| *** earlephilhower has quit IRC | 00:02 | |
| *** diogogmt has joined #openstack-nova | 00:03 | |
| *** zzzeek has quit IRC | 00:04 | |
| *** zhangjn has quit IRC | 00:04 | |
| *** sirushti has quit IRC | 00:07 | |
| *** aysyd has quit IRC | 00:11 | |
| *** angdraug has quit IRC | 00:11 | |
| *** thorst has joined #openstack-nova | 00:23 | |
| *** thorst has quit IRC | 00:26 | |
| *** toabctl has quit IRC | 00:32 | |
| *** kgalanov_ has quit IRC | 00:32 | |
| *** kgalanov_ has joined #openstack-nova | 00:33 | |
| *** frickler has quit IRC | 00:33 | |
| *** _fortis has quit IRC | 00:33 | |
| *** achanda has joined #openstack-nova | 00:33 | |
| *** ctrath has joined #openstack-nova | 00:34 | |
| *** frickler has joined #openstack-nova | 00:34 | |
| *** toabctl has joined #openstack-nova | 00:35 | |
| *** _fortis has joined #openstack-nova | 00:37 | |
| openstackgerrit | Ken'ichi Ohmichi proposed openstack/nova: Remove unused _check_string_length() https://review.openstack.org/227059 | 00:38 |
|---|---|---|
| *** achanda has quit IRC | 00:38 | |
| *** tsekiyam_ has joined #openstack-nova | 00:41 | |
| *** kfox1111 is now known as kfox1111_away | 00:42 | |
| *** tpeoples has quit IRC | 00:43 | |
| *** tsekiyama has quit IRC | 00:44 | |
| *** tsekiyam_ has quit IRC | 00:45 | |
| *** mtanino has quit IRC | 00:46 | |
| *** edmondsw has quit IRC | 00:46 | |
| *** jecarey has joined #openstack-nova | 00:48 | |
| *** boris-42 has quit IRC | 00:49 | |
| *** hitalia has quit IRC | 00:50 | |
| *** zhangjn has joined #openstack-nova | 00:52 | |
| *** annegentle has joined #openstack-nova | 00:54 | |
| *** leeantho has quit IRC | 00:57 | |
| *** haomaiwang has joined #openstack-nova | 00:57 | |
| *** annegentle has quit IRC | 00:59 | |
| *** haomaiwang has quit IRC | 01:00 | |
| *** haomaiwa_ has joined #openstack-nova | 01:00 | |
| *** haomaiwa_ has quit IRC | 01:01 | |
| *** haomaiwang has joined #openstack-nova | 01:01 | |
| *** asettle has quit IRC | 01:03 | |
| *** ctrath has quit IRC | 01:12 | |
| *** haomaiwang has quit IRC | 01:12 | |
| *** su_zhang has quit IRC | 01:13 | |
| *** penick has quit IRC | 01:22 | |
| *** Apoorva has quit IRC | 01:24 | |
| *** gongysh has joined #openstack-nova | 01:26 | |
| *** Marga__ has joined #openstack-nova | 01:29 | |
| *** otter768 has joined #openstack-nova | 01:30 | |
| *** Marga_ has quit IRC | 01:32 | |
| *** Marga__ has quit IRC | 01:33 | |
| *** atuvenie has quit IRC | 01:34 | |
| *** vilobhmm11 has quit IRC | 01:34 | |
| *** otter768 has quit IRC | 01:34 | |
| *** baoli has joined #openstack-nova | 01:35 | |
| *** baoli has quit IRC | 01:40 | |
| *** zhenguo has joined #openstack-nova | 01:43 | |
| *** otter768 has joined #openstack-nova | 01:47 | |
| *** kiran-r has joined #openstack-nova | 01:49 | |
| *** kiran-r has quit IRC | 01:50 | |
| *** kiran-r has joined #openstack-nova | 01:51 | |
| *** tojuvone_ has joined #openstack-nova | 01:52 | |
| *** kiranr has joined #openstack-nova | 01:52 | |
| *** tojuvone has quit IRC | 01:55 | |
| *** haomaiwang has joined #openstack-nova | 01:57 | |
| *** rushil has joined #openstack-nova | 01:58 | |
| *** kiran-r has quit IRC | 01:59 | |
| *** ijw has quit IRC | 02:00 | |
| *** haomaiwang has quit IRC | 02:01 | |
| *** rushil_ has joined #openstack-nova | 02:01 | |
| *** haomaiwa_ has joined #openstack-nova | 02:01 | |
| *** achanda has joined #openstack-nova | 02:01 | |
| *** rushil has quit IRC | 02:04 | |
| *** yamahata has quit IRC | 02:17 | |
| *** baoli has joined #openstack-nova | 02:17 | |
| *** mwagner_lap has joined #openstack-nova | 02:18 | |
| *** pixelbeat has joined #openstack-nova | 02:21 | |
| *** pixelb has quit IRC | 02:23 | |
| *** rushil_ has quit IRC | 02:23 | |
| *** unicell has quit IRC | 02:23 | |
| *** rushil has joined #openstack-nova | 02:24 | |
| *** asettle has joined #openstack-nova | 02:31 | |
| *** r-daneel has quit IRC | 02:33 | |
| *** rushil has quit IRC | 02:33 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova-specs: Add project id and user id information in os-server-group API https://review.openstack.org/209917 | 02:35 |
| *** annegentle has joined #openstack-nova | 02:43 | |
| *** ericksonsantos has quit IRC | 02:47 | |
| *** gabriel-bezerra has quit IRC | 02:47 | |
| *** annegentle has quit IRC | 02:48 | |
| openstackgerrit | Jian LI proposed openstack/nova: Sync vms az with the host when using aggregate https://review.openstack.org/226683 | 02:48 |
| *** ericksonsantos has joined #openstack-nova | 02:53 | |
| *** gabriel-bezerra has joined #openstack-nova | 02:53 | |
| *** salv-orl_ has joined #openstack-nova | 02:59 | |
| *** haomaiwa_ has quit IRC | 03:01 | |
| *** ferest has joined #openstack-nova | 03:01 | |
| *** yamahata has joined #openstack-nova | 03:01 | |
| *** haomaiwang has joined #openstack-nova | 03:01 | |
| *** salv-orlando has quit IRC | 03:02 | |
| *** diogogmt has quit IRC | 03:04 | |
| *** ferest has quit IRC | 03:05 | |
| openstackgerrit | venkatesh proposed openstack/nova: make the driver.Scheduler as abstract class https://review.openstack.org/212016 | 03:07 |
| *** otter768 has quit IRC | 03:14 | |
| *** garyk has quit IRC | 03:16 | |
| *** oomichi has joined #openstack-nova | 03:26 | |
| *** ctrath has joined #openstack-nova | 03:26 | |
| *** deepthi has joined #openstack-nova | 03:26 | |
| *** su_zhang has joined #openstack-nova | 03:27 | |
| *** sirushti has joined #openstack-nova | 03:27 | |
| *** pixelbeat has quit IRC | 03:29 | |
| *** kiran-r has joined #openstack-nova | 03:31 | |
| *** shuquan has joined #openstack-nova | 03:34 | |
| *** sirushti has quit IRC | 03:35 | |
| *** sirushti has joined #openstack-nova | 03:35 | |
| *** kiran-r has quit IRC | 03:39 | |
| *** fawadkhaliq has joined #openstack-nova | 03:41 | |
| *** garyk has joined #openstack-nova | 03:46 | |
| *** baoli has quit IRC | 03:47 | |
| *** shuquan has quit IRC | 03:51 | |
| *** achanda has quit IRC | 03:55 | |
| *** shuquan has joined #openstack-nova | 03:56 | |
| *** haomaiwang has quit IRC | 04:01 | |
| *** haomaiwang has joined #openstack-nova | 04:01 | |
| *** kiranr has quit IRC | 04:02 | |
| *** Piet has quit IRC | 04:06 | |
| openstackgerrit | Giridhar Jayavelu proposed openstack/nova: VMware: Append disk type to image cache folder name https://review.openstack.org/227104 | 04:10 |
| *** mcarden has quit IRC | 04:15 | |
| *** ctrath has quit IRC | 04:16 | |
| *** VikasC has joined #openstack-nova | 04:16 | |
| *** mc has joined #openstack-nova | 04:17 | |
| *** mc is now known as mcarden | 04:17 | |
| *** rajesht__ has quit IRC | 04:21 | |
| *** rajesht has joined #openstack-nova | 04:21 | |
| *** asettle has quit IRC | 04:24 | |
| *** jwcroppe has joined #openstack-nova | 04:26 | |
| *** annegentle has joined #openstack-nova | 04:32 | |
| *** annegentle has quit IRC | 04:36 | |
| *** Marga_ has joined #openstack-nova | 04:42 | |
| *** Marga_ has quit IRC | 04:43 | |
| *** Marga_ has joined #openstack-nova | 04:43 | |
| openstackgerrit | Andrei V. Ostapenko proposed openstack/nova: Fixes bug with notify_decorator bad getattr default value https://review.openstack.org/223890 | 04:43 |
| *** fifieldt has joined #openstack-nova | 04:52 | |
| *** achanda has joined #openstack-nova | 04:55 | |
| *** dave-mccowan has quit IRC | 05:00 | |
| *** haomaiwang has quit IRC | 05:01 | |
| *** haomaiwang has joined #openstack-nova | 05:01 | |
| *** shuquan has quit IRC | 05:03 | |
| *** shuquan has joined #openstack-nova | 05:03 | |
| *** kiran-r has joined #openstack-nova | 05:08 | |
| *** ircuser-1 has quit IRC | 05:09 | |
| *** kiran-r has quit IRC | 05:10 | |
| *** fawadkhaliq has quit IRC | 05:10 | |
| *** otter768 has joined #openstack-nova | 05:15 | |
| *** _gryf has quit IRC | 05:17 | |
| *** shuquan has quit IRC | 05:18 | |
| openstackgerrit | Moshe Levi proposed openstack/nova: Fix pci_stats logging in resource tracker https://review.openstack.org/227116 | 05:18 |
| *** otter768 has quit IRC | 05:20 | |
| *** hitalia has joined #openstack-nova | 05:23 | |
| *** ankit_ag has joined #openstack-nova | 05:31 | |
| *** _gryf has joined #openstack-nova | 05:38 | |
| *** Murali has joined #openstack-nova | 05:39 | |
| *** asettle has joined #openstack-nova | 05:42 | |
| *** Sree has joined #openstack-nova | 05:45 | |
| *** rajesht has quit IRC | 05:48 | |
| *** sahid has joined #openstack-nova | 05:53 | |
| *** Sree has quit IRC | 05:58 | |
| *** Sree has joined #openstack-nova | 05:58 | |
| *** jwcroppe has quit IRC | 05:59 | |
| *** jwcroppe has joined #openstack-nova | 06:00 | |
| *** nkrinner has joined #openstack-nova | 06:00 | |
| *** haomaiwang has quit IRC | 06:01 | |
| *** haomaiwang has joined #openstack-nova | 06:01 | |
| *** rajesht has joined #openstack-nova | 06:03 | |
| *** jwcroppe has quit IRC | 06:04 | |
| *** pkoniszewski has joined #openstack-nova | 06:05 | |
| *** rdopiera has joined #openstack-nova | 06:08 | |
| *** _gryf has quit IRC | 06:11 | |
| *** cfriesen__ has quit IRC | 06:11 | |
| *** su_zhang has quit IRC | 06:12 | |
| *** ircuser-1 has joined #openstack-nova | 06:16 | |
| *** bkopilov has quit IRC | 06:17 | |
| *** rotbeard has joined #openstack-nova | 06:18 | |
| *** bkopilov has joined #openstack-nova | 06:20 | |
| *** annegentle has joined #openstack-nova | 06:20 | |
| *** Sree_ has joined #openstack-nova | 06:23 | |
| *** Sree_ is now known as Guest78363 | 06:23 | |
| *** annegentle has quit IRC | 06:25 | |
| *** nihilifer1 has quit IRC | 06:26 | |
| *** nihilifer has joined #openstack-nova | 06:26 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Imported Translations from Zanata https://review.openstack.org/227138 | 06:27 |
| *** Sree has quit IRC | 06:27 | |
| *** markvoelker has quit IRC | 06:27 | |
| *** _gryf has joined #openstack-nova | 06:32 | |
| gmann | alex_xu: johnthetubaguy : sdague : kilo backport of server name relax patch - https://review.openstack.org/#/c/227135/1 | 06:32 |
| alex_xu | gmann: thanks! | 06:33 |
| gmann | alex_xu: sorry for putting this up late | 06:33 |
| alex_xu | gmann: it's fine | 06:33 |
| *** boris-42 has joined #openstack-nova | 06:34 | |
| *** irenab has quit IRC | 06:35 | |
| *** hitalia has quit IRC | 06:38 | |
| *** bkopilov has quit IRC | 06:39 | |
| *** pradipta has joined #openstack-nova | 06:43 | |
| *** deepakcs has joined #openstack-nova | 06:46 | |
| *** ildikov has quit IRC | 06:46 | |
| *** flaper87 has quit IRC | 06:47 | |
| *** flaper87 has joined #openstack-nova | 06:47 | |
| *** bkopilov has joined #openstack-nova | 06:54 | |
| *** salv-orl_ has quit IRC | 06:55 | |
| *** salv-orlando has joined #openstack-nova | 06:56 | |
| openstackgerrit | huan proposed openstack/nova: Do not rely on XenAPI to guess device names https://review.openstack.org/213112 | 06:57 |
| *** chenzeng has joined #openstack-nova | 06:58 | |
| openstackgerrit | Wen Zhi Yu proposed openstack/nova-specs: Spec for nova API blueprint correct-volume-attachment-id-in-nova-api-volumes https://review.openstack.org/214021 | 06:59 |
| *** haomaiwang has quit IRC | 07:01 | |
| *** haomaiwa_ has joined #openstack-nova | 07:01 | |
| *** scheuran has joined #openstack-nova | 07:02 | |
| *** romainh has joined #openstack-nova | 07:02 | |
| *** irenab has joined #openstack-nova | 07:04 | |
| *** Guest78363 has quit IRC | 07:04 | |
| *** Sree has joined #openstack-nova | 07:05 | |
| *** browne has quit IRC | 07:07 | |
| *** asettle has quit IRC | 07:07 | |
| *** sc68cal has quit IRC | 07:08 | |
| *** ptm_ has joined #openstack-nova | 07:09 | |
| *** sc68cal has joined #openstack-nova | 07:10 | |
| *** Sree has quit IRC | 07:10 | |
| *** eliqiao has left #openstack-nova | 07:13 | |
| *** rotbeard has quit IRC | 07:15 | |
| *** otter768 has joined #openstack-nova | 07:16 | |
| *** abhishekk has joined #openstack-nova | 07:16 | |
| *** jichen has joined #openstack-nova | 07:19 | |
| *** mpaolino has joined #openstack-nova | 07:19 | |
| *** stevemar has quit IRC | 07:20 | |
| *** otter768 has quit IRC | 07:21 | |
| *** stevemar has joined #openstack-nova | 07:21 | |
| *** alex_klimov has joined #openstack-nova | 07:22 | |
| *** stevemar has quit IRC | 07:24 | |
| *** markvoelker has joined #openstack-nova | 07:28 | |
| *** jwcroppe has joined #openstack-nova | 07:28 | |
| *** achanda has quit IRC | 07:28 | |
| *** karimb has joined #openstack-nova | 07:33 | |
| *** markvoelker has quit IRC | 07:33 | |
| *** kiranr has joined #openstack-nova | 07:34 | |
| openstackgerrit | Moshe Levi proposed openstack/nova: pci utils: fix the check if pci device is physical_function https://review.openstack.org/227160 | 07:39 |
| bauzas | good morning Nova | 07:40 |
| *** moshele has joined #openstack-nova | 07:41 | |
| *** jlanoux has joined #openstack-nova | 07:41 | |
| alex_xu | sdague: johnthetubaguy gmann, backport for hostname fix https://review.openstack.org/#/c/227159/ | 07:41 |
| *** kiran-r has joined #openstack-nova | 07:47 | |
| *** jamielennox is now known as jamielennox|away | 07:53 | |
| *** moshele has quit IRC | 07:54 | |
| *** haomaiwa_ has quit IRC | 08:01 | |
| *** haomaiwang has joined #openstack-nova | 08:01 | |
| *** markus_z has joined #openstack-nova | 08:04 | |
| gmann | alex_xu: looks fine, found utils.py has more changes than master one. Do not know that is picked by cherry-pick ? | 08:09 |
| *** annegentle has joined #openstack-nova | 08:09 | |
| alex_xu | gmann: thanks, let me check | 08:12 |
| *** annegentle has quit IRC | 08:14 | |
| *** StevenK has quit IRC | 08:14 | |
| *** ptm_ has quit IRC | 08:15 | |
| alex_xu | gmann: indeed, there is something more, it should be a mistake by conflict | 08:16 |
| gmann | alex_xu: ohk | 08:17 |
| gmann | alex_xu: I also find little bit hard on cherry pick due to file structure change :) | 08:17 |
| alex_xu | gmann: heh :) | 08:18 |
| *** StevenK has joined #openstack-nova | 08:21 | |
| *** pixelbeat has joined #openstack-nova | 08:21 | |
| alex_xu | gmann: just correct https://review.openstack.org/227176 | 08:25 |
| alex_xu | sdague: johnthetubaguy please ignore previous backport patch, this one is right https://review.openstack.org/227176 | 08:25 |
| * johnthetubaguy nods | 08:25 | |
| BobBall | Morning all :) | 08:26 |
| * johnthetubaguy waves | 08:28 | |
| BobBall | Howdy Mr Garbutt. | 08:29 |
| *** lucasagomes has joined #openstack-nova | 08:30 | |
| *** pradipta has quit IRC | 08:31 | |
| *** MarcF has joined #openstack-nova | 08:32 | |
| gmann | alex_xu: looks good. Thanks | 08:32 |
| *** zhangjn has quit IRC | 08:35 | |
| *** zhangjn has joined #openstack-nova | 08:38 | |
| *** shoutm has joined #openstack-nova | 08:41 | |
| bauzas | oh man, I wish we'd have a world without OpenStack AZs | 08:42 |
| *** FL1SK has quit IRC | 08:43 | |
| johnthetubaguy | bauzas: totally need to fix those bloody crazy things | 08:44 |
| bauzas | johnthetubaguy: I just opened a can of worms | 08:44 |
| bauzas | I mean, mriedem_away did too | 08:45 |
| bauzas | historical things | 08:45 |
| *** lpetrut has joined #openstack-nova | 08:46 | |
| *** Venkatesh has joined #openstack-nova | 08:48 | |
| *** lifeless has quit IRC | 08:49 | |
| Venkatesh | Hi johnthetubaguy | 08:49 |
| Venkatesh | Can You review the https://review.openstack.org/#/c/212016/6 | 08:49 |
| Venkatesh | I had a small confusion | 08:50 |
| Venkatesh | should I remove the implementation of abstract method | 08:50 |
| Venkatesh | ? | 08:50 |
| *** derekh has joined #openstack-nova | 08:51 | |
| bauzas | johnthetubaguy: now that the Mitaka change merged, I guess you need to provide the SHA to cut the branch ? | 08:51 |
| *** rotbeard has joined #openstack-nova | 08:53 | |
| bauzas | Venkatesh: hey | 08:54 |
| bauzas | Venkatesh: hans had a quite good argument | 08:54 |
| Venkatesh | Hi bauzas | 08:54 |
| bauzas | Venkatesh: so, I'm not super opiniated either way | 08:54 |
| bauzas | Venkatesh: but since his argument is valid, could you just respin a new patchset by taking in account his comment ? | 08:55 |
| Venkatesh | sure | 08:55 |
| bauzas | Venkatesh: thanks | 08:55 |
| Venkatesh | welcome | 08:56 |
| *** rotbeard has quit IRC | 08:57 | |
| *** haomaiwang has quit IRC | 09:01 | |
| *** haomaiwang has joined #openstack-nova | 09:01 | |
| *** alex_klimov has quit IRC | 09:02 | |
| *** lifeless has joined #openstack-nova | 09:03 | |
| openstackgerrit | Masaki Matsushita proposed openstack/nova: Fix logging_sample.conf to use oslo_log formatter https://review.openstack.org/227186 | 09:05 |
| *** rotbeard has joined #openstack-nova | 09:11 | |
| *** atuvenie has joined #openstack-nova | 09:15 | |
| *** MarcF has quit IRC | 09:15 | |
| *** otter768 has joined #openstack-nova | 09:17 | |
| *** MarcF has joined #openstack-nova | 09:17 | |
| johnthetubaguy | Venkatesh: I think there is something nice about defining the method you need to over-ride, but others may dissagree | 09:20 |
| Venkatesh | johnthetubaguy: Ok, I will submit a new patch | 09:20 |
| Venkatesh | Thanks | 09:20 |
| *** otter768 has quit IRC | 09:22 | |
| *** mpaolino has quit IRC | 09:24 | |
| *** mpaolino has joined #openstack-nova | 09:25 | |
| *** ndipanov has joined #openstack-nova | 09:27 | |
| *** dtantsur has joined #openstack-nova | 09:27 | |
| dtantsur | morning folks! | 09:27 |
| dtantsur | is it a know failure: http://logs.openstack.org/05/226805/2/check/gate-ironic-inspector-dsvm/5cd5071/logs/screen-n-cond.txt.gz?level=WARNING ? | 09:27 |
| *** markvoelker has joined #openstack-nova | 09:29 | |
| *** aix has quit IRC | 09:29 | |
| johnthetubaguy | bauzas: I have a feeling we have an issue with your metrics patch, did you see the above link from dtantsur ? | 09:30 |
| bauzas | johnthetubaguy: looking | 09:31 |
| bauzas | f*** | 09:31 |
| johnthetubaguy | yeah | 09:31 |
| johnthetubaguy | different host manager | 09:31 |
| johnthetubaguy | damm it | 09:31 |
| bauzas | so a dict comprehension is not liking a None | 09:31 |
| bauzas | I think it's a python problem | 09:31 |
| dtantsur | maybe something like (None or ()) would work... | 09:32 |
| johnthetubaguy | oh... host_state is None manybe? | 09:32 |
| bauzas | johnthetubaguy: nope, HostState.metrics = None | 09:32 |
| bauzas | by default | 09:32 |
| johnthetubaguy | oh, no, right, host_state.metrics is none | 09:32 |
| johnthetubaguy | yeah | 09:32 |
| openstackgerrit | Bartosz Fic proposed openstack/nova: Live migration failure in API doesn't revert task_state of vm https://review.openstack.org/168916 | 09:32 |
| bauzas | unless you call update_from_cn() where it updates it to have a metricslist | 09:32 |
