| openstackgerrit | Ken'ichi Ohmichi proposed openstack/nova: Add error handling for delete-volume API https://review.openstack.org/382660 | 00:00 |
|---|---|---|
| *** rama_y has quit IRC | 00:00 | |
| *** jamesden_ has quit IRC | 00:07 | |
| *** dikonoor has joined #openstack-nova | 00:09 | |
| *** ijw has joined #openstack-nova | 00:11 | |
| *** ducttape_ has quit IRC | 00:12 | |
| *** kaisers has joined #openstack-nova | 00:13 | |
| *** baoli has quit IRC | 00:15 | |
| *** ijw has quit IRC | 00:16 | |
| *** kaisers has quit IRC | 00:17 | |
| *** dikonoor has quit IRC | 00:18 | |
| *** pradk has joined #openstack-nova | 00:19 | |
| *** markvoelker has joined #openstack-nova | 00:22 | |
| *** stewie925 has quit IRC | 00:24 | |
| *** markvoelker has quit IRC | 00:27 | |
| *** ducttape_ has joined #openstack-nova | 00:37 | |
| *** thorst_ has joined #openstack-nova | 00:44 | |
| *** hoangcx has joined #openstack-nova | 00:45 | |
| oomichi | alex_xu: mikal: maybe now is your tz. Can you take a look at https://review.openstack.org/#/c/351962/ ? | 00:46 |
| oomichi | I am not sure why this happens.. | 00:46 |
| *** rfolco has quit IRC | 00:48 | |
| *** Pavo has quit IRC | 00:51 | |
| *** anush has quit IRC | 00:52 | |
| *** baoli has joined #openstack-nova | 00:54 | |
| *** dave-mcc_ has joined #openstack-nova | 00:54 | |
| *** zul has joined #openstack-nova | 00:55 | |
| *** dave-mccowan has quit IRC | 00:55 | |
| *** Pavo has joined #openstack-nova | 00:56 | |
| *** Hosam has quit IRC | 00:57 | |
| *** Hosam has joined #openstack-nova | 00:57 | |
| *** Apoorva_ has joined #openstack-nova | 00:58 | |
| *** phuongnh has joined #openstack-nova | 01:00 | |
| *** Apoorva has quit IRC | 01:01 | |
| *** Hosam has quit IRC | 01:01 | |
| *** Apoorva_ has quit IRC | 01:03 | |
| *** raunak has quit IRC | 01:04 | |
| *** browne has quit IRC | 01:04 | |
| *** zul has quit IRC | 01:08 | |
| *** kberger has joined #openstack-nova | 01:09 | |
| *** zul has joined #openstack-nova | 01:12 | |
| *** ijw has joined #openstack-nova | 01:13 | |
| *** code-R has joined #openstack-nova | 01:14 | |
| *** Sukhdev has quit IRC | 01:15 | |
| *** ijw has quit IRC | 01:18 | |
| *** rfolco has joined #openstack-nova | 01:21 | |
| *** TravT has quit IRC | 01:22 | |
| *** unicell has quit IRC | 01:25 | |
| *** dave-mccowan has joined #openstack-nova | 01:28 | |
| *** dave-mcc_ has quit IRC | 01:29 | |
| *** suresh12 has quit IRC | 01:32 | |
| *** suresh12 has joined #openstack-nova | 01:33 | |
| *** suresh12 has quit IRC | 01:33 | |
| *** code-R_ has joined #openstack-nova | 01:33 | |
| *** code-R has quit IRC | 01:36 | |
| *** hemanthm is now known as hemanthm|away | 01:54 | |
| *** tuanluong has joined #openstack-nova | 01:55 | |
| openstackgerrit | sean mooney proposed openstack/nova: convert libvirt driver to use os-vif for vhost-user with ovs. https://review.openstack.org/334048 | 02:00 |
| *** BobBall has joined #openstack-nova | 02:01 | |
| *** dave-mccowan has quit IRC | 02:01 | |
| *** BobBall is now known as Guest69093 | 02:01 | |
| *** ducnc has quit IRC | 02:01 | |
| *** kaisers has joined #openstack-nova | 02:02 | |
| *** code-R_ has quit IRC | 02:03 | |
| *** code-R has joined #openstack-nova | 02:03 | |
| *** suresh12 has joined #openstack-nova | 02:04 | |
| *** suresh12 has quit IRC | 02:05 | |
| *** baoli has quit IRC | 02:05 | |
| *** kaisers has quit IRC | 02:06 | |
| *** Guest69093 has quit IRC | 02:06 | |
| *** gouthamr has quit IRC | 02:07 | |
| *** code-R_ has joined #openstack-nova | 02:10 | |
| *** amotoki has joined #openstack-nova | 02:12 | |
| *** ducnc has joined #openstack-nova | 02:13 | |
| *** code-R has quit IRC | 02:14 | |
| *** ijw has joined #openstack-nova | 02:14 | |
| *** mriedem has quit IRC | 02:18 | |
| *** ijw has quit IRC | 02:20 | |
| *** BobBall_ has joined #openstack-nova | 02:20 | |
| *** BobBall_ has quit IRC | 02:24 | |
| *** ducnc has quit IRC | 02:24 | |
| *** rfolco has quit IRC | 02:25 | |
| *** thorst_ has quit IRC | 02:25 | |
| *** BobBall_ has joined #openstack-nova | 02:26 | |
| *** thorst_ has joined #openstack-nova | 02:26 | |
| *** winston-d_ has joined #openstack-nova | 02:29 | |
| *** BobBall_ has quit IRC | 02:31 | |
| openstackgerrit | sean mooney proposed openstack/os-vif: add support for vhost-user reconnect https://review.openstack.org/357555 | 02:31 |
| *** BobBall_ has joined #openstack-nova | 02:32 | |
| *** ducnc has joined #openstack-nova | 02:33 | |
| *** thorst_ has quit IRC | 02:35 | |
| *** code-R has joined #openstack-nova | 02:37 | |
| *** code-R_ has quit IRC | 02:37 | |
| *** kylek3h has quit IRC | 02:37 | |
| *** BobBall_ has quit IRC | 02:37 | |
| *** haplo37_ has quit IRC | 02:39 | |
| *** kylek3h has joined #openstack-nova | 02:39 | |
| *** BobBall_ has joined #openstack-nova | 02:39 | |
| *** haplo37_ has joined #openstack-nova | 02:41 | |
| *** tongli has quit IRC | 02:43 | |
| *** tiendc has joined #openstack-nova | 02:43 | |
| *** thorst_ has joined #openstack-nova | 02:47 | |
| *** thorst_ has quit IRC | 02:47 | |
| *** thorst_ has joined #openstack-nova | 02:48 | |
| *** Pavo has quit IRC | 02:51 | |
| *** BobBall_ has quit IRC | 02:51 | |
| *** ducttape_ has quit IRC | 02:51 | |
| *** BobBall has joined #openstack-nova | 02:52 | |
| *** BobBall is now known as Guest54540 | 02:52 | |
| *** thorst_ has quit IRC | 02:53 | |
| *** hieulq has joined #openstack-nova | 02:53 | |
| *** maticue has quit IRC | 02:55 | |
| *** Pavo has joined #openstack-nova | 02:55 | |
| *** diana-test has quit IRC | 02:56 | |
| *** diana-test has joined #openstack-nova | 02:57 | |
| *** Guest54540 has quit IRC | 03:02 | |
| *** phschwartz has quit IRC | 03:03 | |
| *** BobBall_ has joined #openstack-nova | 03:03 | |
| *** mordred has quit IRC | 03:03 | |
| *** hieulq has quit IRC | 03:04 | |
| *** mordred has joined #openstack-nova | 03:06 | |
| *** phschwartz has joined #openstack-nova | 03:07 | |
| *** BobBall_ has quit IRC | 03:12 | |
| *** yamahata has quit IRC | 03:12 | |
| *** sdake has joined #openstack-nova | 03:14 | |
| *** BobBall_ has joined #openstack-nova | 03:14 | |
| *** ijw has joined #openstack-nova | 03:16 | |
| *** hieulq has joined #openstack-nova | 03:18 | |
| *** ijw has quit IRC | 03:21 | |
| *** hieulq has quit IRC | 03:38 | |
| *** BobBall_ has quit IRC | 03:38 | |
| *** armax has quit IRC | 03:41 | |
| *** annegentle has quit IRC | 03:42 | |
| *** code-R has quit IRC | 03:43 | |
| *** sdake_ has joined #openstack-nova | 03:45 | |
| *** sdake has quit IRC | 03:47 | |
| *** BobBall has joined #openstack-nova | 03:47 | |
| *** armax has joined #openstack-nova | 03:47 | |
| *** links has joined #openstack-nova | 03:47 | |
| *** BobBall is now known as Guest642 | 03:48 | |
| *** kaisers has joined #openstack-nova | 03:50 | |
| *** thorst_ has joined #openstack-nova | 03:51 | |
| *** code-R has joined #openstack-nova | 03:52 | |
| *** Guest642 has quit IRC | 03:53 | |
| *** code-R_ has joined #openstack-nova | 03:53 | |
| *** BobBall_ has joined #openstack-nova | 03:55 | |
| *** kaisers has quit IRC | 03:55 | |
| *** rama_y has joined #openstack-nova | 03:55 | |
| *** code-R has quit IRC | 03:57 | |
| *** thorst_ has quit IRC | 03:58 | |
| *** BobBall_ has quit IRC | 04:00 | |
| *** armax has quit IRC | 04:02 | |
| *** armax has joined #openstack-nova | 04:02 | |
| *** armax has quit IRC | 04:03 | |
| *** mtanino has quit IRC | 04:03 | |
| *** John341 has quit IRC | 04:06 | |
| *** haplo37 has quit IRC | 04:07 | |
| *** GB21 has joined #openstack-nova | 04:11 | |
| *** agireud has quit IRC | 04:16 | |
| *** haplo37 has joined #openstack-nova | 04:17 | |
| *** ijw has joined #openstack-nova | 04:17 | |
| *** ijw has quit IRC | 04:23 | |
| *** agireud has joined #openstack-nova | 04:24 | |
| *** Sukhdev has joined #openstack-nova | 04:25 | |
| *** markvoelker has joined #openstack-nova | 04:25 | |
| *** jaosorior has joined #openstack-nova | 04:25 | |
| *** GB21 has quit IRC | 04:26 | |
| *** markvoelker has quit IRC | 04:30 | |
| *** unicell has joined #openstack-nova | 04:32 | |
| *** unicell has quit IRC | 04:33 | |
| *** unicell has joined #openstack-nova | 04:36 | |
| *** sudswas__ has joined #openstack-nova | 04:36 | |
| *** sudipto has joined #openstack-nova | 04:36 | |
| *** code-R has joined #openstack-nova | 04:37 | |
| *** code-R_ has quit IRC | 04:37 | |
| *** GB21 has joined #openstack-nova | 04:38 | |
| *** code-R_ has joined #openstack-nova | 04:41 | |
| *** salv-orlando has joined #openstack-nova | 04:41 | |
| *** sudipto has quit IRC | 04:42 | |
| *** sudswas__ has quit IRC | 04:42 | |
| *** code-R has quit IRC | 04:44 | |
| *** liverpooler has quit IRC | 04:45 | |
| *** salv-orlando has quit IRC | 04:46 | |
| *** zz_dimtruck is now known as dimtruck | 04:47 | |
| *** ratailor has joined #openstack-nova | 04:51 | |
| *** Pavo has quit IRC | 04:51 | |
| *** psachin has joined #openstack-nova | 04:51 | |
| *** unicell has quit IRC | 04:51 | |
| *** unicell1 has joined #openstack-nova | 04:51 | |
| *** yamahata has joined #openstack-nova | 04:52 | |
| *** thorst_ has joined #openstack-nova | 04:55 | |
| *** Pavo has joined #openstack-nova | 04:55 | |
| *** winston-d_ has quit IRC | 04:56 | |
| *** adisky has joined #openstack-nova | 04:56 | |
| *** narayrak has joined #openstack-nova | 04:56 | |
| *** narayrak has quit IRC | 04:59 | |
| *** sridharg has joined #openstack-nova | 04:59 | |
| *** thorst_ has quit IRC | 05:02 | |
| *** code-R has joined #openstack-nova | 05:07 | |
| *** code-R_ has quit IRC | 05:07 | |
| *** salv-orlando has joined #openstack-nova | 05:07 | |
| *** takashin has joined #openstack-nova | 05:10 | |
| *** pbandark has joined #openstack-nova | 05:12 | |
| *** code-R has quit IRC | 05:12 | |
| *** haplo37_ has quit IRC | 05:15 | |
| *** takashin has quit IRC | 05:16 | |
| *** takashin has joined #openstack-nova | 05:17 | |
| *** haplo37_ has joined #openstack-nova | 05:17 | |
| *** TravT has joined #openstack-nova | 05:18 | |
| *** ijw has joined #openstack-nova | 05:19 | |
| *** david-lyle has joined #openstack-nova | 05:19 | |
| *** prateek has joined #openstack-nova | 05:21 | |
| *** ijw has quit IRC | 05:24 | |
| *** markvoelker has joined #openstack-nova | 05:26 | |
| *** markvoelker has quit IRC | 05:31 | |
| *** ldeptula has joined #openstack-nova | 05:38 | |
| *** John341 has joined #openstack-nova | 05:39 | |
| *** kaisers has joined #openstack-nova | 05:39 | |
| *** oanson has joined #openstack-nova | 05:41 | |
| *** dimtruck is now known as zz_dimtruck | 05:43 | |
| *** alex_xu has quit IRC | 05:48 | |
| *** kaisers has quit IRC | 05:48 | |
| *** sahid has joined #openstack-nova | 05:50 | |
| *** sahid has quit IRC | 05:51 | |
| *** alex_xu has joined #openstack-nova | 05:51 | |
| *** sahid has joined #openstack-nova | 05:51 | |
| openstackgerrit | sean mooney proposed openstack/os-vif: remove use of contextlib and with nested https://review.openstack.org/382725 | 05:51 |
| *** Sukhdev has quit IRC | 05:51 | |
| *** kaisers has joined #openstack-nova | 05:55 | |
| *** jaosorior has quit IRC | 05:55 | |
| *** suresh12 has joined #openstack-nova | 05:55 | |
| *** jaosorior has joined #openstack-nova | 05:57 | |
| *** diga has joined #openstack-nova | 05:58 | |
| *** thorst_ has joined #openstack-nova | 06:00 | |
| *** liverpooler has joined #openstack-nova | 06:02 | |
| *** david-lyle_ has joined #openstack-nova | 06:04 | |
| *** david-lyle has quit IRC | 06:04 | |
| *** nkrinner_afk is now known as nkrinner | 06:04 | |
| *** thorst_ has quit IRC | 06:07 | |
| *** liverpooler has quit IRC | 06:07 | |
| *** liverpooler has joined #openstack-nova | 06:07 | |
| *** paul-carlton2 has joined #openstack-nova | 06:10 | |
| *** sdake_ has quit IRC | 06:11 | |
| *** david-lyle_ has quit IRC | 06:12 | |
| *** code-R has joined #openstack-nova | 06:14 | |
| *** moshele has joined #openstack-nova | 06:15 | |
| *** rcernin has joined #openstack-nova | 06:15 | |
| *** code-R_ has joined #openstack-nova | 06:16 | |
| *** hussain_c has quit IRC | 06:17 | |
| *** code-R has quit IRC | 06:19 | |
| *** ijw has joined #openstack-nova | 06:20 | |
| *** lpetrut has joined #openstack-nova | 06:22 | |
| sahid | jaypipes: hello, any chance you ACK that one https://review.openstack.org/#/c/378560/ | 06:22 |
| sahid | it's real issue related to live migrate instances with ephemeral disks | 06:22 |
| openstackgerrit | Divya K Konoor proposed openstack/nova-specs: Allow admin users to pass key_name corresponding to another user and deploy a VM successfully. https://review.openstack.org/382588 | 06:25 |
| *** ijw has quit IRC | 06:26 | |
| *** markvoelker has joined #openstack-nova | 06:27 | |
| *** salv-orlando has quit IRC | 06:29 | |
| *** markvoelker has quit IRC | 06:32 | |
| *** hshiina has joined #openstack-nova | 06:32 | |
| *** andreas_s has joined #openstack-nova | 06:33 | |
| *** salv-orlando has joined #openstack-nova | 06:36 | |
| *** salv-orlando has quit IRC | 06:41 | |
| *** zenoway has joined #openstack-nova | 06:41 | |
| *** rama_y has quit IRC | 06:43 | |
| *** haplo37 has quit IRC | 06:51 | |
| *** Pavo has quit IRC | 06:51 | |
| *** lpetrut has quit IRC | 06:51 | |
| *** haplo37 has joined #openstack-nova | 06:51 | |
| *** jichen has joined #openstack-nova | 06:51 | |
| *** Pavo has joined #openstack-nova | 06:56 | |
| *** suresh12 has quit IRC | 06:57 | |
| *** unicell has joined #openstack-nova | 06:57 | |
| *** unicell1 has quit IRC | 06:58 | |
| *** ekuris has joined #openstack-nova | 07:03 | |
| *** code-R_ has quit IRC | 07:03 | |
| *** thorst_ has joined #openstack-nova | 07:05 | |
| *** hoangcx has quit IRC | 07:06 | |
| *** lpetrut has joined #openstack-nova | 07:07 | |
| *** ihrachys has joined #openstack-nova | 07:08 | |
| *** ecelik has joined #openstack-nova | 07:08 | |
| *** markus_z has joined #openstack-nova | 07:09 | |
| *** tesseract- has joined #openstack-nova | 07:11 | |
| *** psachin has quit IRC | 07:12 | |
| *** thorst_ has quit IRC | 07:12 | |
| *** ecelik has left #openstack-nova | 07:12 | |
| *** Dinesh_Bhor has joined #openstack-nova | 07:13 | |
| *** code-R has joined #openstack-nova | 07:13 | |
| *** salv-orlando has joined #openstack-nova | 07:14 | |
| *** psachin has joined #openstack-nova | 07:17 | |
| *** ihrachys has quit IRC | 07:18 | |
| *** lpetrut has quit IRC | 07:21 | |
| *** ijw has joined #openstack-nova | 07:22 | |
| *** jlanoux has joined #openstack-nova | 07:24 | |
| *** jlanoux has quit IRC | 07:25 | |
| *** jlanoux has joined #openstack-nova | 07:26 | |
| *** pcaruana has joined #openstack-nova | 07:26 | |
| *** code-R has quit IRC | 07:27 | |
| *** ijw has quit IRC | 07:28 | |
| *** markvoelker has joined #openstack-nova | 07:28 | |
| *** code-R has joined #openstack-nova | 07:28 | |
| openstackgerrit | Divya K Konoor proposed openstack/nova-specs: Allow admin users to pass key_name corresponding to another user and deploy a VM successfully. https://review.openstack.org/382588 | 07:30 |
| *** ralonsoh has joined #openstack-nova | 07:32 | |
| *** markvoelker has quit IRC | 07:32 | |
| *** code-R has quit IRC | 07:34 | |
| *** danpawlik has joined #openstack-nova | 07:43 | |
| danpawlik | Hello folks, can you check PS https://review.openstack.org/#/c/314928/ ? | 07:43 |
| *** danpb has joined #openstack-nova | 07:44 | |
| *** alexchadin has joined #openstack-nova | 07:46 | |
| *** mvk has quit IRC | 07:51 | |
| *** unicell has quit IRC | 07:54 | |
| *** ihrachys has joined #openstack-nova | 07:56 | |
| *** salv-orl_ has joined #openstack-nova | 07:56 | |
| *** ducttape_ has joined #openstack-nova | 07:56 | |
| *** unicell has joined #openstack-nova | 07:56 | |
| *** salv-orlando has quit IRC | 07:58 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** paul-carlton2 has quit IRC | 08:00 | |
| *** ducttape_ has quit IRC | 08:01 | |
| *** zzzeek has joined #openstack-nova | 08:01 | |
| *** haplo37_ has quit IRC | 08:01 | |
| *** hoangcx has joined #openstack-nova | 08:03 | |
| *** haplo37_ has joined #openstack-nova | 08:03 | |
| *** thorst_ has joined #openstack-nova | 08:10 | |
