20:10 | |
| mriedem | dansmith: a walk cleared something up for me here https://review.openstack.org/#/c/403821/1/nova/objects/build_request.py@87 | 20:12 |
| mriedem | i think we literally can't set the instance.deleted field based on the serialized instance object in the build request, we'd have to get the instance from the db based on it's uuid, which we probably don't care to do | 20:12 |
| mriedem | so i think bauzas' change is probably the best we can do | 20:12 |
| *** avolkov has quit IRC | 20:13 | |
| dansmith | I don't understand why we're not just setting that to False (like the 0 now) | 20:13 |
| dansmith | we know it's not deleted yet because it's not even created | 20:13 |
| dansmith | and we handle deleting the instance after creation, if the build request is deleted from underneath us | 20:14 |
| dansmith | using the set_defaults() there isn't even right, in case that later means something else (unlikely, but..) | 20:14 |
| dansmith | I'm confused about why there's so much fuss over this seemingly simple thing, I guess :) | 20:14 |
| dansmith | and I'm not sure I understand your second comment about soft-deletedness | 20:14 |
| *** victorhe has quit IRC | 20:14 | |
| mriedem | well this morning we were talking about doing logic like https://github.com/openstack/nova/blob/646151ff1b8d481b70884d201a1b395d904e35b0/nova/objects/instance.py#L334 in buildrequest, which is why i was on a goose chase | 20:15 |
| dansmith | instance.deleted = True means "soft deleted" | 20:15 |
| mriedem | my latest comment is that we can't do ^ | 20:15 |
| mriedem | because we don't have a db instance primitive, we have a serialized instance object | 20:15 |
| dansmith | who was talking about that? | 20:15 |
| mriedem | so deleted isn't the same thing | 20:15 |
| mriedem | you, me and bauzas | 20:15 |
| mriedem | (9:11:02 AM) dansmith: mriedem: definitely | 20:15 |
| dansmith | well, right, but I mean.. I don't recall discussing the integerness | 20:16 |
| dansmith | mriedem: I was saying "definitely use instance.deleted=False there instead of =0" | 20:16 |
| *** tbachman has quit IRC | 20:16 | |
| *** victorhe has joined #openstack-nova | 20:16 | |
| dansmith | mriedem: maybe we should hangout to sync up? | 20:16 |
| mriedem | if it's just a matter of setting self.instance.deleted = False in the build request then i'll just do that | 20:17 |
| dansmith | yeah, that's all | 20:17 |
| mriedem | as i said, i was on a different goose chase | 20:17 |
| dansmith | I dunno why bauzas wanted the set_defaults() thing because there's no reason for that | 20:17 |
| dansmith | note that we could also set all these things in the get_new_instance() helper thing and not rewrite them every time we do a load | 20:18 |
| dansmith | if we persist the instance with deleted set, it'll come out with deleted set | 20:18 |
| *** jgriffith_away is now known as jgriffith | 20:18 | |
| mriedem | that's closer to this then https://review.openstack.org/#/c/403035/2/nova/objects/build_request.py | 20:19 |
| dansmith | mriedem: ah, maybe he wants that because it will not override the value if it's stored with one | 20:19 |
| dansmith | if we were setting this in get_new_instance() or whatever, we'd definitely want to just force it =False | 20:19 |
| dansmith | but maybe there's a reasonable point to be made to use the defaults thing there just so we're not always loading it out and un-deleting it, even though that will never happen | 20:20 |
| *** victorhe has quit IRC | 20:20 | |
| dansmith | mriedem: what is your specific complaint with the defaults thing? | 20:20 |
