| mtreinish | mriedem: oh I didn't know about that method. That means I can fix https://github.com/openstack-infra/subunit2sql/blob/master/subunit2sql/tests/migrations/test_model_sync.py#L86 now | 00:00 |
|---|---|---|
| mtreinish | thanks | 00:00 |
| *** Apoorva has quit IRC | 00:03 | |
| *** devananda is now known as devananda|dinner | 00:04 | |
| mriedem | "" | 00:07 |
| mriedem | oops | 00:07 |
| mriedem | "In addition, PostgreSQL provides the text type, which stores strings of any length. Although the type text is not in the SQL standard, several other SQL database management systems have it as well." | 00:07 |
| *** tjones has left #openstack-nova | 00:07 | |
| *** vishwanathj is now known as vishwanathj_zzz | 00:08 | |
| *** vishwanathj_zzz is now known as vishwanathj_ | 00:08 | |
| *** markvoelker has joined #openstack-nova | 00:13 | |
| *** zul has quit IRC | 00:14 | |
| *** BjoernT has quit IRC | 00:16 | |
| mriedem | dansmith: does an alter table change for a column need to be checked in the migration? i'm wondering if a 2nd alter table statement when the type is already set causes a failure | 00:17 |
| mriedem | or if it's just a noop | 00:17 |
| dansmith | mriedem: unsure.. there is a checkfirst=True option which might check and make it a noop if nothing would happen, but ... you'd have to check | 00:17 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Increase BDM column in build_requests table https://review.openstack.org/366955 | 00:22 |
| *** rama_y has quit IRC | 00:23 | |
| mriedem | i'm too busy with this carrot cake to test it atm | 00:24 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: DNM: test to see if column alter is idempotent https://review.openstack.org/367703 | 00:25 |
| *** lixiaoy1 has joined #openstack-nova | 00:26 | |
| *** lixiaoy11 has joined #openstack-nova | 00:27 | |
| *** lixiaoy1 has quit IRC | 00:27 | |
| *** gyee has quit IRC | 00:37 | |
| *** wasmum has joined #openstack-nova | 00:37 | |
| *** marst has joined #openstack-nova | 00:39 | |
| *** mlavalle has quit IRC | 00:39 | |
| *** kaisers has joined #openstack-nova | 00:40 | |
| *** kaisers1 has quit IRC | 00:40 | |
| *** mlavalle has joined #openstack-nova | 00:40 | |
| *** rloo has quit IRC | 00:42 | |
| *** hoangcx has joined #openstack-nova | 00:42 | |
| *** gongysh has quit IRC | 00:45 | |
| melwitt | mriedem: on the glance image metadata thing, from https://review.openstack.org/#/c/303658/ it looks like one could argue that nova is the source of truth for the numa feature and that glance should conform to that. I don't know that much about it | 00:48 |
| *** thorst has joined #openstack-nova | 00:54 | |
| mriedem | melwitt: i'm totally good with glance making the change | 00:55 |
| mriedem | and nikhil is +1 on it | 00:55 |
| *** browne has quit IRC | 00:56 | |
| melwitt | mriedem: yeah. it's a matter of who is supposed to be the source of truth, I think. initially I was thinking, glance controls image metadata properties and nova should use what they define | 00:57 |
| *** artom has quit IRC | 00:57 | |
| nikhil | mriedem: melwitt : those etc/ metadefs are placeholders and upstream defaults. glance doesn't rely on the keys or values when metadefs are concerned. so we've those k, v sourced in from other projects, in this case nova. | 00:57 |
| melwitt | but now I'm more confused about it | 00:57 |
| melwitt | nikhil: I see | 00:58 |
| *** mtreinish has quit IRC | 01:01 | |
| *** thorst has quit IRC | 01:02 | |
| *** artom has joined #openstack-nova | 01:03 | |
| *** mtreinish has joined #openstack-nova | 01:05 | |
| *** tonytan4ever has joined #openstack-nova | 01:05 | |
| *** jamesdenton has quit IRC | 01:07 | |
| *** jamesdenton has joined #openstack-nova | 01:08 | |