| johnthetubaguy | bauzas: did that get into RC1 that we just cut? | 09:32 |
| bauzas | johnthetubaguy: nope, lemme check | 09:32 |
| bauzas | johnthetubaguy: I think it was merged *after* | 09:33 |
| * johnthetubaguy crosses fingers | 09:33 | |
| bauzas | johnthetubaguy: I mean after the Mitaka patch | 09:33 |
| johnthetubaguy | bauzas: cool, let me know when you have a none check added, and we can push that through | 09:33 |
| dtantsur | thanks a lot! | 09:33 |
| *** markvoelker has quit IRC | 09:34 | |
| johnthetubaguy | dtantsur: thanks for letting us know | 09:34 |
| bauzas | johnthetubaguy: yup | 09:34 |
| bauzas | dtantsur: what puzzles me is why you have this exception | 09:34 |
| dtantsur | me too... | 09:34 |
| bauzas | dtantsur: since when calling the hostmanager you call update_from_cn() which updates host_state.metrics to a metricslist | 09:34 |
| dtantsur | we don't do anything particularly tricky in gate: we use default ironic devstack setup, then do 'nova boot' | 09:35 |
| bauzas | dtantsur: yeah I know... | 09:35 |
| bauzas | dtantsur: mmmm | 09:35 |
| bauzas | dtantsur: wondering if it's not due to the IronicHostManager | 09:35 |
| *** danpb has joined #openstack-nova | 09:36 | |
| bauzas | given there is another HostState | 09:36 |
| dtantsur | that's what we do in gate, if it helps: https://github.com/openstack/ironic-inspector/blob/master/devstack/exercise.sh#L188-L192 | 09:37 |
| dtantsur | (yes, we don't have a tempest plugin yet) | 09:37 |
| bauzas | johnthetubaguy: so, it was merged after the Mitaka update | 09:37 |
| bauzas | [sbauza@sbauza nova]$ git log --oneline | egrep "(Open Mitaka development|Fix MetricWeigher to use MonitorMetricList)" | 09:37 |
| bauzas | 14b6d8e Merge "Fix MetricWeigher to use MonitorMetricList" | 09:37 |
| bauzas | 6d736946 Merge "Open Mitaka development" | 09:37 |
| bauzas | fd2b868 Fix MetricWeigher to use MonitorMetricList | 09:37 |
| *** kiran-r has quit IRC | 09:37 | |
| bauzas | d3a4797 Open Mitaka development | 09:37 |
| *** mpaolino has quit IRC | 09:38 | |
| bauzas | dtantsur: fixing the dict comprehension first and then will see why it failed | 09:38 |
| *** eglynn has joined #openstack-nova | 09:39 | |
| dtantsur | yeah, thanks. I'll be back in ~ 1 hour, if you don't mind. | 09:39 |
| *** alex_klimov has joined #openstack-nova | 09:40 | |
| *** mleroy has joined #openstack-nova | 09:45 | |
| *** atuvenie has quit IRC | 09:48 | |
| *** sambetts has joined #openstack-nova | 09:49 | |
| openstackgerrit | Wei Jiangang proposed openstack/nova: wsgi: removing semicolon https://review.openstack.org/227197 | 09:50 |
| *** mpaolino has joined #openstack-nova | 09:50 | |
| *** dtantsur is now known as dtantsur|brb | 09:50 | |
| openstackgerrit | venkatesh proposed openstack/nova: make the driver.Scheduler as abstract class https://review.openstack.org/212016 | 09:54 |
| *** IanGovett has joined #openstack-nova | 09:55 | |
| *** rotbeard has quit IRC | 09:55 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Fix NoneType error when calling MetricsWeigher https://review.openstack.org/227199 | 09:56 |
| *** VikasC has quit IRC | 09:57 | |
| *** Venkatesh has quit IRC | 09:57 | |
| bauzas | johnthetubaguy: ^ | 09:57 |
| bauzas | johnthetubaguy: I marked the bug as critical to make sure we land it asap | 09:57 |
| bauzas | johnthetubaguy: since it's now a gate problem... | 09:57 |
| bauzas | I sincerely apologize for having missed that :/ | 09:58 |
| *** annegentle has joined #openstack-nova | 09:58 | |
| johnthetubaguy | you should blame the reviewers too | 09:58 |
| bauzas | nah, it was an obvious one, I took the decision to update the CachingScheduler instead of just adding an or [] | 09:59 |
| bauzas | it was a bad decision | 09:59 |
| bauzas | johnthetubaguy: btw. has the release team cut RC1 with the help of the SHA you provided ? | 10:00 |
| bauzas | johnthetubaguy: haven't yet seen a ttx email about Nova | 10:00 |
| *** haomaiwang has quit IRC | 10:01 | |
| bauzas | johnthetubaguy: just wondering how to backport the above bug to Liberty (I mean 2 changes or one squashed) | 10:01 |
| bauzas | given the previous commit was buggy | 10:01 |
| *** haomaiwang has joined #openstack-nova | 10:01 | |
| *** kiranr has quit IRC | 10:02 | |
| johnthetubaguy | bauzas: unsure what is normal, two commits squashed makes sense to me, with both commit messages in the git commit message | 10:02 |
| *** aix has joined #openstack-nova | 10:02 | |
| bauzas | yeah with a big fat commit msg explaining why | 10:03 |
| *** annegentle has quit IRC | 10:03 | |
| johnthetubaguy | bauzas: lets wait for jenkins to vote, and then we can rally the troops to get that merged | 10:03 |
| johnthetubaguy | bauzas: yeah | 10:03 |
| bauzas | johnthetubaguy: totally agreed | 10:03 |
| *** rotbeard has joined #openstack-nova | 10:04 | |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: rebuild: Apply migration context before calling the driver https://review.openstack.org/226923 | 10:05 |
| ndipanov | johnthetubaguy, ^ not sure you saw that one - anyway if we are doing rc2 might be good to get it in than | 10:07 |
| ndipanov | then* | 10:07 |
| ttx | bauzas: it's out now | 10:07 |
| *** ildikov has joined #openstack-nova | 10:07 | |
| ttx | I made a combo with trove that was on its way | 10:07 |
| bauzas | ttx: oh cool ! | 10:07 |
| *** atuvenie has joined #openstack-nova | 10:07 | |
| *** claudiub has joined #openstack-nova | 10:08 | |
| *** scheuran has quit IRC | 10:13 | |
| *** shyu has quit IRC | 10:14 | |
| johnthetubaguy | ndipanov: we will need an RC2 for the translations, just use the bug tag to make sure its considered, I think we agreed on liberty-rc-potential (might become liberty-backport-potential, but I think thats more a stable branch thing) | 10:20 |
| *** takedakn has joined #openstack-nova | 10:20 | |
| ndipanov | johnthetubaguy, I did add that I think let me check | 10:21 |
| *** zhangjn has quit IRC | 10:21 | |
| johnthetubaguy | I know a lot were cleared out yesterday, so we can start again | 10:21 |
| johnthetubaguy | oh dear, has gerrit stopped again :( | 10:21 |
| *** takedakn has quit IRC | 10:22 | |
| *** kiran-r has joined #openstack-nova | 10:22 | |
| ndipanov | johnthetubaguy, yep still has the tag so | 10:22 |
| ndipanov | it's not critical for non-numa stuff - that's still fixed and will do RT-ing | 10:23 |
| ndipanov | but would be good to get in if possible - it's tiny and dansmith likes it | 10:23 |
| ndipanov | :) | 10:23 |
| *** e0ne has joined #openstack-nova | 10:23 | |
| *** e0ne has quit IRC | 10:24 | |
| *** sudipto has joined #openstack-nova | 10:28 | |
| johnthetubaguy | ndipanov: seems like a worth while backport to the stable branch before RC2 goes out | 10:28 |
| johnthetubaguy | ndipanov: just getting my head around it | 10:28 |
| ndipanov | johnthetubaguy, thanks | 10:28 |
| *** rdopiera has quit IRC | 10:29 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Fix NoneType error when calling MetricsWeigher https://review.openstack.org/227199 | 10:29 |
| bauzas | johnthetubaguy: so I verified and the unittest was not cool because when you have only one host, it doesn't call the weigher... | 10:29 |
| bauzas | johnthetubaguy: so I needed to respin another one | 10:29 |
| bauzas | johnthetubaguy: it was not covering my change | 10:29 |
| johnthetubaguy | bauzas: bummer, yes, thats true | 10:30 |
| *** Venkatesh has joined #openstack-nova | 10:32 | |
| claudiub | johnthetubaguy: hello. if you have some time, can you take a look at the host capabilities spec? https://review.openstack.org/#/c/222200/ | 10:33 |
| claudiub | johnthetubaguy: also, this is the os-win in nova patch, hyper-v ci is happy with it: https://review.openstack.org/#/c/212102/ | 10:34 |
| johnthetubaguy | claudiub: yeah, will be getting that on my list soon, just looking at the ones that have been waiting a month or two first | 10:34 |
| johnthetubaguy | claudiub: you need the blueprint approved first right? | 10:34 |
| claudiub | johnthetubaguy: true. | 10:34 |
| *** EinstCrazy has joined #openstack-nova | 10:34 | |
| *** chenzeng has quit IRC | 10:35 | |
| *** ndipanov has quit IRC | 10:35 | |
| claudiub | johnthetubaguy: https://blueprints.launchpad.net/nova/+spec/add-os-win-library | 10:35 |
| johnthetubaguy | claudiub: I need to send an email about that, to kick off things for mitaka | 10:35 |
| claudiub | johnthetubaguy: also, you previously had a +2 on the uefi secure boot spec: https://review.openstack.org/#/c/190997/ | 10:35 |
| johnthetubaguy | probably post nova-meeting, I suspect | 10:35 |
| claudiub | johnthetubaguy: ok, cool. :) | 10:36 |
| johnthetubaguy | yep, thats on the list, trying to work out whats there first, then start tackling them | 10:36 |
| bauzas | dtantsur|brb: johnthetubaguy: shit, I got the problem | 10:37 |
| johnthetubaguy | we have 45 specs wanting a review, I know i average around 30 mins per spec review | 10:37 |
| *** irenab has quit IRC | 10:37 | |
| johnthetubaguy | bauzas: I was going to say, did you reproduce the traceback in your unit test yet? | 10:38 |
| bauzas | dtantsur|brb: johnthetubaguy: IronicHostManager reimplements update_from_cn() which was setting the right host_state.metrics | 10:38 |
| claudiub | johnthetubaguy: i see. no worries, take your time. :) | 10:38 |
| johnthetubaguy | bauzas: yeah, different host manager is often an issue here | 10:38 |
| bauzas | johnthetubaguy: yeah that's why I respined the change | 10:38 |
| *** vlaza has joined #openstack-nova | 10:38 | |
| bauzas | johnthetubaguy: I was wondering why the new unittest I added wasn't failing when I was reverting my change | 10:38 |
| johnthetubaguy | bauzas: totally need to get some functional tests around all the filters and weights, which both host managers | 10:39 |
| bauzas | johnthetubaguy: hell yeah | 10:39 |
| johnthetubaguy | bauzas: do you fancy that as a follow on patch? | 10:39 |
| bauzas | you mean Tempest gating ? | 10:39 |
| johnthetubaguy | bauzas: I think that caching_scheduler patch might do it | 10:39 |
| johnthetubaguy | bauzas: no, in tree tests | 10:39 |
| *** rdopiera has joined #openstack-nova | 10:39 | |
| bauzas | johnthetubaguy: that's doable, I'm pretty more concerned about the coverage we could get | 10:39 |
| *** salv-orlando has quit IRC | 10:40 | |
| bauzas | johnthetubaguy: since lots of filters are behaviouring differently given the provided resources | 10:40 |
| bauzas | johnthetubaguy: but at least we can add just a default check | 10:40 |
| *** salv-orlando has joined #openstack-nova | 10:40 | |
| bauzas | johnthetubaguy: so we could improve that later on | 10:41 |
| johnthetubaguy | bauzas: yeah, I was thinking just something that touches them all, so people can build on it | 10:41 |
| johnthetubaguy | bauzas: yeah, what you just said :) | 10:41 |
| bauzas | johnthetubaguy: yeah the cachingscheduler performance test is a first step | 10:41 |
| *** amotoki has joined #openstack-nova | 10:41 | |
| bauzas | johnthetubaguy: so I could create some new in-tree module for leaving room to improvements | 10:41 |
| johnthetubaguy | yeah, adding one for ironic would be a good first step | 10:42 |
| *** zhenguo has quit IRC | 10:42 | |
| bauzas | although it wouldn't be a real functional test, just a big unittest | 10:42 |
| bauzas | covering a 2-level depth of calls | 10:42 |
| bauzas | johnthetubaguy: I can certainly write that, please stay on the line :) | 10:43 |
| *** FL1SK has joined #openstack-nova | 10:44 | |
| bauzas | johnthetubaguy: the other way I was thinking about functionally testing the scheduler was to do that way : https://review.openstack.org/#/c/221779/4/nova/tests/functional/test_servers.py,cm (but modifying the flags so that all filters would be called) | 10:47 |
| bauzas | johnthetubaguy: but that's just a basic coverage that doesn't allow to refine more tests | 10:47 |
| *** haomaiwang has quit IRC | 10:47 | |
| *** haomaiwang has joined #openstack-nova | 10:48 | |
| *** haomaiwa_ has joined #openstack-nova | 10:48 | |
| *** haomaiwang has quit IRC | 10:48 | |
| johnthetubaguy | bauzas: yeah, I think more isolation is a good idea to start with | 10:48 |
| johnthetubaguy | bauzas: something nice and simple to extend | 10:49 |
| bauzas | okay, I have something in mind | 10:49 |
| *** mpaolino has quit IRC | 10:51 | |
| *** stevemar has joined #openstack-nova | 10:52 | |
| *** dtantsur|brb is now known as dtantsur | 10:53 | |
| * dtantsur is back | 10:55 | |
| *** stevemar has quit IRC | 10:55 | |
| *** gongysh has quit IRC | 10:57 | |
| *** dave-mccowan has joined #openstack-nova | 10:57 | |
| dtantsur | bauzas, so https://review.openstack.org/#/c/227199 is the fix, right? If so, I'll try to add Depends-On for one of our patches | 10:58 |
| bauzas | dtantsur: yup, like I said, the problem is that Ironic is overriding the method responsible for setting host_state.metrics | 10:59 |
| dtantsur | oh.. | 10:59 |
| *** tomoiaga has joined #openstack-nova | 10:59 | |
| openstackgerrit | Zhao Lei proposed openstack/nova-specs: Fix a value typo for hw_numa_mempolicy https://review.openstack.org/227222 | 11:00 |
| *** haomaiwa_ has quit IRC | 11:01 | |
| *** rook has quit IRC | 11:01 | |
| *** haomaiwa_ has joined #openstack-nova | 11:01 | |
| *** irenab has joined #openstack-nova | 11:03 | |
| dtantsur | https://review.openstack.org/#/c/227193/ now Depends-On this change, we'll see how it goes | 11:03 |
| tomoiaga | I wonder what happened to compute.instance.create start/end notifications. Are those replaced by compute.instance.update ? | 11:04 |
| johnthetubaguy | tomoiaga: should have both I thought | 11:05 |
| johnthetubaguy | tomoiaga: problem is the notification are totally untested, and un versioned right now, so its really as stable as log messages right now | 11:06 |
| johnthetubaguy | tomoiaga: there is a plan to fix all that, but no one working on it right now | 11:06 |
| tomoiaga | johnthetubaguy: I am still looking through the code. I can see the compute node sending create.start/end but I am not receiving them throufh ceilometer and a second queue I setup for duplicates. | 11:06 |
| tomoiaga | johnthetubaguy: than you, I'll investigate further, maybe there is a bug or something else wrong here | 11:07 |
| johnthetubaguy | yeah, would be interesting to know whats happening inside ceilomter to those, I am stacktach user, I think we are still getting those, last time I checked | 11:07 |
| *** rook has joined #openstack-nova | 11:09 | |
| tomoiaga | johnthetubaguy: I was going to try out stacktach but I can see ceilometer supporting events and what I need for now. I'm curious anyway if stacktach receives those events :) | 11:09 |
| *** njirap has joined #openstack-nova | 11:09 | |
| openstackgerrit | ZhuChunzhan proposed openstack/nova: Improve the performance of get instances https://review.openstack.org/214555 | 11:11 |
| *** moshele has joined #openstack-nova | 11:14 | |
| *** otter768 has joined #openstack-nova | 11:18 | |
| markus_z | sahid: johnthetubaguy: I answered to your feedback in https://review.openstack.org/#/c/215102/12 | 11:19 |
| markus_z | It could be that johnthetubaguy found something, but I'm not sure if I see this right. | 11:19 |
| johnthetubaguy | markus_z: we have to upgrade live between any commit, not just release to release | 11:20 |
| markus_z | johnthetubaguy: Right, I remember the statement. | 11:23 |
| *** otter768 has quit IRC | 11:23 | |
| markus_z | johnthetubaguy: I think the example I gave in the review still holds. | 11:23 |
| markus_z | johnthetubaguy: ... oh... one sec... | 11:24 |
| * johnthetubaguy bravely runs off to get some lunch | 11:24 | |
| *** zhangjn has joined #openstack-nova | 11:27 | |
| *** tpeoples has joined #openstack-nova | 11:28 | |
| *** moshele has quit IRC | 11:28 | |
| garyk | johnthetubaguy: danpb: added missing libvirt tests for the designer - https://review.openstack.org/225975 | 11:29 |
| *** achanda has joined #openstack-nova | 11:29 | |
| *** exploreshaifali has joined #openstack-nova | 11:30 | |
| *** markvoelker has joined #openstack-nova | 11:30 | |
| *** ociuhandu has quit IRC | 11:30 | |
| *** exploreshaifali has quit IRC | 11:32 | |
| *** markvoelker has quit IRC | 11:34 | |
| *** Kevin_Zheng has quit IRC | 11:35 | |
| *** tomoiaga has left #openstack-nova | 11:37 | |
| *** alexpilotti has joined #openstack-nova | 11:41 | |
| *** achanda has quit IRC | 11:41 | |
| *** bswartz has quit IRC | 11:43 | |
| *** takedakn has joined #openstack-nova | 11:44 | |
| *** thorst has joined #openstack-nova | 11:46 | |
| *** annegentle has joined #openstack-nova | 11:47 | |
| *** annegentle has quit IRC | 11:52 | |
| *** takedakn has quit IRC | 11:52 | |
| *** sudipto has quit IRC | 11:52 | |
| *** njirap has quit IRC | 11:54 | |
| *** Sree has joined #openstack-nova | 11:54 | |
| *** yuner027 has joined #openstack-nova | 11:55 | |
| *** baoli has joined #openstack-nova | 11:56 | |
| *** takedakn has joined #openstack-nova | 11:56 | |
| *** haomaiwa_ has quit IRC | 12:01 | |
| openstackgerrit | ZhuChunzhan proposed openstack/nova: Improve the performance of get instances https://review.openstack.org/214555 | 12:01 |
| *** haomaiwang has joined #openstack-nova | 12:01 | |
| *** rfolco has joined #openstack-nova | 12:02 | |
| *** aysyd has joined #openstack-nova | 12:02 | |
| *** bkopilov has quit IRC | 12:03 | |
| *** exploreshaifali has joined #openstack-nova | 12:03 | |
| openstackgerrit | jichenjc proposed openstack/nova: Remove duplicate VALID_NAME_REGEX https://review.openstack.org/218489 | 12:04 |
| *** ildikov has quit IRC | 12:05 | |
| *** ankit_ag has quit IRC | 12:06 | |
| *** haomaiwang has quit IRC | 12:09 | |
| markus_z | Is anyone noticing a slow down of Gerrit too? | 12:10 |
| dtantsur | very slow | 12:10 |
| vlaza | +1 | 12:12 |
| *** artom has joined #openstack-nova | 12:12 | |
| claudiub | ++ | 12:13 |
| dtantsur | bauzas, johnthetubaguy, my patch depending on the gate fix succeeded \o/ could you please merge it? | 12:13 |
| *** gszasz has joined #openstack-nova | 12:14 | |
| garyk | yeah, terrible at the moment | 12:15 |
| *** exploreshaifali has quit IRC | 12:16 | |
| *** sayali has quit IRC | 12:17 | |
| Venkatesh | Hi all, | 12:19 |
| Venkatesh | for the following patch | 12:19 |
| Venkatesh | https://review.openstack.org/#/c/212016/ | 12:19 |
| Venkatesh | jenkins is giving -1 | 12:19 |
| *** dims__ has quit IRC | 12:19 | |
| Venkatesh | how to recheck? | 12:19 |
| *** sayali has joined #openstack-nova | 12:20 | |