| dansmith | or just because it's different? | 20:20 |
| mriedem | i don't have one, you had the problem with using obj_set_defaults :) | 20:20 |
| dansmith | okay, you commented initially though | 20:21 |
| dansmith | ANYWAY | 20:21 |
| mriedem | wait, who's on first? | 20:21 |
| dansmith | the smallest fix is to just s/0/False/ | 20:21 |
| *** bvanhav has quit IRC | 20:21 | |
| dansmith | we could then argue that we should move several of these things to get_new_instance() | 20:21 |
| dansmith | but, smallest change wins, IMHO | 20:21 |
| mriedem | dansmith: the simplest thing for me is just +2ing this https://review.openstack.org/#/c/403821/ | 20:22 |
| mriedem | as-is | 20:22 |
| *** rajathagasthya has joined #openstack-nova | 20:22 | |
| dansmith | I guess | 20:23 |
| *** yuanyuan has joined #openstack-nova | 20:25 | |
| *** slaweq_ has joined #openstack-nova | 20:25 | |
| *** yamahata has quit IRC | 20:25 | |
| *** victorhe has joined #openstack-nova | 20:25 | |
| *** lpetrut has joined #openstack-nova | 20:26 | |
| *** victorhe has quit IRC | 20:27 | |
| *** victorhe has joined #openstack-nova | 20:28 | |
| rajathagasthya | Hello! Relative newbie to working with nova. I was wondering if nova generates notifications/events if a hypervisor is down? I took a look at https://wiki.openstack.org/wiki/SystemUsageData, but doesn't look like it has hypervisor events. | 20:29 |
| *** ociuhandu has quit IRC | 20:30 | |
| *** anush has joined #openstack-nova | 20:31 | |
| mriedem | rajathagasthya: that wiki is old | 20:31 |
| mriedem | see https://github.com/openstack/nova/blob/master/nova/objects/service.py#L319 | 20:31 |
| *** claudiub has quit IRC | 20:33 | |
| jaypipes | gibi: you around by any chance? | 20:33 |
| mriedem | dansmith: +2 on your change. the ceilometer pg change is passing now with both fixes https://review.openstack.org/#/c/403101/ | 20:34 |
| mriedem | jaypipes: want to get this in? https://review.openstack.org/#/c/403756/ | 20:34 |
| jaypipes | mriedem: on it. | 20:35 |
| rajathagasthya | mriedem: Thanks! Is there a list of all the event types in nova somewhere? | 20:35 |
| *** oanson has quit IRC | 20:36 | |
| jaypipes | rajathagasthya: those aren't the events you're looking for... | 20:37 |
| mriedem | there are existing versioned notifications here http://docs.openstack.org/developer/nova/notifications.html#existing-versioned-notifications | 20:37 |
| mriedem | but that's missing the service notifications pointed out above, which is a bug | 20:37 |
| jaypipes | rajathagasthya: if I had to guess, you are actually looking for an event bus that notifies guests when the hypervisor/host experiences failures/faults, not when a hypervisor is manually marked "down for maintenance". | 20:37 |
| mriedem | legacy notifications are here https://github.com/openstack/nova/blob/master/nova/rpc.py#L211 | 20:37 |
| rajathagasthya | Ah, thanks. It looks like most of them are instance related notifications. I'd like to generate notifications based on hypervisor availability and send it to syslog so it can be handled by the syslog consumer. So far I thought of checking if nova-compute service is up and then pinging hypervisors to see if they're up. If nova already somehow provides such events, I thought I'd just subscribe to them. | 20:43 |
| jaypipes | yup, just as I thought :) | 20:43 |
| rajathagasthya | But if someone has suggestions on the best way to do this, I'd appreciate it! | 20:43 |
| jaypipes | rajathagasthya: when you say "hypervisor availability" what precisely are you referring to? | 20:44 |
