Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-03-23
18:42:47 sdague it's kind of suspicious that after a bunch of crash dumping rabbit goes sideways
18:42:58 sdague I'd start with disabling all those crashdump tests first
18:43:34 mriedem i seem to remember andreykurilin fixing something where the crash dump tests started taking a long time
18:43:44 mriedem andreykurilin: if you're around, do you remember something like that?
18:45:13 openstackgerrit Sean Dague proposed openstack/python-novaclient master: set test timout to 300 seconds https://review.openstack.org/449211
18:45:31 sdague that's got the skip in there
18:45:35 mriedem we can use bug test_trigger_crash_dump_in_shutoff_state
18:45:39 mriedem oops
18:45:45 openstack bug 1675526 in python-novaclient "Trigger crash dump functional tests are timing out and hanging" [High,Triaged] https://launchpad.net/bugs/1675526
18:45:45 mriedem bug 1675526
18:45:54 openstackgerrit Eric Fried proposed openstack/nova master: PowerVM Driver: spawn/destroy #3: TaskFlow https://review.openstack.org/438729
18:45:59 sdague ok, feel free to update if you like, got to run away from keyboard for a bit
18:46:04 mriedem ok
18:55:52 openstackgerrit Matt Riedemann proposed openstack/python-novaclient master: Set test timout to 300 seconds https://review.openstack.org/449211
19:09:37 openstackgerrit Rick Bartra proposed openstack/nova master: Add policy granularity to the Flavor API https://review.openstack.org/449288
19:12:01 openstackgerrit Rick Bartra proposed openstack/nova master: Add policy granularity to the Flavor API https://review.openstack.org/449288
19:16:32 openstackgerrit Eric Fried proposed openstack/nova master: PowerVM Driver: spawn/destroy #4: full flavor https://review.openstack.org/391288
19:20:09 openstackgerrit Sujitha proposed openstack/nova master: Adding tags field to InstancePayload https://review.openstack.org/407228
19:26:51 openstackgerrit Eric Fried proposed openstack/nova master: PowerVM Driver: console https://review.openstack.org/409402
20:02:23 openstackgerrit Matt Riedemann proposed openstack/nova master: api-ref: fix description of volumeAttachment for attach/swap-volume https://review.openstack.org/449302
20:33:24 openstackgerrit Eric Fried proposed openstack/nova master: PowerVM Driver: SSP emphemeral disk support https://review.openstack.org/443189
20:42:05 mriedem sdague: in the api-ref for server details we have os-extended-volumes:volumes_attached which just says it's an array of attached volumes. at 2.1 it's a list of dicts with a single 'id' key which is the volume id. after 2.3 the list of dicts can also contain the 'delete_on_termination' key.
20:42:14 mriedem we don't describe the actual contents of that dict in the list,
20:42:20 mriedem i.e. the id and delete_on_termination keys,
20:42:27 mriedem i'm trying to figure out what's the best way to represent that,
20:42:32 mriedem would the name be os-extended-volumes:volumes_attached.id?
20:42:37 mriedem os-extended-volumes:volumes_attached.delete_on_termination?
20:42:49 sdague mriedem: that has been the pattern we were using
20:44:26 openstackgerrit Anusha Unnam proposed openstack/nova master: Add description to policies in suspend_server.py https://review.openstack.org/447722
20:44:32 mriedem sdague: ok, alternatively to avoid the turducken of nested parameters, i could just describe those keys in the description for the os-extended-volumes:volumes_attached parameter?