| *** ihrachys has quit IRC | 08:11 | |
| *** code-R has joined #openstack-nova | 08:12 | |
| *** code-R_ has joined #openstack-nova | 08:13 | |
| *** jichen has quit IRC | 08:16 | |
| *** yassine has joined #openstack-nova | 08:16 | |
| *** code-R has quit IRC | 08:17 | |
| *** yassine is now known as Guest39109 | 08:17 | |
| *** thorst_ has quit IRC | 08:17 | |
| *** jaosorior is now known as jaosorior_lunch | 08:17 | |
| *** code-R_ has quit IRC | 08:18 | |
| *** mgoddard has joined #openstack-nova | 08:19 | |
| *** mvk has joined #openstack-nova | 08:23 | |
| *** ijw has joined #openstack-nova | 08:24 | |
| *** ijw has quit IRC | 08:29 | |
| *** lpetrut has joined #openstack-nova | 08:31 | |
| *** Matti_ has quit IRC | 08:35 | |
| *** derekh has joined #openstack-nova | 08:38 | |
| *** ihrachys has joined #openstack-nova | 08:39 | |
| *** suresh12 has joined #openstack-nova | 08:43 | |
| *** suresh12 has quit IRC | 08:48 | |
| *** sridharg has quit IRC | 08:50 | |
| *** Pavo has quit IRC | 08:51 | |
| *** Pavo has joined #openstack-nova | 08:56 | |
| *** ociuhandu has joined #openstack-nova | 08:56 | |
| *** ducttape_ has joined #openstack-nova | 08:57 | |
| *** alexchad_ has joined #openstack-nova | 08:57 | |
| *** alexchadin has quit IRC | 09:00 | |
| *** takashin has left #openstack-nova | 09:01 | |
| *** ducttape_ has quit IRC | 09:02 | |
| mdbooth | kashyap: Did you get anywhere interesting with the rebase thing last night? | 09:02 |
| kashyap | mdbooth: After 2 hours of last night, ended up with this - https://bugzilla.redhat.com/show_bug.cgi?id=1382165 | 09:02 |
| openstack | bugzilla.redhat.com bug 1382165 in libvirt "virDomainGetBlockJobInfo: Adjust job reporting based on QEMU stats & the "ready" field of `query-block-jobs`" [Unspecified,New] - Assigned to libvirt-maint | 09:02 |
| kashyap | mdbooth: I wrote an email to - http://lists.openstack.org/pipermail/openstack-dev/2016-October/105158.html | 09:02 |
| mdbooth | kashyap: Reading, thanks. Ah... just saw your mail. | 09:02 |
| *** sridharg has joined #openstack-nova | 09:04 | |
| *** ralonsoh_ has joined #openstack-nova | 09:04 | |
| kashyap | mdbooth: Also read the small IRC snippet w/ Eric, on the above bug. | 09:05 |
| mdbooth | kashyap: That's an extremely thorough bz | 09:05 |
| *** ralonsoh has quit IRC | 09:05 | |
| kashyap | mdbooth: Yeah, spent way too much time than I care to admit last night :P | 09:05 |
| kashyap | Thx. | 09:06 |
| csatari | Hi, can someone help me in api-ref tox-ing? It has problems with the sidebar_mode. | 09:06 |
| kashyap | It was nice how Eric thought out loud about "first I need to be convinced that this is a real race" | 09:06 |
| kashyap | Then we reached the conclusion :-) | 09:06 |
| mdbooth | kashyap: I'm not a huge fan of libvirt fudging the progress numbers, I have to say. | 09:07 |
| kashyap | mdbooth: Hmm, I thought about it too, briefly. But wondered what else options are there | 09:08 |
| mdbooth | Yeah, I guess the issue is not breaking existing users who made assumptions which were valid at the time. | 09:09 |
| kashyap | mdbooth: Exactly. If we know that QEMU is new enough, then, we could wait for that event rather than polling | 09:09 |
| mdbooth | kashyap: So, if we switch to listening for events, would it fix this? | 09:09 |
| mdbooth | The event is generated by qemu, not libvirt, right? | 09:09 |
| mdbooth | Incidentally, what do other hypervisors do here? | 09:10 |
| kashyap | mdbooth: Yes, the event BLOCK_JOB_READY is a QMP event | 09:10 |
| mdbooth | We can see that event in the libvirt logs from the failed test run, btw | 09:11 |
| mdbooth | But 200ms *after* we tried to pivot | 09:11 |
| kashyap | mdbooth: Exactly. That's what I pointed out in step (6) | 09:11 |
| kashyap | In the email | 09:11 |
| * mdbooth has only read the bz so far | 09:11 | |
| mdbooth | kashyap: Sorry, let me catch up with you :) | 09:11 |
| kashyap | mdbooth: Yeah, please read the email analysis too. I wrote that just for you | 09:12 |
| *** karthiks has joined #openstack-nova | 09:13 | |
| *** thorst_ has joined #openstack-nova | 09:14 | |
| *** sridharg has quit IRC | 09:15 | |
| *** sridharg has joined #openstack-nova | 09:16 | |
| *** Matti_ has joined #openstack-nova | 09:18 | |
| *** ihrachys has quit IRC | 09:19 | |
| mdbooth | kashyap: That's excellent, thanks. | 09:19 |
| mdbooth | kashyap: So, are there any downsides if we switch to handling completion events? | 09:20 |
| mdbooth | Would it work with non-qemu hypervisors? | 09:20 |
| mdbooth | danpb: ^^^ | 09:20 |
| kashyap | mdbooth: I don't know of any downsides at the moment. | 09:20 |
| kashyap | Not sure about the non-QEMU hypervisors, good question. | 09:20 |
| danpb | mdbooth: completion events for what ? | 09:21 |
| mdbooth | kashyap: polling for completion just looks like a quagmire of workarounds | 09:21 |
| mdbooth | danpb: in this case, block jobs | 09:21 |
| danpb | mdbooth: non-qemu is irrelevant then - qemu is the only thing doing block jobs | 09:21 |
| mdbooth | kashyap: So, sounds like we should switch to handling completion events | 09:22 |
| *** thorst_ has quit IRC | 09:22 | |
| mdbooth | danpb: How about support from Ubuntu's paleolithic libvirt/qemu versions? | 09:22 |
| *** zenoway has quit IRC | 09:22 | |
| kashyap | mdbooth: Yeah. That seems to be the clearest indication | 09:22 |
| kashyap | Hehe, "paleolithic" :D | 09:23 |
| danpb | mdbooth: no idea | 09:23 |
| *** mnestratov has joined #openstack-nova | 09:23 | |
| mdbooth | danpb: Are we likely to hit issues with old versions which didn't generate certain events? | 09:23 |
| danpb | no idea | 09:23 |
| mdbooth | danpb: Ok, thanks. | 09:24 |
| kashyap | mdbooth: Here's what you're looking for -- the QAPI schema doc for that event says, it is available since QEMU 1.3 | 09:24 |
| kashyap | See line 2670 -- http://git.qemu.org/?p=qemu.git;a=blob;f=qapi/block-core.json;h=9d797b8fe0050443eb6393cbda77e126ac974ed3;hb=HEAD#l2651 | 09:24 |
| mdbooth | kashyap: I think we only need COMPLETED, FAILED, CANCELLED, and READY, right? | 09:24 |
| mdbooth | kashyap: Your git-fu is very strong, btw. | 09:25 |
| *** ijw has joined #openstack-nova | 09:25 | |
| *** code-R has joined #openstack-nova | 09:25 | |
| kashyap | mdbooth: Don't let it give such an impression, I can't do Octopus merges | 09:25 |
| * mdbooth neither, but I do know what one is ;) | 09:26 | |
| kashyap | mdbooth: About what we need. Incidentally John Snow, from QEMU, was just debating yesterday | 09:26 |
| kashyap | Trying to answer the question: "Whether jobs should always emit events?" | 09:26 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: add event lifecycle details https://review.openstack.org/373943 | 09:27 |
| *** zenoway has joined #openstack-nova | 09:27 | |
| mdbooth | kashyap: There needs to be an api we can rely on. | 09:27 |
| kashyap | Yes | 09:28 |
| mdbooth | Right now, I'm convinced polling job info isn't it. | 09:28 |
| mdbooth | Well, almost convinced. | 09:28 |
| * kashyap is on a coffee shop n/w, and every half-an-hour this has some stupid captive-portal nonsense. Wonder how I could subvert it | 09:28 | |
| mdbooth | I'm slightly nervous about events, because a missing event would be disasterous. | 09:28 |
| *** mvk has quit IRC | 09:29 | |
| *** code-R_ has joined #openstack-nova | 09:29 | |
| kashyap | mdbooth: Indeed, maybe we could have a "two-phased approach" | 09:29 |
| mdbooth | That also lacks simplicity. | 09:30 |
| mdbooth | Likely we'd get the union of both sets of bugs :) | 09:30 |
| kashyap | LOL, union of both sets | 09:30 |
| *** ociuhandu has quit IRC | 09:31 | |
| *** ijw has quit IRC | 09:31 | |
| kashyap | mdbooth: Currently, I see these possible QMP events: BLOCK_IO_ERROR, BLOCK_JOB_CANCELLED, BLOCK_JOB_COMPLETED, BLOCK_JOB_ERROR, BLOCK_JOB_READY | 09:31 |
| mdbooth | kashyap: I'm just trying to decide if we need to handle IO_ERROR | 09:31 |
| mdbooth | # Note: If action is "stop", a STOP event will eventually follow the | 09:32 |
| mdbooth | 2564 # BLOCK_IO_ERROR event | 09:32 |
| *** code-R has quit IRC | 09:32 | |
| *** hshiina has quit IRC | 09:33 | |
| kashyap | Yeah, that's certainly useful - what action to take when read / write IO errors occur | 09:33 |
| mdbooth | kashyap: So, it looks like fatal IO_ERROR will be followed by BLOCK_JOB_ERROR | 09:34 |
| mdbooth | IoW, we wouldn't want to handle IO_ERROR directly | 09:35 |
| kashyap | mdbooth: Yes | 09:35 |
| mdbooth | Except maybe to log it | 09:35 |
| *** jaosorior_lunch has quit IRC | 09:36 | |
| *** garyk has joined #openstack-nova | 09:36 | |
| *** jaosorior_lunch has joined #openstack-nova | 09:36 | |
| kashyap | Definitely. mdbooth -- I'm also wondering if / how it interacts with the QEMU command-line 'werror/rerror' | 09:37 |
| kashyap | werror=action,rerror=action | 09:37 |
| kashyap | Specify which action to take on write and read errors. Valid actions are: "ignore" (ignore the error and try to continue), "stop" (pause QEMU), "report" (report the error | 09:37 |
| kashyap | to the guest), "enospc" (pause QEMU only if the host disk is full; report the error to the guest otherwise). The default setting is werror=enospc and rerror=report. | 09:37 |
| *** gszasz has joined #openstack-nova | 09:37 | |
| * kashyap just pasted that from QEMU man page | 09:37 | |
| mdbooth | kashyap: So, fixing this is going to require a significant reworking of this code, tricky dev setups, completely new unit tests, etc. | 09:38 |
| garyk | danpb: mdbooth: maybe you guys can help out here. i am running stable/newton with 3 kvm nodes and have hit the following: https://bugs.launchpad.net/nova/+bug/1630912 and https://bugs.launchpad.net/nova/+bug/1630918 | 09:38 |
| openstack | Launchpad bug 1630912 in OpenStack Compute (nova) "stable/newton: unable to spin up a kvm instance" [Critical,New] | 09:38 |
| openstack | Launchpad bug 1630918 in OpenStack Compute (nova) "stable/new: resource tracker exception" [Critical,New] | 09:38 |
| kashyap | mdbooth: Yes, probably a spec proposal required for this? | 09:38 |
| mdbooth | kashyap: I was just looking for a way to get the immediate issue off our plate. | 09:38 |
| kashyap | mdbooth: Is that gate-blocking? | 09:38 |
| mdbooth | kashyap: Nah, this is a bug without a wider impact. Spec is total overkill imho. | 09:39 |
| mdbooth | Anyway, the way to do that would be to handle the Ready flag, I think | 09:39 |
| mdbooth | BUT.... | 09:39 |
| mdbooth | libvirt doesn't make it available to us | 09:39 |
| kashyap | Sure. I'm not eager for it. Just checking the criteria | 09:39 |
| mdbooth | It's not just the python bindings, either | 09:39 |
| mdbooth | Look at struct _virDomainBlockJobInfo | 09:40 |
| mdbooth | It's just type, bandwidth, cur, end. | 09:40 |
| *** mvk has joined #openstack-nova | 09:41 | |
| kashyap | mdbooth: Yes, it's not. I just asked on #virt, OFTC to a libvirt dev if they think exposing the 'ready' flag is a reasonable idea | 09:41 |
| mdbooth | So at first glance, I don't think it's possible to fix this at all with the current interface. | 09:41 |
| mdbooth | Except by moving to handling events. | 09:41 |
| kashyap | mdbooth: You know libvirt internally seem to use VIR_DOMAIN_BLOCK_JOB_READY (which is just a translation of QEMU's BLOCK_JOB_READY) I guess | 09:43 |
| mdbooth | And given that we need to 'support' the unsupported Ubuntu virt stack, which will never receive bug fixes, I think we need to JFDI. | 09:43 |
| kashyap | mdbooth: There we go: | 09:44 |
| kashyap | 11:41 < kashyap> mprivozn: When you get a moment, is it reasonble to request libvirt to expose the the "ready" flag (from QMP `query-block-jobs`) via its bindings? | 09:44 |
| kashyap | 11:42 * kashyap should probably bring this up on the mailing list | 09:44 |
| kashyap | 11:42 < mprivozn> kashyap: yeah; but my answer here is that it's not possible with the current API because of ABI stability :( | 09:44 |
| kashyap | 11:43 < mprivozn> kashyap: but please do initiate discussion on the list | 09:44 |
| *** jaosorior_lunch is now known as jaosorior | 09:44 | |
| mdbooth | kashyap: ABI stability is moot when you've got total stability because the stack is unmaintained. | 09:44 |
| mdbooth | I don't see any option other than to switch to the other API | 09:45 |
| mdbooth | garyk: Looking | 09:46 |
| kashyap | mdbooth: Yeah. mdbooth Also: The immediate Nova issue at hand, is it a Gate-blocking one? (I suppose so, if it's enabled in Tempest?) | 09:46 |
| mdbooth | kashyap: Nah, it's a new test, right? | 09:46 |
| mdbooth | So this has always been broken, we're just discovering it now. | 09:47 |
| kashyap | mdbooth: Oh, right. It's a new test, reading Matt R's commit message | 09:47 |
| mdbooth | garyk: Fun! | 09:48 |
| mdbooth | garyk: Are they related, btw? | 09:48 |
| *** aarefiev has quit IRC | 09:50 | |
| kashyap | "04709c94-0ba3-4312-ad1c-5739e5600bf4] IOError: [Errno 13] Permission denied: '/opt/stack/data/nova/instances/locks/nova-273d5645757056cdd056de4cfe9f121b9eee6ae3'" | 09:50 |
| *** thorst_ has joined #openstack-nova | 09:50 | |
| *** ralonsoh_ is now known as ralonsoh | 09:50 | |
| *** roeyc has joined #openstack-nova | 09:50 | |
| *** aarefiev has joined #openstack-nova | 09:51 | |
| *** salv-orl_ has quit IRC | 09:51 | |
| mdbooth | kashyap: Looks like it's in spawn trying to lock presumably the image in the image cache | 09:51 |
| openstackgerrit | jichenjc proposed openstack/nova-specs: Add host-command-with-POST https://review.openstack.org/303388 | 09:52 |
| *** kaisers_ has joined #openstack-nova | 09:54 | |
| *** kaisers_ has quit IRC | 09:55 | |
| *** ducttape_ has joined #openstack-nova | 09:57 | |
| *** thorst_ has quit IRC | 09:58 | |
| *** kaisers has quit IRC | 09:58 | |
| *** kaisers has joined #openstack-nova | 09:58 | |
| *** cdent has joined #openstack-nova | 09:59 | |
| *** Margarita_S has joined #openstack-nova | 09:59 | |
| *** woodster_ has quit IRC | 10:00 | |
| *** kaisers has quit IRC | 10:00 | |
| *** ihrachys has joined #openstack-nova | 10:02 | |
| openstackgerrit | jichenjc proposed openstack/nova-specs: Add host-command-with-POST https://review.openstack.org/303388 | 10:02 |
| *** ducttape_ has quit IRC | 10:02 | |
| *** ociuhandu has joined #openstack-nova | 10:04 | |
| cdent | rpodolyaka: in response to your most recent comment on the placement aggregates: https://github.com/openstack/nova/blob/master/nova/db/sqlalchemy/migrate_repo/versions/318_resource_provider_name_aggregates.py#L33 | 10:06 |
| mdbooth | garyk: Responded to the first with a brain dump | 10:06 |
| cdent | so yeah, messy | 10:06 |
| rpodolyaka | cdent: ah, I see. Thanks! Yeah, kind of messy. But I guess it's fine as long as we do not need to change the charset of existing data in tables | 10:09 |
| cdent | the expectation is that that table will never change since it just mapping id to uuid but ... the future is foggy | 10:10 |
| *** kaisers has joined #openstack-nova | 10:10 | |
| *** Matti_ has quit IRC | 10:10 | |
| *** hoangcx has quit IRC | 10:11 | |
| garyk | mdbooth: thanks! ithink that they are unrelated but in the same env… it will take a while for stable to actaully become stable | 10:12 |
| *** hoangcx has joined #openstack-nova | 10:12 | |
| openstackgerrit | jichenjc proposed openstack/nova-specs: Add host-command-with-POST https://review.openstack.org/303388 | 10:12 |
| mdbooth | garyk: Just taking a quick look at the second. | 10:12 |
| mdbooth | Looks like bdms doesn't contain the instance uuid | 10:12 |
| *** hoangcx has quit IRC | 10:12 | |
| mdbooth | So... we've got an instance with no bdms? Or a race. | 10:12 |
| *** kaisers has quit IRC | 10:14 | |
| mdbooth | So... | 10:14 |
| danpawlik | Hello, pls check https://bugs.launchpad.net/nova/+bug/1630930 and if you have some free time just take a look on other PS https://review.openstack.org/#/c/314928/ | 10:14 |
| openstack | Launchpad bug 1630930 in OpenStack Compute (nova) "Affinity instance slot reservation" [Undecided,New] | 10:14 |
| mdbooth | bdms = objects.BlockDeviceMappingList.bdms_by_instance_uuid( | 10:14 |
| mdbooth | ctx, instance_uuids) | 10:14 |