| *** ijw has quit IRC | 20:45 | |
| *** rtheis has quit IRC | 20:45 | |
| rajathagasthya | jaypipes: Just checking if the hypervisor is up or down for starting instances on it. If my ping to the hypervisor doesn't work, it's not available and I'd like to know that to fix it | 20:45 |
| *** moshele has joined #openstack-nova | 20:46 | |
| jaypipes | rajathagasthya: is "my ping" a tenant-level communication or an administrative/management-level communication? | 20:47 |
| *** Sukhdev_ has quit IRC | 20:47 | |
| rajathagasthya | jaypipes: Admin level communication | 20:47 |
| diana_clarke | python-novaclient has 4 different CLIs that say something like: "if limit == -1, all keypairs|flavors|servers|hypervisors will be returned", but I can't for the life of me find the magic -1 case code. Anyone know where it is? | 20:48 |
| diana_clarke | I'm asking because I was thinking of doing something similar for simple tenant usage, but I'm not sure if I should push this magic case up into nova or the nova client. | 20:48 |
| *** andrey-mp has quit IRC | 20:48 | |
| *** beagles is now known as beagles_biab | 20:49 | |
| jaypipes | rajathagasthya: I think what you're looking for there is something like an npre check script that monitors the results of something like `virsh list` then. Remember that the hypervisor's state != the state of the nova-compute daemon. The former is the thing that launches guests on a host. The latter is the control plane for OpenStack. what mriedem described earlier is the event notification for | 20:49 |
| jaypipes | the latter, not the former. | 20:49 |
| mriedem | diana_clarke: not sure how that would get past this https://github.com/openstack/nova/blob/master/nova/api/openstack/common.py#L205 | 20:49 |
| mriedem | unless the -1 is purely novaclient sugar that does the paging for you | 20:49 |
| diana_clarke | mriedem: I don't see it there either. Maybe it's just lies... perhaps I should manually try the -1 cases and see if they even work. | 20:50 |
| diana_clarke | I'm usually pretty good at grepping my way to answers, but I can't find this magic -1 code anywhere. | 20:51 |
| mriedem | diana_clarke: i see it | 20:51 |
| diana_clarke | nice! | 20:51 |
| mriedem | https://github.com/openstack/python-novaclient/blob/6.0.0/novaclient/v2/servers.py#L813 | 20:51 |
| mriedem | https://github.com/openstack/python-novaclient/blob/6.0.0/novaclient/v2/servers.py#L839 | 20:51 |
| *** slaweq_ has quit IRC | 20:51 | |
| diana_clarke | mriedem: I did find that one already... but not the other three. I should have mentioned that. | 20:52 |
| mriedem | i'm not really sure how ^ works if you have over 1000 instances and it's not using a marker | 20:52 |
| mriedem | oh it sets it's own marker | 20:52 |
| diana_clarke | mriedem: and like you said, I think common is going to trump that with osapi_max_limit anyway, so it's still lying, I think. | 20:53 |
| rajathagasthya | jaypipes: Thanks! Yeah, I think that's a better way. I thought if I check nova-compute, I could also make sure things are okay with openstack as well. I guess that's additional check I could still make | 20:53 |
| mriedem | well, the -1 for server is novaclient doing the paging for you | 20:53 |
| mriedem | it sets its own marker here https://github.com/openstack/python-novaclient/blob/6.0.0/novaclient/v2/servers.py#L841 | 20:53 |
| rajathagasthya | But it would've been great if nova also gave me events on whether the hypervisor is reachable for it to schedule instances on it :) | 20:54 |