| *** dims_ has joined #openstack-nova | 12:20 | |
| *** annegentle has joined #openstack-nova | 12:22 | |
| *** alex_klimov has quit IRC | 12:24 | |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: rebuild: Apply migration context before calling the driver https://review.openstack.org/226923 | 12:25 |
| markus_z | Venkatesh: hans already did it. Your change is in the queue again: http://status.openstack.org/zuul/ | 12:27 |
| Venkatesh | OK, Thanks | 12:27 |
| *** chenzeng has joined #openstack-nova | 12:28 | |
| *** Daisy_ has joined #openstack-nova | 12:29 | |
| *** raildo-afk is now known as raildo | 12:29 | |
| *** markvoelker has joined #openstack-nova | 12:31 | |
| *** Venkatesh has quit IRC | 12:32 | |
| *** edmondsw has joined #openstack-nova | 12:32 | |
| *** ociuhandu has joined #openstack-nova | 12:33 | |
| dtantsur | core folks, could you please merge https://review.openstack.org/#/c/227199/ ? it blocks the gate for ironic-inspector. thanks! | 12:33 |
| dtantsur | danpb, ^^ | 12:34 |
| *** annegentle has quit IRC | 12:34 | |
| *** jamielennox|away is now known as jamielennox | 12:37 | |
| *** burgerk has joined #openstack-nova | 12:37 | |
| openstackgerrit | Jianghua Wang proposed openstack/nova: XenAPI: Add support for vGPU https://review.openstack.org/223426 | 12:39 |
| *** leitan has joined #openstack-nova | 12:40 | |
| *** kiran-r has quit IRC | 12:43 | |
| *** tpeoples has quit IRC | 12:44 | |
| *** zhangjn has quit IRC | 12:46 | |
| *** EinstCrazy has quit IRC | 12:46 | |
| *** Daisy_ has quit IRC | 12:46 | |
| *** Daisy has joined #openstack-nova | 12:46 | |
| *** kodokuu has joined #openstack-nova | 12:47 | |
| *** mnaser has quit IRC | 12:47 | |
| *** vladikr has joined #openstack-nova | 12:48 | |
| *** ccarmack has joined #openstack-nova | 12:50 | |
| *** Daisy has quit IRC | 12:50 | |
| *** deepakcs has quit IRC | 12:51 | |
| *** alex_klimov has joined #openstack-nova | 12:53 | |
| *** stevemar has joined #openstack-nova | 12:53 | |
| *** mnaser has joined #openstack-nova | 12:54 | |
| *** takedakn has quit IRC | 12:56 | |
| *** rlrossit has joined #openstack-nova | 12:56 | |
| *** stevemar has quit IRC | 12:57 | |
| *** takedakn has joined #openstack-nova | 12:59 | |
| *** ildikov has joined #openstack-nova | 12:59 | |
| *** julien_ZTE has joined #openstack-nova | 13:00 | |
| *** tongli has joined #openstack-nova | 13:00 | |
| *** mdrabe has joined #openstack-nova | 13:03 | |
| *** chenzeng has quit IRC | 13:04 | |
| *** mdrabe has quit IRC | 13:04 | |
| *** garyk has quit IRC | 13:06 | |
| *** Sree has quit IRC | 13:07 | |
| *** larsks has quit IRC | 13:07 | |
| openstackgerrit | jichenjc proposed openstack/nova-specs: Add change-return-code-api-20x.rst https://review.openstack.org/227274 | 13:08 |
| *** ctrath has joined #openstack-nova | 13:09 | |
| *** sayali has quit IRC | 13:10 | |
| *** thumpba has joined #openstack-nova | 13:11 | |
| *** EinstCrazy has joined #openstack-nova | 13:12 | |
| openstackgerrit | Bartosz Fic proposed openstack/nova: Block migrate with attached volumes copies volumes to themselves https://review.openstack.org/227278 | 13:12 |
| *** sayali has joined #openstack-nova | 13:13 | |
| *** shoutm has quit IRC | 13:13 | |
| *** deepthi has quit IRC | 13:14 | |
| *** alejandrito has joined #openstack-nova | 13:14 | |
| *** ijuwang has joined #openstack-nova | 13:16 | |
| *** mriedem_away is now known as mriedem | 13:18 | |
| *** otter768 has joined #openstack-nova | 13:19 | |
| *** READ10 has joined #openstack-nova | 13:21 | |
| mriedem | alaski: bauzas: DuncanT: did anyone propose a revert yet for the cinder.cross_az_attach deprecation change? | 13:22 |
| mriedem | my interweb is crawling | 13:22 |
| openstackgerrit | Zhao Lei proposed openstack/nova: Fix some spelling typo in manual https://review.openstack.org/227294 | 13:22 |
| alaski | I haven't, but I was about to go look | 13:22 |
| mriedem | ok, i'll propose it, | 13:22 |
| mriedem | but wanted to reply to the ML first | 13:22 |
| alaski | gerrit is really slow today | 13:22 |
| *** takedakn has quit IRC | 13:23 | |
| DuncanT | mriedem: I didn't - I'm happy to sort it after tokyo, since we might well end up deprecating that opetion and fixing up cinder to make AZ decisions rather than nova needing to know anything about cinder deployment | 13:23 |
| mriedem | as part of the revert, i wanted to point out that (1) this all needs to be better documented, and (2) i'm interested in how godaddy is handling boot from volume with this, because that needs some work wrt cross_az_attach | 13:23 |
| *** otter768 has quit IRC | 13:23 | |
| mriedem | DuncanT: yeah, that's the problem i'm afraid of, is in boot from volume nova has to figure out what cinder has setup for the tenant and then make decisions | 13:23 |
| mriedem | plus there are some big usability issues here, because this is really api policy defined in config options that the user has no idea about | 13:24 |
| DuncanT | mriedem: Why? It just asks cinder to create a volume, telling cinder what AZ the compute host is in... cinder will know its AZ config, and can therefore just do the right thing | 13:24 |
| mriedem | DuncanT: unless that az isn't defined in cinder | 13:25 |
| mriedem | which is the bug that started all of this | 13:25 |
| *** VW has joined #openstack-nova | 13:25 | |
| *** tpeoples has joined #openstack-nova | 13:25 | |
| DuncanT | mriedem: That's cinder's problem, not nova's IMO. We need to fix cinder to handle that correctly, not add work arounds in nova | 13:25 |
| DuncanT | mriedem: I'm very much of the opinion that we need to get knowledge of cinder's internals out of nova, not add more | 13:26 |
| *** jichen has quit IRC | 13:26 | |
| mriedem | there was a workaround option added to cinder in liberty to fallback to a default cinder az if the one requested doesn't exist, but that's a config option and it's false by default | 13:26 |
| mriedem | DuncanT: i'd agree with that :) | 13:26 |
| mriedem | wrt attach volume cases, cross_az_attach is pretty clear | 13:26 |
| *** Venkatesh has joined #openstack-nova | 13:26 | |
| mriedem | it's just the boot from volume case that is buggy and not very usable imo | 13:26 |
| mriedem | i'm pretty sure when this fails on a boot from volume, the user is just getting back NoValidHost | 13:26 |
| *** rotbeard has quit IRC | 13:27 | |
| DuncanT | mriedem: again, I don't think nova should be making decisions about what can and can't be attached, that's a cinder decision, just make sure cinder knows the compute AZ as part of the attach method and it can decide if that makes sense or not | 13:27 |
| mriedem | literal attach or volume create? | 13:27 |
| DuncanT | mriedem: Both | 13:28 |
| mriedem | or just both | 13:28 |
| mriedem | yeah | 13:28 |
| alaski | yep, agreed | 13:28 |
| mriedem | ok, so i'll propose the revert once gerrit is working and i'm out of my next meeting | 13:28 |
| *** xyang1 has joined #openstack-nova | 13:29 | |
| mriedem | i plan on updating the help text for that option also as part of this, and maybe writing up some devref on it | 13:29 |
| DuncanT | mriedem: We need to re-do the nova<->cinder API to fix a whole bunch of issues, we'll fix this there as well. Still leaves the legacy code in a bit of a mess, but that's harder to fix. The default API for the M release should be the new one, and we can do that right | 13:29 |
| bauzas | mriedem: DuncanT: alaski: sorry was afk | 13:29 |
| bauzas | scrolling back | 13:29 |
| DuncanT | mriedem: I'm certainly happy to review (or find people with a good understanding to review) and docs. Thanks | 13:29 |
| *** mriedem is now known as mriedem_away | 13:30 | |
| *** mriedem_away is now known as mriedem_meeting | 13:30 | |
| alaski | DuncanT: what's the best way to log a feature request against Cinder to track this? | 13:31 |
| bauzas | DuncanT: mriedem_meeting: alaski: so given our discussion, I prefer to see the Cinder session during Tokyo before doing anything | 13:31 |
| mriedem_meeting | bauzas: you mean before reverting the deprecation? | 13:32 |
| DuncanT | alaski: A spec normally.... though I plan on writing one to cover the 2 usecases I've already got (detailed in the email thread) so maybe giving me a couple of days to get that done then reviewing that might be easier? | 13:32 |
| bauzas | mriedem_meeting: no no, sorry | 13:32 |
| bauzas | mriedem_meeting: I mean, let's revert the change and discuss the AZ with the Cinder guys anytime | 13:33 |
| alaski | DuncanT: that would be great | 13:33 |
| mriedem_meeting | bauzas: yeah, working on that | 13:33 |
| *** takedakn has joined #openstack-nova | 13:33 | |
| bauzas | DuncanT: MHO is that I would really know why Nova AZs are useful in Cinder - if this is just for knowing | 13:34 |
| *** su_zhang has joined #openstack-nova | 13:34 | |
| bauzas | DuncanT: I mean, like I said to you, I know that operators use the AZs, but that's really a workaround IMHO | 13:34 |
| alaski | bauzas: so far the only use I've seen is to do rough cross project scheduling | 13:35 |
| bauzas | possibly | 13:35 |
| bauzas | alaski: I mean, I wanted to discuss about x-project scheduling | 13:35 |
| bauzas | alaski: but last times, I've been told that's only for Nova vs. Cinder vs. Neutron, not a really x-project discussion | 13:36 |
| *** dustins has joined #openstack-nova | 13:36 | |
| bauzas | but sure, we can try to propose a discussion about that | 13:36 |
| alaski | it would probably have to take a slot in one of those projects tracks | 13:37 |
| *** romainh has left #openstack-nova | 13:37 | |
| *** tpeoples has quit IRC | 13:37 | |
| alaski | but it seems a bit early to be discussing that, right? | 13:38 |
| *** njirap has joined #openstack-nova | 13:38 | |
| DuncanT | bauzas: Nova AZs are useful to operators who choose to group their storage in the same way as their nova AZs are grouped (which is a logical thing to do) | 13:38 |
| *** burgerk has quit IRC | 13:38 | |
| alaski | we still have a good bit of work to do before we can consider splitting a scheduler to do cross project work | 13:38 |
| DuncanT | bauzas: Examples of reasons why you'd do it include thinning down the cross-az storage traffic | 13:38 |
| dtantsur | folks, sorry for nagging you, but ironic-inspector plans on liberty release today, and we're blocked by https://review.openstack.org/#/c/227199/ | 13:39 |
| dtantsur | could you please merge it? | 13:39 |
| bauzas | argh, I had some IRC problem so my last phrases were out | 13:39 |
| *** thumpba has quit IRC | 13:39 | |
| *** thumpba has joined #openstack-nova | 13:40 | |
| bauzas | alaski: no, I was not saying to split the scheduler | 13:41 |
| smcginnis | DuncanT: AZ's could also be useful for folks using fibre channel. Don't try to access storage on a different fabric. | 13:42 |
| bauzas | alaski: just having a discussion with the other projects to see what could be the requirements for providing to the nova scheduler the list of resources that the scheduler could consume for having affinity | 13:42 |
| DuncanT | smcginnis: ++ | 13:42 |
| bauzas | alaski: but I'm not very opiniated on that | 13:42 |
| *** garyk has joined #openstack-nova | 13:43 | |
| *** haomaiwang has joined #openstack-nova | 13:43 | |
| garyk | johnthetubaguy: when can all of the patches that are blocked be unblocked? when will rc1 be cut? | 13:43 |
| johnthetubaguy | garyk: when the blueprints are approved for mitaka | 13:44 |
| johnthetubaguy | need to discuss how we do that at the meeting later, or this afternoon in prep for that, will send a mail on tomorrow saying how to do that | 13:44 |
| garyk | can you approve them? you approved them in liberty | 13:44 |
| johnthetubaguy | for the previously approved specs, I have a few of those in the pipe already | 13:45 |
| DuncanT | bauzas: AZs are different to the general use affinity directives as far as I'm concerned. Volume affinity can mean different things to different people | 13:45 |
| garyk | johnthetubaguy: and those that do not require specs? | 13:45 |
| *** snikitin has quit IRC | 13:45 | |
| *** zhangjn has joined #openstack-nova | 13:45 | |
| *** zhangjn has quit IRC | 13:46 | |
| johnthetubaguy | garyk: not decided the process yet, want to discuss ideas with folks first | 13:46 |
| openstackgerrit | Daniel Berrange proposed openstack/nova-specs: Libvirt real time instances https://review.openstack.org/139688 | 13:46 |
| garyk | ok. sure | 13:46 |
| *** zhangjn has joined #openstack-nova | 13:47 | |
| *** zhangjn has quit IRC | 13:47 | |
| *** zhangjn has joined #openstack-nova | 13:48 | |
| *** zhangjn has quit IRC | 13:49 | |
| johnthetubaguy | garyk: so, for the moment, feel free to add them into here, under specless blueprint: https://etherpad.openstack.org/p/mitaka-nova-spec-review-tracking | 13:49 |
| *** njirap has quit IRC | 13:49 | |
| johnthetubaguy | garyk: if we change how it works, I will move them for you | 13:49 |
| *** sudipto has joined #openstack-nova | 13:50 | |
| garyk | ok, will do | 13:50 |
| bauzas | DuncanT: agreed, I'm just concerned about the overall design of that | 13:50 |
| *** njirap has joined #openstack-nova | 13:51 | |
| bauzas | DuncanT: like I said, operators used the AZs in some way that we didn't formalized, and before moving on, I want to make sure that it matches with our goals | 13:51 |
| *** karimb has quit IRC | 13:52 | |
| *** thumpba has quit IRC | 13:52 | |
| DuncanT | bauzas: Well, there'll be a good writeup in a cinder spec in the near future, and I'll add a link to it on the mailing list thread. It covers the two ways I know operators want AZs to work. Feedback very welcome. | 13:53 |
| bauzas | DuncanT: okay, I don't want to cycle more than that | 13:53 |
| DuncanT | bauzas: If somebody else is doing something different, I'd like to hear about it. If not, we formalise these as the two supported configs and move on | 13:53 |
| bauzas | DuncanT: so IIUC, Cinder AZs are more metatadate applied to volumes, right ? nothing really mapped with some logical components, like we do with aggregates? | 13:54 |
| DuncanT | They're metadata applied to backends at the moment | 13:55 |
| DuncanT | bauzas: a backend can only be in one AZ | 13:55 |
| bauzas | when you say backend, you say cinder-volume services ? | 13:56 |
| *** eharney has joined #openstack-nova | 13:56 | |
| *** EinstCrazy has quit IRC | 13:57 | |
| *** achanda has joined #openstack-nova | 13:57 | |
| *** sudipto has quit IRC | 13:57 | |
| *** zzzeek has joined #openstack-nova | 13:58 | |
| *** achanda has quit IRC | 13:59 | |
| *** zhangjn has joined #openstack-nova | 13:59 | |
| *** baoli has quit IRC | 14:00 | |
| *** haomaiwang has quit IRC | 14:01 | |
| *** njirap has quit IRC | 14:01 | |
| *** haomaiwang has joined #openstack-nova | 14:01 | |
| *** baoli has joined #openstack-nova | 14:01 | |
| bauzas | DuncanT: could you please point me out some place where the AZ enforcement is done in Cinder ? I need to ramp-up on Cinder before actually being opiniated enough | 14:01 |
| *** karimb has joined #openstack-nova | 14:01 | |
| openstackgerrit | Daniel Berrange proposed openstack/nova-specs: Libvirt driver emulator threads placement policy https://review.openstack.org/225893 | 14:02 |
| DuncanT | bauzas: At the moment it is ill defined, inconsistent and possibly broken | 14:02 |
| bauzas | DuncanT: ack | 14:02 |
| DuncanT | bauzas: And yes, backend == cinder volume service | 14:03 |
| *** njirap has joined #openstack-nova | 14:06 | |
| mriedem_meeting | is sam morrison ever on irc? | 14:07 |
| *** mriedem_meeting is now known as mriedem | 14:07 | |
| mriedem | does godaddy allow boot from volume where source != volume? | 14:07 |
| *** diogogmt has joined #openstack-nova | 14:09 | |
| mriedem | danpb: dansmith: has anyone from red hat looked at this? https://bugs.launchpad.net/nova/+bug/1415087 - there are several places in nova where qemu-img convert is called w/o specifying -f, so it seems like it'd be an issue in nova too | 14:09 |
| openstack | Launchpad bug 1415087 in OpenStack Compute (nova) "[OSSA 2015-011] Format-guessing and file disclosure in image convert (CVE-2015-1850, CVE-2015-1851)" [High,Triaged] - Assigned to Tony Breeds (o-tony) | 14:09 |
| openstackgerrit | Chuck Carmack proposed openstack/nova: Disable security group tests in tempest-dsvm-cells-rc using a config option https://review.openstack.org/226043 | 14:10 |
| bauzas | mriedem: AFAIK, sam morrison is a NECTAR guy so possibly Australian | 14:10 |
| *** breitz has quit IRC | 14:10 | |
| bauzas | mriedem: that's why his comment is useful, because he probably maps AZs with cells | 14:11 |
| mriedem | yeah he's in australia | 14:11 |
| *** breitz has joined #openstack-nova | 14:11 | |
| dansmith | mriedem: I thought we had a scrub of some of those in the last year, but no, I dunno | 14:11 |
| *** dims__ has joined #openstack-nova | 14:11 | |
| dansmith | kashyap could also probably look too | 14:11 |
| *** njirap has quit IRC | 14:11 | |
| bauzas | mriedem: https://www.openstack.org/community/members/profile/7544 | 14:11 |
| kashyap | dansmith: Sorry, what is it? | 14:11 |
| * kashyap reads a bit context | 14:11 | |
| mriedem | kashyap: https://bugs.launchpad.net/nova/+bug/1415087 | 14:11 |
| openstack | Launchpad bug 1415087 in OpenStack Compute (nova) "[OSSA 2015-011] Format-guessing and file disclosure in image convert (CVE-2015-1850, CVE-2015-1851)" [High,Triaged] - Assigned to Tony Breeds (o-tony) | 14:11 |
| mriedem | kashyap: trying to sort out if nova is affected by that bug | 14:11 |
| *** j_king has quit IRC | 14:12 | |
| mriedem | from looking at the code i'd think it was, but tonyb wasn't able to reproduce | 14:12 |
| *** rdopiera has quit IRC | 14:12 | |
| bauzas | FWIW, just discovered https://www.openstack.org/community/members/ (was pointing from https://wiki.openstack.org/wiki/People), very handy | 14:12 |
| kashyap | mriedem: Ah, this one, the key bit is that if Nova is creating backing files, then we should explcitly specify the backing format | 14:12 |
| kashyap | s/backing format/backing file format/ | 14:13 |
| bauzas | I sometimes suck at knowing what is the nick of the gerrit owner | 14:13 |
| *** j_king has joined #openstack-nova | 14:13 | |
| tonyb | kashyap: Yeah the only place we to that is in create_cow or similar but that was a secondary issue to being able to call convert on an image under user control | 14:14 |
| tonyb | kashyap: It's trivial to do this with volumes but images as too hard for me | 14:14 |
| *** dims_ has quit IRC | 14:14 | |