| mdbooth | bdms[guest.uuid] | 10:14 |
| mdbooth | Can we do indexed access to BlockDeviceMappingList? | 10:15 |
| mdbooth | Ah, nm: base.obj_make_dict_of_lists | 10:17 |
| kashyap | Wow, this has to be the "Month of Race Conditions" or something | 10:19 |
| garyk | mdbooth: i am not attaching any volumes | 10:19 |
| mdbooth | kashyap: Race conditions are always in season. | 10:20 |
| mdbooth | garyk: I think there should be a BDM for at least the root disk | 10:21 |
| mdbooth | garyk: Also, this code has been around a while, so perhaps something changed elsewhere... | 10:22 |
| mdbooth | garyk: Do you have a slave configured? | 10:22 |
| *** salv-orlando has joined #openstack-nova | 10:23 | |
| *** code-R_ has quit IRC | 10:24 | |
| garyk | mdbooth: not aware of any slave configured | 10:24 |
| mdbooth | That looks irrelevant anyway. | 10:24 |
| *** code-R has joined #openstack-nova | 10:24 | |
| mdbooth | Are these running in a CI, btw? | 10:24 |
| mdbooth | Can you link to full logs? | 10:24 |
| *** riccardo_ has joined #openstack-nova | 10:25 | |
| garyk | mdbooth: no, its not in CI. its an internal test enviroment that we build. 3 KVM nodes and one VC node. | 10:26 |
| openstackgerrit | Chris Dent proposed openstack/nova: A demo script for updating resource inventory https://review.openstack.org/382613 | 10:26 |
| mdbooth | garyk: Anything weird going on? Concurrent deletes, maybe? | 10:26 |
| garyk | mdbooth: prblem started when we saw that some KVM instances had 2 or 3 IP's. | 10:26 |
| *** ijw has joined #openstack-nova | 10:26 | |
| garyk | mdbooth: no. also happens when i spin up one VM | 10:27 |
| riccardo_ | hi guys, where in the nova code can I find the place where resource allocation is done? | 10:27 |
| *** kaisers has joined #openstack-nova | 10:27 | |
| garyk | mdbooth: its pretty random. still trying to iron it down | 10:27 |
| mdbooth | garyk: Not the second one though, right? | 10:27 |
| garyk | mdbooth: problem is also when there is a rechedule the instance ports are not cleand up | 10:27 |
| garyk | mdbooth: nop - only one vnic | 10:28 |
| mdbooth | Second bug is in get_available_resource() | 10:28 |
| mdbooth | That's isn't called during spawn(), is it? | 10:28 |
| *** zenoway has quit IRC | 10:29 | |
| *** zenoway has joined #openstack-nova | 10:31 | |
| *** ijw has quit IRC | 10:32 | |
| openstackgerrit | garyk proposed openstack/nova: libvirt: ensure that BDM exists https://review.openstack.org/382833 | 10:32 |
| garyk | mdbooth: correct - that is related to the resource tracker. those exceptions are showing up randomly | 10:32 |
| garyk | i posted something above - still need to figure out how to unit test this | 10:33 |
| mdbooth | garyk: Are you using NFS, btw? | 10:33 |
| garyk | mdbooth: not that i am aware of | 10:35 |
| garyk | it should all be local | 10:35 |
| *** tiendc has quit IRC | 10:35 | |
| mdbooth | Justing wondering about the first thing. If you can get on there in happy hacker mode, might be interesting to get a 'ls -l' from the locks directory immediately before the lock fails. | 10:36 |
| garyk | mdbooth: ok, i will try | 10:36 |
| mdbooth | garyk: To get as close as possible, I'd hack the library :) | 10:37 |
| *** alexchad_ is now known as alexchadin | 10:37 | |
| garyk | mdbooth: my bad on the NFS - the KVM nodes are usning NFS | 10:38 |
| garyk | that may be the root cause. | 10:38 |
| mdbooth | garyk: Well we should still work on NFS, however it is a huge additional vector for bugs and/or misconfiguration. | 10:38 |
| openstackgerrit | Stephen Finucane proposed openstack/nova-specs: Add extra-specs-image-meta-conflicts spec https://review.openstack.org/382382 | 10:38 |
| garyk | mdbooth: yeah, i hear you. | 10:39 |
| *** kaisers has quit IRC | 10:39 | |
| *** kaisers has joined #openstack-nova | 10:40 | |
| garyk | danpb: mdbooth: any change you guys can please look at https://review.openstack.org/140045. this is nearly 2 years in review | 10:40 |
| garyk | johnthetubaguy: ^ | 10:40 |
| garyk | i thought that the vmware drivers had a rough time, but this is taking the piss | 10:40 |
| *** sdague has joined #openstack-nova | 10:41 | |
| garyk | mdbooth: dumb question - when using NFS, is there a way in which we configure nova to be aware of that. i am really not familiar with this | 10:42 |
| mdbooth | garyk: Nova isn't aware of it. It's just a quirky filesystem. | 10:43 |
| garyk | mdbooth: ok, thanks | 10:43 |
| johnthetubaguy | os-vif was blocking that, should be clear to land now that was ironed out last cycle | 10:43 |
| johnthetubaguy | there are plans to make resource tracker "shared storage" aware this cycle | 10:43 |
| *** alexchadin has quit IRC | 10:44 | |
| mdbooth | garyk: Are you trying to spawn multiple instances simultaneously, btw? | 10:45 |
| garyk | mdbooth: no, one at a time | 10:46 |
| garyk | if i shut down all but one nova compute then all works well. when there are 3 running everything goes pear shaped | 10:47 |
| garyk | may be a setup issue | 10:47 |
| mdbooth | garyk: I was just thinking around concurrent access to the locks directory | 10:47 |
| mdbooth | But again, we support that, so if there was anything architecturally broken around it I'd expect it to have come up before. | 10:48 |
| garyk | agree | 10:48 |
| mdbooth | garyk: Does it work with any single nova compute, or 1 specific one? | 10:48 |
| garyk | mdbooth: it works with a single nova compute | 10:48 |
| *** tbachman has quit IRC | 10:48 | |
| mdbooth | So I'd guess probably a misconfiguration on the other 2? | 10:49 |
| sfinucan | sahid: Any reason not to resuscitate this patch? https://review.openstack.org/#/c/172079/ | 10:49 |
| openstackgerrit | Feodor Tersin proposed openstack/nova: libvirt: Improve _is_booted_from_volume implementation https://review.openstack.org/382024 | 10:49 |
| openstackgerrit | Feodor Tersin proposed openstack/nova: libvirt: Remove unused disk_info parameter https://review.openstack.org/382025 | 10:49 |
| mdbooth | Or on the NFS host for the other 2. | 10:49 |
| garyk | i just terminated 2 nova computes and it works | 10:50 |
| garyk | when all 3 are running it does not | 10:50 |
| johnthetubaguy | is that the image cache going crazy? not sure how you configure that on NFS | 10:50 |
| mdbooth | kashyap: Incidentally, these docs: | 10:50 |
| mdbooth | http://libvirt.org/html/libvirt-libvirt-domain.html#virDomainBlockRebase | 10:50 |
| mdbooth | kashyap: Say to do exactly what we're doing | 10:50 |
| mdbooth | kashyap: So if the outcome is we need to do something else, we also need to fix the docs. | 10:50 |
| *** Pavo has quit IRC | 10:51 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova-specs: Enable cold migration with target host https://review.openstack.org/334286 | 10:51 |
| *** slaweq has quit IRC | 10:52 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova-specs: Enable cold migration with target host https://review.openstack.org/334286 | 10:53 |
| *** takashin has joined #openstack-nova | 10:53 | |
| *** pjm6 has quit IRC | 10:54 | |
| *** pjm6 has joined #openstack-nova | 10:54 | |
| *** thorst_ has joined #openstack-nova | 10:54 | |
| *** kaisers has quit IRC | 10:56 | |
| *** Pavo has joined #openstack-nova | 10:56 | |
| *** mjura has joined #openstack-nova | 10:57 | |
| openstackgerrit | Emilien Macchi proposed openstack/nova: Document experimental pipeline in Nova CI https://review.openstack.org/381838 | 10:58 |
| *** ducttape_ has joined #openstack-nova | 10:58 | |
| *** thorst_ has quit IRC | 11:02 | |
| *** ducttape_ has quit IRC | 11:03 | |
| *** mjura has quit IRC | 11:04 | |
| *** hshiina has joined #openstack-nova | 11:05 | |
| kashyap | mdbooth: (Was AFK for lunch) True, the API docs ought to be fixed, if the outcome is diff. | 11:05 |
| *** roeyc has quit IRC | 11:06 | |
| *** rtheis has joined #openstack-nova | 11:08 | |
| *** roeyc has joined #openstack-nova | 11:09 | |
| *** paul-carlton2 has joined #openstack-nova | 11:10 | |
| *** Margarita_S has quit IRC | 11:10 | |
| openstackgerrit | Gergely Csatari proposed openstack/nova: Checking the parameters of servers-actions.inc https://review.openstack.org/327112 | 11:10 |
| *** code-R has quit IRC | 11:13 | |
| openstackgerrit | Paul Carlton proposed openstack/nova: Update nova network info when doing rebuild for evacuate operation https://review.openstack.org/382853 | 11:13 |
| *** bvanhav has joined #openstack-nova | 11:16 | |
| *** anthonyper has quit IRC | 11:17 | |
| *** anthonyper has joined #openstack-nova | 11:17 | |
| *** mjura has joined #openstack-nova | 11:17 | |
| *** ralonsoh has quit IRC | 11:17 | |
| *** ralonsoh has joined #openstack-nova | 11:18 | |
| *** ftersin has quit IRC | 11:20 | |
| *** ftersin has joined #openstack-nova | 11:21 | |
| *** code-R has joined #openstack-nova | 11:26 | |
| *** ijw has joined #openstack-nova | 11:28 | |
| openstackgerrit | JABAR ALI proposed openstack/python-novaclient: Chnages for bug 1613105 https://review.openstack.org/382864 | 11:31 |
| openstack | bug 1613105 in python-novaclient "functional tests leak cinder volumes" [Undecided,Confirmed] https://launchpad.net/bugs/1613105 - Assigned to JABAR ALI (jabbar-ali) | 11:31 |
| *** baoli has joined #openstack-nova | 11:31 | |
| *** hussain_c has joined #openstack-nova | 11:32 | |
| *** code-R has quit IRC | 11:32 | |
| *** nkrinner has quit IRC | 11:34 | |
| *** ijw has quit IRC | 11:34 | |
| *** ducnc has quit IRC | 11:34 | |
| *** phuongnh has quit IRC | 11:37 | |
| sahid | sfinucan: it's probably a valid patch, just leak of reviews | 11:37 |
| sahid | let me try to resusciatete it | 11:38 |
| *** claudiub has quit IRC | 11:45 | |
| *** thorst_ has joined #openstack-nova | 11:45 | |
| *** nkrinner has joined #openstack-nova | 11:46 | |
| *** alexchadin has joined #openstack-nova | 11:49 | |
| *** gcb has joined #openstack-nova | 11:49 | |
| *** ociuhandu has quit IRC | 11:52 | |
| *** rfolco has joined #openstack-nova | 11:55 | |
| *** ratailor has quit IRC | 11:56 | |
| *** Margarita_S has joined #openstack-nova | 11:58 | |
| *** ducttape_ has joined #openstack-nova | 11:59 | |
| *** thorst__ has joined #openstack-nova | 12:00 | |
| *** prateek has quit IRC | 12:02 | |
| cdent | johnthetubaguy, sdague, jaypipes : a series of bug fixes on placement welcome your attention if you've got the time: https://review.openstack.org/#/c/374800/ https://review.openstack.org/#/c/374833/ https://review.openstack.org/#/c/374870/ | 12:02 |
| *** gouthamr has joined #openstack-nova | 12:03 | |
| *** ducttape_ has quit IRC | 12:04 | |
| *** thorst_ has quit IRC | 12:04 | |
| *** tuanluong has quit IRC | 12:04 | |
| *** thorst__ has quit IRC | 12:05 | |
| *** thorst_ has joined #openstack-nova | 12:06 | |
| *** lucas-afk is now known as lucasagomes | 12:07 | |
| *** thorst__ has joined #openstack-nova | 12:07 | |
| *** mgoddard_ has joined #openstack-nova | 12:08 | |
| *** thorst_ has quit IRC | 12:11 | |
| *** ducttape_ has joined #openstack-nova | 12:11 | |
| *** mgoddard has quit IRC | 12:11 | |
| *** alex_xu has quit IRC | 12:12 | |
| *** thorst__ has quit IRC | 12:12 | |
| *** thorst_ has joined #openstack-nova | 12:13 | |
| *** thorst__ has joined #openstack-nova | 12:14 | |
| cdent | thanks sdague | 12:14 |
| *** alex_xu has joined #openstack-nova | 12:14 | |
| *** mgoddard_ has quit IRC | 12:15 | |
| *** thorst_ has quit IRC | 12:17 | |
| sahid | we have an "easy" fix here related to live migrate instances with ephemeral disks https://review.openstack.org/#/c/378560/ | 12:18 |
| openstackgerrit | Gábor Antal proposed openstack/nova: Removed unnecessary parantheses in return statements https://review.openstack.org/370549 | 12:19 |
| sahid | it has already been +2ed so if we can have the final +2+W it would be nice | 12:19 |
| *** tblakes has joined #openstack-nova | 12:19 | |
| *** thorst__ has quit IRC | 12:19 | |
| *** Jeffrey4l_ has quit IRC | 12:20 | |
| *** Jeffrey4l_ has joined #openstack-nova | 12:21 | |
| *** prateek has joined #openstack-nova | 12:21 | |
| *** mgoddard has joined #openstack-nova | 12:22 | |
| *** kaisers has joined #openstack-nova | 12:24 | |
| *** salv-orlando has quit IRC | 12:24 | |
| *** lamt has quit IRC | 12:25 | |
| *** maticue has joined #openstack-nova | 12:25 | |
| *** edmondsw has joined #openstack-nova | 12:26 | |
| *** omnipresent has joined #openstack-nova | 12:27 | |
| *** zenoway has quit IRC | 12:27 | |
| *** alex_xu has quit IRC | 12:27 | |
| *** esberglu has joined #openstack-nova | 12:29 | |
| *** alex_xu has joined #openstack-nova | 12:29 | |
| *** ijw has joined #openstack-nova | 12:29 | |
| *** ducttape_ has quit IRC | 12:30 | |
| *** markvoelker has joined #openstack-nova | 12:31 | |
| *** code-R has joined #openstack-nova | 12:34 | |
| garyk | mdbooth: seems like the file permission on the directory were screwed up. | 12:34 |
| garyk | when changing them manually to the devstack user then everthing worked | 12:34 |
| mdbooth | garyk: On the locks directory itself? | 12:34 |
| garyk | mdbooth: yeah. | 12:34 |
| mdbooth | Wonder why it worked on one node but not the others | 12:34 |
| garyk | just not sure why they were messed up in the first place | 12:34 |
| mdbooth | NFS thing? | 12:35 |
| mdbooth | What mode did you have to set? | 12:35 |
| *** ijw has quit IRC | 12:35 | |
| mdbooth | If it was 777 you might have a uid mapping issue | 12:35 |
| garyk | not sure. what is concerning is that when there are failures that nova does not free the correct neutron resources. that is the real bug | 12:35 |
| garyk | i will try and look into that as that is very concerning | 12:35 |
| garyk | let me check | 12:36 |
| *** diana-test has joined #openstack-nova | 12:37 | |
| garyk | mdbooth: it 755 | 12:37 |
| *** diana-test has quit IRC | 12:40 | |
| mdbooth | garyk: So interestingly I can't see what creates the locks directory | 12:40 |
| mdbooth | I wonder if it has be created manually... | 12:40 |
| *** peter-hamilton has joined #openstack-nova | 12:40 | |
| kashyap | mdbooth: When you about later, turns out that 'ready' state is already exposed via guest XML...See Peter's response on that upstream thread. | 12:41 |
| openstackgerrit | Sivasathurappan Radhakrishnan proposed openstack/nova: Added CLI args for nova-policy check command https://review.openstack.org/382915 | 12:41 |
| *** jaosorior has quit IRC | 12:41 | |
| sdague | sahid: can you pull that comment outside of the parameter list and put it before the function? That's really odd style as it is, and makes it a little hard to realize all those parameters are part of the same function | 12:41 |
| sdague | the fix seems fine otherwise | 12:42 |
| *** omnipresent has quit IRC | 12:42 | |
| sahid | sdague: yes sure, let me do that and thanks to have reviewed it | 12:42 |
| *** zenoway has joined #openstack-nova | 12:42 | |
| garyk | mdbooth: no idea. i am trying to figure out where this is done… | 12:43 |
| *** ducttape_ has joined #openstack-nova | 12:43 | |
| *** anush has joined #openstack-nova | 12:45 | |
| *** zenoway has quit IRC | 12:47 | |
| *** ducttape_ has quit IRC | 12:47 | |
| *** diana-test has joined #openstack-nova | 12:49 | |
| *** code-R_ has joined #openstack-nova | 12:49 | |
| *** links has quit IRC | 12:50 | |
| *** Pavo has quit IRC | 12:51 | |
| *** zenoway has joined #openstack-nova | 12:51 | |
| *** diana-test has quit IRC | 12:51 | |
| *** kaisers has quit IRC | 12:51 | |
| *** diana-test has joined #openstack-nova | 12:52 | |
| *** zenoway has quit IRC | 12:52 | |
| *** zenoway has joined #openstack-nova | 12:52 | |
| *** code-R has quit IRC | 12:53 | |
| openstackgerrit | sahid proposed openstack/nova: libvirt: fix DiskSmallerThanImage when block migrate ephemerals https://review.openstack.org/378560 | 12:53 |
| *** diana-test has quit IRC | 12:55 | |
| *** Pavo has joined #openstack-nova | 12:55 | |
| *** liverpooler has quit IRC | 12:56 | |
| *** diana-test has joined #openstack-nova | 12:56 | |
| *** anush has quit IRC | 12:56 | |
| *** diana-test has quit IRC | 12:56 | |
| *** ociuhandu has joined #openstack-nova | 12:56 | |