| mriedem | diana_clarke: i don't see any limit parameter to nova keypair-list | 20:54 |
| mriedem | or for flavors or hypervisors | 20:55 |
| diana_clarke | mriedem: it was added in microversion 2.35 | 20:55 |
| mriedem | diana_clarke: i mean the -1 | 20:55 |
| diana_clarke | mriedem: ah, yes. sorry | 20:55 |
| mriedem | there is no limit=-1 support for novaclient for those other resources | 20:55 |
| diana_clarke | mriedem: Okay, thanks. That's the conclusion I was coming to too. Lies! | 20:56 |
| jaypipes | rajathagasthya: ya, you will want to have npre (or similar monitoring) checks for both nova-compute events *and* the hypervisor failures. | 20:56 |
| *** Sukhdev has joined #openstack-nova | 20:57 | |
| mriedem | diana_clarke: oh i was looking at old code, from liberty, sorry | 20:57 |
| mriedem | so nova keypair-list has a limit option https://github.com/openstack/python-novaclient/blob/6.0.0/novaclient/v2/shell.py#L3319 | 20:57 |
| mriedem | diana_clarke: but you're right, the -1 isn't implemented for keypairs in novaclient https://github.com/openstack/python-novaclient/blob/6.0.0/novaclient/v2/keypairs.py#L191 | 20:57 |
| mriedem | the help text was just copy pasta | 20:58 |
| mriedem | that's a bug | 20:58 |
| rajathagasthya | jaypipes: Got it! Thanks for your help :) | 20:58 |
| *** slaweq_ has joined #openstack-nova | 20:58 | |
| diana_clarke | mriedem: I think it's 3 copy pasta bugs. I'll add them to my list. | 20:58 |
| mriedem | yeah, easy docs fix there though | 20:59 |
| mriedem | ping me when you have a change up | 20:59 |
| mriedem | rajathagasthya: i've also updated https://wiki.openstack.org/wiki/SystemUsageData at the top to try and redirect people now | 20:59 |
| mriedem | that's been a long time coming | 20:59 |
| mriedem | gibi: fyi https://bugs.launchpad.net/nova/+bug/1645466 | 21:00 |
| openstack | Launchpad bug 1645466 in OpenStack Compute (nova) "ServiceStatusNotification is not in docs for existing versioned notifications" [Medium,Confirmed] | 21:00 |
| diana_clarke | mriedem: will do, thanks! Now I have to figure out how to actually implement a -1 case (or something similar). It's hard to test in the client side. | 21:00 |
| rajathagasthya | Perfect. Thanks mriedem | 21:00 |
| mriedem | diana_clarke: for which resource? simple tenant usage? | 21:00 |
| *** penick has quit IRC | 21:00 | |
| mriedem | diana_clarke: you thought you were just going to remove the -1 lie from the help strings? | 21:00 |
| mriedem | that's the simple fix | 21:01 |
| mriedem | *i thought you | 21:01 |
| diana_clarke | mriedem: yup, just a doc fix for 3 of the 4 existing -1 cases, but I also need a -1 case for usage (or something, perhaps not a -1) | 21:01 |
| diana_clarke | mriedem: because 'nova usage' CLI needs to loops over the next links and still provide the full summary | 21:02 |
| *** penick has joined #openstack-nova | 21:03 | |
| *** yamahata has joined #openstack-nova | 21:04 | |
| *** vishwana_ has joined #openstack-nova | 21:04 | |
| mriedem | diana_clarke: does it? can't we just expose limit but not handle -1 right now? | 21:05 |
| mriedem | by default you get the os_api_max_limit or whatever | 21:05 |
| mriedem | i guess i'd add the -1 support in later if you feel the need | 21:05 |
| mriedem | as it's really a novaclient side feature | 21:05 |
| diana_clarke | mriedem: 'nova usage' only returns totals, making it impossible for the user to figure out what the marker should be to get the next page of data. Example: http://paste.openstack.org/show/589771/ | 21:06 |
| *** slaweq_ has quit IRC | 21:06 | |