20:45:35 sdague you could, but it sort of breaks the flow
20:45:42 sdague I would honestly do the nested thing
20:45:44 mriedem i'll just do the normal pattern
20:45:45 mriedem ok
20:45:55 mriedem danke
20:46:09 mriedem writing functional regression tests that actually use our api makes you have to actually read the api docs :)
20:46:10 mriedem it's amazing
20:46:49 sdague mriedem: yeh, and it's close enough now that it's worth fixing when you find a weird issue we missed
20:46:59 mriedem yeah that's true
20:47:11 mriedem if it were a needle in a haystack i'd be much less motivated
20:58:16 cfriesen I'm looking at the "realtime" handling of emupcpus in _get_guest_numa_config() and I'm confused about the case where the instance isn't pinned
20:58:34 mriedem cfriesen: sahid and sfinucan are your people
20:58:38 mriedem but aren't around probably
20:59:16 cfriesen yeah...I'll ask tomorrow
21:36:05 mriedem dansmith: melwitt: just verified that local delete regression when deleting a shelved offloaded server with volumes attached
21:36:10 mriedem will push up the test to show it shortly
21:36:13 mriedem writing the bug report
21:36:21 dansmith uh yeahokay
21:36:24 dansmith upps
21:36:26 dansmith wtf
21:37:17 melwitt regression :(
21:38:33 mriedem mfing local delete
21:38:38 openstack Launchpad bug 1675570 in OpenStack Compute (nova) "Volumes are not detached when deleting shelved offloaded servers" [High,Triaged]
21:38:38 mriedem https://bugs.launchpad.net/nova/+bug/1675570
21:39:36 bauzas mriedem: dansmith: hola, FYI https://review.openstack.org/#/c/391060/ is ready for reviews :)
21:41:17 melwitt mriedem: we have the _local_delete function that does all the volume detach stuff, but we never incorporated calling it into the cells v2 short circuits of delete. I think because we were only thinking of the "not yet scheduled" case
21:41:25 mriedem yup
21:41:31 mriedem hold gentle caller,
21:41:33 mriedem for the test
21:42:18 openstackgerrit Matt Riedemann proposed openstack/nova master: Regression test for local delete with an attached volume https://review.openstack.org/449334
21:42:18 melwitt thanks for the writing the test. *cues karate kid montage again*
21:42:19 mriedem melwitt: dansmith: ^
21:42:32 mriedem Kevin_Zheng gets all of the credit for reporting this to me this morning
21:42:50 melwitt *plays montage for Kevin_Zheng too*
21:42:55 mriedem we've never done any decent volume-related functional testing because our cinder fixture wasn't good enough
21:43:02 mriedem that patch builds up the cinder fixture too
21:43:06 mriedem to actually track attach/detach
21:43:32 openstackgerrit Anusha Unnam proposed openstack/nova master: Add description to policies in used_limits https://review.openstack.org/447710
21:44:10 melwitt mriedem: ah, yeah. I remember adding some stuff to that in my famous boot-from-volume patch. and I kept thinking to break it out into a separate thing but didn't
21:44:32 melwitt probably because it would have been adding stuff that nothing was using
21:45:30 openstackgerrit Matt Riedemann proposed openstack/nova master: Regression test for local delete with an attached volume https://review.openstack.org/449334
21:45:32 mriedem is that like melwitt's famous bbq?
21:45:38 melwitt nevermind, what you did is different than what I was thinking. I had done this https://review.openstack.org/#/c/355091/8/nova/tests/unit/volume/fake.py
21:45:49 melwitt haha
21:49:36 openstackgerrit Anusha Unnam proposed openstack/nova master: Add description to policies to availability_zone https://review.openstack.org/440836
21:51:32 openstackgerrit Matt Riedemann proposed openstack/nova master: Regression test for local delete with an attached volume https://review.openstack.org/449334
21:52:34 mriedem melwitt: yeah btw if we would have called that _local_delete method, we could have gotten the instance.delete.start/end notifications too
21:52:39 mriedem which was the other regression
21:52:59 melwitt /facepalm
21:53:25 mriedem yuup
21:53:42 Apoorva_ NFS based Cinder connections seem to be broken by Nova changes. Our CI logs http://openstack-ci.tintri.com/tintri/refs-changes-59-448159-4/logs/screen-n-cpu.log.txt
21:55:48 Apoorva_ http://paste.openstack.org/show/603992/
21:59:23 openstackgerrit Sujitha proposed openstack/nova master: [WIP] Add BDM to InstancePayload https://review.openstack.org/448779
21:59:25 openstackgerrit Sujitha proposed openstack/nova master: Adding auto_disk_config field to InstancePayload https://review.openstack.org/419185
21:59:25 openstackgerrit Sujitha proposed openstack/nova master: Add keypairs field to InstancePayload https://review.openstack.org/419730
21:59:27 openstackgerrit Sujitha proposed openstack/nova master: Adding tags field to InstancePayload https://review.openstack.org/407228
22:00:04 openstackgerrit Anusha Unnam proposed openstack/nova master: Add descripiton to policies in virtual_interfaces.py https://review.openstack.org/445652
22:00:31 Apoorva_ Can someone help me with this? http://paste.openstack.org/show/603992/
22:00:54 tjones so i was sitting in openstack-meeting today at 2pm and no one was there… but i see i missed the meeting. so - did it move???
22:01:21 mriedem tjones: wrong week, today was the 7pm PT
22:01:26 mriedem *7am
22:07:32 openstackgerrit Sujitha proposed openstack/nova master: Adding tags field to InstancePayload https://review.openstack.org/407228
22:08:26 mriedem Apoorva_: have you diffed the _connect_volume method to see what changed the signature?
22:08:43 Apoorva_ lyarwood mriedem would this https://review.openstack.org/#/c/437665/ cause Cinder nfs based driver connections to fail?
22:08:53 mriedem on
22:08:55 mriedem *no
22:09:09 openstackgerrit Sujitha proposed openstack/nova master: Adding tags field to InstancePayload https://review.openstack.org/407228
22:09:23 openstackgerrit Sujitha proposed openstack/nova master: Change tags to default field in Instance object. https://review.openstack.org/415298
22:09:40 mriedem Apoorva_: we have an NFS CI job running in the community infra, are you running a 3rd party Cinder NFS driver CI?
22:09:50 mriedem this is the failure

Earlier   Later