| *** zenoway has quit IRC | 12:57 | |
| *** burgerk has joined #openstack-nova | 12:57 | |
| *** ociuhandu has quit IRC | 12:59 | |
| *** garyk has quit IRC | 13:02 | |
| *** jheroux has joined #openstack-nova | 13:02 | |
| *** thorst_ has joined #openstack-nova | 13:02 | |
| *** alexchadin has quit IRC | 13:02 | |
| *** ociuhandu has joined #openstack-nova | 13:04 | |
| *** dave-mccowan has joined #openstack-nova | 13:04 | |
| *** mdrabe has joined #openstack-nova | 13:04 | |
| *** ayoung has joined #openstack-nova | 13:04 | |
| *** moshele has quit IRC | 13:04 | |
| *** moshele has joined #openstack-nova | 13:05 | |
| *** links has joined #openstack-nova | 13:05 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova-specs: Enable cold migration with target host https://review.openstack.org/334286 | 13:05 |
| *** namnh has joined #openstack-nova | 13:05 | |
| *** ducttape_ has joined #openstack-nova | 13:06 | |
| *** vishwanathj has joined #openstack-nova | 13:07 | |
| *** tbachman has joined #openstack-nova | 13:07 | |
| *** agireud has quit IRC | 13:10 | |
| *** tongli has joined #openstack-nova | 13:11 | |
| *** whenry has joined #openstack-nova | 13:11 | |
| *** roeyc has quit IRC | 13:11 | |
| namnh | Hi everyone, I am trying to upgrade Nova from Mitaka to Newton, I finished all step but in the final step: (nova-manage db online_data_migrations). It had error like this: ERROR nova.db.sqlalchemy.api [req-8328271d-7d1c-4aca-bbcd-5651cccba6ca None None] Data migrations for PciDevice are not safe, likely becausest version . do you encounter this bug? | 13:13 |
| *** Dinesh_Bhor has quit IRC | 13:13 | |
| *** gcb has quit IRC | 13:15 | |
| *** hferenc_ has joined #openstack-nova | 13:16 | |
| *** hemanthm|away is now known as hemanthm | 13:16 | |
| *** vladikr has joined #openstack-nova | 13:16 | |
| *** mriedem has joined #openstack-nova | 13:16 | |
| *** roeyc has joined #openstack-nova | 13:17 | |
| *** agireud has joined #openstack-nova | 13:18 | |
| namnh | avolkov, Hi Mr.Volkow, are you free? | 13:19 |
| avolkov | namnh: hi, yes | 13:20 |
| edleafe | namnh: Yes, that was reported yesterday: https://bugs.launchpad.net/nova/+bug/1630732 | 13:21 |
| openstack | Launchpad bug 1630732 in OpenStack Compute (nova) newton "online_data_migrations does not ignore deleted services" [High,In progress] - Assigned to Dan Smith (danms) | 13:21 |
| namnh | avolkov, thanks for your reply. :) Could you please see above my message. | 13:22 |
| namnh | avolkov, I saw a patch set which you were co-author: https://review.openstack.org/#/c/290793/ | 13:23 |
| openstackgerrit | Drew Thorstensen (thorst) proposed openstack/nova-specs: PowerVM Initial Driver Work https://review.openstack.org/381772 | 13:23 |
| namnh | avolkov, oh. you mean that, currenlty this is a bug | 13:23 |
| *** prateek has quit IRC | 13:25 | |
| *** prateek has joined #openstack-nova | 13:25 | |
| *** moshele has quit IRC | 13:26 | |
| *** mkoderer has joined #openstack-nova | 13:27 | |
| *** ducttape_ has quit IRC | 13:28 | |
| *** haplo37_ has quit IRC | 13:28 | |
| *** riccardo_ has quit IRC | 13:29 | |
| *** zul has quit IRC | 13:30 | |
| *** haplo37_ has joined #openstack-nova | 13:30 | |
| *** ijw has joined #openstack-nova | 13:31 | |
| markus_z | mriedem: Would you remove the -2 from https://review.openstack.org/#/c/323765/ please? I'm moving the spec to Ocata with https://review.openstack.org/#/c/382526/ | 13:31 |
| *** baoli_ has joined #openstack-nova | 13:32 | |
| mriedem | markus_z: i will after the bp is approved | 13:33 |
| *** zul has joined #openstack-nova | 13:33 | |
| openstackgerrit | Julian proposed openstack/nova: Transform instance.power_off notifications https://review.openstack.org/330761 | 13:34 |
| mriedem | markus_z: commented on the spec patch | 13:34 |
| markus_z | It was approved for Newton but got blocked due to a qemu bug. The qemu bug is fixed (I'm currently testing that). | 13:34 |
| openstackgerrit | Julian proposed openstack/nova: Transform instance.snapshot notifications https://review.openstack.org/330128 | 13:34 |
| *** baoli has quit IRC | 13:35 | |
| openstackgerrit | Maciej Szankin proposed openstack/nova: config options: improve libvirt imagebackend https://review.openstack.org/378507 | 13:35 |
| markus_z | thanks for the link to the process, will do | 13:35 |
| *** bnemec has joined #openstack-nova | 13:36 | |
| openstackgerrit | Maciej Szankin proposed openstack/nova: Improve help text for network options. https://review.openstack.org/341203 | 13:36 |
| *** ijw has quit IRC | 13:37 | |
| *** vladikr has quit IRC | 13:38 | |
| *** vladikr has joined #openstack-nova | 13:38 | |
| openstackgerrit | Maciej Szankin proposed openstack/nova: Transform instance.reboot notifications https://review.openstack.org/382959 | 13:41 |
| *** prateek_ has joined #openstack-nova | 13:41 | |
| *** mlavalle has joined #openstack-nova | 13:42 | |
| *** prateek has quit IRC | 13:43 | |
| *** burgerk has quit IRC | 13:45 | |
| *** thorst__ has joined #openstack-nova | 13:46 | |
| *** eharney has joined #openstack-nova | 13:47 | |
| *** roeyc has quit IRC | 13:48 | |
| *** akshai has joined #openstack-nova | 13:49 | |
| *** claudiub has joined #openstack-nova | 13:49 | |
| *** thorst_ has quit IRC | 13:49 | |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova-specs: Libvirt: Use the virtlogd deamon https://review.openstack.org/382526 | 13:50 |
| openstackgerrit | Takashi NATSUME proposed openstack/nova-specs: Check the destination host in cold migration https://review.openstack.org/334725 | 13:50 |
| *** cmartin has joined #openstack-nova | 13:50 | |
| *** akshai_ has joined #openstack-nova | 13:50 | |
| *** hshiina has quit IRC | 13:51 | |
| mriedem | nova meetingathon in 7 minutes | 13:52 |
| *** GB21 has quit IRC | 13:53 | |
| claudiub | danpb: hello. if you have some spare time, can you take a look at the hyper-v ovs vif plug thingy? Thanks! https://review.openstack.org/#/c/140045/ | 13:54 |
| *** akshai has quit IRC | 13:54 | |
| *** mtanino has joined #openstack-nova | 13:56 | |
| openstackgerrit | Stephen Finucane proposed openstack/nova: live-mig: Add claims and proper resource tracking https://review.openstack.org/244489 | 13:57 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: libvirt: pass libvirt reference methods to migration https://review.openstack.org/328997 | 13:57 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: libvirt: live-migrate updates NUMA and cpus in the XML https://review.openstack.org/286744 | 13:57 |
| jroll | mriedem: some folks are interested in a x-p meetup with nova on line 66 here https://etherpad.openstack.org/p/ironic-ocata-summit | 13:57 |
| jroll | think we can get some of y'all on wednesday evening when there's no sessions? | 13:57 |
| jroll | no nova sessions* | 13:57 |
| *** ducttape_ has joined #openstack-nova | 13:58 | |
| mriedem | jroll: we have a 5pm wed nova/neutron session in the neutron room | 13:58 |
| mriedem | is there a slot before that? | 13:59 |
| jroll | mriedem: we have a 5:55-6:35 | 13:59 |
| dansmith | oof | 13:59 |
| jroll | ikr | 13:59 |
| *** thorst__ is now known as thorst_ | 13:59 | |
| jroll | it affects api and probably placement so I expect everybody will have opinions | 13:59 |
| *** tonytan4ever has joined #openstack-nova | 13:59 | |
| mriedem | 6pm is tapas time | 13:59 |
| jroll | not for me :( | 13:59 |
| *** sdake has joined #openstack-nova | 13:59 | |
| jroll | I'll bug you in the meeting and see what we can figure out | 14:00 |
| mriedem | ok | 14:00 |
| mriedem | if it's placement related i'd want dansmith and jaypipes there | 14:00 |
| *** anush has joined #openstack-nova | 14:01 | |
| *** bvanhav has quit IRC | 14:01 | |
| *** bvanhav has joined #openstack-nova | 14:03 | |
| mriedem | mdbooth: when do you head out from the summit? | 14:04 |
| mdbooth | mriedem: I saw that comment. Friday mid-afternoon. | 14:04 |
| mdbooth | That slot is fine with me. | 14:04 |
| mriedem | 11am is ok? good. | 14:04 |
| mriedem | if that doesn't work we can move it | 14:04 |
| namnh | edleafe, Wow, thanks for your information, it's really useful to me. Now I have to go out my office. Good bye. | 14:05 |
| mdbooth | mriedem: Nope, that's fine. | 14:05 |
| *** namnh has quit IRC | 14:06 | |
| *** tonytan4ever has quit IRC | 14:06 | |
| openstackgerrit | Andrew Laski proposed openstack/nova-specs: Proposal for cells v2 scheduling interaction https://review.openstack.org/381275 | 14:06 |
| *** tonytan4ever has joined #openstack-nova | 14:06 | |
| *** bvanhav_ has joined #openstack-nova | 14:07 | |
| openstackgerrit | Lee Yarwood proposed openstack/nova: Do not reserve device before reserving the volume https://review.openstack.org/290793 | 14:07 |
| *** READ10 has joined #openstack-nova | 14:08 | |
| *** bvanhav has quit IRC | 14:08 | |
| openstackgerrit | sahid proposed openstack/nova: virt: remove small pages request https://review.openstack.org/172079 | 14:09 |
| *** anush has quit IRC | 14:10 | |
| *** zhenguo has joined #openstack-nova | 14:12 | |
| *** burgerk has joined #openstack-nova | 14:13 | |
| *** kaisers has joined #openstack-nova | 14:14 | |
| *** prometheanfire has left #openstack-nova | 14:14 | |
| *** sdake has quit IRC | 14:15 | |
| *** rama_y has joined #openstack-nova | 14:15 | |
| *** pbandark has quit IRC | 14:17 | |
| *** bvanhav_ is now known as bvanhav | 14:18 | |
| *** moshele has joined #openstack-nova | 14:19 | |
| *** jamesdenton has joined #openstack-nova | 14:22 | |
| *** parora has joined #openstack-nova | 14:22 | |
| *** liverpooler has joined #openstack-nova | 14:23 | |
| *** zenoway has joined #openstack-nova | 14:24 | |
| *** jlanoux has quit IRC | 14:25 | |
| *** prateek_ has quit IRC | 14:25 | |
| openstackgerrit | Maciej Szankin proposed openstack/nova: Remove config option snapshot_name_template https://review.openstack.org/371203 | 14:27 |
| openstackgerrit | Drew Thorstensen (thorst) proposed openstack/nova-specs: PowerVM Initial Driver Work https://review.openstack.org/381772 | 14:28 |
| *** zenoway has quit IRC | 14:28 | |
| *** gouthamr has quit IRC | 14:29 | |
| *** mgoddard_ has joined #openstack-nova | 14:29 | |
| *** gouthamr has joined #openstack-nova | 14:31 | |
| *** ekuris has quit IRC | 14:31 | |
| *** mgoddard has quit IRC | 14:31 | |
| *** ijw has joined #openstack-nova | 14:33 | |
| *** tanvirt has joined #openstack-nova | 14:34 | |
| sdague | dansmith: http://logs.openstack.org/60/378560/4/check/gate-tempest-dsvm-neutron-full-ubuntu-xenial/b18e638/logs/devstacklog.txt.gz#_2016-10-06_13_19_15_279 - is kind of odd | 14:35 |
| *** READ10 has quit IRC | 14:35 | |
| sdague | that seems to be happening in regular gate runs | 14:35 |
| dansmith | sdague: that's regular db sync complaining about the api db? | 14:36 |
| alaski | looks like that | 14:36 |
| *** anush has joined #openstack-nova | 14:36 | |
| dansmith | sdague: I wonder if some change recently in a decorator or something is initializing the api db connection early or something | 14:37 |
| dansmith | but it also looks unfatal? | 14:38 |
| dansmith | really strange | 14:38 |
| *** ijw has quit IRC | 14:38 | |
| sdague | dansmith: it's unfatal | 14:39 |
| sdague | but ugly | 14:39 |
| sdague | and scary | 14:39 |
| dansmith | sdague: I know, but I'm trying to figure out why it would be doing that at all, and why it does't seem to care that it failed | 14:39 |
| *** zenoway has joined #openstack-nova | 14:40 | |
| *** Hosam has joined #openstack-nova | 14:40 | |
| dansmith | I imagine this is the thread to follow:" ERROR: could not access cell mapping database - has api db been created? | 14:40 |
| *** Hosam_ has joined #openstack-nova | 14:40 | |
| *** tanvirt has quit IRC | 14:41 | |
| openstackgerrit | John Garbutt proposed openstack/nova-specs: WIP: API: rework security group APIs https://review.openstack.org/382414 | 14:41 |
| dansmith | ah yeah | 14:41 |
| alaski | it's looking up a cellmapping | 14:41 |
| alaski | since syncing cell0 was added | 14:41 |
| *** sridharg has quit IRC | 14:41 | |
| *** annegentle has joined #openstack-nova | 14:42 | |
| alaski | and the nova_api db is created by devstack right after running the cell db sync | 14:42 |
| *** anush has quit IRC | 14:43 | |
| dansmith | alaski: so with that ordering we're just ignoring the mapping and going straight to the config'd db I guess? | 14:43 |
| dansmith | not sure I see why we're getting a stack trace though | 14:44 |
| alaski | yes. thge cell0 db never gets synced | 14:44 |
| mriedem | danpb: can you jump into the nova meeting? | 14:44 |
| *** Hosam has quit IRC | 14:44 | |
| danpb | what channel ? | 14:44 |
| mriedem | #openstack-meeting | 14:45 |
| dansmith | alaski: I guess the decorator is logging that traceback for us | 14:45 |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova-specs: Add spec for lower case metadata keys only https://review.openstack.org/350843 | 14:45 |
| auggy | sdague: thanks for your input on the lc-metadata spec, i've just pushed an update to address some of the comments | 14:45 |
| auggy | and to give mtreinish credit for his help ;) | 14:45 |
| sdague | auggy: great, thanks | 14:45 |
| alaski | dansmith: the db context manager? | 14:45 |
| dansmith | alaski: yeah | 14:46 |
| *** annegentle has quit IRC | 14:47 | |
| mtreinish | auggy: hah, it's just a single patch. You didn't have to rev the spec for that | 14:47 |
| dansmith | alaski: so I guess we need to reverse those orderings.. we could run it again after api, but we won't be able to avoid that error log I suppose | 14:47 |
| *** kaisers has quit IRC | 14:47 | |
| auggy | mtreinish: :P it was a big help | 14:48 |
| alaski | dansmith: we would need to sync the api db first as well or it will fail on the table lookup | 14:48 |
| dansmith | alaski: yeah I meant sync cell0 after we're done with the api, but... | 14:49 |
| dansmith | alaski: we can just reverse the ordering and do api first I assume right? | 14:49 |
| alaski | dansmith: the enginefacade code is horrible to try to trace, but I'm not seeing where it would log the exception yet | 14:49 |
| dansmith | alaski: just in the cell mapping lookup | 14:49 |
| *** lpetrut has quit IRC | 14:49 | |
| dansmith | alaski: I know, the traceback is useless and I don't know why | 14:49 |
| *** rrecio has joined #openstack-nova | 14:49 | |
| alaski | dansmith: creating and syncing the api db first should be fine | 14:50 |
| dansmith | alaski: I think we never even run our db code, because when we enter the decorator it tries to connect and start a transaction against api_db, fails, and we never run our code | 14:50 |
| dansmith | alaski: okay | 14:50 |
| *** Pavo has quit IRC | 14:51 | |
| dansmith | alaski: sdague I'd like to make this less of an explosion for people that happen to reverse these steps as well, but I'm not sure there is a practical way to do that because of how enginefacade works | 14:51 |
| openstackgerrit | Gábor Antal proposed openstack/nova: Removed unnecessary parantheses in return statements https://review.openstack.org/370549 | 14:51 |
| sdague | dansmith: that makes me kind of a sad panda if that's true | 14:51 |
| alaski | ahh, I found the log line. still not sure how it gets hit yet | 14:52 |
| sdague | dansmith: I guess we could register a high priority bug against oslo.db for that and see if that team can take a look? | 14:52 |
| *** rrecio_ has joined #openstack-nova | 14:52 | |
| alaski | http://git.openstack.org/cgit/openstack/oslo.db/tree/oslo_db/sqlalchemy/exc_filters.py#n373 | 14:53 |
| dansmith | sdague: it's kinda fundamental | 14:53 |
| dansmith | sdague: unless we decorate one call with a "don't log errors please" thing and just handle this path because of how we use it | 14:53 |
| *** annegentle has joined #openstack-nova | 14:53 | |
| alaski | it's also due to our attempt to implicitly do the right thing here without asking much of users | 14:54 |
| *** erlon has quit IRC | 14:54 | |
| dansmith | sdague: we could try to connect to api_db ourselves quickly to validate that it's been setup and then not call this if we think it's going to fail | 14:54 |