| diana_clarke | mriedem: Once I realized that flaw, I marked the client patch as a WIP. | 21:06 |
| openstackgerrit | Pushkar Umaranikar proposed openstack/nova: [WIP] Placement api: Add json_error_formatter to defaults https://review.openstack.org/395194 | 21:07 |
| diana_clarke | mriedem: My current plan is to dog food the next links and basically remove paging from 'nova usage'. | 21:07 |
| yuanyuan | Hi, I have a question about updating aggregate, if I run the CLI commands: nova aggregate-update 4 ha-2 --availability-zone "", how the parser handle the empty string of availability zone name | 21:07 |
| diana_clarke | mriedem: but the client test hardness isn't really set up for wackiness like that. | 21:08 |
| diana_clarke | harness | 21:08 |
| *** vishwanathj has quit IRC | 21:08 | |
| mriedem | diana_clarke: hmm, we have the functional tests which run in a live devstack, | 21:08 |
| mriedem | but you'd need to hack CONF.osapi_max_limit i think | 21:08 |
| *** slaweq_ has joined #openstack-nova | 21:08 | |
| mriedem | and you can't do that from the client side, it would have to be from the server side when setting up devstack... | 21:09 |
| mriedem | idk, there are probably ways you could stub out the limit set in the functional tests and still test the paging code in the client | 21:10 |
| mriedem | like always send limit=1 but somehow fake out the client side code to think that the user provided limit=-1 | 21:10 |
| mriedem | s/set/sent/ | 21:10 |
| diana_clarke | mriedem: but the looping over next links logic should go in the client right? or do you think I can get away with pushing it up into nova with a special flag? | 21:11 |
| mriedem | the paging should happen client-side | 21:11 |
| diana_clarke | (it would still page under the hood in nova) | 21:11 |
| mriedem | i'm just saying i think we could be creative in testing that in the novaclient functional tests | 21:11 |
| *** whenry has joined #openstack-nova | 21:13 | |
| diana_clarke | mriedem: Anyhoo, I'll think about it a bit some more, and come back with more fully formed thoughts. I don't want to eat up anymore of your time. Mostly, I just popped by for a -1 sanity check. Thanks! | 21:13 |
| *** anush has quit IRC | 21:13 | |
| *** victorhe has quit IRC | 21:14 | |
| *** whenry has quit IRC | 21:18 | |
| *** swamireddy has quit IRC | 21:18 | |
| *** moshele has quit IRC | 21:20 | |
| *** slaweq_ has quit IRC | 21:21 | |
| *** kristia__ has joined #openstack-nova | 21:22 | |
| *** kristian__ has quit IRC | 21:22 | |
| *** slaweq_ has joined #openstack-nova | 21:22 | |
| mriedem | dansmith: this has to be totally busted doesn't it? https://github.com/openstack/nova/blob/4e747092bcb015303efc2ab13da98ef5ce575ec8/nova/api/openstack/compute/server_external_events.py#L72 | 21:23 |
| mriedem | expected_attrs should be a list | 21:23 |
| mriedem | that ends up coming in https://github.com/openstack/nova/blob/4e747092bcb015303efc2ab13da98ef5ce575ec8/nova/objects/instance.py#L73 | 21:23 |
| dansmith | yeah totally | 21:23 |
| mriedem | and i think simple_cols just ends up being [] | 21:23 |
| dansmith | I think I reviewed that too | 21:24 |
| mriedem | because 'm', 'i', 'g' aren't in tehre | 21:24 |
| mriedem | *there | 21:24 |
| mriedem | ok, me fixy | 21:24 |
| *** tbachman has joined #openstack-nova | 21:24 | |
| dansmith | oh, I guess it doesn't fail because it's just expecting an iterable there | 21:24 |
| dansmith | ugh | 21:24 |
| mriedem | yeah | 21:25 |
| dansmith | artom: we suck | 21:25 |
| mriedem | also backported that... | 21:25 |
| mriedem | oops! | 21:25 |