| kashyap | Bad: qemu-img create -f qcow2 -b ./base.img ./overlay1.qcow2 | 14:14 |
| kashyap | Good: qemu-img create -f qcow2 -b ./base.img -F raw ./overlay1.qcow2 | 14:14 |
| markus_z | bauzas: Uh, fancy :) | 14:14 |
| *** takedakn has quit IRC | 14:15 | |
| *** dave-mccowan has quit IRC | 14:15 | |
| *** shoutm has joined #openstack-nova | 14:16 | |
| tonyb | kashyap: where -F is the same as -o,backing_fmt=xxx ? | 14:16 |
| kashyap | tonyb: Yep, was about to type that | 14:16 |
| kashyap | It's same as: qemu-img create -f qcow2 -b ./base.img -o backing_fmt=raw ./overlay1.qcow2 | 14:16 |
| tonyb | kashyap: okay I have a patch that does that part but It languished while I worked on the convert issue | 14:17 |
| mriedem | kashyap: so this is probably an issue in nova then right? | 14:17 |
| mriedem | even if not recreated? | 14:17 |
| tonyb | kashyap, mriedem: I'm working on a patch that raises an exception if we're converting to raw and the source has a backign_file | 14:17 |
| tonyb | mriedem: there are 2 related issues. | 14:17 |
| kashyap | mriedem: Without code auditing I can't certainly say -- I'm trying to look. . . | 14:18 |
| *** njirap has joined #openstack-nova | 14:18 | |
| kashyap | If Nova is creating backing files w/o explicitly specifying the format, then yes, it is affected. | 14:18 |
| *** jaypipes has joined #openstack-nova | 14:19 | |
| *** tonytan4ever has joined #openstack-nova | 14:19 | |
| kashyap | mriedem: tonyb: From my minimal grepping in the source, seems like the occurances I've seen so far _do_ specify the format. . . | 14:19 |
| tonyb | kashyap: check nova/virt/libvirt/utils.py | 14:20 |
| mriedem | alaski: DuncanT: bauzas: my reply on the ML https://www.mail-archive.com/openstack-dev@lists.openstack.org/msg64179.html - i'll include that when i revert that deprecation change | 14:20 |
| kashyap | mriedem: tonyb: Err, I spoke too soon. What I saw is source_fmt | 14:20 |
| kashyap | tonyb: Yeah, that's what I'm looking at | 14:20 |
| tonyb | kashyap: function at line 75 | 14:21 |
| kashyap | Yeah, create_cow_image() | 14:21 |
| tonyb | kashyap: Yup. | 14:21 |
| tonyb | that one needs the o,backing_fmt=xxx fix which is pretty simple | 14:21 |
| kashyap | Yep | 14:22 |
| kashyap | tonyb: You said you already have a patch, is submitted? | 14:22 |
| tonyb | kashyap: no. | 14:23 |
| mriedem | kashyap: tonyb: so we can at least mark the nova part of the bug as in progress right? | 14:23 |
| mriedem | and affected | 14:23 |
| tonyb | kashyap: I wrote it months ago but it sat there while tried and failed to reproduce the convert bug | 14:23 |
| *** diogogmt has quit IRC | 14:23 | |
| kashyap | mriedem: Yes, we've identified the fix | 14:23 |
| tonyb | mriedem: Well the topic kashyap and I are talkign about it closley related but seperate to the bug at hand IIUC | 14:24 |
| tonyb | kashyap: but that's not the whole fix right? | 14:24 |
| kashyap | tonyb: You mean, for 1415087? | 14:25 |
| sdague | markus_z: was there a question about bug tags the other day? | 14:25 |
| tonyb | kashyap: yes, that's the one | 14:25 |
| kashyap | tonyb: I still have to load full context in my brain of that convert bug. Allow me to read a bit. | 14:26 |
| tonyb | kashyap: okay. | 14:26 |
| *** marzif has joined #openstack-nova | 14:26 | |
| *** jamielennox is now known as jamielennox|away | 14:26 | |
| *** ndipanov has joined #openstack-nova | 14:26 | |
| markus_z | sdague: kind of. I wanted to label bugs with more fine grained tags. | 14:27 |
| markus_z | *want | 14:27 |
| markus_z | sdague: I wanted to make an analysis based on tags were we have bigger pain points from a use-case point of view | 14:28 |
| *** jwcroppe has quit IRC | 14:28 | |
| markus_z | The current tags seem to be very subteam/component oriented, which is fine for organizing but don't give a good direction where to focus on. | 14:29 |
| *** jecarey has quit IRC | 14:29 | |
| garyk | johnthetubaguy: i have updated the list. i guess i need to repost a spec that was half merged. what do you suggest there or do i need to wait? | 14:29 |
| *** VW_ has joined #openstack-nova | 14:29 | |
| *** READ10 has quit IRC | 14:30 | |
| bauzas | markus_z: MHO is that any tag we could provide would need to get some people looking at it | 14:30 |
| markus_z | sdague: I tried to reason it here: http://lists.openstack.org/pipermail/openstack-dev/2015-September/073376.html | 14:30 |
| markus_z | bauzas: why? | 14:31 |
| *** VW__ has joined #openstack-nova | 14:31 | |
| bauzas | markus_z: I mean, we could have more tags, but I wonder what could be the benefits | 14:31 |
| johnthetubaguy | garyk: got lots of specs up for mitaka already, they are starting to get reviews now, so I would get the specs you want for mitaka up ASAP | 14:31 |
| *** VW_ has quit IRC | 14:31 | |
| *** VW__ has quit IRC | 14:31 | |
| sdague | markus_z: so, I think my only comment is maybe we need a prefix or something to those tags to ensure that people realize we're not using them as triage categories | 14:31 |
| *** VW_ has joined #openstack-nova | 14:32 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Revert "Deprecate cinder.cross_az_attach option" https://review.openstack.org/227340 | 14:32 |
| *** VW has quit IRC | 14:32 | |
| mriedem | jaypipes: alaski: dansmith: sdague: bauzas: DuncanT: smcginnis: ^ | 14:32 |
| bauzas | mriedem: still reading your email :) | 14:33 |
| bauzas | mriedem: for the moment, using http://translate.cinder.com | 14:33 |
| garyk | johnthetubaguy: ok | 14:33 |
| *** Murali has quit IRC | 14:33 | |
| garyk | johnthetubaguy: if the spec was approved for liberty then what do we do? | 14:33 |
| sdague | mriedem: it does seem a little crazy that the default is True on that | 14:33 |
| mriedem | sdague: on cinder.cross_az_attach? | 14:34 |
| sdague | mriedem: yeh | 14:34 |
| mriedem | i'm assuming it was done that way for backward compat in grizzly | 14:34 |
| markus_z | sdague: So the assumption right now is really "tag xy" == "subteam abc"? | 14:34 |
| *** marzif_ has joined #openstack-nova | 14:34 | |
| mriedem | but it predated specs and didn't have a blueprint so who the hell knows | 14:34 |
| johnthetubaguy | garyk: same as before, propose for mitaka, and add the previously approved tag in the git commit message: https://github.com/openstack/nova-specs#previously-approved-specifications | 14:34 |
| garyk | johnthetubaguy: ok, will do | 14:35 |
| mriedem | sdague: markus_z: +1 to "I think my only comment is maybe we need a prefix or something to those tags to ensure that people realize we're not using them as triage categories" | 14:35 |
| mriedem | b/c spawn as a tag is not a subteam | 14:35 |
| mriedem | which was why i removed it yesterday | 14:35 |
| sdague | markus_z: not always, but the agressive addition of tagging was to try to create functional silos where we could make sure we could get folks to look at some slice | 14:35 |
| *** electrocucaracha has joined #openstack-nova | 14:36 | |
| *** dave-mccowan has joined #openstack-nova | 14:36 | |
| mriedem | like neutron is not an officail bug tag in nova, but i use it for network (official nova tag) bugs which are for the neutron backend | 14:36 |
| *** marzif has quit IRC | 14:37 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Revert "Deprecate cinder.cross_az_attach option" https://review.openstack.org/227340 | 14:37 |
| mriedem | alaski: sdague: ^ fixed a typo in the commit message | 14:38 |
| DuncanT | sdague: cross-az attach is a close approximation of what most people seem to want | 14:38 |
| markus_z | I wasn't aware that it was this tightly coupled. I'll think about the prefix approach. | 14:38 |
| sdague | DuncanT: really? I thought that people wanted AZ locality | 14:38 |
| mriedem | markus_z: https://wiki.openstack.org/wiki/Nova/BugTriage#Step_1:_Tagging | 14:39 |
| sdague | or is this just because AZ get different uses in different places | 14:39 |
| DuncanT | sdague: Informal survey says not | 14:39 |
| markus_z | mriedem: I've read that... | 14:39 |
| DuncanT | sdague: Nova azs mapping to a pan-az cinder seems the most common case | 14:39 |
| *** zhangjn has quit IRC | 14:39 | |
| bauzas | DuncanT: mostly because jgriffith told us that Cinder AZs are currently not really in use ? | 14:40 |
| *** EinstCrazy has joined #openstack-nova | 14:40 | |
| mriedem | sdague: DuncanT: it all feels very terrible for the end user experience trying to write apps against multiple clouds | 14:40 |
| mriedem | which is my major beef with this | 14:40 |
| bauzas | DuncanT: but rather an historical piece of nova-volume code that was kept | 14:40 |
| mriedem | mmmm, major beef | 14:40 |
| *** zhangjn has joined #openstack-nova | 14:40 | |
| DuncanT | bauzas: I've found two operators using them in anger, whough I don't (yet) know the exact code they're using | 14:40 |
| markus_z | I applied the "triage tags" and added "analysis tags". | 14:40 |
| *** zhangjn has quit IRC | 14:40 | |
| *** obedmr has joined #openstack-nova | 14:40 | |
| bauzas | mriedem: my biggest concern is that we even don't have a concept of multi-clouds in OpenStack but regions and cells | 14:41 |
| *** salv-orlando has quit IRC | 14:41 | |
| *** vlaza has left #openstack-nova | 14:41 | |
| *** su_zhang has quit IRC | 14:41 | |
| mriedem | bauzas: i guess we should have done fractals then | 14:41 |
| *** zhangjn has joined #openstack-nova | 14:41 | |
| mriedem | they were right all along | 14:41 |
| bauzas | regions are keystone, cells are nova | 14:41 |
| *** nic has joined #openstack-nova | 14:41 | |
| *** zhangjn has quit IRC | 14:41 | |
| alaski | mriedem: +1 | 14:41 |
| DuncanT | mriedem: I'm not sure how to square the interoperability circle, but I've a suggestion in the spec | 14:41 |
| *** salv-orlando has joined #openstack-nova | 14:41 | |
| mriedem | yeah, except people also isolate different hypervisors by region | 14:41 |
| dansmith | mriedem: lol | 14:42 |
| *** zhangjn has joined #openstack-nova | 14:42 | |
| *** fawadkhaliq has joined #openstack-nova | 14:42 | |
| *** zhangjn has quit IRC | 14:42 | |
| mriedem | lol to which? you know that's a thing right? | 14:42 |
| bauzas | FWIW, fractal is a language I learnt at university :p | 14:42 |
| *** zhangjn has joined #openstack-nova | 14:43 | |
| *** zhangjn has quit IRC | 14:43 | |
| bauzas | well, rather a component model based on EJBs but I'm diverting | 14:43 |
| dansmith | mriedem: lol to the fractals | 14:43 |
| *** zhangjn has joined #openstack-nova | 14:44 | |
| mriedem | DuncanT: well i think i could at least add some validation to the compute api in the boot from volume case such that if cinder.cross_az_attach is False, and the instance.az isn't in the list of azs from cinder, we return a 400 | 14:44 |
| openstackgerrit | Alexis Lee proposed openstack/nova-specs: Get valid server state https://review.openstack.org/192246 | 14:44 |
| mriedem | DuncanT: but allow_availability_zone_fallback in cinder messes that up | 14:44 |
| *** danpb has left #openstack-nova | 14:45 | |
| *** cfriesen__ has joined #openstack-nova | 14:45 | |
| *** danpb has joined #openstack-nova | 14:45 | |
| *** mtanino has joined #openstack-nova | 14:45 | |
| *** VW_ has quit IRC | 14:45 | |
| DuncanT | mriedem: I'd rather nova just ask cinder if it can satisfy the request | 14:45 |
| mriedem | DuncanT: new api? | 14:45 |
| *** VW has joined #openstack-nova | 14:45 | |
| DuncanT | mriedem: Nova validating things messes up some of what we can do in cinder, and is the cause of much of our current problems | 14:45 |
| DuncanT | mriedem: Very much a new API | 14:45 |
| bauzas | -1 to that | 14:46 |
| bauzas | I'd rather prefer to leave Nova give the AZ information if really Cinder needs | 14:46 |
| bauzas | but that doesn't necessarly mean that Cindre would have AZs or not | 14:46 |
| DuncanT | bauzas: Nova doesn't know the cinder AZ setup, or whether cinder is enforcing the AZs - that's the current problem | 14:47 |
| bauzas | how Cinder matches the AZ is just an operator decision | 14:47 |
| mriedem | bauzas: well, if cross_az_attach is False, then we're supposed to be making sure they land in the same AZ | 14:47 |
| mriedem | DuncanT: right | 14:47 |
| bauzas | mriedem: so, get rid of nova creating the volume IMHO | 14:47 |
| bauzas | I know I'm fool | 14:47 |
| mriedem | bauzas: heh, but then you enrage mgagne | 14:47 |
| DuncanT | bauzas: Cinder making a decision about whether to do anything with the AZ field from nova is a cinder decision | 14:47 |
| bauzas | mriedem: I know | 14:48 |
| mriedem | let's just see how many operators we can piss off here :) | 14:48 |
| DuncanT | It'd be nice to move orchestration somewhere otehr than the core compute API for sure | 14:48 |
| bauzas | mriedem: I always said that I was about to send an email saying "AZs are NOT failure domains, period" and leave the operators yell | 14:48 |
| alaski | DuncanT: yes! | 14:48 |
| bauzas | DuncanT: +1 to that | 14:49 |
| mriedem | alaski: you had a valiant effort trying to bring heat into this :) | 14:49 |
| sdague | well, isn't this just placement, so this is the thing that should go into the scheduler layer? | 14:49 |
| bauzas | sdague: I'm not that sure | 14:49 |
| alaski | mriedem: heh, no one took the bait | 14:49 |
| DuncanT | alaski: While keeping basically the same behaviour from an external PoV.... you can even leave a proxy in nova-api like the volume proxy, at least for a while - then nobody needs to know the dirty secrets | 14:49 |
| bauzas | sdague: given cinder-volumes are attached to only one AZ | 14:49 |
| erlon | ndipanov: ping | 14:49 |
| DuncanT | bauzas: Sometimes yes, sometimes no | 14:50 |
| bauzas | sdague: there is no real placement decision from a nova standpoint | 14:50 |
| openstackgerrit | Alexis Lee proposed openstack/nova-specs: Get valid server state https://review.openstack.org/192246 | 14:50 |
| *** VW has quit IRC | 14:50 | |
| sdague | bauzas: right, but eventually that's placement policy from the cloud perspective | 14:50 |
| alaski | DuncanT: yeah, I think that's the direction this needs to go. an internal separation of responsibilities | 14:50 |
| DuncanT | bauzas: That's my point. Nova doesn't know that info, cinder does | 14:50 |
| bauzas | sdague: I see your point | 14:50 |
| sdague | we should probably stop calling it the scheduler and call it the landscaper or something | 14:50 |
| bauzas | sdague: you mean then NoValidHost if Nova can't place an instance with the same Cinder AZ | 14:50 |
| bauzas | sdague: so, that's then probably Cinder should provide to the scheduler, yes | 14:51 |
| mriedem | bauzas: that's what we already have today | 14:51 |
| mriedem | except novalidhost comes from the compute rather than scheduler | 14:51 |
| mriedem | to the end user it sucks either way | 14:51 |
| *** kiran-r has joined #openstack-nova | 14:51 | |
| bauzas | mriedem: I know DuncanT is against that wording, but call it volume affinity placement | 14:51 |
| bauzas | mriedem: I see the usecase where | 14:51 |
| bauzas | an user could say "okay, I have this volume, I want to create an instance close to that volume" | 14:52 |
| *** apuimedo has quit IRC | 14:52 | |
| johnthetubaguy | not to but in here, but it is a general problem, there are physical limitations that causes dependencies, that you really want to schedule based on | 14:52 |
| bauzas | mriedem: if that's a filter, that's an hard requirement, if that's a weighter, that's a soft requirement | 14:52 |
| johnthetubaguy | networks have similar issues here, I think network segments is trying to model some of those | 14:52 |
| *** thumpba has joined #openstack-nova | 14:52 | |
| johnthetubaguy | bauzas: yeah, I keep wanting to unify those into a weigher with a threshold | 14:52 |
| bauzas | so, here is the thing | 14:53 |
| *** yamahata has quit IRC | 14:53 | |
| bauzas | what if Cinder was able to report to the scheduler : take that, here are the list of volumes and AZs | 14:53 |
| bauzas | and then, when booting using --volume=foo, then map that in a weigher and try to find the closest locality | 14:54 |
| mriedem | well we can get the tenant AZs from the volume API today | 14:54 |
| kashyap | tonyb: So, the backing_file of the base image (in nova/_instances) can always be raw I guess. IIRC, an image uploaded is always first convert to raw. | 14:54 |
| mriedem | the question is if they are used in cinder | 14:54 |
| *** stevemar has joined #openstack-nova | 14:54 | |
| bauzas | mriedem: as per jgriffith, not really atm | 14:54 |
| *** ssurana has joined #openstack-nova | 14:54 | |
| alaski | as johnthetubaguy is saying, this is really the same issue that cells is facing with networks | 14:54 |
| bauzas | mriedem: what I'm a bit not fan of is that defining AZs is really a yes or no decision for placement | 14:54 |
| bauzas | alaski: yup that too | 14:55 |
| alaski | given a resource <here> how do I place a different resource <there> so that here is close to there | 14:55 |
| johnthetubaguy | alaski: yeah, although its a hard limit with cells (usually) and a soft limit with volumes (usually) | 14:55 |
| bauzas | johnthetubaguy: not if we're talking about AZs for volumes, it's an hard limit then | 14:55 |
| johnthetubaguy | alaski: yeah, that | 14:55 |
| tonyb | kashyap: no they can be qcow2 IIUC | 14:55 |
| alaski | johnthetubaguy: well for sam it seemed like a hard limit | 14:55 |
| johnthetubaguy | ah, OK, so thats simpler | 14:55 |
| bauzas | alaski: yup | 14:55 |
| mriedem | yeah i was thinking the compute and volume hosts are the same, so they are in the same AZ for nova and cinder | 14:56 |
| mriedem | tightly coupled | 14:56 |
| bauzas | AZs are hard limits by definition, you're in or you're out | 14:56 |
| *** tsekiyama has joined #openstack-nova | 14:56 | |
| mriedem | in sam's case i mean | 14:56 |
| johnthetubaguy | most folks I heard rack of servers with storage array at the bottom, with some flexibility, but yeah, fiberchannel means limited connectivity between boxes | 14:56 |
| mriedem | so if you want an instance with volume in nova az 1 you also have to have htat volume in cinder az 1 | 14:56 |
| johnthetubaguy | anyways, hard limits, got it | 14:56 |
| alaski | mriedem: that very well could be | 14:57 |
| bauzas | mriedem: yup | 14:57 |
| bauzas | but that means that the scheduler knows that | 14:57 |
| johnthetubaguy | so we need a mapping of "locality" between the services | 14:57 |
| johnthetubaguy | AZ is one way to do that | 14:57 |
| bauzas | which is not an HostState information | 14:57 |