| *** ildikov has quit IRC | 14:54 | |
| alaski | we're really calling the sync method incorrectly since the db isn't setup | 14:54 |
| dansmith | sdague: but that's kinda going round | 14:54 |
| *** rrecio has quit IRC | 14:54 | |
| *** armax has joined #openstack-nova | 14:54 | |
| dansmith | alaski: which is why hiding the error here doesn't make sense normally, only here because we're in manage | 14:54 |
| alaski | yeah. I agree with that | 14:55 |
| *** DuncanT has quit IRC | 14:55 | |
| sdague | right, I think this is one of the cases where the error handling all being buried makes for bad ops UX, and in this specific case we should be able to not stack trace | 14:55 |
| *** mkoderer has quit IRC | 14:55 | |
| *** geekinutah has quit IRC | 14:55 | |
| *** anush has joined #openstack-nova | 14:55 | |
| dansmith | sdague: well, it probably makes for good ops UX normally, just not right here | 14:55 |
| sdague | dansmith: yeh, is there any existing way to disable logging for calls? | 14:55 |
| *** jlanoux has joined #openstack-nova | 14:55 | |
| dansmith | sdague: this is a special case | 14:55 |
| *** csatari has quit IRC | 14:55 | |
| *** macsz has quit IRC | 14:55 | |
| *** boris-42 has quit IRC | 14:55 | |
| *** dougwig has quit IRC | 14:55 | |
| *** cargonza has quit IRC | 14:55 | |
| *** ediardo has quit IRC | 14:55 | |
| *** andrewbogott has quit IRC | 14:55 | |
| *** coolsvap has quit IRC | 14:55 | |
| *** patrickeast has quit IRC | 14:55 | |
| sdague | dansmith: yeh | 14:55 |
| *** tomhambleton has quit IRC | 14:55 | |
| *** fyxim has quit IRC | 14:55 | |
| *** Pavo has joined #openstack-nova | 14:55 | |
| dansmith | sdague: well, that's what I mean.. we *could* ask them for a special flag, it just seems kinda like a weird request | 14:56 |
| dansmith | a "disable logging in this one method" flag | 14:56 |
| *** hieulq__ has joined #openstack-nova | 14:56 | |
| dansmith | but then again, it would disable logging for this path at runtime, which we don't want | 14:56 |
| dansmith | unless we have some way to toggle | 14:56 |
| dansmith | I dunno | 14:56 |
| *** tanvirt has joined #openstack-nova | 14:56 | |
| dansmith | I have to be on a call in four minutes | 14:57 |
| johnthetubaguy | mriedem: I was planning on attending the PTG | 14:57 |
| mriedem | ok | 14:57 |
| sdague | dansmith: yeh, it seems like we honestly would like to have a context manager that shushs the logging during a block where we call this | 14:58 |
| dansmith | mriedem: I expect I will be there | 14:58 |
| sdague | and returns to normal outside of it | 14:58 |
| sdague | mriedem: I expect to be at the PTG | 14:58 |
| dansmith | sdague: yeah, but.. I guess.. | 14:58 |
| dansmith | sdague: we could just disable error logging before this call and restore the level after :) | 14:58 |
| sdague | dansmith: we could | 14:58 |
| *** haplo37 has quit IRC | 14:58 | |
| *** kaisers has joined #openstack-nova | 14:59 | |
| sdague | ok, how about I file a bug, and we try to see what oslo.db folks have to say before completely working around it on the nova side :) | 14:59 |
| *** salv-orlando has joined #openstack-nova | 14:59 | |
| *** anush has quit IRC | 14:59 | |
| *** anush_ is now known as anush | 15:00 | |
| dansmith | sdague: sure, you do that :) | 15:00 |
| *** anush_ has joined #openstack-nova | 15:00 | |
| *** takashin has left #openstack-nova | 15:00 | |
| alaski | mriedem: I expect I'll be at the PTG. I can actually drive to this one | 15:00 |
| *** haplo37 has joined #openstack-nova | 15:01 | |
| mriedem | hferenc: so i think the image meta / flavor extra specs format thing is a known issue | 15:01 |
| mriedem | hferenc: as i said in the meeting, melwitt recently triaged a similar bug that turned out to be a problem in glance | 15:01 |
| mriedem | hferenc: so she probably has more details | 15:01 |
| edleafe | jaypipes: dansmith: bauzas: quick reminder that you need to update the etherpad with the long-term objective and path to get there for the placement DB: https://etherpad.openstack.org/p/placement-optional-db-spec | 15:01 |
| sfinucan | mriedem: hferenc I don't think the issue melwitt was looking at was quite the same | 15:02 |
| melwitt | I was looking at a mismatch between what nova uses and what glance uses for certain image meta keys | 15:03 |
| *** mjura has quit IRC | 15:03 | |
| melwitt | https://bugs.launchpad.net/glance/+bug/1568191 | 15:03 |
| openstack | Launchpad bug 1568191 in Glance "Pre-defined VirtCPUTopology metadata should have the same name as flavor extra specs in nova" [Medium,In progress] - Assigned to Ying Zuo (yingzuo) | 15:03 |
| *** bdeschenes has joined #openstack-nova | 15:03 | |
| hferenc_ | i see | 15:04 |
| kashyap | mdbooth: Not sure if you're following that upstream thread (both on openstack-dev & libvir-list), Eric made interesting replies, and even responds to your concern of "Why should libvirt munge the reporting" | 15:04 |
| mriedem | hferenc: as far as i can tell, the image metadata format is os_foo_bar | 15:05 |
| mriedem | with the os_ prefix | 15:05 |
| sdague | https://bugs.launchpad.net/oslo.db/+bug/1631033 | 15:05 |
| openstack | Launchpad bug 1631033 in oslo.db "scary traceback if nova-manage db sync is run before nova-manage api_db sync" [Undecided,New] | 15:05 |
| mriedem | flavor extra specs is os:foo_bar | 15:05 |
| mriedem | with the os: prefix | 15:05 |
| mdbooth | kashyap: Yes, I saw that. I don't disagree with him as I said earlier, but it's messy. | 15:05 |
| mriedem | e.g. https://review.openstack.org/#/c/209581/49/releasenotes/notes/hyperv-uefi-secure-boot-a2a617ac2c313afd.yaml | 15:05 |
| mdbooth | 'uncomfortable' rather than 'opposed' | 15:05 |
| *** oanson has quit IRC | 15:05 | |
| mdbooth | Still better to use the other api, I think | 15:05 |
| *** diogogmt has joined #openstack-nova | 15:06 | |
| hferenc_ | mriedem yeah, the strange thing is that horizon uses os_foo_bar even for flavor extra specs | 15:06 |
| openstackgerrit | Takashi NATSUME proposed openstack/nova: Stop using mox stubs in nova/tests/unit/cells https://review.openstack.org/323287 | 15:08 |
| openstackgerrit | Takashi NATSUME proposed openstack/nova: Remove mox in test_block_device_mapping_v1.py https://review.openstack.org/308746 | 15:08 |
| melwitt | I think that's okay as long as the key (after the prefix) matches what's in nova. because IIRC nova strips the prefixes off before accessing keys | 15:08 |
| *** marst has quit IRC | 15:08 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova: api-ref: Fix wrong parameters in os-volumes.inc https://review.openstack.org/347112 | 15:08 |
| openstackgerrit | Takashi NATSUME proposed openstack/nova: Add a hacking rule for string interpolation at logging https://review.openstack.org/334844 | 15:09 |
| *** jlanoux_ has joined #openstack-nova | 15:09 | |
| *** annegentle has quit IRC | 15:09 | |
| *** jlanoux has quit IRC | 15:10 | |
| openstackgerrit | Radoslav Gerganov proposed openstack/nova: Fix API doc for os-console-auth-tokens https://review.openstack.org/340969 | 15:10 |
| *** armax has quit IRC | 15:10 | |
| kashyap | mdbooth: What do is the "other" API? | 15:10 |
| kashyap | s/do// | 15:11 |
| hferenc_ | i'd not agree, because in my experience the vm fails to boot if i specify flavor metadata using the os_foo_bar format | 15:11 |
| *** armax has joined #openstack-nova | 15:11 | |
| *** fyxim has joined #openstack-nova | 15:11 | |
| hferenc_ | see https://bugs.launchpad.net/nova/+bug/1582693 for different usecases | 15:12 |
| openstack | Launchpad bug 1582693 in OpenStack Compute (nova) "Image and flavor metadata for libvirt watchdog is handled erroneously" [Undecided,In progress] - Assigned to Richil Bhalerao (richil-bhalerao) | 15:12 |
| cmartin | Hey I've had a bugfix up for a while now on resizing vm's and would really appreciate some reviews https://review.openstack.org/#/c/370861/ | 15:12 |
| mriedem | hferenc: seems like a horizon bug | 15:13 |
| mriedem | hferenc: i just left a comment in the bug | 15:13 |
| *** Hosam_ has quit IRC | 15:13 | |
| *** Hosam has joined #openstack-nova | 15:14 | |
| *** lucasagomes is now known as lucas-hungry | 15:15 | |
| mriedem | hferenc: the docs also use <capability>: for extra_specs http://docs.openstack.org/admin-guide/compute-flavors.html#extra-specs | 15:15 |
| mriedem | i'm not really sure why nova doesn't start enforcing that behavior though with a microversoin | 15:15 |
| mriedem | so that flavor extra specs require a <capability>: prefix | 15:16 |
| mriedem | danpb: dansmith: jaypipes: sdague: alaski: johnthetubaguy: ^ any reason we couldn't/shouldn't do that? | 15:16 |
| johnthetubaguy | mriedem: I was hoping we would go that way in a microversion, unify the image and extra specs stuff using the capabilities/traits stuff | 15:17 |
| *** r-daneel has joined #openstack-nova | 15:17 | |
| mriedem | we don't control image meta anymore | 15:18 |
| *** code-R_ has quit IRC | 15:18 | |
| mriedem | well, after alex_xu's bp is done to 404 on image meta proxy | 15:18 |
| alaski | it's a little odd because there are many extra specs already out there that don't follow that pattern. so we'll force deployments to switch between microversions to set what they need | 15:18 |
| melwitt | hferenc_: oh, I see, so horizon is using the glance prefixed name on a nova flavor extra spec, which is a problem | 15:18 |
| mriedem | it's certainly easier to validate the flavor extra specs format given the : | 15:18 |
| *** READ10 has joined #openstack-nova | 15:18 | |
| johnthetubaguy | mriedem: I am on about what we interpret in provided image metadata | 15:18 |
| danpb | mriedem: sorr, do what ? | 15:18 |
| melwitt | it has to be glance prefix + image meta or nova prefix + flavor extra spec | 15:18 |
| *** dougwig has joined #openstack-nova | 15:19 | |
| *** marst has joined #openstack-nova | 15:19 | |
| *** Hosam has quit IRC | 15:19 | |
| mriedem | alaski: nova under the covers could still handle flavor extra specs with both formats, | 15:19 |
| alaski | I would think a better path would be to put these new capabilities things into something different in the api | 15:19 |
| mriedem | but with a microversion we could start enforcing the new format | 15:19 |
| johnthetubaguy | yeah, need to handle the old ones, and map it across if possible I guess | 15:20 |
| sdague | mriedem: so, the idea being that we filter the setting of specs to only allow setting the namespaced versions with a new microversion? | 15:20 |
| danpb | mriedem: we can't use underscore syntax in flavour extra specs without clashing with other code | 15:20 |
| openstackgerrit | Gábor Antal proposed openstack/nova: Removed unnecessary parantheses in return statements https://review.openstack.org/370549 | 15:20 |
| johnthetubaguy | I guess we are saying, drop extra specs and add a new thing? | 15:20 |
| danpb | one of the schedular filters matches on everything that doesn't have a xxx: prefix | 15:20 |
| mriedem | sdague: yes | 15:20 |
| alaski | johnthetubaguy: we would be implicitly saying extra specs are a new thing now | 15:20 |
| sdague | mriedem: that seems fine, then some legacy mapping for things we know | 15:20 |
| *** mkoderer has joined #openstack-nova | 15:21 | |
| sdague | mriedem: given the degree of confusion here, probably worth writing a quick spec on it to make sure we're all clear on that | 15:21 |
| *** kfarr has joined #openstack-nova | 15:21 | |
| sdague | danpb: it's actually the opposite idea, make sure everything comes through with a : | 15:21 |
| sdague | so that we don't get people confusedly pushing _ versions of things and being confused they don't work | 15:22 |
| sdague | which is what has happened in horizon apparently | 15:22 |
| danpb | what would actually be nice is if we actually had an object defining extra specs like we did for image meta, so we could actually validate them sensibly | 15:22 |
| mriedem | the aggregate_image_properties_isolation_separator config option kind of breaks this | 15:23 |
| danpb | as even if you have the right prefix : vs _, doesn't mean you've got the data right, or that this version of nova supports it | 15:23 |
| sdague | mriedem: well, start with deprecating that | 15:23 |
| * hferenc_ is sorry for starting a tsunami | 15:24 | |
| *** boris-42 has joined #openstack-nova | 15:24 | |
| *** mdrabe has quit IRC | 15:24 | |
| sdague | danpb: given that the field is currently free form, how would that work with other services / users setting arbitrary things? | 15:24 |
| alaski | my only issue is that both extra specs and scheduler filters are open for use/extension and some deployers have taken advantage of that to do their own things. This would be changing things up on them. But I guess as long as the prefix isn't constrained it's just a small change for those deployers. | 15:24 |
| alaski | so I think I've convinced myself there's no reason not to do it | 15:24 |
| mriedem | yeah AggregateInstanceExtraSpecsFilter expects the flavor extra specs to have the : format | 15:24 |
| sdague | alaski: is it only deployers changing this because of scheduler filters? | 15:25 |
| *** diga has quit IRC | 15:25 | |
| sdague | are there any ways in which other services use this data? | 15:25 |
| *** zenoway has quit IRC | 15:25 | |
| alaski | sdague: no. I did some stuff with policy using extra specs as well | 15:25 |
| alaski | gut if the only requirement is that it now needs a foo: appended to it I think it'll be a small change for folks | 15:26 |
| *** zenoway has joined #openstack-nova | 15:26 | |
| alaski | s/gut/but/ | 15:26 |
| mriedem | alaski: as far as i can tell, AggregateInstanceExtraSpecsFilter and ComputeCapabilitiesFilter both require the extra spec to have a <cap>: prefix | 15:26 |
| mriedem | else they don't work | 15:26 |
| mriedem | ti's hard-coded | 15:26 |
| sdague | alaski: I wonder if making a registry in nova, which includes an entrypoint where you can, as a deployer, register your own things for validation would be ok | 15:26 |
| mriedem | AggregateImagePropertiesIsolation is configurable | 15:26 |
| alaski | mriedem: it's not required that those are used though | 15:26 |
| auggy | mriedem: i'm planning to crash the PTG and contribute many bad jokes | 15:26 |
| mriedem | alaski: true | 15:26 |
| sdague | just looking for a : helps with one class of errors | 15:26 |
| sdague | but actually validating properties against a known list, like danpb suggested, catches a whole lot more | 15:27 |
| mriedem | so we can't enforce : in flavor extra specs b/c of out of tree scheduler filters? | 15:27 |
| alaski | sdague: this is why I suggested we add a new API thing that's not flavor extra specs to set these things | 15:27 |
| mriedem | i'd argue that any out of tree scheduler filter should support : | 15:27 |
| mriedem | in addition to whatever other format | 15:27 |
| sdague | mriedem: sure, I think that's fine | 15:27 |
| sdague | I was just thinking about the hard validation | 15:28 |
| sdague | which is probably also fine, as long as there is an easy mechanism for people to add register their things | 15:28 |
| sdague | alaski: so... you are thinking flavor extra specs registration API? | 15:28 |
| *** hieulq__ has quit IRC | 15:28 | |
| sdague | or some other thing entirely? | 15:29 |
| alaski | well, I would stop calling them extra_specs. even if they're stored that way on flavors | 15:29 |
| sdague | alaski: sure | 15:30 |
| alaski | but yes, some other thing with strict validation on what's added | 15:30 |
| sdague | I guess the question is, if we have hard strict validation, how do you whitelist in new values | 15:30 |
| sdague | and it could be an API call | 15:30 |
| sdague | or it could be an entry point with definitions | 15:30 |
| sdague | or something else | 15:30 |
| *** zenoway has quit IRC | 15:31 | |
| *** bvanhav has quit IRC | 15:31 | |
| alaski | do we need to whitelist in values at that point? extra specs would still be open and free form | 15:31 |
| *** mdrabe has joined #openstack-nova | 15:31 | |
| sdague | alaski: if we wanted hard strict validation, yes | 15:31 |