| dansmith | yeah | 21:25 |
| mriedem | and tony just rammed it in https://review.openstack.org/#/c/392219/ | 21:25 |
| *** victorhe has joined #openstack-nova | 21:25 | |
| mriedem | anyway, fixing :) | 21:25 |
| dansmith | well, I guess it doesn't break, | 21:25 |
| dansmith | it just doesn't give us the optimization | 21:25 |
| mriedem | yeah i need to add info_cache in there, that's what got me looking | 21:26 |
| dansmith | so I guess it's obscure enough for us to retain a shred of dignity | 21:26 |
| mriedem | but i'll always know | 21:26 |
| dansmith | the horror! | 21:26 |
| dansmith | that's okay, I know things no human should know about tonyb | 21:26 |
| dansmith | mutually assured destruction | 21:26 |
| mriedem | but not me | 21:26 |
| *** victorhe has quit IRC | 21:26 | |
| dansmith | true. | 21:26 |
| mriedem | except, maybe that public bath in budapest | 21:27 |
| dansmith | heh | 21:27 |
| openstackgerrit | Sivasathurappan Radhakrishnan proposed openstack/nova: Initiate all network changes from source during live migration https://review.openstack.org/369423 | 21:27 |
| openstackgerrit | Sivasathurappan Radhakrishnan proposed openstack/nova: Don't call setup_networks_on_host(dest) multiple times https://review.openstack.org/370089 | 21:27 |
| mriedem | nice https://review.openstack.org/#/c/371048/10/nova/tests/unit/api/openstack/compute/test_server_external_events.py | 21:28 |
| *** tblakes has quit IRC | 21:28 | |
| *** kristia__ is now known as kristian__ | 21:28 | |
| *** smatzek has quit IRC | 21:28 | |
| *** anush has joined #openstack-nova | 21:29 | |
| openstackgerrit | Diana Clarke proposed openstack/python-novaclient: Correct copy/paste errors in help https://review.openstack.org/403913 | 21:32 |
| *** rajathagasthya has quit IRC | 21:33 | |
| *** vishwana_ has quit IRC | 21:35 | |
| *** Jeffrey4l has quit IRC | 21:35 | |
| *** vishwanathj has joined #openstack-nova | 21:35 | |
| *** yuanyuan has quit IRC | 21:36 | |
| *** thorst_ has quit IRC | 21:37 | |
| *** thorst_ has joined #openstack-nova | 21:37 | |
| *** anush has quit IRC | 21:37 | |
| *** tblakes has joined #openstack-nova | 21:38 | |
| *** thorst_ has quit IRC | 21:42 | |
| mriedem | diana_clarke: i'd like to backport that so please associate a bug with it and then i'm +2 | 21:44 |
| artom | dansmith, what? | 21:45 |
| artom | dansmith, I mean, yeah, we do | 21:45 |
| mriedem | artom: bug 1645479 | 21:45 |
| openstack | bug 1645479 in OpenStack Compute (nova) "ServerExternalEventsController doesn't properly pre-load migration_context" [Low,Triaged] https://launchpad.net/bugs/1645479 - Assigned to Matt Riedemann (mriedem) | 21:45 |
| artom | But in this specific case, why do we? | 21:45 |
| *** matrohon has quit IRC | 21:45 | |
| diana_clarke | mriedem: Sure, will do. First, I need to torture my family by feeding them beets for dinner. back in a bit | 21:45 |
| artom | Ah, we passed a string when we should have passed an array | 21:46 |
| artom | I mean, strictly speaking, we're right | 21:46 |
| artom | A string is iterable ;) | 21:46 |
| artom | It's trying to load m, i, g, etc :D | 21:46 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Fix expected_attrs kwarg in server_external_events https://review.openstack.org/403917 | 21:47 |
| mriedem | artom: dansmith: ^ | 21:47 |
| *** Jeffrey4l has joined #openstack-nova | 21:47 | |
| melwitt | dansmith: transport_url is owned by oslo.messaging, it defines it https://github.com/openstack/oslo.messaging/blob/ea8fad4/oslo_messaging/transport.py#L42-L46 | 21:48 |