| mriedem | now let's talk about cinder multi-attach and how that's all going to work with AZs :) | 14:57 |
| bauzas | johnthetubaguy: right, that's my point, I'm not sure that AZs are the best model for that | 14:57 |
| *** emagana has joined #openstack-nova | 14:58 | |
| johnthetubaguy | mriedem: a certain unicode character comes to mind | 14:58 |
| mriedem | pile of poo? | 14:58 |
| johnthetubaguy | yup | 14:58 |
| bauzas | mriedem: well, if that's a scheduler filter, we could consider that the success criteria is the sum of all verifications, nope ? | 14:58 |
| *** tjones has joined #openstack-nova | 14:58 | |
| *** browne1 has joined #openstack-nova | 14:59 | |
| johnthetubaguy | so the thing is, I see this as needing a claim system in the scheduler, but I could be overthinking it | 14:59 |
| bauzas | johnthetubaguy: right and a way to get those metrics to the scheduler | 14:59 |
| johnthetubaguy | so reserve a volume, then try get the compute space, if thats not possible, drop claims, try another one, etc | 14:59 |
| DuncanT | mriedem: Multi-attach? If cinder is enforcing AZs then all instances you attach a volume to must be in the same AZ. If cinder isn't enforcing AZs then you can attach anything anywhere | 14:59 |
| mriedem | what if there were a scheduler project that was outside of nova that could be used by both nova and cinder.....call it like, gunt or something....? | 14:59 |
| bauzas | mriedem: nah nah nah | 15:00 |
| mriedem | DuncanT: yeah | 15:00 |
| mriedem | bauzas: joking | 15:00 |
| *** abhishekk has left #openstack-nova | 15:00 | |
| bauzas | mriedem: given Cinder can report stuff to nova, I'm fine with keeping the scheduler in a dark place | 15:00 |
| bauzas | mriedem: don't awake the best | 15:00 |
| bauzas | beast | 15:00 |
| *** haomaiwang has quit IRC | 15:01 | |
| mriedem | DuncanT: so i guess we need a new volume api to tell is if cinder is enforcing AZs | 15:01 |
| mriedem | or i guess what else do we need? if the az we have in nova is also available in cinder? seems we already have that | 15:01 |
| *** haomaiwang has joined #openstack-nova | 15:01 | |
| bauzas | so, ideally, my Christmas gift list would be : a Cinder ResourceTracker providing metrics to scheduler, and a scheduler able to claim on those metrics | 15:01 |
| *** dtantsur has left #openstack-nova | 15:01 | |
| DuncanT | mriedem: I was thinking of just overloading the cinder az list api... if that returns 'none' then cinder isn't enforcing AZs, otehrwise it is | 15:01 |
| kashyap | tonyb: Hmm, how would you guess the format of the "Existing image" (that's mentoined on line 80)? | 15:01 |
| *** annegentle has joined #openstack-nova | 15:01 | |
| mriedem | DuncanT: i was thinking along the same lines | 15:02 |
| bauzas | DuncanT: mriedem: who would call the Cinder API ? | 15:02 |
| DuncanT | mriedem: I'll make sure to lay that out in the spec | 15:02 |
| mriedem | DuncanT: well, none could be confusing | 15:02 |
| DuncanT | mriedem: Why? | 15:02 |
| mriedem | DuncanT: was thinking more like 'any' or 'n/a' or something | 15:02 |
| mriedem | b/c none could just mean there are no AZs defined yet right? | 15:02 |
| mriedem | or is the default 'cinder' az always returned | 15:02 |
| danpb | jaypipes: mriedem dansmith johnthetubaguy would appreciate feedback on this spec sometime in the next few weeks, in case it turns out to be something needing summit discussion https://review.openstack.org/#/c/195662/ | 15:02 |
| DuncanT | mriedem: Ok, we can make it very unambiguous, sure | 15:02 |
| tonyb | kashyap: sorry 'line 80'? | 15:02 |
| *** haomaiwang has quit IRC | 15:03 | |
| bauzas | mriedem: you're thinking of Nova calling the cinder API, right? | 15:03 |
| kashyap | tonyb: Line 80 in utils.py - the one we were referring to a few mins ago | 15:03 |
| dansmith | danpb: hmm, okay | 15:03 |
| mriedem | bauzas: yeah | 15:03 |
| bauzas | mriedem: if so, I guess the compute manager? | 15:03 |
| *** marzif_ has quit IRC | 15:03 | |
| johnthetubaguy | danpb: nod | 15:03 |
| mriedem | bauzas: no, api | 15:03 |
| kashyap | danpb: While you're here, how does one guess the format of the base image here -- https://github.com/openstack/nova/blob/master/nova/virt/libvirt/utils.py#L87 | 15:03 |
| bauzas | mriedem: mmm | 15:03 |
| mriedem | bauzas: in the case that the request is to boot an instance in a specific az | 15:04 |
| mriedem | we validate that az against cinder api | 15:04 |
| kashyap | danpb: It doesn't have -o backing_fmt=FOO mentioned there. | 15:04 |
| tonyb | kashyap: Well we can call qemu-img on it, to get the format | 15:04 |
| bauzas | mriedem: so the cinder AZ information would be attached as a new RequestSpec field or something like that ? | 15:04 |
| mriedem | DuncanT: if you're going to change the semantics of the cinder az list API, i guess microversion? | 15:04 |
| kashyap | tonyb: Yeah | 15:04 |
| *** whenry_ has joined #openstack-nova | 15:04 | |
| tonyb | kashyap: I understand there are some issues around that but it's all we have | 15:04 |
| *** fifieldt has quit IRC | 15:05 | |
| bauzas | mriedem: ideally, I'd like to see the scheduler responsible for raising the NoValidHost, not the compute manager, hence my question | 15:05 |
| johnthetubaguy | isn't it more, you boot from volume in an AZ, that means you need a volume in the same AZ? | 15:05 |
| danpb | kashyap: well it should be setting a backing_fmt parameter - in general it is insecure to not set it | 15:05 |
| mriedem | bauzas: i'm not sure why we'd have cinder az info in the request spec for the scheduler - nothing in the scheduler is consuming that right? | 15:05 |
| *** shoutm has quit IRC | 15:05 | |
| mriedem | johnthetubaguy: right | 15:05 |
| danpb | you'd need to look at the callers who hopefully know what the backing format ought to be | 15:05 |
| bauzas | mriedem: well, the scheduler is giving you a node | 15:05 |
| mriedem | we should be able to determine that in nova-api if we can get the proper info from cinder | 15:05 |
| kashyap | danpb: Indeed, you've mentioned it previously too on this -- https://bugs.launchpad.net/nova/+bug/1415087 | 15:05 |
| openstack | Launchpad bug 1415087 in OpenStack Compute (nova) "[OSSA 2015-011] Format-guessing and file disclosure in image convert (CVE-2015-1850, CVE-2015-1851)" [High,Triaged] - Assigned to Tony Breeds (o-tony) | 15:05 |
| tonyb | danpb: that's what we're talking about adding. | 15:05 |
| tonyb | danpb: as part of the bigger issue of that ^^ bug | 15:06 |
| bauzas | mriedem: I'm supposing that if you can't have a node that matches the Cinder AZ requested and the Nova AZ requested, then it's a novalidhost | 15:06 |
| DuncanT | mriedem: Hopefully microversions will land in time, yes | 15:06 |
| *** erhudy has joined #openstack-nova | 15:06 | |
| johnthetubaguy | so I could see the scheduler being told it needs a compute resources and cinder resources, so it makes sure it puts it somewhere thats possible | 15:06 |
| johnthetubaguy | just like, putting it somewhere you still have IPs left | 15:06 |
| bauzas | johnthetubaguy: I feel the same direction | 15:06 |
| *** marzif_ has joined #openstack-nova | 15:07 | |
| mriedem | somewhere you still have IPs left - for nova-network only? | 15:07 |
| danpb | kashyap: tonyb ok, well you'll just have to work backwards through the call path to identify places where the fix neds to be plumbed in | 15:07 |
| mriedem | or does that work for neutron? | 15:07 |
| kashyap | danpb: Hmm, my feeble brain guessed as much. | 15:07 |
| danpb | as create_cow() will have to be explicitly passed in the backing format | 15:07 |
| alaski | mriedem: there's a neutron spec under discussion for providing that info, then the scheduler could use it | 15:07 |
| tonyb | danpb: can't we just do http://paste.openstack.org/show/473900/? | 15:07 |
| mriedem | johnthetubaguy: bauzas: i'm failing to see how nova's scheduler can predict where cinder is going to place the volume when the compute code creates the volume | 15:07 |
| tonyb | danpb: to detect the format of the backing_file? | 15:08 |
| alaski | mriedem: however it does work for us internally (IP capacity scheduling w/ Neutron) | 15:08 |
| *** electrocucaracha has quit IRC | 15:08 | |
| mriedem | the nova scheduler will pick the compute node and on there we'll try creating the volume with a given az | 15:08 |
| bauzas | mriedem: IMHO that shouldn't be supported | 15:08 |
| DuncanT | mriedem: Surely you just tell cinder 'I need a volume in this AZ' and cinder either creates it or says no? | 15:08 |
| tonyb | danpb: or is that just papering over the whole format guessing issues | 15:08 |
| bauzas | mriedem: not saying removed from code, just considering that it could work - or not | 15:08 |
| mriedem | DuncanT: yeah | 15:08 |
| danpb | tonyb: well it depends on the context in which this code is called really | 15:09 |
| mriedem | DuncanT: but i'd like to know if that's going to work in the nova api layer, not hte compute | 15:09 |
| bauzas | mriedem: well, by thinking about that, we do the same with computes | 15:09 |
| bauzas | mriedem: using claims | 15:09 |
| *** annegentle has quit IRC | 15:09 | |
| bauzas | mriedem: we know that the compute manager can fail to reserve the resource | 15:09 |
| danpb | tonyb: running qemu-img info on the backing file, may, or may not be safe depending on use context | 15:09 |
| DuncanT | mriedem: so call the create in nova-api and the attach in compute? | 15:09 |
| mriedem | i guess i'm not seeing how this would work in the nova scheduler or RT | 15:09 |
| mriedem | DuncanT: no, just validate in nova-api | 15:09 |
| *** zhangjn_ has joined #openstack-nova | 15:09 | |
| *** annegentle has joined #openstack-nova | 15:09 | |
| bauzas | mriedem: so, given that you BFM, you only verify the Nova AZ | 15:09 |
| *** ZZelle has quit IRC | 15:09 | |
| bauzas | mriedem: if Cidner is failing, that's a retry | 15:10 |
| danpb | to be absolutely guaranteed safe we'd want the caller to have passed in backingformat, having got it from some trusted source | 15:10 |
| tonyb | danpb: okay I'll look again. | 15:10 |
| *** ZZelle has joined #openstack-nova | 15:10 | |
| bauzas | BFV btw. :) | 15:10 |
| johnthetubaguy | mriedem: so I was thinking of a picking a good spot for compute, picks the AZ get claim, then ask cinder for that AZ, if it fails, try next AZ, etc | 15:10 |
| bauzas | for the sake of clarity :) | 15:10 |
| *** yassine has joined #openstack-nova | 15:10 | |
| tonyb | danpb: assuming we can find a trusted source ;P | 15:10 |
| DuncanT | mriedem: I'd say just call create... that call does a quick sanity check (including the AZ) then returns anyway | 15:10 |
| bauzas | johnthetubaguy: yeah that | 15:10 |
| danpb | tonyb: yep, indeed | 15:10 |
| DuncanT | mriedem: The attach would be called in compute | 15:10 |
| johnthetubaguy | DuncanT: yeah, that would work (in my head) | 15:11 |
| *** nic1 has joined #openstack-nova | 15:11 | |
| danpb | tonyb: if we're lucky, that backing file will be one we just downloaded from glance, in which case glance will have told us the expected format | 15:11 |
| mriedem | DuncanT: the reason i don't like it from the nova api is that if we can't place the instance to build on a host, we have to rollback that volume we created, but that's not a huge deal i guess, the user would have to do that too probably | 15:11 |
| DuncanT | mriedem: Consider a create to be just a reservation that can fail later | 15:11 |
| bauzas | DuncanT: mriedem: question, what happens when an instance is not spawning but the volume is created ? | 15:11 |
| johnthetubaguy | DuncanT: assuming we get an error if that AZ is out of capacity, I guess | 15:11 |
| *** nic has quit IRC | 15:11 | |
| bauzas | DuncanT: mriedem: is this removing the volume ? | 15:11 |
| *** nic1 is now known as nic | 15:11 | |
| mriedem | bauzas: you mean boot from volume where source is volume (the user already created the volume)? | 15:12 |
| mriedem | we just try to attach it during spawn | 15:12 |
| bauzas | mriedem: no the nova creation thing | 15:12 |
| johnthetubaguy | ah, true, if you have a volume, you need the compute near it | 15:12 |
| DuncanT | bauzas: if nova created a volume, nova can delete it at the point it realises it has a problem. It's no different to cleaning up after some other failure | 15:12 |
| johnthetubaguy | works both ways | 15:12 |
| *** yassine has quit IRC | 15:12 | |
| DuncanT | If you've already got a volume, nova can ask cinder what AZ it is in and therefore know the limitations, no problem | 15:13 |
| bauzas | DuncanT: that I consider it the expected behaviour, I'm more interested in getting the actual implementation :) | 15:13 |
| *** zhangjn has quit IRC | 15:13 | |
| mriedem | if we're provided the volume, then yeah it's just the regular check attach case | 15:13 |
| mriedem | the implementation gets a bit gorpy | 15:13 |
| mriedem | i mean if we move the volume create from the compute to the api | 15:14 |
| kashyap | Hmm, the image handling code is a bit complex, isn't it. . . :P | 15:14 |
| DuncanT | bauzas: Well, check the spec (that will appear once I can get away from discussing it ;-) ) and then we can start on the implementation | 15:14 |
| bauzas | mriedem: that's my worries there | 15:14 |
| mriedem | ok, let's just all shut up about this and let DuncanT write the spec | 15:14 |
| DuncanT | mriedem: To a certain degree, cinder is going to be proposing the API it can best work with, and asking nova to use it - doing the reverse has proved a disaster so far | 15:15 |
| bauzas | mriedem: I know mgagne wouldn't be super happy, but I seriously think that creating a volume in Nova is just error-prone | 15:15 |
| ndipanov | erlon, hey | 15:15 |
| mriedem | bauzas: i agree | 15:15 |
| mriedem | most people agree i think | 15:15 |
| mriedem | except ops and users :) | 15:15 |
| mriedem | but what do they know?! | 15:15 |
| bauzas | heh, what ? | 15:15 |
| bauzas | you need to have users to write stuff ? | 15:16 |
| bauzas | that's insane | 15:16 |
| alaski | in my experience with users they don't care that Nova creates the volume, they care that it can be done in one step | 15:17 |
| bauzas | so, okay, we can wait for DuncanT's spec, but the more I think, the more I guess we need to provide the AZ information to the scheduler, not the other way | 15:17 |
| bauzas | alaski: heat stack-create ? | 15:17 |
| alaski | bauzas: something like that :) | 15:17 |
| kashyap | tonyb: The file_format there is what Glance provides us? | 15:17 |
| bauzas | -f mysupercloud.yaml | 15:17 |
| mriedem | stack@fixedips:~/devstack$ cinder availability-zone-list | 15:17 |
| mriedem | +------+-----------+ | 15:17 |
| mriedem | | Name | Status | | 15:17 |
| mriedem | +------+-----------+ | 15:17 |
| mriedem | | nova | available | | 15:17 |
| mriedem | +------+-----------+ | 15:17 |
| mriedem | i wasn't expecting that | 15:17 |
| tonyb | danpb, kashyap: okay the setting of backing_fmt is harder than I thought but I don't think it's impossible. | 15:17 |
| jaypipes | danpb: will do. sorry for delays. been in meeting hell for a couple weeks :( | 15:17 |
| bauzas | mriedem: uh funny | 15:18 |
| danpb | jaypipes: no prob, its not urgent - just wanted to make sure it was on peoples radar | 15:18 |
| bauzas | the default Cinder AZ is 'nova' | 15:18 |
| mriedem | jaypipes: i think we should have a meeting to talk about the number of meetings you're in | 15:18 |
| *** Venkatesh has quit IRC | 15:18 | |
| jaypipes | mriedem: :P | 15:18 |
| tonyb | danpb, kashyap: any thoughts on how to exploit the convert case to get an information disclousure? | 15:18 |
| *** marzif_ has quit IRC | 15:18 | |
| kashyap | tonyb: Yeah, highly context dependent, as we can see. | 15:18 |
| mriedem | bauzas: i guess that makes sense given cross_az_attach | 15:18 |
| bauzas | jaypipes: still not trained your pugs ? c'mon | 15:18 |
| kashyap | tonyb: Also, this is probably worth a write-up the upstream list | 15:19 |
| kashyap | Since other disk image experts (like pbrady) might also respond. | 15:19 |
| bauzas | mriedem: well, I could say something very opiniated about what I think of cross_az_attach, but that's not fair | 15:19 |
| danpb | tonyb: for the most part I think openstack is /not/ in fact vulnerable in practice | 15:20 |
| danpb | since it uses qcow2 images almost exclusively | 15:20 |
| *** otter768 has joined #openstack-nova | 15:20 | |
| *** artom has quit IRC | 15:20 | |
| bauzas | that's not like if French people would always argue | 15:20 |
| danpb | it only really becomes a problem if you give guests raw images | 15:20 |
| tonyb | danpb: right, which we give people a nice easy way to do ;P | 15:20 |
| *** armax has joined #openstack-nova | 15:21 | |
| tonyb | danpb: well operators but still .... | 15:21 |
| danpb | yeah, but i figure most people just have cow turned on all the time | 15:21 |
| danpb | as using raw just sucks down disk space | 15:21 |
| *** diogogmt has joined #openstack-nova | 15:21 | |
| *** artom has joined #openstack-nova | 15:21 | |
| kashyap | But Nova converts the base image (the one that lives in nova/instance/_base/) from qcow2 to raw, IIRC? | 15:23 |
| openstackgerrit | Chuck Carmack proposed openstack/nova: Disable security group tests in tempest-dsvm-cells-rc using a config option https://review.openstack.org/226043 | 15:23 |
| *** neiljerram_bb has joined #openstack-nova | 15:23 | |
| *** alexpilotti has quit IRC | 15:23 | |
| tonyb | danpb: so the dom xml would be a trusted source right? | 15:23 |
| *** otter768 has quit IRC | 15:24 | |
| tonyb | kashyap: I *think* that only happens if you ask for it not by default. | 15:24 |
| *** sahid has quit IRC | 15:25 | |
| danpb | kashyap: even that would be ok, as long as nova never then gave that raw image to a guest | 15:25 |
| kashyap | tonyb: Grepping in my n-cpu.log of my just setup DevStack instance: | 15:25 |
| kashyap | "qemu-img convert -O raw /home/stack/src/cloud/data/nova/instances/_base/7eb0bb2db79ba82008d71f9d6a295006c45db4ea.part /home/stack/src/cloud/data/nova/instances _base/7eb0bb2db79ba82008d71f9d6a295006c45db4ea.convert" | 15:25 |
| kashyap | danpb: Ah, okay | 15:26 |
| *** tonytan4ever has quit IRC | 15:26 | |
| *** kfox1111_away is now known as kfox1111 | 15:26 | |
| *** zhangjn_ has quit IRC | 15:26 | |
| *** zhangjn has joined #openstack-nova | 15:27 | |
| *** zhangjn has quit IRC | 15:27 | |
| *** zhangjn has joined #openstack-nova | 15:28 | |
| *** zhangjn has quit IRC | 15:28 | |
| *** electrocucaracha has joined #openstack-nova | 15:28 | |