| *** bvanhav has joined #openstack-nova | 15:31 | |
| sdague | to defend against | 15:32 |
| sdague | quota:cpu_peroid | 15:32 |
| alaski | okay. I thought we had a definitive list like I see at http://docs.openstack.org/admin-guide/compute-flavors.html#extra-specs | 15:32 |
| *** tonytan_brb has joined #openstack-nova | 15:32 | |
| danpb | alaski: lol, no that docs is just someones best guess as to what we use | 15:32 |
| alaski | if we still need an extension mechanism then keeping it in the extra specs call seems better to me | 15:32 |
| alaski | danpb: gotcha | 15:32 |
| danpb | until we have a formal versioned object defining this, there is no definitive list | 15:33 |
| danpb | it took us ages to track down all the secret image properties nova code was relying on | 15:33 |
| *** geekinutah has joined #openstack-nova | 15:33 | |
| danpb | and flavour extra specs is an even bigger mess than image properties was | 15:33 |
| alaski | it seems like that's the starting point to this then. Have a definitive list we can use to guard with | 15:34 |
| *** ijw has joined #openstack-nova | 15:34 | |
| *** suresh12 has joined #openstack-nova | 15:34 | |
| *** peter-hamilton has quit IRC | 15:34 | |
| *** tonytan4ever has quit IRC | 15:35 | |
| *** Matti_ has joined #openstack-nova | 15:35 | |
| *** andrewbogott has joined #openstack-nova | 15:36 | |
| *** erlon has joined #openstack-nova | 15:36 | |
| alaski | mriedem: back to your original question, restricting on ':' seems fine to me. A longer term plan may be to add a new strictly validating api call if we have a way to do that strict validation. | 15:36 |
| *** DuncanT has joined #openstack-nova | 15:36 | |
| *** ducttape_ has quit IRC | 15:37 | |
| *** andrewbogott has quit IRC | 15:37 | |
| *** andrewbogott has joined #openstack-nova | 15:37 | |
| *** patrickeast has joined #openstack-nova | 15:37 | |
| *** ildikov has joined #openstack-nova | 15:38 | |
| *** andreas_s has quit IRC | 15:38 | |
| *** suresh12 has quit IRC | 15:39 | |
| *** ijw has quit IRC | 15:40 | |
| *** cargonza has joined #openstack-nova | 15:41 | |
| *** macsz has joined #openstack-nova | 15:41 | |
| *** moshele has quit IRC | 15:42 | |
| mriedem | alaski: sdague: ok i'll give myself a todo to try and start a spec on the : validation for extra specs, seems like a no-brainer | 15:42 |
| *** ediardo has joined #openstack-nova | 15:43 | |
| mriedem | hferenc: ^ | 15:43 |
| mriedem | hferenc: i'd probably take that to the horizon team though that flavor extra specs are assumed to be prefixed with the : | 15:43 |
| mriedem | and they should start doing that | 15:43 |
| hferenc_ | i think i managed to keep up with you guys | 15:44 |
| hferenc_ | and since is started all this please let me know if i can help | 15:44 |
| hferenc_ | *i | 15:45 |
| *** csatari has joined #openstack-nova | 15:46 | |
| mriedem | i need to fire up a horizon and see what it shows | 15:46 |
| *** tomhambleton has joined #openstack-nova | 15:46 | |
| *** coolsvap has joined #openstack-nova | 15:47 | |
| *** paul-carlton2 has quit IRC | 15:47 | |
| *** bvanhav has quit IRC | 15:48 | |
| *** ftersin has left #openstack-nova | 15:50 | |
| *** ducttape_ has joined #openstack-nova | 15:52 | |
| *** diogogmt_ has joined #openstack-nova | 15:53 | |
| *** diogogmt has quit IRC | 15:53 | |
| *** diogogmt_ is now known as diogogmt | 15:53 | |
| *** kaisers has quit IRC | 15:53 | |
| *** akshai_ has quit IRC | 15:55 | |
| *** haplo37_ has quit IRC | 15:57 | |
| *** haplo37_ has joined #openstack-nova | 15:59 | |
| *** rcernin has quit IRC | 16:01 | |
| *** mdrabe has quit IRC | 16:01 | |
| *** mdrabe has joined #openstack-nova | 16:01 | |
| *** david-lyle has joined #openstack-nova | 16:01 | |
| *** jheroux has quit IRC | 16:02 | |
| *** bvanhav has joined #openstack-nova | 16:02 | |
| *** avolkov has quit IRC | 16:03 | |
| openstackgerrit | JABAR ALI proposed openstack/python-novaclient: commit cc963fcdde86c6b5724bd8fa80cd4a2eb7d4b02c Author: Jabar Ali <jabbar.ali@gmail.com> Date: Thu Oct 6 16:59:37 2016 +0530 https://review.openstack.org/382864 | 16:05 |
| *** kberger has quit IRC | 16:05 | |
| *** kberger has joined #openstack-nova | 16:05 | |
| *** jheroux has joined #openstack-nova | 16:06 | |
| *** avolkov has joined #openstack-nova | 16:08 | |
| *** avolkov` has joined #openstack-nova | 16:09 | |
| *** nkrinner is now known as nkrinner_afk | 16:09 | |
| *** zenoway has joined #openstack-nova | 16:09 | |
| *** akshai has joined #openstack-nova | 16:11 | |
| *** lucas-hungry is now known as lucasagomes | 16:12 | |
| *** zenoway has quit IRC | 16:14 | |
| *** jistr is now known as jistr|afk | 16:14 | |
| vladikr | I've written some functional tests to cover the VF/PF passthrough that we've submitted a while back. I'd be happy to get more reviews please :) : https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:sriov_numa_func_test | 16:14 |
| *** raunak has joined #openstack-nova | 16:16 | |
| *** marst has quit IRC | 16:20 | |
| *** marst has joined #openstack-nova | 16:22 | |
| *** penick has joined #openstack-nova | 16:24 | |
| *** nic has joined #openstack-nova | 16:25 | |
| *** markus_z has quit IRC | 16:26 | |
| *** jlanoux_ has quit IRC | 16:26 | |
| *** annegentle has joined #openstack-nova | 16:27 | |
| *** mvk has quit IRC | 16:27 | |
| *** Swami has joined #openstack-nova | 16:28 | |
| *** ducttape_ has quit IRC | 16:30 | |
| *** danpb has quit IRC | 16:30 | |
| *** tesseract- has quit IRC | 16:31 | |
| *** bvanhav has quit IRC | 16:31 | |
| *** xyang1 has joined #openstack-nova | 16:32 | |
| *** zhenguo has quit IRC | 16:33 | |
| *** mat128 is now known as mat128|lunch | 16:34 | |
| *** ijw has joined #openstack-nova | 16:36 | |
| mriedem | anyone want to clean this up? https://review.openstack.org/#/c/345162/4 | 16:40 |
| mriedem | it's a duplicate original of the bug/patch that garyk reported which we talked about in the nova meeting | 16:40 |
| *** ijw has quit IRC | 16:40 | |
| sfinucan | mriedem: Isn't garyk going to work on that? | 16:42 |
| mriedem | sfinucan: on updating the original fix? | 16:42 |
| mriedem | idk | 16:42 |
| mriedem | he's not around | 16:42 |
| sfinucan | mriedem: bdm isn't my area of expertise but I'll take a look | 16:43 |
| mriedem | it's just a matter of fixing the tests | 16:43 |
| sfinucan | cool | 16:43 |
| mriedem | and probably adding a new test for the bug case | 16:43 |
| sfinucan | Also, I'd like to help out on that extra_specs thing when it comes around. Had talked with johnthetubaguy about doing something there some time back | 16:43 |
| mriedem | ok | 16:43 |
| sfinucan | cool cool | 16:43 |
| johnthetubaguy | yeah, you had good insight into the NUMA things, so sounded like a good fit | 16:44 |
| *** bvanhav has joined #openstack-nova | 16:47 | |
| *** sahid has quit IRC | 16:48 | |
| dansmith | mriedem: so how is that happening? | 16:48 |
| dansmith | mriedem: no bdms for an instance.. | 16:48 |
| mriedem | i felt like i'd actually triaged / fixed something like this before already b/c it happened in the gate | 16:49 |
| mriedem | it's a race between server create and the periodic task running i think | 16:49 |
| dansmith | I don't actually understand how bdms[uuid] works either | 16:49 |
| dansmith | mriedem: ah, that race makes sense, but the existing code doesn't | 16:50 |
| dansmith | bdms is a BlockDeviceMappingList, which is a .. list, indexed by integers, not uuids | 16:50 |
| *** Pavo has quit IRC | 16:51 | |
| dansmith | oh, that one is special | 16:51 |
| mriedem | that's where you're wrong buddy | 16:51 |
| mriedem | yeah | 16:51 |
| dansmith | gotcha | 16:51 |
| mriedem | obj_make_dict_of_lists | 16:51 |
| mriedem | wah wah | 16:51 |
| dansmith | yeah | 16:51 |
| mriedem | so the guest is on the host | 16:51 |
| dansmith | sfinucan: you're doing this? | 16:51 |
| mriedem | the instance is in the db | 16:51 |
| mriedem | but the bdm isn't | 16:51 |
| dansmith | aye | 16:52 |
| dansmith | sfinucan: otherwise I'll be glad to | 16:52 |
| sfinucan | dansmith: It's 5:50pm, so it's all yours :) | 16:52 |
| dansmith | sfinucan: ack | 16:52 |
| mriedem | oh come on | 16:52 |
| mriedem | 6pm psh | 16:52 |
| dansmith | mriedem: sfinucan is a lightweight | 16:53 |
| mriedem | i guess | 16:53 |
| mriedem | i'm particularly offended because i know he's young | 16:53 |
| dansmith | mriedem: also, he has a government-mandated pub commitment on the horizon | 16:53 |
| mriedem | oh yeah i guess that's true | 16:53 |
| mriedem | gotta get that 16oz in | 16:53 |
| dansmith | right | 16:53 |
| mriedem | or else the queen will nag you | 16:53 |
| dansmith | don't want him getting in trouble | 16:53 |
| sfinucan | sfinucan has an Ireland match to watch | 16:53 |
| sfinucan | and maybe a sup of beer... | 16:54 |
| *** tonytan_brb is now known as tonytan4ever | 16:54 | |
| *** ducttape_ has joined #openstack-nova | 16:54 | |
| *** Apoorva has joined #openstack-nova | 16:54 | |
| *** Pavo has joined #openstack-nova | 16:55 | |
| *** derekh has quit IRC | 16:56 | |
| *** jistr|afk is now known as jistr | 16:57 | |
| *** ralonsoh has quit IRC | 16:59 | |
| *** Apoorva_ has joined #openstack-nova | 17:01 | |
| *** yamahata has quit IRC | 17:01 | |
| *** rama_y has quit IRC | 17:01 | |
| *** sdake has joined #openstack-nova | 17:01 | |
| *** yamahata has joined #openstack-nova | 17:01 | |
| *** slaweq_ has joined #openstack-nova | 17:01 | |
| *** jamielennox|away has quit IRC | 17:02 | |
| *** david-lyle has quit IRC | 17:03 | |
| cfriesen_ | sfinucan: just reviewed https://review.openstack.org/#/c/244489...we need to sort out the async setting of migration status to "finished" vs "completed" | 17:04 |
| *** Apoorva has quit IRC | 17:04 | |
| *** CustosLimen has left #openstack-nova | 17:05 | |
| *** woodster_ has joined #openstack-nova | 17:05 | |
| *** Apoorva_ has quit IRC | 17:05 | |
| *** unicell has quit IRC | 17:07 | |
| *** mvk has joined #openstack-nova | 17:10 | |
| *** Apoorva has joined #openstack-nova | 17:12 | |
| *** jamielennox|away has joined #openstack-nova | 17:16 | |
| *** jamielennox|away is now known as jamielennox | 17:16 | |
| *** slaweq_ has quit IRC | 17:21 | |
| *** TravT_ has joined #openstack-nova | 17:22 | |
| *** slaweq_ has joined #openstack-nova | 17:23 | |
| *** TravT_ has quit IRC | 17:24 | |
| *** zz_dimtruck is now known as dimtruck | 17:24 | |
| *** bnemec has quit IRC | 17:25 | |
| *** david-lyle has joined #openstack-nova | 17:25 | |
| *** TravT has quit IRC | 17:25 | |
| openstackgerrit | STEW TY proposed openstack/nova: Transform instance.unshelve notifications https://review.openstack.org/382565 | 17:27 |
| *** jvgrant_ has quit IRC | 17:28 | |
| mriedem | hferenc: so i was playing around with flavor extra specs in horizon, and it looks like the flavor extra specs you can choose come from the glance metadata catalog | 17:29 |
| mriedem | and some of the items have the namespace: prefix and some have namespace_ | 17:29 |
| diana | dansmith: For simple tenant usage paging, were you thinking of paging the tenant list (tenant_usages) or the nested servers list (server_usages)? Current example response: http://paste.openstack.org/show/584548/ | 17:29 |
| mriedem | hferenc: so it seems to be an issue in the glance metadata catalog? | 17:30 |
| dansmith | diana: it's just one query to the db, right? | 17:30 |
| diana | dansmith: yup | 17:30 |
| dansmith | diana: paginating that is the only thing that will help | 17:31 |
| *** suresh12 has joined #openstack-nova | 17:31 | |
| *** unicell has joined #openstack-nova | 17:31 | |
| *** haplo37 has quit IRC | 17:32 | |
| *** slaweq_ has quit IRC | 17:33 | |
| *** annegentle has quit IRC | 17:34 | |
| *** tblakes has quit IRC | 17:34 | |
| *** haplo37 has joined #openstack-nova | 17:34 | |
| diana | dansmith: But from an API response perspective are you expecting to page on the tenant_usages? Which won't help much in the "few tenats, many instances" case. | 17:35 |
| *** slaweq_ has joined #openstack-nova | 17:35 | |
| dansmith | diana: I don't understand.. are you asking what the client should use as a marker? | 17:35 |
| *** annegentle has joined #openstack-nova | 17:36 | |
| *** ijw has joined #openstack-nova | 17:36 | |
| *** sdague has quit IRC | 17:37 | |
| diana | dansmith: Ignore the limit & marker concept for now. I'm asking at an API level. We currently return a list of tenant usages that each contain a list of server usages. | 17:37 |
| *** hferenc_ has quit IRC | 17:37 | |
| dansmith | um, I guess I don't know how we can ignore the limit/marker if we're talking about pagination.. we need to limit the results to N instances, which will be the total length of all server_usages combined, and which may end up with N tenants or one tenant, depending on the order in which they come out | 17:38 |
| *** bnemec has joined #openstack-nova | 17:38 | |
| *** psachin has quit IRC | 17:39 | |
| mriedem | hferenc: see https://bugs.launchpad.net/nova/+bug/1582693/comments/16 | 17:39 |
| openstack | Launchpad bug 1582693 in OpenStack Dashboard (Horizon) "Image and flavor metadata for libvirt watchdog is handled erroneously" [Undecided,New] | 17:39 |
| mriedem | hferenc: it's a glance metadefs bug i htink | 17:39 |
| mriedem | *think | 17:39 |
| mriedem | melwitt: ^ | 17:39 |
| *** Margarita_S has quit IRC | 17:40 | |
| diana | dansmith: I originally assumed you wanted the paging at the tenant_usages level. That is, you could limit it to just the first 5 tenants, with their nested server_usages (but those could still be really large lists). | 17:40 |
| dansmith | diana: yeah that doesn't help | 17:40 |
| dansmith | and it's not how we're querying the db anyway, so it would be harder to do that I think, even if it made sense | 17:41 |
| diana | dansmith: right, because you can already do that with show filtered on tenant_id | 17:41 |
| *** ijw has quit IRC | 17:41 | |
| *** lpetrut has joined #openstack-nova | 17:43 | |
| dansmith | mriedem: the test that fails in that bdms patch is so effed | 17:44 |
| mriedem | language please | 17:44 |
| mriedem | sorry, | 17:44 |
| mriedem | watch your effing language please | 17:44 |
| diana | dansmith: If there are 2 tenats, each with 10 instances, and the page size is 5: what should the response look like? I think that answer will clarify intent for me. | 17:46 |
| dansmith | diana: depends on the order that comes out of the db | 17:47 |
| dansmith | diana: might be one tenant with five instances, two tenants with 2/3 instances, etc | 17:47 |
| *** bnemec has quit IRC | 17:47 | |
| diana | dansmith: But then the aggregate counts are off... | 17:47 |
| *** roeyc has joined #openstack-nova | 17:48 | |
| dansmith | diana: you mean like the start/stop/total_hours etc? | 17:48 |
| diana | dansmith: the values in the tenants_usages dict are aggregates of the values in the servers_usages list | 17:49 |
| *** artom has quit IRC | 17:49 | |
| diana | dansmith: yes, the aggregate totals | 17:49 |
| dansmith | diana: right, so this is pivoting from purely time-based to time-and-count based, so you'll have to correct those numbers, right? | 17:50 |
| diana | dansmith: those aggregates are supposed to reflect all the instances that are in that window for a given tenant though... correct? Not just some arbitrary group of them based on the order returned by the database. | 17:52 |
| dansmith | diana: they're supposed to represent the instances in the server_usages list I would say | 17:52 |
| dansmith | diana: obviously if you're a client and you're passing a limit, you're going to have to stitch things together from the pieces, just like any paginated result right? | 17:53 |
| diana | dansmith: so then horizon would need to stitch them all back together? | 17:53 |
| diana | (we typed at the same time, that wasn't in response) | 17:53 |
| *** artom has joined #openstack-nova | 17:54 | |
| *** david-lyle has quit IRC | 17:55 | |
| diana | dansmith: I'm starting to think that show should be paged instead of index then... | 17:55 |
| diana | dansmith: anyhoo, thanks for your time, I'll think on this some more | 17:55 |