| mriedem | melwitt: i think you should totally jump into the glance channel and raise all hell | 01:09 |
| mriedem | or just ping flaper87 | 01:09 |
| *** thorst has joined #openstack-nova | 01:09 | |
| *** gongysh has joined #openstack-nova | 01:11 | |
| *** mlavalle has quit IRC | 01:11 | |
| *** ociuhandu has quit IRC | 01:13 | |
| *** READ10 has joined #openstack-nova | 01:14 | |
| *** Apoorva has joined #openstack-nova | 01:14 | |
| melwitt | mriedem: yeah ready to raise some hell! \o/ | 01:14 |
| melwitt | I'll see if I can catch him tomorrow | 01:16 |
| mriedem | ha | 01:17 |
| mriedem | excellent | 01:17 |
| *** wangqun has joined #openstack-nova | 01:22 | |
| *** kaisers_ has joined #openstack-nova | 01:22 | |
| *** jamesdenton has quit IRC | 01:25 | |
| *** kaisers_ has quit IRC | 01:27 | |
| *** READ10 has quit IRC | 01:32 | |
| *** yamahata has joined #openstack-nova | 01:38 | |
| mgagne | so I'm trying to debug an issue with power state sync periodic task failing with Ironic driver. I know that this might be Ironic team's "problem" but I'm now looking for more broader input. The sync fails with BadStatusLine exception. The HTTP response from Ironic API looks to be empty, like no HTP status line, no body, nada. | 01:38 |
| mgagne | I'm trying to reproduce the issue in a standalone script and so far, I only managed to bring the load-balancer to its knees without reproducing the BadStatusLine exception... | 01:39 |
| mgagne | so anyone had this issue? | 01:40 |
| mgagne | I also tried to bypass the load-balancer, same problem. | 01:41 |
| *** gongysh has quit IRC | 01:41 | |
| *** artom has quit IRC | 01:44 | |
| *** Apoorva has quit IRC | 01:45 | |
| *** artom has joined #openstack-nova | 01:45 | |
| mgagne | now I managed to reproduce the issue! bypassing network, installed nova code on ironic api node, ran the script and got BadStatusLine | 01:48 |
| openstackgerrit | Ha Van Tu proposed openstack/nova: [api-ref] Update configuration file https://review.openstack.org/367712 | 01:49 |
| melwitt | mgagne: that sounds bizarre. like ironic had closed the connection before responding. do you see any errors in ironic logs? | 01:53 |
| mgagne | no error =) | 01:54 |
| mgagne | melwitt: trying to debug atm. reproduced on ironic-api box. trying to see which instance fails to power sync to see if api received (and logged) the request | 01:55 |
| melwitt | what about "returning response ..." type thing in the ironic api log? | 01:55 |
| melwitt | okay | 01:55 |
| mgagne | if not, will tcpdump or strace and see what happens | 01:55 |
| mgagne | it fails in httplib which can't parse the response | 01:55 |
| melwitt | I haven't heard of this before | 01:55 |
| mgagne | yea, really curious to find what's the problem =) | 01:56 |
| *** liuyulong|away is now known as liuyulong | 01:57 | |
| mgagne | I suspect there is something running within Ironic that might sometimes fails the request. can no longer reproduce for the moment | 01:57 |
| mgagne | I will sure let you know if I find something | 01:57 |
| *** gongysh has joined #openstack-nova | 02:01 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: Trival fix a typo https://review.openstack.org/365938 | 02:01 |
| *** akshai has joined #openstack-nova | 02:04 | |
| *** yamahata has quit IRC | 02:07 | |
| *** EinstCrazy has joined #openstack-nova | 02:08 | |
| gmann | mriedem: oomichi alex_xu johnthetubaguy : can you review this. last one on user_id things for tests correction- https://review.openstack.org/#/c/354690/ | 02:11 |
| alex_xu | gmann: I'm a little confuse for the comment said let's go with default policy tests | 02:12 |
| alex_xu | gmann: the default policy means admin_or_owner? | 02:13 |
| *** unicell has quit IRC | 02:14 | |
| *** artom has quit IRC | 02:16 | |