| *** salv-orlando has quit IRC | 15:28 | |
| tonyb | It's 1:30am here and I'm not going to get this finished tonight. | 15:28 |
| *** zhangjn has joined #openstack-nova | 15:28 | |
| *** shoutm has joined #openstack-nova | 15:28 | |
| *** angdraug has joined #openstack-nova | 15:29 | |
| kashyap | tonyb: Get some rest! | 15:29 |
| tonyb | danpb, kashyap: Thanks for the help. Feel free to fix this while I'm sleeping ;D | 15:29 |
| bauzas | tonyb: consider that you're on the EU TZ, that's only 5.30pm | 15:29 |
| tonyb | bauzas: woot! | 15:29 |
| tonyb | bauzas: I feel fresh and ready for beer | 15:30 |
| bauzas | tonyb: zactly my point | 15:30 |
| bauzas | it's beer-o-clock | 15:30 |
| tonyb | bauzas: :) | 15:30 |
| *** tjones1 has joined #openstack-nova | 15:30 | |
| bauzas | tonyb: enough kidding, go to bed ! | 15:30 |
| *** tjones1 has left #openstack-nova | 15:30 | |
| tonyb | See some of you in 5ish hours for the meeting. | 15:30 |
| bauzas | you're mad | 15:31 |
| *** lsmola has quit IRC | 15:32 | |
| lxsli | johnthetubaguy: is it necessary to open a blueprint for a backlog spec? | 15:33 |
| johnthetubaguy | lxsli: its seems silly to enforce that I guess | 15:33 |
| johnthetubaguy | lxsli: not really thought about that till now | 15:34 |
| lxsli | ok, throwing up a clay pigeon | 15:34 |
| *** zhangjn_ has joined #openstack-nova | 15:35 | |
| openstackgerrit | Alexis Lee proposed openstack/nova-specs: Request logs on failure https://review.openstack.org/227379 | 15:36 |
| *** armax has quit IRC | 15:36 | |
| *** earlephilhower has joined #openstack-nova | 15:36 | |
| lxsli | bauzas: damnit it's only 4:30 here, you had me thinking of home | 15:37 |
| bauzas | lxsli: I said EU, since when UK is in EU ? :p | 15:37 |
| *** abalutoiu has joined #openstack-nova | 15:37 | |
| *** zhangjn has quit IRC | 15:37 | |
| lxsli | touche | 15:37 |
| *** hemnafk is now known as hemna | 15:37 | |
| *** julien_ZTE has quit IRC | 15:37 | |
| *** sahid has joined #openstack-nova | 15:39 | |
| *** krtaylor has quit IRC | 15:41 | |
| *** kodokuu has quit IRC | 15:44 | |
| *** Bent0 has joined #openstack-nova | 15:45 | |
| Bent0 | Anyone here who would give me some pointers on how to ake openstack talk to my 2 XenServer machines? | 15:45 |
| BobBall | Bent0: http://docs.openstack.org/juno/config-reference/content/introduction-to-xen.html | 15:46 |
| BobBall | Run nova-compute in a VM on each XenServer machine and configure it according to that doc | 15:46 |
| *** zhangjn has joined #openstack-nova | 15:46 | |
| Bent0 | So each xenserver host should run a compute VM? | 15:47 |
| dansmith | Bent0: also, please see topic | 15:47 |
| Bent0 | the main channel no one answers | 15:47 |
| dansmith | still | 15:47 |
| BobBall | yes Bent0 - needed because the compute VM needs to attach the disk (e.g. to resize it) | 15:47 |
| openstackgerrit | garyk proposed openstack/nova-specs: Propose VMware limits, reservation and shares https://review.openstack.org/227386 | 15:47 |
| *** zhangjn_ has quit IRC | 15:49 | |
| Bent0 | Ok. I thought the compute service on my controller would just talk to xen over ssh and XAPI | 15:49 |
| openstackgerrit | Moshe Levi proposed openstack/nova: Fix pci_stats logging in resource tracker https://review.openstack.org/227116 | 15:49 |
| *** burgerk has joined #openstack-nova | 15:51 | |
| *** obedmr has quit IRC | 15:52 | |
| openstackgerrit | Moshe Levi proposed openstack/nova: Fix pci_stats logging in resource tracker https://review.openstack.org/227116 | 15:53 |
| *** nic1 has joined #openstack-nova | 15:53 | |
| *** nic has quit IRC | 15:53 | |
| *** nic1 is now known as nic | 15:54 | |
| *** krtaylor has joined #openstack-nova | 15:54 | |
| *** obedmr has joined #openstack-nova | 15:54 | |
| *** diogogmt has quit IRC | 15:55 | |
| *** diogogmt has joined #openstack-nova | 15:56 | |
| *** kiran-r has quit IRC | 15:58 | |
| BobBall | Bent0: You need one compute service per hypervisor - same as you need for KVM | 15:59 |
| mriedem | new live migration bug https://bugs.launchpad.net/nova/+bug/1499405 | 16:00 |
| Bent0 | ok thanks. That wasn't very clear in the manual :P | 16:00 |
| openstack | Launchpad bug 1499405 in OpenStack Compute (nova) "Live-migration fails for instances with LVM ephemeral devices" [Undecided,New] | 16:00 |
| *** neiljerram_bb has quit IRC | 16:00 | |
| *** fawadkhaliq has quit IRC | 16:01 | |
| *** fawadkhaliq has joined #openstack-nova | 16:02 | |
| *** marzif has joined #openstack-nova | 16:03 | |
| *** fawadkhaliq has quit IRC | 16:05 | |
| *** jlanoux has quit IRC | 16:06 | |
| *** sahid has quit IRC | 16:06 | |
| ndipanov | an easy one for someone trying to get their openstack contributor game on https://bugs.launchpad.net/nova/+bug/1499449 | 16:06 |
| openstack | Launchpad bug 1499449 in OpenStack Compute (nova) "libvirt live-migration: Monitoring task does not track progress watermark correctly" [Low,New] | 16:06 |
| *** derekh has quit IRC | 16:09 | |
| *** pixelbeat has quit IRC | 16:09 | |
| *** pixelbeat has joined #openstack-nova | 16:11 | |
| *** ndipanov has quit IRC | 16:11 | |
| *** lucasagomes has quit IRC | 16:12 | |
| *** pumaranikar has joined #openstack-nova | 16:15 | |
| mriedem | ctrath: ^ | 16:18 |
| ctrath | mriedem: https://bugs.launchpad.net/nova/+bug/1499449? | 16:18 |
| openstack | Launchpad bug 1499449 in OpenStack Compute (nova) "libvirt live-migration: Monitoring task does not track progress watermark correctly" [Low,Confirmed] | 16:18 |
| mriedem | yeah | 16:18 |
| ctrath | thx.. I'll fix it | 16:19 |
| mriedem | markus_z: we should get an official nova tag for pci | 16:19 |
| *** emagana has quit IRC | 16:20 | |
| *** markvoelker_ has joined #openstack-nova | 16:21 | |
| *** yamahata has joined #openstack-nova | 16:21 | |
| *** gszasz has quit IRC | 16:21 | |
| *** nkrinner has quit IRC | 16:23 | |
| *** ociuhandu has quit IRC | 16:24 | |
| *** afazekas has quit IRC | 16:24 | |
| *** markvoelker has quit IRC | 16:24 | |
| *** oomichi has quit IRC | 16:24 | |
| *** gabriel-bezerra has quit IRC | 16:24 | |
| *** wolsen has quit IRC | 16:24 | |
| *** leifz has quit IRC | 16:24 | |
| *** oomichi_ has joined #openstack-nova | 16:24 | |
| *** x3k` has joined #openstack-nova | 16:24 | |
| *** x3k has quit IRC | 16:24 | |
| *** kbyrne has quit IRC | 16:24 | |
| *** adreznec has quit IRC | 16:24 | |
| *** alaski has quit IRC | 16:24 | |
| *** leeantho has joined #openstack-nova | 16:24 | |
| *** leeantho has quit IRC | 16:24 | |
| *** afranc202 has quit IRC | 16:24 | |
| *** iraw- has joined #openstack-nova | 16:24 | |
| *** afazekas has joined #openstack-nova | 16:24 | |
| johnthetubaguy | markus_z: yeah, do let us know you need to add offical tags, I think mriedem has permissions for that | 16:24 |
| *** alaski_ has joined #openstack-nova | 16:24 | |
| *** wolsen has joined #openstack-nova | 16:24 | |
| *** unicell has joined #openstack-nova | 16:24 | |
| *** kbyrne has joined #openstack-nova | 16:24 | |
| *** leifz has joined #openstack-nova | 16:24 | |
| *** gabriel-bezerra has joined #openstack-nova | 16:25 | |
| markus_z | mriedem: sure, I'll add it | 16:25 |
| markus_z | mriedem: The networking subteam will have a look at it? | 16:25 |
| *** adreznec has joined #openstack-nova | 16:25 | |
| openstackgerrit | Mark Doffman proposed openstack/nova: libvirt: Don't attempt to convert initrd images. https://review.openstack.org/225289 | 16:26 |
| markus_z | johnthetubaguy: I have permissions too. | 16:26 |
| johnthetubaguy | markus_z: ah, cool | 16:26 |
| *** dustins is now known as dustins|afk | 16:26 | |
| *** pushkaru has joined #openstack-nova | 16:27 | |
| *** pumarani__ has joined #openstack-nova | 16:27 | |
| markus_z | mriedem: We have 6 other bugs tagged with "pci-passthrough". Switch them to "pci"? https://bugs.launchpad.net/nova/+bugs/?field.tag=pci-passthrough | 16:27 |
| *** pushkaru has quit IRC | 16:27 | |
| kashyap | mriedem: I'll test that thing you asked on the live mig bug, I'm preparing a local setup (hopefully it works). | 16:28 |
| *** pumarani__ has quit IRC | 16:28 | |
| *** VW has joined #openstack-nova | 16:30 | |
| kashyap | Does anyone recall why live mig fails with 'force_config_drive = False'? | 16:31 |
| *** burt has joined #openstack-nova | 16:31 | |
| *** ociuhandu has joined #openstack-nova | 16:31 | |
| kashyap | In my local DevStack setup (w/o LVM) I have to disable config_drive to be able to successfully do live block migration. | 16:31 |
| johnthetubaguy | kashyap: isn't that a bug about qemu and read only devices, I can't remember the details now | 16:32 |
| *** danpb has quit IRC | 16:32 | |
| *** abalutoiu has quit IRC | 16:32 | |
| kashyap | johnthetubaguy: Oh, there's actually selective block device migration feature that was recently added to libvirt | 16:32 |
| kashyap | That should fix this I guess. johnthetubaguy Yes, there's a bug about this which I can't immediately find | 16:33 |
| kashyap | johnthetubaguy: This one - https://bugs.launchpad.net/nova/+bug/1246201 | 16:34 |
| openstack | Launchpad bug 1246201 in OpenStack Compute (nova) "Live migration fails when the instance has a config_drive_format=iso9660" [High,In progress] - Assigned to Pavel Boldin (pboldin) | 16:34 |
| *** pratikmallya has joined #openstack-nova | 16:34 | |
| mriedem | markus_z: pci is simpler | 16:34 |
| mriedem | so i'd just use that | 16:34 |
| *** EinstCrazy has quit IRC | 16:34 | |
| pratikmallya | hello, what is the difference between an interface and a virtual interface on a nova server? | 16:34 |
| *** unicell has quit IRC | 16:35 | |
| markus_z | mriedem: ack | 16:35 |
| johnthetubaguy | kashyap: yeah, I that sounds right, I remember danpb saying there was something coming to fix that | 16:35 |
| mriedem | doesn't the code check for vfat now or something? | 16:36 |
| * mriedem looks | 16:36 | |
| mriedem | yeah here https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L6163 | 16:36 |
| kashyap | johnthetubaguy: Yep, this http://libvirt.org/git/?p=libvirt.git;a=commit;h=93a19e283e5a147d147d84383aaaa8be63b6a68d | 16:37 |
| markus_z | baoli: It seems that mostly you used the "pci-passthrough" tag. We will use the official tag "pci" from now on which includes pci-passthrough. | 16:37 |
| *** julien_ZTE has joined #openstack-nova | 16:38 | |
| markus_z | mriedem: baoli: 9 pci bugs https://bugs.launchpad.net/nova/+bugs?field.tag=pci | 16:39 |
| markus_z | mriedem: I'll keep an eye open in the future. | 16:39 |
| mriedem | thanks | 16:39 |
| *** julien_ZTE has quit IRC | 16:40 | |
| *** karimb has quit IRC | 16:40 | |
| *** armax has joined #openstack-nova | 16:41 | |
| *** nic has quit IRC | 16:42 | |
| *** mdrabe has joined #openstack-nova | 16:43 | |
| mriedem | garyk: is there any reason why we couldn't move nova.mks to nova.console.mks? | 16:43 |
| *** raildo is now known as raildo-afk | 16:45 | |
| *** eglynn has quit IRC | 16:46 | |
| *** nic has joined #openstack-nova | 16:46 | |
| garyk | mriedem: i think it it should be possible. let me ask rado about it and get back to you | 16:46 |
| mriedem | jcook: i think this is hitting some rax forks for quota classes https://bugs.launchpad.net/nova/+bug/1495755 | 16:47 |
| openstack | Launchpad bug 1495755 in OpenStack Compute (nova) "test_show_policy_failed fails (depending on another test to create db?)" [Undecided,Invalid] | 16:47 |
| *** alex_klimov has quit IRC | 16:47 | |
| *** raildo-afk is now known as raildo | 16:48 | |
| pratikmallya | hello, what is the difference between an interface and a virtual interface on a nova server? | 16:48 |
| *** browne1 has quit IRC | 16:49 | |
| mriedem | pratikmallya: the virtual interface in the data model holds the mac address | 16:50 |
| mriedem | created here for nova-network https://github.com/openstack/nova/blob/master/nova/network/manager.py#L750 | 16:50 |
| mriedem | 'interface' is generic so depends on context | 16:50 |
| mriedem | the attach interface API deals with ports | 16:51 |
| mriedem | https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/attach_interfaces.py | 16:51 |
| mriedem | i.e. fixed ips | 16:51 |
| mriedem | https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L4870 | 16:52 |
| *** atuvenie has quit IRC | 16:52 | |
| *** VW has quit IRC | 16:53 | |
| mriedem | pratikmallya: also check out the network model here https://github.com/openstack/nova/blob/master/nova/network/model.py | 16:53 |
| *** VW has joined #openstack-nova | 16:54 | |
| ccarmack | mriedem: question about the setting the Description field of the instance on nova-boot. Would there also need to be a CLI change to edit the field later? | 16:55 |
| pratikmallya | ahh, thanks mriedem | 16:56 |
| mriedem | ccarmack: yeah i think whatever applies to display_name (name in the api) would also apply to description | 16:56 |
| mriedem | as in specify on create and be able to update | 16:57 |
| ccarmack | mriedem: I couldn't find a nova CLI that updates server properties | 16:57 |
| mriedem | i guess there isn't one | 16:58 |
| ccarmack | mriedem: I do see an update server API | 16:59 |
| *** su_zhang has joined #openstack-nova | 16:59 | |
| mriedem | ccarmack: yes, that's all that really matters | 16:59 |
| mriedem | https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/servers.py#L772 | 16:59 |
| mriedem | people don't have to use the CLI | 16:59 |
| *** zhangjn_ has joined #openstack-nova | 16:59 | |
| ccarmack | mriedem: ok | 16:59 |
| *** zhangjn has quit IRC | 16:59 | |
| *** VW has quit IRC | 17:00 | |
| *** achanda has joined #openstack-nova | 17:00 | |
| *** VW has joined #openstack-nova | 17:00 | |
| *** VW_ has joined #openstack-nova | 17:01 | |
| *** ctrath has quit IRC | 17:01 | |
| mriedem | alaski_: this looks super fun https://bugs.launchpad.net/nova/+bug/1497076 | 17:02 |
| openstack | Launchpad bug 1497076 in OpenStack Compute (nova) "Unable to delete an instance due to foreign key violation" [Undecided,New] | 17:02 |
| *** ccarmack has quit IRC | 17:02 | |
| mriedem | partial upgraded cells with busted db's and foreign key constraints | 17:02 |
| *** electrocucaracha has quit IRC | 17:03 | |
| *** markus_z has quit IRC | 17:03 | |
| *** mriedem has quit IRC | 17:03 | |
| *** electrocucaracha has joined #openstack-nova | 17:04 | |
| *** VW has quit IRC | 17:04 | |
| *** marzif has quit IRC | 17:05 | |
| *** su_zhang has quit IRC | 17:05 | |
| *** marzif has joined #openstack-nova | 17:05 | |
| *** su_zhang has joined #openstack-nova | 17:05 | |
| *** eglynn has joined #openstack-nova | 17:08 | |
| *** su_zhang has quit IRC | 17:08 | |
| *** su_zhang has joined #openstack-nova | 17:08 | |
| *** zhangjn has joined #openstack-nova | 17:08 | |
| *** zhangjn_ has quit IRC | 17:09 | |
| *** su_zhang has quit IRC | 17:09 | |
| *** su_zhang has joined #openstack-nova | 17:09 | |
| *** ccarmack has joined #openstack-nova | 17:10 | |
| *** ctrath has joined #openstack-nova | 17:10 | |
| *** unicell has joined #openstack-nova | 17:10 | |
| *** penick has joined #openstack-nova | 17:11 | |
| *** shoutm has quit IRC | 17:11 | |
| *** mriedem has joined #openstack-nova | 17:13 | |
| alaski_ | ugh, looking now | 17:14 |
| openstackgerrit | Claudiu Belu proposed openstack/nova: WIP: Hyper-V: determine vm_com_pipe error during spawn https://review.openstack.org/213720 | 17:15 |
| alaski_ | hah. "Steps to reproduce: ... 3.2: Use it for years" | 17:15 |
| penick | Nice | 17:16 |
| mriedem | alaski_: yeah, i laughed at that | 17:16 |
| melwitt | mriedem, alaski_: I had this patch to handle the FK constraint thing https://review.openstack.org/#/c/188261/ but it went into liberty. when an instance exists in API DB but not a cell DB, during a broadcast "delete everywhere" | 17:17 |
| mriedem | honestly i haven't gotten through comment 2 on that bug | 17:18 |
| dansmith | alaski_: just say "cool, I've started working on this bug, I'll let you know in years" | 17:18 |
| melwitt | I saw it happen when I did a delete request quickly after a create, before the instance had been scheduled | 17:18 |
| alaski_ | dansmith: yeah. "allowing time to pass per reproduce steps, please hang on" | 17:20 |
| *** zxyf has joined #openstack-nova | 17:20 | |
| *** otter768 has joined #openstack-nova | 17:21 | |
| * dansmith nods | 17:21 | |
| openstackgerrit | Stephen Gordon proposed openstack/nova-specs: Add thread policy note to virt-driver-cpu-pinning https://review.openstack.org/227428 | 17:21 |
| dansmith | or "following instructions, stuck on step 3.2" | 17:21 |
| mriedem | so the instance is deleted and gone from the instances table in the child cell, but still showing up in the parent cell | 17:21 |
| mriedem | but still in the instance_mappings table in both | 17:21 |
| openstackgerrit | Stephen Gordon proposed openstack/nova-specs: Add thread policy note to virt-driver-cpu-pinning https://review.openstack.org/227428 | 17:22 |
| mriedem | *instance_id_mappings | 17:22 |
| dansmith | alaski_: being serious, isn't that bug talking about running multiple versions of things in cells? | 17:22 |
| alaski_ | I'd like to confirm if it's the thing melwitt fixed, but I'm not seeing service logs anywhere | 17:22 |
| mriedem | alaski_: yeah, he just posted console output for novaclient and db query dumps | 17:23 |
| *** zhangjn has quit IRC | 17:23 | |
| alaski_ | dansmith: it says "partial upgrade" but it's not clear what's upgraded and what's not | 17:23 |
| mriedem | dansmith: he said he was on a partial upgrade from juno to kilo | 17:23 |
| mriedem | yeah | 17:23 |
| dansmith | right | 17:23 |
| melwitt | alaski_: there's an excerpt of a cells log as a attachment on comment #2 | 17:23 |
| alaski_ | melwitt: good eye. missed the attachment | 17:24 |
| *** raildo is now known as raildo-afk | 17:24 | |
| melwitt | I'm thinking about the instance_id_mappings and whether that would be different than the "delete unscheduled instance" scenario | 17:24 |
| melwitt | that is, if that means the instance used to exist in the cell DB and got deleted from there and not the top at some point | 17:25 |
| *** otter768 has quit IRC | 17:25 | |
| *** pratikmallya has quit IRC | 17:26 | |
| mriedem | hrm, instance_id_mappings doesn't have a fkey to anything else | 17:26 |
| mriedem | the failure is between instances and instance_extra https://launchpadlibrarian.net/218238506/nova-cells.log | 17:27 |
| alaski_ | oh man, I just realized that means they're using the ec2 api as well | 17:27 |
| *** njirap has quit IRC | 17:27 | |
| mriedem | yup | 17:27 |
| alaski_ | going for the trifecta of unsupported configs | 17:27 |