| *** hferenc has quit IRC | 17:56 | |
| *** bnemec has joined #openstack-nova | 17:57 | |
| *** Matti_ has quit IRC | 17:57 | |
| *** rama_y has joined #openstack-nova | 17:57 | |
| *** slaweq_ has quit IRC | 17:58 | |
| *** priteau has joined #openstack-nova | 18:01 | |
| *** tonytan4ever has quit IRC | 18:01 | |
| *** gszasz has quit IRC | 18:01 | |
| *** tonytan4ever has joined #openstack-nova | 18:01 | |
| *** dimtruck is now known as zz_dimtruck | 18:02 | |
| *** tanvirt has quit IRC | 18:02 | |
| *** ihrachys has quit IRC | 18:03 | |
| *** bnemec has quit IRC | 18:03 | |
| *** lpetrut has quit IRC | 18:05 | |
| *** lpetrut has joined #openstack-nova | 18:05 | |
| *** david-lyle has joined #openstack-nova | 18:06 | |
| *** cmartin has quit IRC | 18:06 | |
| jaypipes | lbeliveau: reviewed. | 18:07 |
| *** ldeptula has quit IRC | 18:08 | |
| *** cmartin has joined #openstack-nova | 18:08 | |
| *** bnemec has joined #openstack-nova | 18:08 | |
| *** annegentle has quit IRC | 18:08 | |
| lbeliveau | jaypipes: thanks ! much appreciated | 18:09 |
| jaypipes | lbeliveau: thank *you* for the fix/patch! :) | 18:09 |
| lbeliveau | jaypipes: will retest everything and look at the race conditions, will most likely have results at summit | 18:10 |
| jaypipes | lbeliveau: excellent work. | 18:10 |
| *** artom has quit IRC | 18:11 | |
| *** artom has joined #openstack-nova | 18:11 | |
| *** whenry has quit IRC | 18:12 | |
| *** artom has quit IRC | 18:12 | |
| *** artom has joined #openstack-nova | 18:12 | |
| *** hemanthm is now known as hemanthm|lunch | 18:13 | |
| openstackgerrit | Ludovic Beliveau proposed openstack/nova: Fix drop_move_claim() on revert resize https://review.openstack.org/349060 | 18:14 |
| *** tblakes has joined #openstack-nova | 18:14 | |
| lbeliveau | jaypipes: ^ fixed commit message | 18:15 |
| *** yamahata has quit IRC | 18:15 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Fix exception due to BDM race in get_available_resource() https://review.openstack.org/345162 | 18:16 |
| openstackgerrit | Andreas Jaeger proposed openstack/nova: Enable release notes translation https://review.openstack.org/383189 | 18:17 |
| *** bnemec has quit IRC | 18:17 | |
| *** tbachman has quit IRC | 18:17 | |
| jaypipes | lbeliveau: thx. re-added my +2. | 18:19 |
| *** slaweq_ has joined #openstack-nova | 18:20 | |
| *** mkoderer has quit IRC | 18:22 | |
| parora | mriedem, can you please take a look at https://review.openstack.org/#/c/294513/ , it had a +2 but had some nits | 18:22 |
| *** danpawlik has quit IRC | 18:23 | |
| *** roeyc has quit IRC | 18:23 | |
| *** sdague has joined #openstack-nova | 18:25 | |
| mriedem | parora: yeah i'll try | 18:26 |
| *** mat128|lunch is now known as mat128 | 18:27 | |
| *** bnemec has joined #openstack-nova | 18:27 | |
| *** kfarr has quit IRC | 18:30 | |
| *** bnemec has quit IRC | 18:33 | |
| *** bnemec has joined #openstack-nova | 18:34 | |
| *** Sukhdev has joined #openstack-nova | 18:36 | |
| *** slaweq_ has quit IRC | 18:36 | |
| *** TravT has joined #openstack-nova | 18:37 | |
| *** ijw has joined #openstack-nova | 18:37 | |
| *** bnemec has quit IRC | 18:41 | |
| openstackgerrit | Andreas Jaeger proposed openstack/os-vif: Enable release notes translation https://review.openstack.org/383325 | 18:43 |
| *** haplo37_ has quit IRC | 18:45 | |
| *** ijw has quit IRC | 18:45 | |
| *** ociuhandu has quit IRC | 18:46 | |
| *** haplo37_ has joined #openstack-nova | 18:47 | |
| *** krtaylor has joined #openstack-nova | 18:48 | |
| *** harlowja has quit IRC | 18:50 | |
| *** Pavo has quit IRC | 18:51 | |
| openstackgerrit | Andreas Jaeger proposed openstack/python-novaclient: Enable release notes translation https://review.openstack.org/383381 | 18:51 |
| *** hwoarang has quit IRC | 18:51 | |
| *** hwoarang has joined #openstack-nova | 18:52 | |
| *** yamahata has joined #openstack-nova | 18:53 | |
| *** bnemec has joined #openstack-nova | 18:54 | |
| *** tbachman has joined #openstack-nova | 18:55 | |
| sdague | mriedem: https://review.openstack.org/#/c/351608/18 + W on that lands 4 conf cleanups | 18:55 |
| *** Pavo has joined #openstack-nova | 18:56 | |
| mriedem | looking at this thing that dan cleaned up | 18:56 |
| *** david-lyle has quit IRC | 18:57 | |
| *** kswiatek has quit IRC | 18:57 | |
| *** hwoarang has quit IRC | 18:57 | |
| *** hwoarang has joined #openstack-nova | 18:58 | |
| *** kswiatek has joined #openstack-nova | 18:58 | |
| *** raunak has quit IRC | 18:59 | |
| mriedem | dansmith: the test in https://review.openstack.org/#/c/345162/ needs work | 18:59 |
| *** TravT_ has joined #openstack-nova | 19:01 | |
| *** slaweq_ has joined #openstack-nova | 19:01 | |
| parora | mriedem, thanks | 19:01 |
| *** parora has quit IRC | 19:01 | |
| *** TravT has quit IRC | 19:02 | |
| *** TravT_ has quit IRC | 19:02 | |
| *** TravT has joined #openstack-nova | 19:02 | |
| *** jamielennox has quit IRC | 19:03 | |
| *** yamahata has quit IRC | 19:05 | |
| *** jamielennox|away has joined #openstack-nova | 19:05 | |
| *** jamielennox|away is now known as jamielennox | 19:06 | |
| *** ociuhandu has joined #openstack-nova | 19:06 | |
| *** Matti_ has joined #openstack-nova | 19:07 | |
| *** tonytan4ever has quit IRC | 19:07 | |
| *** yamahata has joined #openstack-nova | 19:08 | |
| *** ducttape_ has quit IRC | 19:09 | |
| *** mgoddard has joined #openstack-nova | 19:10 | |
| *** mgoddard_ has quit IRC | 19:10 | |
| *** bnemec has quit IRC | 19:10 | |
| *** rtheis_ has joined #openstack-nova | 19:11 | |
| *** bnemec has joined #openstack-nova | 19:11 | |
| *** artom has quit IRC | 19:11 | |
| *** cmartin has quit IRC | 19:11 | |
| *** rtheis has quit IRC | 19:12 | |
| *** salv-orlando has quit IRC | 19:13 | |
| slaweq_ | melwitt: hello | 19:13 |
| *** artom has joined #openstack-nova | 19:13 | |
| slaweq_ | melwitt: some time ago we was talking about patch https://review.openstack.org/#/c/371592/2 (https://bugs.launchpad.net/nova/+bug/1623809), do You remember maybe? | 19:13 |
| openstack | Launchpad bug 1623809 in OpenStack Compute (nova) "Quota exceeded when spawning instances in server group" [Undecided,In progress] - Assigned to Slawek Kaplonski (slaweq) | 19:13 |
| *** annegentle has joined #openstack-nova | 19:13 | |
| slaweq_ | can You check in meantime my answear in this bug and check this patch? thx in advance | 19:14 |
| *** mgoddard has quit IRC | 19:15 | |
| openstackgerrit | Andrew Laski proposed openstack/nova-specs: Conductor deployment options in cellsv2 https://review.openstack.org/383419 | 19:16 |
| *** hemanthm|lunch is now known as hemanthm | 19:16 | |
| *** coolsvap has quit IRC | 19:16 | |
| *** agireud has quit IRC | 19:21 | |
| *** READ10 has quit IRC | 19:23 | |
| *** tbachman has quit IRC | 19:24 | |
| mriedem | mtreinish: sdague: we should get this into stable/newton as it impacts grenade jobs on master https://review.openstack.org/#/c/380645/ | 19:25 |
| *** hwoarang has quit IRC | 19:25 | |
| sdague | mriedem: looking... | 19:25 |
| *** hamzy has quit IRC | 19:26 | |
| sdague | mriedem: lgtm | 19:26 |
| mtreinish | mriedem: +2 | 19:26 |
| *** cmartin has joined #openstack-nova | 19:26 | |
| *** annegentle has quit IRC | 19:28 | |
| *** rama_y has quit IRC | 19:28 | |
| *** agireud has joined #openstack-nova | 19:29 | |
| *** sileht has quit IRC | 19:31 | |
| *** sileht has joined #openstack-nova | 19:32 | |
| *** hwoarang has joined #openstack-nova | 19:32 | |
| *** agireud has quit IRC | 19:33 | |
| *** penick has quit IRC | 19:34 | |
| *** jichen has joined #openstack-nova | 19:36 | |
| *** cmartin has quit IRC | 19:37 | |
| *** hamzy has joined #openstack-nova | 19:37 | |
| *** ijw has joined #openstack-nova | 19:37 | |
| *** ducttape_ has joined #openstack-nova | 19:39 | |
| jichen | mriedem: hi, do you know if we want to create a non-voting 3rd party CI for nova drivers ,what kind of process or request should we make ? | 19:40 |
| *** lpetrut has quit IRC | 19:40 | |
| jichen | mriedem: we already have something like http://extbasicopstackcilog01.w3-969.ibm.com/test_logs/ , we want to make it non-voting job, but don't know where to start | 19:40 |
| *** bvanhav has quit IRC | 19:40 | |
| mriedem | jichen: i think here http://docs.openstack.org/infra/system-config/third_party.html | 19:42 |
| jichen | mriedem: yes, we consulted with CI folks and they said they focused on CI, and for nova specific task, we need ask inside nova | 19:43 |
| *** agireud has joined #openstack-nova | 19:43 | |
| jichen | so if we follow that page, there is nothing we need to do against nova itself? | 19:44 |
| mriedem | jichen: i don't believe so | 19:45 |
| mriedem | you can comment on changes, just make sure they don't vote | 19:45 |
| *** tonytan4ever has joined #openstack-nova | 19:45 | |
| mriedem | https://review.openstack.org/#/admin/groups/511,members is the list of voting jobs | 19:46 |
| jichen | ok, I see, maybe need some request to add us into nova-ci group and non-voting , I will see whether that info included in http://docs.openstack.org/infra/system-config/third_party.html or not | 19:47 |
| *** penick has joined #openstack-nova | 19:47 | |
| *** dave-mccowan has quit IRC | 19:49 | |
| mriedem | i don't believe your account needs to be in the nova-ci group | 19:49 |
| mriedem | there are several 3rd party accounts that run jobs on nova changes which aren't in that group, like quobyte, virtuozzo, mellanox, etc | 19:49 |
| *** Sukhdev has quit IRC | 19:51 | |
| jichen | ok, I see, I will go back check that again , thanks a lot for that info | 19:51 |
| *** suresh12 has quit IRC | 19:51 | |
| *** ijw has quit IRC | 19:52 | |
| anteaya | jichen: yes, just read the information on that page about your gerrit account | 19:52 |
| jichen | anteaya: ok, will read that again, appreciate your info~ | 19:52 |
| anteaya | jichen: if you don't want to vote on patches just comment then you don't need to follow any process other than what is on that page | 19:52 |
| *** zz_dimtruck is now known as dimtruck | 19:53 | |
| *** penick has quit IRC | 19:53 | |
| anteaya | jichen: if you have questions about that page you can find me in the #openstack-infra channel and I'll do the best I can to help you | 19:53 |
| jichen | anteaya: ok, thanks for offerings and will contact you when we have any issue :) , thanks | 19:54 |
| *** penick has joined #openstack-nova | 19:54 | |
| anteaya | welcome, and good luck | 19:54 |
| mriedem | sdague: +W on johns thing | 19:55 |
| *** rama_y has joined #openstack-nova | 19:56 | |
| openstackgerrit | Slawek Kaplonski proposed openstack/nova: Fix quota calculation for instances in server_group https://review.openstack.org/371592 | 19:57 |
| mriedem | i've also gone through the list of stable/newton backports now that it's open https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:stable/newton | 19:58 |
| *** dave-mccowan has joined #openstack-nova | 19:58 | |
| *** dave-mcc_ has joined #openstack-nova | 20:01 | |
| *** haplo37 has quit IRC | 20:02 | |
| *** dimtruck is now known as zz_dimtruck | 20:02 | |
| *** dave-mccowan has quit IRC | 20:04 | |
| *** agireud has quit IRC | 20:06 | |
| *** david-lyle has joined #openstack-nova | 20:06 | |
| *** sdake has quit IRC | 20:09 | |
| *** haplo37 has joined #openstack-nova | 20:11 | |
| melwitt | mriedem: nice find on the hw_watchdog_action thing | 20:14 |
| *** tonytan4ever has quit IRC | 20:16 | |
| *** harlowja has joined #openstack-nova | 20:17 | |
| mriedem | \o/ | 20:17 |
| mriedem | all sorts of terrible | 20:17 |
| *** devananda|afk is now known as devananda | 20:18 | |
| *** code-R has joined #openstack-nova | 20:22 | |
| *** ihrachys has joined #openstack-nova | 20:23 | |
| openstackgerrit | Dan Prince proposed openstack/nova: Fix API _workers options to be backwards compat https://review.openstack.org/383440 | 20:24 |
| *** annegentle has joined #openstack-nova | 20:24 | |
| *** zz_dimtruck is now known as dimtruck | 20:24 | |
| *** anush_ has quit IRC | 20:25 | |
| *** salv-orlando has joined #openstack-nova | 20:27 | |
| openstackgerrit | Ken'ichi Ohmichi proposed openstack/nova: Add error handling for delete-volume API https://review.openstack.org/382660 | 20:29 |
| *** jichen has quit IRC | 20:30 | |
| *** cdent has quit IRC | 20:31 | |
| *** hemanthm is now known as hemanthm|bbl | 20:31 | |
| *** code-R_ has joined #openstack-nova | 20:31 | |
| *** suresh12 has joined #openstack-nova | 20:32 | |
| *** dimtruck is now known as zz_dimtruck | 20:34 | |
| *** code-R has quit IRC | 20:34 | |
| *** rtheis_ has quit IRC | 20:38 | |
| *** slaweq_ has quit IRC | 20:38 | |
| *** browne has joined #openstack-nova | 20:40 | |
| mriedem | oomichi: so deleting a volume via the compute API proxy is deprecated | 20:41 |
| *** dave-mcc_ has quit IRC | 20:41 | |
| mriedem | oomichi: i'm not sure how much we should care to fix that | 20:41 |
| oomichi | mriedem: yeah, I know that. I was a little concern about that. | 20:41 |
| mriedem | of course you could do it it with microversion<2.36 | 20:41 |
| oomichi | mriedem: but some tests of Tempest depends on the deprecated API now | 20:42 |
| oomichi | mriedem: My original purpose is https://review.openstack.org/#/c/367358/ | 20:43 |
| oomichi | mriedem: that is not deprecated API | 20:43 |
| oomichi | mriedem: when trying to reproduce the bug on the gate with https://review.openstack.org/#/c/382083/ , I found the proglem of https://review.openstack.org/#/c/382660/ | 20:44 |
| oomichi | s/proglem/problem/ | 20:44 |
| *** cmartin has joined #openstack-nova | 20:45 | |
| *** baoli_ has quit IRC | 20:46 | |
| mriedem | oomichi: i thought tempest was getting rid of negative tests? | 20:48 |
| mriedem | i haven't been following that discussion really | 20:48 |
| *** baoli has joined #openstack-nova | 20:48 | |
| oomichi | mriedem: that is a nice point. The conclusion is we can add some negative tests based on https://review.openstack.org/#/c/379013/ | 20:49 |
| oomichi | mriedem: I am thinking this case is integrated tests(between nova and cinder) and has value to avoid internal error. In addition, this is first negative test around this API | 20:50 |
| *** Pavo has quit IRC | 20:51 | |
| oomichi | mriedem: thanks for nice review anyways :) | 20:52 |
| oomichi | for both nova and tempest | 20:52 |
| mriedem | oomichi: just posted anohther comment | 20:52 |
| mriedem | sure :) | 20:52 |
| mriedem | since i had to add those create/attach volume common methods in tempest i have to -1 people for not using them now | 20:52 |
| *** lucasagomes is now known as lucas-afk | 20:53 | |
| mriedem | oomichi: so on https://review.openstack.org/#/c/382660/ i'd be ok with it if you didn't introduce the new exception type | 20:53 |
| mriedem | just handle InvalidInput in the delete-volume REST API and we're good - it's much smaller | 20:53 |
| oomichi | hehe, I know. maybe I also would get -1 on current one even if you don't point it | 20:53 |
| mriedem | this 25 LOC patch is just way too big for me :) | 20:53 |
| *** suresh12 has quit IRC | 20:55 | |
| *** Pavo has joined #openstack-nova | 20:55 | |
| *** FL1SK has quit IRC | 20:55 | |
| oomichi | mriedem: heh, yeah 25 LoC patch is huge ;) | 20:57 |
| *** suresh12 has joined #openstack-nova | 20:57 | |
| *** TravT has quit IRC | 20:58 | |
| bauzas | mriedem: about https://review.openstack.org/#/c/380068/ | 20:59 |
| bauzas | mriedem : I think we can merge it now | 20:59 |
| *** code-R_ has quit IRC | 20:59 | |
| bauzas | So the operators will see it | 20:59 |
| mriedem | bauzas: seems kind of odd since it will show up in the 14.0.1 release | 21:00 |
| *** whenry has joined #openstack-nova | 21:00 | |
| mriedem | but it mentions 14.0.0 | 21:00 |
| mriedem | i'm not totally opposed to doing it now though | 21:00 |
| bauzas | Mmm, I'll see | 21:01 |
| *** dave-mccowan has joined #openstack-nova | 21:02 | |
| bauzas | You're right, it will be probably mentioning 14.0.-something | 21:02 |
| *** pilgrimstack1 has joined #openstack-nova | 21:02 | |
| *** ayoung has quit IRC | 21:02 | |