| mriedem | btw, i have to use the gross ass assert because it's a classmethod | 21:48 |
| melwitt | dansmith: as for the presence of it, we had a change in the past to set it in devstack here https://github.com/openstack-dev/devstack/blob/dec1211/lib/rpc_backend#L122 so I'm not yet sure what's happening in that trace | 21:50 |
| *** victorhe has joined #openstack-nova | 21:50 | |
| *** masuberu has quit IRC | 21:51 | |
| *** erlon-airlong has quit IRC | 21:52 | |
| *** jheroux has quit IRC | 21:52 | |
| *** victorhe has quit IRC | 21:53 | |
| openstackgerrit | Danil Akhmetov proposed openstack/nova: Use proper user and tenant in the owner section of libvirt.xml. https://review.openstack.org/399679 | 21:53 |
| jaypipes | guh, finally figured out the source of this pesky bug.. | 21:56 |
| *** kylek3h has quit IRC | 21:58 | |
| tonyb | dansmith: ummm you said you wouldn't tell somethign about mutually assured destruction ;P | 21:59 |
| *** sreeram_v has joined #openstack-nova | 22:01 | |
| tonyb | mriedem that was a clear mistake. No idea why I hit +W | 22:01 |
| *** tblakes has quit IRC | 22:01 | |
| melwitt | dansmith: I commented on the review | 22:02 |
| *** rajathagasthya has joined #openstack-nova | 22:03 | |
| *** ijw has joined #openstack-nova | 22:04 | |
| dansmith | tonyb: :P | 22:04 |
| *** ljxiash_ has quit IRC | 22:05 | |
| *** lamt has quit IRC | 22:05 | |
| *** ljxiash has joined #openstack-nova | 22:06 | |
| dansmith | melwitt: but it's registered in here too right? https://github.com/openstack/oslo.messaging/blob/ea8fad4/oslo_messaging/transport.py#L258 | 22:06 |
| *** sreeram_v has quit IRC | 22:06 | |
| dansmith | oh, but not before parse maybe | 22:06 |
| dansmith | cripes. | 22:06 |
| melwitt | dansmith: oh, I didn't see that other call. so yeah, two ways to register the options ... this is odd | 22:07 |
| dansmith | but still, not in the path we need | 22:07 |
| melwitt | looks like a bug that it doesn't try to register the opts before access in the parse being that it's a classmethod | 22:08 |
| dansmith | right | 22:09 |
| dansmith | cooking up something now | 22:09 |
| *** ljxiash has quit IRC | 22:10 | |
| *** lamt has joined #openstack-nova | 22:10 | |
| *** ljxiash has joined #openstack-nova | 22:10 | |
| melwitt | okay. the change that added 'url or conf.transport_url' was recently done as part of the changes we made in devstack to move to transport_url. so it was just an oversight at that time | 22:11 |
| dansmith | no tests for that stuff? | 22:11 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Pre-load info_cache when handling external events and handle NotFound https://review.openstack.org/403925 | 22:11 |
| melwitt | or rather, when I went poking around the rpc options it caused the deprecation of all the backend-specific options in oslo.messaging and so some changes were made to use oslo.messaging | 22:11 |
| dansmith | added you to the o.msg review | 22:12 |
| melwitt | *were made to oslo.messaging | 22:12 |
| dansmith | I guess we're kinda stuck for the moment then | 22:12 |
| melwitt | okay. I think there's some testing but obviously not for that I guess. a TransportURL.parse right off the bat | 22:12 |
| *** vladikr has quit IRC | 22:12 | |
| *** thorst_ has joined #openstack-nova | 22:12 | |
| dansmith | well, nothing that checks opts registering in test_transport that I can see | 22:13 |
| openstackgerrit | Diana Clarke proposed openstack/python-novaclient: Correct copy/paste errors in help https://review.openstack.org/403913 | 22:14 |
| *** thorst__ has joined #openstack-nova | 22:15 | |