| mriedem | ha | 17:27 |
| melwitt | I think delete from cell but not top could happen actually by the "instance can be undeleted at the top" bug, where a "save state" from the cell arrives at the top after the destroy goes to the top | 17:27 |
| *** salv-orlando has joined #openstack-nova | 17:29 | |
| *** su_zhang has quit IRC | 17:29 | |
| melwitt | mriedem: that's at least consistent with https://bugs.launchpad.net/nova/+bug/1462128 | 17:30 |
| openstack | Launchpad bug 1462128 in OpenStack Compute (nova) "Cells: DBReferenceError possible deleting unscheduled instance" [Low,Fix released] - Assigned to melanie witt (melwitt) | 17:30 |
| *** ijw has joined #openstack-nova | 17:32 | |
| alaski_ | melwitt: yeah, it's looking like a similar issue. though how they got there is unclear | 17:33 |
| *** exploreshaifali has joined #openstack-nova | 17:33 | |
| *** tongli has quit IRC | 17:34 | |
| *** raildo-afk is now known as raildo | 17:34 | |
| mriedem | years of use | 17:35 |
| mriedem | what's not clear about that | 17:35 |
| *** tongli has joined #openstack-nova | 17:35 | |
| mriedem | melwitt: maybe it's worth backporting https://review.openstack.org/#/c/188261/ to stable/kilo anyway | 17:35 |
| mriedem | since they are moving to kilo | 17:36 |
| *** achanda has quit IRC | 17:36 | |
| melwitt | alaski_: yeah. both of the ways I know it can happen involve deleting an unscheduled instance. which I think would be sort of rare, so I wonder how many undeletable instances the reporter has | 17:39 |
| *** jasondotstar has joined #openstack-nova | 17:39 | |
| *** tongli has quit IRC | 17:39 | |
| *** pumaranikar has quit IRC | 17:39 | |
| *** eharney has quit IRC | 17:39 | |
| alaski_ | melwitt: we have found that it becomes more common when capacity is low | 17:40 |
| *** browne has joined #openstack-nova | 17:40 | |
| alaski_ | in conjunction with bursts of builds, so the cells state tracking doesn't keep up for the cells scheduler | 17:41 |
| jasondotstar | hi all - qq re: the openstack-ios project: https://launchpad.net/openstack-ios - not sure where to ask abt it but I figured here might be a place to start. | 17:41 |
| jasondotstar | is this essentially dead? or is has this worked picked up somewhere else? | 17:41 |
| melwitt | mriedem: yeah, I was thinking that. it's a simple backport and would help people clean up instances | 17:41 |
| jasondotstar | s/is// | 17:41 |
| *** vilobhmm11 has joined #openstack-nova | 17:41 | |
| alaski_ | melwitt: mriedem +1. that seems to be all the bug is asking for. not a way to avoid the situation just a way to get out of it | 17:42 |
| *** tonytan4ever has joined #openstack-nova | 17:42 | |
| *** ociuhandu has quit IRC | 17:43 | |
| *** tjones1 has joined #openstack-nova | 17:43 | |
| *** eharney has joined #openstack-nova | 17:43 | |
| openstackgerrit | Mathieu Gagné proposed openstack/nova: Fix evacuate support with Nova cells v1 https://review.openstack.org/227440 | 17:44 |
| *** zxyf has quit IRC | 17:44 | |
| *** jichen has joined #openstack-nova | 17:44 | |
| *** tongli has joined #openstack-nova | 17:46 | |
| *** tongli has quit IRC | 17:49 | |
| *** pratikmallya has joined #openstack-nova | 17:50 | |
| *** Marga_ has quit IRC | 17:50 | |
| *** achanda has joined #openstack-nova | 17:51 | |
| mriedem | jasondotstar: never heard of it, i guess ask the maintainer https://launchpad.net/~mike-mayo | 17:51 |
| jasondotstar | mriedem: i did. no response. | 17:51 |
| jasondotstar | mriedem: well, not yet anyway.... | 17:51 |
| *** pratikma_ has joined #openstack-nova | 17:51 | |
| mriedem | then it's probably dead | 17:52 |
| mriedem | nova activity in 5 years | 17:52 |
| mriedem | *no | 17:52 |
| jasondotstar | mriedem: right.... | 17:52 |
| jasondotstar | mriedem: I was thinking the same thing... | 17:52 |
| *** aix has quit IRC | 17:52 | |
| jasondotstar | maybe the horizon folks have insight on any similar projects/initiatives ? | 17:53 |
| jasondotstar | I can pop over and ask.... | 17:53 |
| *** pratikmallya has quit IRC | 17:54 | |
| *** leseb has quit IRC | 17:54 | |
| *** zxyf has joined #openstack-nova | 17:56 | |
| *** dustins|afk is now known as dustins | 17:56 | |
| *** su_zhang has joined #openstack-nova | 17:56 | |
| *** hitalia has joined #openstack-nova | 17:57 | |
| *** annegentle has quit IRC | 17:57 | |
| *** ctrath has quit IRC | 17:59 | |
| *** pumaranikar has joined #openstack-nova | 18:00 | |
| *** pixelbeat has quit IRC | 18:00 | |
| *** electrocucaracha has quit IRC | 18:01 | |
| *** ctrath has joined #openstack-nova | 18:03 | |
| doffm | ndipanov been around recently? | 18:04 |
| *** pumaranikar has quit IRC | 18:05 | |
| *** ociuhandu has joined #openstack-nova | 18:06 | |
| mriedem | bauzas: -2 on this now https://review.openstack.org/#/c/225119/ | 18:07 |
| mriedem | now that i understand it better | 18:07 |
| *** Piet has joined #openstack-nova | 18:07 | |
| bauzas | mriedem: hell yeah | 18:07 |
| bauzas | mriedem: IMHO we need to pause and wait for DuncanT spec | 18:07 |
| mriedem | yeah, i still plan on updating the help text for the cross_az_attach option | 18:08 |
| mriedem | to at least point out that other option in ciner | 18:08 |
| bauzas | mriedem: to see how we can have Nova AZ having loose relationship with Cinder ones | 18:08 |
| mriedem | *cinder | 18:08 |
| mriedem | loose relationship huh? | 18:08 |
| mriedem | like, no commitments? | 18:08 |
| mriedem | we can see other people? | 18:08 |
| mriedem | s/people/block storage services/ | 18:08 |
| bauzas | mriedem: I mean, just saying no real relationship, just a lazy one | 18:09 |
| mriedem | ha | 18:09 |
| mriedem | laissez faire | 18:09 |
| bauzas | I'm just replying to you, I should s/loose/lazy | 18:09 |
| mriedem | i know, i'm joking | 18:10 |
| *** rlrossit has quit IRC | 18:10 | |
| bauzas | about mgagne's point re: leaving the cinder creation done by nova, I think there is a little misunderstanding | 18:10 |
| *** penick has quit IRC | 18:10 | |
| bauzas | if that's all about having one way for operators to get only one call, then providing the orchestration by a contrib CLI module should be fine | 18:11 |
| bauzas | and we could then deprecate the API | 18:11 |
| *** penick has joined #openstack-nova | 18:11 | |
| *** su_zhang has quit IRC | 18:12 | |
| *** ctrath has quit IRC | 18:13 | |
| *** tonytan4ever has quit IRC | 18:15 | |
| *** salv-orlando has quit IRC | 18:15 | |
| mriedem | yeah, it's an idea | 18:16 |
| dansmith | heh | 18:16 |
| mriedem | that's about as much energy as i have left on deprecating boot from volume | 18:17 |
| *** rlrossit has joined #openstack-nova | 18:18 | |
| *** zxyf has quit IRC | 18:19 | |
| bauzas | mriedem: I'm still feeling young | 18:19 |
| mriedem | i guess | 18:19 |
| *** tonytan4ever has joined #openstack-nova | 18:20 | |
| *** Apoorva has joined #openstack-nova | 18:20 | |
| mriedem | melwitt: i'm backporting that cells fix of yours, there was a conflict in the tests | 18:20 |
| *** ZZelle_ has joined #openstack-nova | 18:21 | |
| *** ctrath has joined #openstack-nova | 18:22 | |
| melwitt | mriedem: okay, cool (that you're backporting it, not that there was a conflict) | 18:24 |
| *** eharney_ has joined #openstack-nova | 18:24 | |
| *** pumaranikar has joined #openstack-nova | 18:25 | |
| *** nelsnelson has quit IRC | 18:25 | |
| *** eharney has quit IRC | 18:26 | |
| *** su_zhang has joined #openstack-nova | 18:29 | |
| *** achanda has quit IRC | 18:29 | |
| bauzas | does anyone know if we have a new tag for asking for a RC2 bug ? | 18:30 |
| *** nelsnelson has joined #openstack-nova | 18:30 | |
| bauzas | or stll liberty-rc-potential ? | 18:30 |
| *** Marga_ has joined #openstack-nova | 18:31 | |
| *** achanda has joined #openstack-nova | 18:31 | |
| *** nelsnelson has quit IRC | 18:32 | |
| *** nelsnelson has joined #openstack-nova | 18:32 | |
| *** r-daneel has joined #openstack-nova | 18:37 | |
| openstackgerrit | Andreas Jaeger proposed openstack/nova: Cleanup of Translations https://review.openstack.org/227461 | 18:37 |
| *** julien_ZTE has joined #openstack-nova | 18:39 | |
| *** rook has quit IRC | 18:40 | |
| *** Marga_ has quit IRC | 18:40 | |
| *** eharney_ is now known as eharney | 18:40 | |
| *** Marga_ has joined #openstack-nova | 18:40 | |
| *** Apoorva has quit IRC | 18:42 | |
| *** Apoorva_ has joined #openstack-nova | 18:42 | |
| *** su_zhang has quit IRC | 18:43 | |
| *** julien_ZTE has quit IRC | 18:44 | |
| *** zxyf has joined #openstack-nova | 18:44 | |
| *** Apoorva_ has quit IRC | 18:47 | |
| *** Apoorva has joined #openstack-nova | 18:47 | |
| mriedem | bauzas: still liberty-rc-potential | 18:47 |
| bauzas | ack | 18:48 |
| *** pumaranikar has quit IRC | 18:49 | |
| openstackgerrit | Tomoki Sekiyama proposed openstack/nova: Fix multipath iSCSI encrypted volume attach failure https://review.openstack.org/196482 | 18:49 |
| *** pumaranikar has joined #openstack-nova | 18:50 | |
| *** hemna is now known as hemnalunch | 18:53 | |
| *** thumpba has quit IRC | 18:57 | |
| *** aix has joined #openstack-nova | 19:01 | |
| *** ijw has quit IRC | 19:01 | |
| *** ctrath has quit IRC | 19:01 | |
| *** electrocucaracha has joined #openstack-nova | 19:02 | |
| *** thangp has joined #openstack-nova | 19:04 | |
| openstackgerrit | Mark Doffman proposed openstack/nova: libvirt: Don't attempt to convert initrd images. https://review.openstack.org/225289 | 19:04 |
| DuncanT | bauzas: mriedem: johnthetubaguy: Problem spec done. https://review.openstack.org/227478 - I'll try to work on the proposed solution section tomorrow | 19:05 |
| mriedem | DuncanT: ok, thanks, i'm just updating the help text on the cross_az_attach option | 19:06 |
| mriedem | andreykurilin: if you're around could you discuss https://bugs.launchpad.net/nova/+bug/1380780 ? | 19:06 |
| openstack | Launchpad bug 1380780 in OpenStack Compute (nova) "Boot from image and create a new volume ignores availability zone" [Low,Fix released] - Assigned to Andrey Kurilin (andreykurilin) | 19:06 |
| DuncanT | mriedem: I'll link it to the mailing list thread once I've got a proposed solution or two done | 19:06 |
| DuncanT | mriedem: REST, doc impact etc will be TBD until after the summit I expect | 19:07 |
| *** electrocucaracha has quit IRC | 19:07 | |
| *** ctrath has joined #openstack-nova | 19:08 | |
| bauzas | DuncanT: ack | 19:09 |
| *** nickchase has joined #openstack-nova | 19:09 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Add more help text to the cinder.cross_az_attach option https://review.openstack.org/227483 | 19:11 |
| mriedem | bauzas: ^ | 19:11 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Add more help text to the cinder.cross_az_attach option https://review.openstack.org/227483 | 19:13 |
| mriedem | i'm going to follow that with a change that only passes the az to volume create if cinder.cross_az_attach is False | 19:13 |
| mriedem | b/c if it's true we don't care what az the cinder volume is in | 19:14 |
| *** zxyf has quit IRC | 19:14 | |
| mriedem | which i think was a mistake made in https://review.openstack.org/#/c/157041/ | 19:14 |
| mgagne | mriedem: +1 on that logic | 19:14 |
| mriedem | mgagne: hi! | 19:14 |
| mriedem | o/ | 19:14 |
| mriedem | mgagne: thanks for the operator POV in that thread | 19:14 |
| mgagne | mriedem: hi there (context switching) | 19:14 |
| *** krtaylor has quit IRC | 19:15 | |
| *** pixelbeat has joined #openstack-nova | 19:15 | |
| mgagne | mriedem: np. I didn't get the chance to contribute much those last months but will make sure to make it happen more, lists and gerrit as well. | 19:15 |
| bauzas | mgagne: heya | 19:18 |
| bauzas | mgagne: given your point of having one call for operators creating both the instance and the volume, what do you think about leaving the nova CLI wrapper doing that for you, like I said ? | 19:19 |
| bauzas | mriedem: is the revert change merged ? | 19:19 |
| mriedem | bauzas: yeah | 19:19 |
| bauzas | okay | 19:20 |
| mgagne | bauzas: I have yet to reply to this proposition. My opinion is that the client shouldn't have to orchestrate that as this means he will be "stuck" booting his instance for several minutes. | 19:20 |
| bauzas | mgagne: not sure I get your point | 19:20 |
| bauzas | mgagne: you're afraid of making 2 REST calls instead of one ? | 19:20 |
| mgagne | bauzas: lets say you issue a nova boot. Does it mean the command will wait until all steps are completed? | 19:21 |
| mgagne | bauzas: you have to understand that it's not just me, I can take care of myself. It's about the UX we offer to our end users. | 19:21 |
| bauzas | mgagne: by design, a REST call is non-blocking | 19:21 |
| bauzas | mgagne: it's a shoot and return | 19:21 |
| *** otter768 has joined #openstack-nova | 19:22 | |
| mgagne | bauzas: lets say I wish to boot from volume. What would be the steps from an end user's perspective? | 19:22 |
| bauzas | mgagne: see an example https://github.com/openstack/python-novaclient/blob/master/novaclient/v2/contrib/host_evacuate.py | 19:22 |
| *** VW_ has quit IRC | 19:23 | |
| *** pixelbeat has quit IRC | 19:23 | |
| mgagne | bauzas: I didn't read backlog so excuse me if I missed something, I know my name got mentioned several times. | 19:23 |
| *** VW has joined #openstack-nova | 19:23 | |
| bauzas | mgagne: sure, I got you're only a messenger of operators's voice :) | 19:23 |
| bauzas | mgagne: and I appreciate your thoughts | 19:24 |
| mgagne | bauzas: I sure am a messenger of my voice in the first place and happen to agree with my opinion :D | 19:24 |
| bauzas | I sometimes disagree with myself :) | 19:24 |
| *** thangp has quit IRC | 19:25 | |
| mgagne | schizophrenia not included =) | 19:25 |
| *** electrocucaracha has joined #openstack-nova | 19:25 | |
| *** VW has quit IRC | 19:25 | |
| *** VW_ has joined #openstack-nova | 19:25 | |
| openstackgerrit | Cale Rath proposed openstack/nova: libvirt live-migration: Monitoring task does not track progress watermark correctly https://review.openstack.org/227501 | 19:25 |
| *** otter768 has quit IRC | 19:26 | |
| mgagne | bauzas: just to clarify, when I referred to boot from volume, I'm referring to boot from volume where source=image. | 19:26 |
| *** krtaylor has joined #openstack-nova | 19:27 | |
| mgagne | bauzas: this has been our biggest selling point and best feature for 3 years now | 19:28 |
| bauzas | mgagne: okay I have to bribe you then | 19:28 |
| bauzas | anyway, need to leave | 19:28 |
| bauzas | back in a few | 19:28 |
| mgagne | bauzas: sure | 19:28 |
| *** lpetrut has quit IRC | 19:29 | |
| *** penick has quit IRC | 19:30 | |
| *** penick has joined #openstack-nova | 19:33 | |
| mriedem | you'd just use --poll when bootting from volume with nova | 19:34 |
| *** nic has quit IRC | 19:34 | |
| mriedem | --poll would include creating the volume and booting the instance with that volume | 19:34 |
| *** bnemec has quit IRC | 19:34 | |
| mriedem | it's essentially what tempest does | 19:34 |
| mriedem | create volume, wait for available, boot instance from volume, wait for active | 19:35 |
| mgagne | mriedem: yea, that's not something we would like to do. | 19:35 |
| *** ctrath has quit IRC | 19:37 | |
| *** pixelbeat has joined #openstack-nova | 19:37 | |
| mriedem | well, i meant that logic could be in novaclient | 19:37 |
| mriedem | or osc | 19:37 |
| mriedem | or whatever | 19:37 |
| mriedem | if you're not using those, then it falls down | 19:38 |
| mgagne | mriedem: the user shouldn't have to orchestrate those tasks. You ask your cloud provider to boot an image on a volume and it should take care of it, not you. It's such a fundamental block that the user shouldn't have to care about it. Other providers are supporting this feature just fine and we should be able to do it too. I know we can make it work. | 19:39 |
| alaski_ | mgagne: my contention with that is that yes we can make this work. But I don't think it should be in Nova | 19:40 |
| alaski_ | users want to make one call. but I don't think they care if that call is to Nova | 19:41 |
| mgagne | alaski_: nova boot is the main entry point. I don't see why people keep suggesting to write Python scripts or use Heat. You don't have to use cloudformation to boot a single instance on a persistent volume. Why should it be the case with OpenStack? It's just bad UX. | 19:41 |
| *** ctrath has joined #openstack-nova | 19:41 | |
| alaski_ | well, I don't really think Heat is the answer either. but something like that | 19:42 |
| alaski_ | but nova boot is the entry point to booting a VM | 19:42 |
| alaski_ | not necessarily to volume or network creation | 19:42 |
| mgagne | alaski_: Not for network but volume, yes. | 19:42 |
| mgagne | alaski_: network, partially true with port creation. | 19:43 |
| alaski_ | then why do we have a cinder API? | 19:43 |
| mgagne | alaski_: why does Amazon still have EBS API? | 19:44 |
| alaski_ | I'm not familiar with what Amazon does | 19:44 |
| jgriffith | mgagne: FWIW, I tend to agree with you; even though I don't like the functionality the way it is your points are quite valid | 19:45 |
| mgagne | I think we should be familiar with what cloud providers and solutions offer. They happen to have a (somewhat) good UX. | 19:45 |
| alaski_ | here's what it comes down to for me. I think we should have a an infrastructure API for compute resources, and then an API for simple user UX and orchestration. I feel that Nova is the former and other think it's the latter | 19:45 |
| openstackgerrit | Cale Rath proposed openstack/nova: libvirt live-migration: Monitoring task tracks progress watermark incorrectly https://review.openstack.org/227501 | 19:46 |
| *** matrohon has joined #openstack-nova | 19:46 | |
| alaski_ | either way, that's something we should agree to so we can design the right things | 19:46 |
| *** emagana has joined #openstack-nova | 19:46 | |
| alaski_ | if the current Nova api is going to be for simple UX and orchestration then my feeling is that we need a split somewhere in the project so we can create a focused compute api | 19:47 |
| alaski_ | but currently it seems that we're trying to do both and it's messy | 19:48 |
| mgagne | alaski_: if that's the case, I feel OpenStack is a hundred years behind alternative solutions as we keep deprecating and moving stuff around "to the new great thing" because we somehow don't feel comfortable with the current state. | 19:48 |
| alaski_ | the current state has known issues. I think we try to improve that situation | 19:49 |