| bauzas | But I'll see how it can be related to 14.0 | 21:02 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Catch DevicePathInUse in attach_volume https://review.openstack.org/367358 | 21:03 |
| mriedem | bauzas: i'm not sure anyone would care | 21:03 |
| mriedem | and it's actually more correct to say 14.0.0 even though the note wouldn't show up until 14.0.1 | 21:03 |
| *** pilgrimstack has quit IRC | 21:04 | |
| *** Apoorva_ has joined #openstack-nova | 21:04 | |
| *** ijw has joined #openstack-nova | 21:04 | |
| mriedem | bauzas: ok +2 | 21:04 |
| *** penick has quit IRC | 21:05 | |
| *** code-R has joined #openstack-nova | 21:07 | |
| *** penick has joined #openstack-nova | 21:07 | |
| *** Apoorva has quit IRC | 21:07 | |
| *** burgerk has quit IRC | 21:08 | |
| *** eharney has quit IRC | 21:09 | |
| *** ijw has quit IRC | 21:10 | |
| openstackgerrit | Ken'ichi Ohmichi proposed openstack/nova: Add error handling for delete-volume API https://review.openstack.org/382660 | 21:10 |
| *** annegentle has quit IRC | 21:13 | |
| *** baoli_ has joined #openstack-nova | 21:13 | |
| *** maticue has quit IRC | 21:14 | |
| *** tonytan4ever has joined #openstack-nova | 21:16 | |
| *** baoli has quit IRC | 21:17 | |
| *** harlowja has quit IRC | 21:17 | |
| *** rfolco has quit IRC | 21:18 | |
| *** tongli has quit IRC | 21:20 | |
| *** suresh12 has quit IRC | 21:20 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Add error handling for delete-volume API https://review.openstack.org/382660 | 21:20 |
| *** annegentle has joined #openstack-nova | 21:20 | |
| *** tonytan4ever has quit IRC | 21:22 | |
| *** baoli_ has quit IRC | 21:23 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Delete traces of in-progress snapshot on VM being deleted https://review.openstack.org/294513 | 21:25 |
| *** thorst_ has quit IRC | 21:27 | |
| *** thorst_ has joined #openstack-nova | 21:27 | |
| *** gouthamr has quit IRC | 21:27 | |
| *** thorst__ has joined #openstack-nova | 21:29 | |
| *** bdeschenes has quit IRC | 21:29 | |
| openstackgerrit | Charles Volzka proposed openstack/nova-specs: Add host-level memory monitor for scheduling https://review.openstack.org/288819 | 21:30 |
| *** annegentle has quit IRC | 21:31 | |
| *** gouthamr has joined #openstack-nova | 21:31 | |
| *** code-R has quit IRC | 21:31 | |
| *** thorst_ has quit IRC | 21:32 | |
| *** gouthamr has quit IRC | 21:32 | |
| *** thorst__ has quit IRC | 21:34 | |
| *** david-lyle_ has joined #openstack-nova | 21:34 | |
| *** david-lyle has quit IRC | 21:34 | |
| *** david-lyle has joined #openstack-nova | 21:35 | |
| *** gouthamr has joined #openstack-nova | 21:35 | |
| *** david-lyle has quit IRC | 21:35 | |
| *** david-lyle has joined #openstack-nova | 21:35 | |
| *** annegentle has joined #openstack-nova | 21:36 | |
| *** gouthamr has quit IRC | 21:37 | |
| *** agireud has joined #openstack-nova | 21:37 | |
| *** esberglu has quit IRC | 21:37 | |
| *** mriedem has quit IRC | 21:38 | |
| *** pradk has quit IRC | 21:39 | |
| *** harlowja has joined #openstack-nova | 21:39 | |
| *** harlowja has quit IRC | 21:39 | |
| *** david-lyle_ has quit IRC | 21:39 | |
| *** harlowja has joined #openstack-nova | 21:39 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Fix exception due to BDM race in get_available_resource() https://review.openstack.org/345162 | 21:39 |
| *** moshele has joined #openstack-nova | 21:41 | |
| dansmith | mriedem_meeting: so the reason I didn't do that ^ in the first place is that I was assuming we'd never get to the case where we had a running domain from libvirt, but no BDMs due to a save race from the API | 21:42 |
| dansmith | mriedem_meeting: which .. I can't really imagine happening, unless it's ramdisk-only domains or something | 21:42 |
| *** krtaylor has quit IRC | 21:42 | |
| dansmith | mriedem_meeting: but you're right, that's clearly where we are in the bug report | 21:42 |
| dansmith | mriedem_meeting: i.e., we can't have defined the domain in libvirt without having seen the BDMs associated with it | 21:43 |
| dansmith | mriedem_meeting: or maybe old instances in the db that have no BDMs? | 21:45 |
| dansmith | the bug report is pretty useless | 21:46 |
| *** rcernin has joined #openstack-nova | 21:47 | |
| *** bnemec has quit IRC | 21:47 | |
| *** jheroux has quit IRC | 21:49 | |
| *** TravT has joined #openstack-nova | 21:50 | |
| *** tblakes has quit IRC | 21:51 | |
| *** marst has quit IRC | 21:54 | |
| *** rfolco has joined #openstack-nova | 21:54 | |
| *** TravT has quit IRC | 21:55 | |
| *** bdeschenes has joined #openstack-nova | 21:55 | |
| *** cmartin has quit IRC | 21:56 | |
| *** david-lyle has quit IRC | 21:56 | |
| *** ducttape_ has quit IRC | 21:58 | |
| *** burt has quit IRC | 21:59 | |
| *** penick has quit IRC | 21:59 | |
| *** rcernin has quit IRC | 21:59 | |
| *** rcernin has joined #openstack-nova | 21:59 | |
| openstackgerrit | Michael Glaser proposed openstack/nova: Transform instance.unpause notifications https://review.openstack.org/380668 | 21:59 |
| *** mdrabe has quit IRC | 22:04 | |
| *** penick has joined #openstack-nova | 22:04 | |
| *** ijw has joined #openstack-nova | 22:06 | |
| jlk | So... | 22:06 |
| jlk | I wonder if I"m seeing a bug here | 22:06 |
| jlk | well | 22:07 |
| *** akshai has quit IRC | 22:07 | |
| jlk | I've made a snapshot of an instance. The instance flavor was an 80 gig flavor. The snapshot itself is 1.2G, but it has specs of min_disks 80. | 22:08 |
| jlk | When nova pulls it down to boot an instance from it, the file in instances/_base/ gets exploded to 80. | 22:09 |
| cfriesen_ | is it actually 80, or is it a sparse file? | 22:09 |
| jlk | it may be a sparse file. What's a good way to test that? | 22:09 |
| jlk | either way, it's counting against the free space on the system | 22:09 |
| cfriesen_ | the file in instances/_base is a raw file, intended to be used as backing store by qcow2 | 22:10 |
| cfriesen_ | stat <file> | 22:10 |
| jlk | Size: 85899345920 Blocks: 2544288 IO Block: 4096 regular file | 22:11 |
| *** moshele has quit IRC | 22:11 | |
| *** ijw has quit IRC | 22:11 | |
| cfriesen_ | it's sparse. "blocks" is the number of disk blocks (512 bytes) actually used by the file. | 22:11 |
| *** Apoorva_ has quit IRC | 22:12 | |
| *** sdake has joined #openstack-nova | 22:13 | |
| *** Apoorva has joined #openstack-nova | 22:13 | |
| *** TravT has joined #openstack-nova | 22:14 | |
| *** jamesdenton has quit IRC | 22:14 | |
| *** bdeschenes has quit IRC | 22:15 | |
| *** jaypipes has quit IRC | 22:16 | |
| jlk | cfriesen_: so something in Nova is sparsing out the file? downloads, sparses it out, then uses it as a backing file? | 22:16 |
| *** suresh12 has joined #openstack-nova | 22:17 | |
| *** ociuhandu has quit IRC | 22:17 | |
| *** sdague has quit IRC | 22:18 | |
| *** ayoung has joined #openstack-nova | 22:21 | |
| *** FL1SK has joined #openstack-nova | 22:21 | |
| cfriesen_ | jlk: I think nova just calls out to qemu-img to convert it to raw for use as the backing file | 22:24 |
| *** zz_dimtruck is now known as dimtruck | 22:25 | |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: Add microversion support to api unit tests https://review.openstack.org/380810 | 22:25 |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: Restrict Metadata to Lower Case Keys https://review.openstack.org/329697 | 22:25 |
| cfriesen_ | jlk: _create_images_and_backing() in nova/virt/libvirt/driver.py | 22:28 |
| jlk | yeah, so even stranger | 22:29 |
| jlk | there is calls in nova to fallocate, if preallocate is turned on, and it makes a call to a disk file to fallocate, such as fallocate -n -l 171798691840 /opt/stack/data/nova/instances/1e2ee91c-8cf3-48d4-9dc5-87920b327936/disk | 22:29 |
| jlk | so that allocates blocks, but the size remains small | 22:30 |
| jlk | almost the opposite of the base image | 22:30 |
| jlk | okay, and disk usage show it that way too | 22:30 |
| *** ntpttr has joined #openstack-nova | 22:31 | |
| *** krtaylor has joined #openstack-nova | 22:31 | |
| *** vishwanathj has quit IRC | 22:32 | |
| jlk | so I wonder which nova uses to sort out sizes | 22:32 |
| cfriesen_ | I think that preallocate defaults to turned off though | 22:33 |
| *** suresh12 has quit IRC | 22:33 | |
| cfriesen_ | if you preallocate I don't think the size will stay small | 22:34 |
| *** suresh12 has joined #openstack-nova | 22:34 | |
| jlk | We preallocate | 22:34 |
| jlk | because we had some customers that would quickly run themselves out of disk | 22:34 |
| cfriesen_ | from the fallocate() man page..." | 22:34 |
| cfriesen_ | The | 22:34 |
| cfriesen_ | file size (as reported by stat(2)) will be changed if offset+len is | 22:34 |
| cfriesen_ | greater than the file size. | 22:34 |
| jlk | so the scenario I'm working with | 22:35 |
| jlk | customer made a snapshot of a 300 gig flavor instance. So the min_disk of the snapshot is 300g, even though it might be small actual size | 22:36 |
| cfriesen_ | oh, wait...that "-n" option tells it to not change the length of the file. that's weird | 22:36 |
| jlk | then they booted a 300g flavor instance from that snapshot | 22:36 |
| jlk | so the instance disk gets fallocated up to 300G, and the image disk gets sparsed up to 300G | 22:36 |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: Restrict Metadata to Lower Case Keys https://review.openstack.org/329697 | 22:37 |
| cfriesen_ | jlk: ouch | 22:38 |
| *** suresh12 has quit IRC | 22:38 | |
| cfriesen_ | jlk: rather than using preallocate we fixed up the "disk_available_least" calculations | 22:39 |
| jlk | to make things worse, I think one of the instances has done something inside the VM so that the copy on write has copied over a TON of the data. So it's really actually sized up to 264 gig | 22:39 |
| mtreinish | auggy: ^^^ is that passing now? | 22:39 |
| jlk | cfriesen_: oh? that's good to know! | 22:39 |
| cfriesen_ | jlk: I don't think they're all upstream yet. it's slow going getting resource tracking fixes pushed. | 22:40 |
| jlk | ah | 22:40 |
| *** TonyXu has quit IRC | 22:41 | |
| *** TonyXu has joined #openstack-nova | 22:42 | |
| auggy | mtreinish: sooooon | 22:43 |
| auggy | mtreinish: i just got everything rebased and (i think?) working after the rebase and then i had to update for some deprecations | 22:43 |
| auggy | mtreinish: once that works i'm moving over to your nova manage patch to get that rebased and updated | 22:44 |
| jlk | dammmit | 22:44 |
| jlk | I set the image min_disk to 5G, and I'm booting a 160G flavor, and yet the downloaded base image file is still getting sparsed up to 80G | 22:45 |
| auggy | jlk: https://review.openstack.org/gitweb?p=openstack/nova.git;a=commitdiff;h=bd982bf2d5f7529f2c61140a11fdf2ac8620971e ?? | 22:48 |
| auggy | not sure if that applies | 22:48 |
| auggy | - sent_meta['min_disk'] = min_disk | 22:49 |
| auggy | + sent_meta['min_disk'] = max(root_gb or 0, min_disk) | 22:49 |
| auggy | oh wait that's in a test | 22:49 |
| auggy | jlk: nm i think i misread this patch | 22:49 |
| *** ducttape_ has joined #openstack-nova | 22:49 | |
| *** annegentle has quit IRC | 22:49 | |
| mtreinish | auggy: heh, I can handle the rebase just tell me when it's a good idea :) | 22:50 |
| auggy | mtreinish: sure, i'll let you know when this works | 22:50 |
| auggy | mtreinish: also there was some discussion in the spec about adding a warning when squashing data if there are conflicts | 22:50 |
| *** Pavo has quit IRC | 22:50 | |
| auggy | i haven't re-reviewed the nova manage stuff since we worked on this a million years ago haha | 22:51 |
| mtreinish | auggy: I don't think there is a unique constraint on any of the columns the manage command touches | 22:51 |
| mtreinish | but it's been a while | 22:51 |
| auggy | aggregate metadata might have one?? | 22:51 |
| *** Pavo has joined #openstack-nova | 22:51 | |
| auggy | but it's more about people just being aware that they have multiple keys with the same name | 22:51 |
| auggy | oh right, i recently put a suggestion to add unit test coverage just to see what happens in that case and figure out what we think should happen | 22:52 |
| mtreinish | auggy: but if that's something we need to add, it's easy enough to do | 22:52 |
| *** ajo_ has quit IRC | 22:53 | |
| auggy | mtreinish: yeah it might be... i'm just pondering the potential impacts | 22:53 |
| auggy | i guess it's just a warn, not a show stopper | 22:54 |
| *** ajo has joined #openstack-nova | 22:58 | |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: Add microversion support to api unit tests https://review.openstack.org/380810 | 23:00 |
| *** salv-orlando has quit IRC | 23:01 | |
| *** gouthamr has joined #openstack-nova | 23:01 | |
| *** ihrachys has quit IRC | 23:03 | |
| *** mlavalle has quit IRC | 23:05 | |
| *** ayoung has quit IRC | 23:05 | |
| *** claudiub has quit IRC | 23:07 | |
| *** ijw has joined #openstack-nova | 23:07 | |
| *** nic has quit IRC | 23:09 | |
| *** portdirect has joined #openstack-nova | 23:12 | |
| *** ijw has quit IRC | 23:14 | |
| *** xyang1 has quit IRC | 23:18 | |
| openstackgerrit | melanie witt proposed openstack/nova: Add quota related tables to the api database. https://review.openstack.org/331810 | 23:23 |
| *** esberglu has joined #openstack-nova | 23:24 | |
| *** amotoki has quit IRC | 23:24 | |
| mtreinish | auggy: hmm, not unique for agg meta either: http://paste.openstack.org/show/584750/ | 23:25 |
| mtreinish | well per aggregate it is | 23:26 |
| mtreinish | but not generally | 23:26 |
| auggy | kk i remember a unique contraint violation being thrown as part of the original bug | 23:26 |
| *** priteau has quit IRC | 23:26 | |
| *** mriedem has joined #openstack-nova | 23:26 | |
| *** dimtruck is now known as zz_dimtruck | 23:27 | |
| *** zz_dimtruck is now known as dimtruck | 23:27 | |
| auggy | https://bugs.launchpad.net/nova/+bug/1538011 | 23:28 |
| openstack | Launchpad bug 1538011 in OpenStack Compute (nova) "Metadata keys are not case sensitive" [Medium,In progress] - Assigned to Augustina Ragwitz (auggy) | 23:28 |
| auggy | ohhh right | 23:28 |
| auggy | UNIQUE KEY `uniq_aggregate_metadata0aggregate_id0key0deleted` (`aggregate_id`,`key`,`deleted`), | 23:28 |
| mtreinish | oh, and flavor extra specs has the same thing. Unique per flavor: http://paste.openstack.org/show/584751/ | 23:28 |
| mtreinish | right, if we violate that sqlalchemy will throw an exception when it tries to do the update | 23:28 |
| auggy | yeah | 23:29 |
| mtreinish | I'm not sure that's something we want to downgrade to a warning though | 23:29 |
| auggy | so would that case be covered by adding a test case to your nova-manage command that attempts to squash Foo and FOO? | 23:29 |
| auggy | yeah.. | 23:29 |
| auggy | yeah i guess let's just get the status quo behavior working, make sure we don't overwrite anything without the user being aware of it and then just get review feedback on it... | 23:30 |
| auggy | we'll need to document this anyways, so we nned to make sure we have a fixing the keys workflow that we know works | 23:32 |
| mtreinish | well I'm not sure how we would want to handle a conflict like that, we really wouldn't want to take an opinion in that case | 23:33 |
| mtreinish | that's something that'll always require manual intervention | 23:33 |
| *** esberglu has quit IRC | 23:35 | |
| *** rcernin has quit IRC | 23:36 | |
| *** dimtruck is now known as zz_dimtruck | 23:36 | |
| auggy | yeah i agree | 23:41 |
| jlk | I think I just need to figure out how to make a snapshot of a VM so that when booting from that snapshot nova doesn't sparse fill the base image file | 23:42 |
| *** ducttape_ has quit IRC | 23:44 | |
| *** ducttape_ has joined #openstack-nova | 23:46 | |
| *** ayoung has joined #openstack-nova | 23:49 | |
| *** Swami has quit IRC | 23:50 | |
| *** esberglu has joined #openstack-nova | 23:51 | |
| *** esberglu has quit IRC | 23:51 | |
| *** thorst_ has joined #openstack-nova | 23:52 | |
| *** rfolco has quit IRC | 23:53 | |
| *** pcaruana has quit IRC | 23:55 | |
| *** thorst_ has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!