| *** jamesdenton has quit IRC | 22:15 | |
| diana_clarke | mriedem: ^ attached to a bug, and also cherry picked for stable/newton. thanks again for the sanity check | 22:16 |
| diana_clarke | mriedem: I need to fix the cherry pick, I used the UI which didn't give me the -x stuff, sorry | 22:17 |
| *** thorst_ has quit IRC | 22:17 | |
| melwitt | dansmith: yeah. I meant a test that just does a TransportURL.parse at the beginning of it that would raise without the option registration | 22:17 |
| dansmith | yeah | 22:17 |
| mriedem | diana_clarke: thanks | 22:18 |
| openstackgerrit | STEW TY proposed openstack/nova: Transform instance.rescue notifications https://review.openstack.org/385644 | 22:18 |
| *** Sukhdev has quit IRC | 22:19 | |
| *** kfarr has quit IRC | 22:21 | |
| pumaranikar | cdent, I have updated https://review.openstack.org/#/c/395194/. Add json_formatter to defaults for placement. Please let me know your suggestions. | 22:22 |
| *** jheroux has joined #openstack-nova | 22:22 | |
| *** jheroux has quit IRC | 22:24 | |
| *** mvk has joined #openstack-nova | 22:25 | |
| *** smatzek has joined #openstack-nova | 22:26 | |
| *** liverpooler has joined #openstack-nova | 22:26 | |
| *** catintheroof has quit IRC | 22:28 | |
| *** zigo has quit IRC | 22:29 | |
| *** xyang1 has quit IRC | 22:29 | |
| *** zigo has joined #openstack-nova | 22:36 | |
| *** kristian__ has quit IRC | 22:36 | |
| *** ircuser-1 has joined #openstack-nova | 22:36 | |
| *** lpetrut has quit IRC | 22:41 | |
| *** nherciu has quit IRC | 22:42 | |
| *** baoli has quit IRC | 22:43 | |
| *** longlq_ has joined #openstack-nova | 22:44 | |
| *** smatzek has quit IRC | 22:45 | |
| *** sam_nowitzki has quit IRC | 22:47 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Make RPCFixture support multiple connections https://review.openstack.org/396417 | 22:48 |
| *** gouthamr has quit IRC | 22:50 | |
| openstackgerrit | Michael Still proposed openstack/nova: libvirt: prefer cinder rbd auth values over nova.conf https://review.openstack.org/389399 | 22:51 |
| *** mnasiadka has quit IRC | 22:51 | |
| *** ljxiash has quit IRC | 22:53 | |
| *** ljxiash has joined #openstack-nova | 22:53 | |
| openstackgerrit | Ed Leafe proposed openstack/nova: placement: REST API for resource classes https://review.openstack.org/391959 | 22:54 |
| *** slaweq_ has quit IRC | 22:56 | |
| *** slaweq_ has joined #openstack-nova | 22:58 | |
| *** ljxiash has quit IRC | 22:58 | |
| *** mriedem has quit IRC | 23:02 | |
| *** eharney has quit IRC | 23:04 | |
| *** Swami has quit IRC | 23:05 | |
| *** tbachman has quit IRC | 23:07 | |
| *** lamt has quit IRC | 23:18 | |
| *** ducttape_ has quit IRC | 23:19 | |
| *** ducttape_ has joined #openstack-nova | 23:20 | |
| *** esberglu has quit IRC | 23:20 | |
| *** mdrabe_ has quit IRC | 23:20 | |
| *** jwcroppe has quit IRC | 23:20 | |
| *** esberglu has joined #openstack-nova | 23:20 | |
| *** thorst__ is now known as thorst_ | 23:21 | |
| *** jwcroppe has joined #openstack-nova | 23:21 | |
| *** thorst_ has quit IRC | 23:23 | |
| *** thorst_ has joined #openstack-nova | 23:24 | |
| *** ducttape_ has quit IRC | 23:24 | |
| *** esberglu has quit IRC | 23:25 | |
| *** jwcroppe has quit IRC | 23:25 | |
| *** esberglu has joined #openstack-nova | 23:32 | |
| *** thorst_ has quit IRC | 23:32 | |
| *** baoli has joined #openstack-nova | 23:33 | |
| *** baoli has quit IRC | 23:37 | |
| *** bnemec has quit IRC | 23:37 | |
| *** ijw has quit IRC | 23:40 | |
| *** Guest89691 has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!