| mgagne | alaski_: in that case, it suggests removing a feature and suggesting a duct tape solution instead. that's how I feel. | 19:50 |
| alaski_ | I'm suggesting moving a feature. there is a laundry list of issues with bfv the way it's currently done with source=image. In order to improve that I think we need to move volume creation to somewhere better suited, since the compute is not the best place | 19:52 |
| mgagne | alaski_: it happens to work very well on our side, thanks to our efforts and the one made by Cinder. | 19:53 |
| *** alejandrito has quit IRC | 19:53 | |
| *** matrohon has quit IRC | 19:53 | |
| alaski_ | it does not work well for me. for the reasons I enumerated in the thread | 19:54 |
| alaski_ | my number one complaint is that there is no feedback to the user on progress or failure | 19:54 |
| mgagne | alaski_: I know we can make it work, we are better than those people from other projects =) | 19:54 |
| mriedem | mgagne: i'm interested in how you tune the CONF.block_device_allocate_retries and CONF.block_device_allocate_retries_interval options for BFV | 19:54 |
| *** matrohon has joined #openstack-nova | 19:55 | |
| *** pixelbeat has quit IRC | 19:55 | |
| mriedem | because by default i think it's 3 minutes for the volume create to complete | 19:55 |
| mgagne | mriedem: I don't have to tune it much, Cinder caches images and clone from a "template volume", speeding up the volume creation | 19:55 |
| alaski_ | mgagne: I very much agree that we can make it work. I just don't hold to the idea that because something is currently in the Nova API that it always has to be there | 19:55 |
| mgagne | alaski_: https://review.openstack.org/#/c/195795/ https://review.openstack.org/#/c/201754/ | 19:56 |
| *** ijw has joined #openstack-nova | 19:56 | |
| mgagne | alaski_: boot from volume takes less than 5 seconds on our side, not minutes. | 19:57 |
| mriedem | we should probably update the docs on CONF.block_device_allocate_retries to point out that you want to either set that high or make sure you have the image cache stuff setup in cinder | 19:57 |
| *** hitalia has quit IRC | 19:57 | |
| mgagne | alaski_: yes, it used to be hell cause SolidFire has QoS and guess what, the image copy was throttled by it :D | 19:58 |
| mriedem | image_volume_cache_enabled | 19:58 |
| mgagne | mriedem: I would be much more receptive to have a way to disable bfv source=image if you know 100% that it will never work on your cloud due to lack of image caching. | 19:59 |
| *** VW_ has quit IRC | 20:00 | |
| melwitt | alaski_, mgagne: fwiw, I've been thinking lately that there needs to be a fronting service that can call out to all the services APIs to give a unified experience for users. and something more friendly than Heat DIY templating | 20:00 |
| *** VW has joined #openstack-nova | 20:01 | |
| alaski_ | mgagne: thanks. I'll pass that on to our Cinder folks. Of course we'll need to upgrade past Icehouse then :) | 20:01 |
| *** ijw has quit IRC | 20:01 | |
| alaski_ | melwitt: I very much agree with that | 20:01 |
| mgagne | melwitt: I agree with this vision. You will understand that I will be opposed to it if the plan happens to be: stop supporting XYZ feature in Nova in favor of that next greatest solution unless it's 100% feature parity. | 20:02 |
| *** VW_ has joined #openstack-nova | 20:03 | |
| *** VW has quit IRC | 20:03 | |
| mgagne | melwitt: we can't afford to break the openstack adoption trend and make it much more difficult by yet moving to another client, API versions or whatever. | 20:04 |
| *** jwcroppe has joined #openstack-nova | 20:05 | |
| mgagne | melwitt: while it has it glitches, Nova happens to (unfortunately?) offer a good part of the solution. | 20:05 |
| mriedem | wow, reading https://review.openstack.org/#/c/201754/ - seems like all of the planets have to align for that to work | 20:06 |
| mriedem | mgagne: ^ | 20:06 |
| *** hemnalunch is now known as hemna | 20:06 | |
| *** su_zhang has joined #openstack-nova | 20:07 | |
| mgagne | mriedem: TBH, I didn't look in all the details of it. I starred the change for future reading. We are using the SolidFire specific image caching system which happens to implement and replace a 90% similar system we had in place for years. thanks jgriffith | 20:07 |
| mriedem | doffm: https://bugs.launchpad.net/nova/+bug/1499518 | 20:08 |
| openstack | Launchpad bug 1499518 in OpenStack Compute (nova) "Docs for CONF.block_device_allocate_retries should be updated to note cinder image caching" [Low,Confirmed] | 20:08 |
| mriedem | mgagne: yeah that's what i suggest to at least note in ^ | 20:08 |
| mriedem | the image caching stuff | 20:08 |
| mriedem | if you find yourself having to tweak CONF.block_device_allocate_retries in nova | 20:08 |
| alaski_ | mgagne: the reality is that we can't pull source=image from the api. the most we could do is add a microversion to stop it but it would still be available on lower versions. But it would be nice to signal that it's not preferred to do it that way. | 20:09 |
| *** alex_klimov has joined #openstack-nova | 20:09 | |
| *** breitz has quit IRC | 20:10 | |
| *** su_zhang has quit IRC | 20:11 | |
| *** breitz has joined #openstack-nova | 20:11 | |
| melwitt | mgagne: yeah, I understand. nova has some things in it to give users a single api call in some cases, and I personally think we need to keep those until we have the end state of a separate single endpoint service that allows nova to deprecate them | 20:12 |
| mgagne | sorry, got context switched | 20:13 |
| mgagne | mriedem: until the image caching system, we had similar issues to alaski_ and we choose to implement orchestration in front of it. It was a dark time tbh and it prevented us from opening the API. | 20:13 |
| *** leseb has joined #openstack-nova | 20:15 | |
| *** vkmc has joined #openstack-nova | 20:16 | |
| mgagne | alaski_: you can pull it with a policy? :D | 20:16 |
| *** alaski_ is now known as alaski | 20:19 | |
| *** salv-orlando has joined #openstack-nova | 20:19 | |
| alaski | mgagne: apparently our issue with the cache is that we don't have fast cloning | 20:20 |
| *** nickchase has quit IRC | 20:20 | |
| mgagne | alaski: yea, that's the secret sauce =) | 20:20 |
| *** Piet has quit IRC | 20:22 | |
| *** jichen has quit IRC | 20:22 | |
| *** salv-orlando has quit IRC | 20:22 | |
| *** su_zhang has joined #openstack-nova | 20:24 | |
| *** bnemec has joined #openstack-nova | 20:24 | |
| *** tpeoples has joined #openstack-nova | 20:25 | |
| *** baoli has quit IRC | 20:28 | |
| *** achanda has quit IRC | 20:30 | |
| *** vladikr has quit IRC | 20:33 | |
| *** gyee has joined #openstack-nova | 20:34 | |
| *** salv-orlando has joined #openstack-nova | 20:35 | |
| *** salv-orlando has quit IRC | 20:35 | |
| *** salv-orlando has joined #openstack-nova | 20:36 | |
| *** julien_ZTE has joined #openstack-nova | 20:40 | |
| *** exploreshaifali has quit IRC | 20:41 | |
| *** tdp has joined #openstack-nova | 20:41 | |
| *** tpeoples has quit IRC | 20:41 | |
| *** julien_ZTE has quit IRC | 20:44 | |
| *** Piet has joined #openstack-nova | 20:47 | |
| * bauzas waves back | 20:49 | |
| bauzas | interesting thread above... | 20:49 |
| *** jgriffith has quit IRC | 20:51 | |
| *** rfolco has quit IRC | 20:52 | |
| mriedem | nova meeting in 7 minutes | 20:53 |
| *** lucasagomes has joined #openstack-nova | 20:54 | |
| *** jgriffith has joined #openstack-nova | 20:55 | |
| bauzas | mriedem: IIUC, --bdm is an old way to get volumes, we now prefer to use --block-device, right? | 20:55 |
| bauzas | mriedem: sorry if you think it's a silly question but it seems I need to improve my volume skills | 20:56 |
| bauzas | (speaking of nova boot) | 20:56 |
| mriedem | bauzas: yeah should be --block-device i think | 20:56 |
| mriedem | http://docs.openstack.org/cli-reference/content/novaclient_commands.html#novaclient_subcommand_boot | 20:56 |
| *** raildo is now known as raildo-afk | 20:56 | |
| *** ijw has joined #openstack-nova | 20:57 | |
| *** rfolco has joined #openstack-nova | 20:57 | |
| mriedem | there is also --boot-volume and --snapshot | 20:57 |
| mriedem | there are 20 ways to skin that cat | 20:57 |
| bauzas | mriedem: that's what I'm looking | 20:57 |
| bauzas | none of them are said deprecated | 20:58 |
| *** dave-mccowan has quit IRC | 20:58 | |
| bauzas | in the client I mean | 20:58 |
| *** VW_ has quit IRC | 20:58 | |
| *** VW has joined #openstack-nova | 20:59 | |
| mriedem | not in the docs i don't think | 20:59 |
| bauzas | ergh, 1 min before a meeting is too short to think about that | 20:59 |
| mriedem | i think there is something in the code though | 20:59 |
| bauzas | yeah I remember something I heard | 20:59 |
| mriedem | we should put something in the option help text though if one is deprecated | 20:59 |
| *** jgriffith has quit IRC | 21:00 | |
| bauzas | anyway, meeting time | 21:00 |
| *** VW_ has joined #openstack-nova | 21:01 | |
| *** VW has quit IRC | 21:01 | |
| *** mtanino has quit IRC | 21:01 | |
| *** hemna is now known as hemnafk | 21:01 | |
| *** edmondsw has quit IRC | 21:02 | |
| *** ijw has quit IRC | 21:02 | |
| *** rfolco has quit IRC | 21:03 | |
| *** achanda has joined #openstack-nova | 21:05 | |
| *** hemna has joined #openstack-nova | 21:05 | |
| *** yamahata has quit IRC | 21:06 | |
| *** ijw has joined #openstack-nova | 21:06 | |
| *** thorst has quit IRC | 21:07 | |
| *** aysyd has quit IRC | 21:07 | |
| *** leitan has quit IRC | 21:08 | |
| *** achanda has quit IRC | 21:09 | |
| *** mtanino has joined #openstack-nova | 21:11 | |
| *** matrohon has quit IRC | 21:12 | |
| *** leitan has joined #openstack-nova | 21:12 | |
| *** jgriffith has joined #openstack-nova | 21:21 | |
| *** derekh has joined #openstack-nova | 21:21 | |
| *** burgerk has quit IRC | 21:21 | |
| *** otter768 has joined #openstack-nova | 21:22 | |
| *** eharney has quit IRC | 21:26 | |
| *** otter768 has quit IRC | 21:27 | |
| *** dims__ has quit IRC | 21:29 | |
| *** hitalia has joined #openstack-nova | 21:30 | |
| *** dustins has quit IRC | 21:31 | |
| *** ccarmack has left #openstack-nova | 21:32 | |
| *** pumaranikar has quit IRC | 21:32 | |
| *** rlrossit has left #openstack-nova | 21:34 | |
| *** changbl has quit IRC | 21:37 | |
| *** armax_ has joined #openstack-nova | 21:37 | |
| mriedem | bauzas: are you going to propose stable/liberty backports for https://review.openstack.org/#/c/227199/ and https://review.openstack.org/#/c/221620/ ? | 21:38 |
| *** changbl has joined #openstack-nova | 21:39 | |
| *** armax has quit IRC | 21:40 | |
| *** leifz has quit IRC | 21:40 | |
| *** armax_ is now known as armax | 21:40 | |
| *** leifz has joined #openstack-nova | 21:40 | |
| *** marzif has quit IRC | 21:40 | |
| *** iraw- has quit IRC | 21:41 | |
| *** julien_ZTE has joined #openstack-nova | 21:41 | |
| bauzas | mriedem: yup, was discussing with johnthetubaguy about the opportunity to squash them | 21:41 |
| bauzas | mriedem: because the first one is buggy | 21:41 |
| *** alejandrito has joined #openstack-nova | 21:41 | |
| bauzas | thanks to a douchebag | 21:41 |
| *** iraw- has joined #openstack-nova | 21:42 | |
| bauzas | mriedem: so I'd prefer to backport only one change by squashing both | 21:42 |
| bauzas | mriedem: thoughts about that? | 21:42 |
| *** derekh has quit IRC | 21:43 | |
| bauzas | (of course, having the 2 commit msgs merged into the same msg) | 21:43 |
| bauzas | I mean leaving the commit msg, merged | 21:43 |
| *** julien_ZTE has quit IRC | 21:45 | |
| mriedem | squashing is ok if it makes sense | 21:46 |
| mriedem | otherwise we just have to resolve merge conflicts | 21:46 |
| bauzas | mriedem: okay, will provide one change by tomorrow morning my time | 21:47 |
| mriedem | it sucks that we have merge conflicts going to stable/liberty within 1 day of having that branch.... | 21:47 |
| mriedem | yeah, np | 21:47 |
| * bauzas says hello to his bed | 21:47 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Only create volumes with instance.az if cinder.cross_az_attach is False https://review.openstack.org/227564 | 21:48 |
| bauzas | hah | 21:48 |
| mriedem | needed to get that up before i leave | 21:48 |
| bauzas | one last before heading off | 21:48 |
| mriedem | which is pretty much now | 21:48 |
| bauzas | food for thoughts for tomorrow | 21:49 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Only create volumes with instance.az if cinder.cross_az_attach is False https://review.openstack.org/227564 | 21:50 |
| *** jlvillal has quit IRC | 21:51 | |
| *** dims_ has joined #openstack-nova | 21:52 | |
| *** mriedem has quit IRC | 21:53 | |
| *** tonytan4ever has quit IRC | 21:54 | |
| *** e0ne has joined #openstack-nova | 21:56 | |
| *** ctrath has quit IRC | 21:56 | |
| *** jlvillal has joined #openstack-nova | 21:57 | |
| *** changbl has quit IRC | 22:00 | |
| *** burt has quit IRC | 22:01 | |
| *** pixelbeat has joined #openstack-nova | 22:03 | |
| *** obedmr_ has joined #openstack-nova | 22:03 | |
| *** obedmr_ has quit IRC | 22:03 | |
| *** ctrath has joined #openstack-nova | 22:04 | |
| *** mdrabe has quit IRC | 22:04 | |
| *** changbl has joined #openstack-nova | 22:04 | |
| *** obedmr_ has joined #openstack-nova | 22:05 | |
| *** thumpba has joined #openstack-nova | 22:06 | |
| *** achanda has joined #openstack-nova | 22:06 | |
| *** obedmr_ has quit IRC | 22:06 | |
| *** obedmr has quit IRC | 22:07 | |
| *** jwcroppe has quit IRC | 22:07 | |
| *** VW has joined #openstack-nova | 22:07 | |
| *** jwcroppe has joined #openstack-nova | 22:07 | |
| *** zhangjn has joined #openstack-nova | 22:09 | |
| *** eglynn has quit IRC | 22:10 | |
| *** VW_ has quit IRC | 22:10 | |
| *** hemna has quit IRC | 22:10 | |
| *** zhangjn has quit IRC | 22:11 | |
| *** VW has quit IRC | 22:12 | |
| *** jwcroppe has quit IRC | 22:12 | |
| *** zhangjn has joined #openstack-nova | 22:12 | |
| *** zhangjn has quit IRC | 22:12 | |
| *** ctrath has quit IRC | 22:13 | |
| *** zhangjn has joined #openstack-nova | 22:13 | |
| *** nic has joined #openstack-nova | 22:14 | |
| *** e0ne has quit IRC | 22:17 | |
| *** thumpba_ has joined #openstack-nova | 22:17 | |
| *** zhangjn has quit IRC | 22:17 | |
| *** tdp has quit IRC | 22:17 | |
| *** leseb_ has quit IRC | 22:19 | |
| *** thumpba has quit IRC | 22:20 | |
| *** hemnafk is now known as hemna | 22:22 | |
| *** artom has quit IRC | 22:24 | |
| *** pixelbeat has quit IRC | 22:24 | |
| *** artom has joined #openstack-nova | 22:26 | |
| *** Guest88 has joined #openstack-nova | 22:26 | |
| *** leseb_ has joined #openstack-nova | 22:27 | |
| *** jhesketh has quit IRC | 22:28 | |
| *** jhesketh has joined #openstack-nova | 22:29 | |
| *** stevemar has quit IRC | 22:30 | |
| *** zhangjn has joined #openstack-nova | 22:36 | |
| *** zhangjn has quit IRC | 22:37 | |
| *** su_zhang has quit IRC | 22:37 | |
| *** zhangjn has joined #openstack-nova | 22:38 | |
| *** penick has quit IRC | 22:40 | |
| *** zzzeek has quit IRC | 22:40 | |
| *** asettle has joined #openstack-nova | 22:41 | |
| *** obedmr has joined #openstack-nova | 22:44 | |
| *** dave-mccowan has joined #openstack-nova | 22:47 | |
| *** obedmr has quit IRC | 22:48 | |
| *** thumpba_ has quit IRC | 22:48 | |
| *** thumpba has joined #openstack-nova | 22:48 | |
| *** pratikma_ has quit IRC | 22:49 | |
| *** obedmr has joined #openstack-nova | 22:51 | |
| *** echoingumesh has joined #openstack-nova | 22:51 | |
| *** jgriffith has quit IRC | 22:55 | |
| *** jgriffith has joined #openstack-nova | 22:55 | |
| *** bswartz has joined #openstack-nova | 22:55 | |
| *** Guest88 has quit IRC | 22:57 | |
| *** jgriffith has quit IRC | 22:57 | |
| *** zhangjn_ has joined #openstack-nova | 22:57 | |
| *** armax has quit IRC | 22:59 | |
| *** jgriffith has joined #openstack-nova | 23:00 | |
| *** IanGovett has quit IRC | 23:00 | |
| *** obedmr has quit IRC | 23:01 | |
| *** zhangjn has quit IRC | 23:01 | |
| *** alejandrito has quit IRC | 23:03 | |
| *** alex_klimov has quit IRC | 23:04 | |
| *** markvoelker_ has quit IRC | 23:04 | |
| *** leitan has quit IRC | 23:04 | |
| *** dims_ has quit IRC | 23:05 | |
| *** thumpba has quit IRC | 23:06 | |
| *** su_zhang has joined #openstack-nova | 23:07 | |
| *** julien_ZTE has joined #openstack-nova | 23:12 | |
| *** ijuwang has quit IRC | 23:15 | |
| *** julien_ZTE has quit IRC | 23:16 | |
| *** diogogmt has quit IRC | 23:16 | |
| *** echoingumesh has quit IRC | 23:17 | |
| *** echoingumesh has joined #openstack-nova | 23:18 | |
| *** echoingumesh has quit IRC | 23:22 | |
| *** otter768 has joined #openstack-nova | 23:23 | |
| *** thumpba has joined #openstack-nova | 23:25 | |
| *** otter768 has quit IRC | 23:28 | |
| *** stevemar has joined #openstack-nova | 23:31 | |
| *** ildikov has quit IRC | 23:32 | |
| *** stevemar has quit IRC | 23:33 | |
| *** tjones1 has quit IRC | 23:38 | |
| *** lucasagomes has quit IRC | 23:38 | |
| *** erhudy has quit IRC | 23:39 | |
| *** nic has quit IRC | 23:41 | |
| *** agireud has quit IRC | 23:42 | |
| *** pratikmallya has joined #openstack-nova | 23:44 | |
| *** shoutm has joined #openstack-nova | 23:45 | |
| *** davideagnello has quit IRC | 23:46 | |
| *** emagana has quit IRC | 23:47 | |
| *** ildikov has joined #openstack-nova | 23:49 | |
| *** ZZelle_ has quit IRC | 23:49 | |
| *** vishwana_ has joined #openstack-nova | 23:54 | |
| *** thumpba_ has joined #openstack-nova | 23:54 | |
| *** electrocucaracha has quit IRC | 23:54 | |
| *** jamielennox|away is now known as jamielennox | 23:56 | |
| *** vishwanathj has quit IRC | 23:56 | |
| *** rex_lee_ has quit IRC | 23:56 | |
| *** jaypipes has quit IRC | 23:56 | |
| *** harlowja has quit IRC | 23:56 | |
| *** thumpba has quit IRC | 23:56 | |
| *** sayali has quit IRC | 23:56 | |
| *** BadCub has quit IRC | 23:56 | |
| *** r-daneel has quit IRC | 23:57 | |
| *** darrenc has quit IRC | 23:57 | |
| *** gus has quit IRC | 23:57 | |
| *** davidlenwell has quit IRC | 23:57 | |
| *** sdague has quit IRC | 23:57 | |
| *** dguitarbite has quit IRC | 23:57 | |
| *** harlowja has joined #openstack-nova | 23:57 | |
| *** darrenc_ has joined #openstack-nova | 23:57 | |
| *** xyang1 has quit IRC | 23:57 | |
| *** darrenc_ is now known as darrenc | 23:58 | |
| *** davidlenwell has joined #openstack-nova | 23:58 | |
| *** dguitarbite has joined #openstack-nova | 23:58 | |
| *** gus has joined #openstack-nova | 23:59 | |
| *** r-daneel has joined #openstack-nova | 23:59 | |
| *** sdague has joined #openstack-nova | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!