| gmann | alex_xu: yea, so i understand by doing the tests with same project id not needed as normal tests does the same | 02:20 |
| gmann | mean tests with default one | 02:20 |
| *** artom has joined #openstack-nova | 02:21 | |
| alex_xu | gmann: I guess johnthetubaguy means we shouldn't override the policy rule, just use the default rule in the test. | 02:22 |
| alex_xu | gmann: but that sounds huge works, a lot of tests need to tune | 02:22 |
| gmann | alex_xu: humm, in that case we have to update each tests | 02:23 |
| alex_xu | gmann: yes, at least remove the override line | 02:23 |
| gmann | alex_xu: and then we need to run with non admin context and with admin so that we cover both for applicable APIs | 02:24 |
| alex_xu | gmann: if i understand johnthetubaguy correct, I prefer we refactor the test in next release | 02:24 |
| *** tojuvone_ has joined #openstack-nova | 02:24 | |
| alex_xu | gmann: yes, we have non_admin tests(enforce the owner), no admin tests now | 02:25 |
| gmann | alex_xu: yea we can do that later but this patch at least fix the overridden rule. | 02:25 |
| alex_xu | gmann: yea, that's true, so....+2 :) | 02:26 |
| gmann | alex_xu: Thanks :) | 02:26 |
| alex_xu | gmann: np | 02:26 |
| gmann | alex_xu: but i am on side to tests the each possible way which API policy support | 02:26 |
| alex_xu | gmann: yea | 02:27 |
| gmann | alex_xu: i mean like we support user_id as of now so we should tests that which we do now | 02:27 |
| alex_xu | gmann: got it | 02:27 |
| gmann | alex_xu: anyways lets discuss this more in O. more policy tests coverage may find more bits we do which we should not like user_id in start server etc | 02:27 |
| *** browne has joined #openstack-nova | 02:28 | |
| *** tojuvone has quit IRC | 02:28 | |
| alex_xu | gmann: yea | 02:29 |
| *** salv-orl_ has joined #openstack-nova | 02:30 | |
| *** devananda|dinner has quit IRC | 02:32 | |
| *** clayton has quit IRC | 02:32 | |
| *** clayton has joined #openstack-nova | 02:33 | |
| *** salv-orlando has quit IRC | 02:33 | |
| *** kevinz has joined #openstack-nova | 02:33 | |
| *** browne has quit IRC | 02:43 | |
| *** phuongnh has joined #openstack-nova | 02:43 | |
| *** mriedem has quit IRC | 02:46 | |
| *** tonytan4ever has quit IRC | 02:46 | |
| *** amotoki has quit IRC | 02:49 | |
| *** TravT has quit IRC | 02:56 | |
| *** devananda|dinner has joined #openstack-nova | 02:59 | |
| *** code-R has joined #openstack-nova | 02:59 | |
| *** code-R_ has joined #openstack-nova | 03:01 | |
| *** fragatina has quit IRC | 03:03 | |
| *** code-R has quit IRC | 03:04 | |
| openstackgerrit | Qiming Teng proposed openstack/nova: Fix api-ref doc for server-rebuild https://review.openstack.org/367730 | 03:07 |
| *** Jeffrey4l has joined #openstack-nova | 03:09 | |
| *** kaisers_ has joined #openstack-nova | 03:11 | |
| *** kaisers_ has quit IRC | 03:16 | |
| *** Apoorva has joined #openstack-nova | 03:21 | |
| *** david-lyle_ has joined #openstack-nova | 03:22 | |
| *** david-lyle_ has quit IRC | 03:22 | |
| *** armax has quit IRC | 03:24 | |
| *** rfolco has quit IRC | 03:25 | |
| *** thomasem has quit IRC | 03:25 | |
| *** ociuhandu has joined #openstack-nova | 03:32 | |
| *** wangqun has quit IRC | 03:35 | |
| *** wangqun has joined #openstack-nova | 03:35 | |
| *** ociuhandu has quit IRC | 03:36 | |
| *** dikonoor has joined #openstack-nova | 03:41 | |
| *** tonytan4ever has joined #openstack-nova | 03:44 | |
| *** sudipto has joined #openstack-nova | 03:47 | |
| *** sudipto_ has joined #openstack-nova | 03:47 | |
| *** TravT has joined #openstack-nova | 03:48 | |
| *** amotoki has joined #openstack-nova | 03:49 | |
| *** tuanla_ has joined #openstack-nova | 03:54 | |
| *** TravT has quit IRC | 03:58 | |
| *** kaisers_ has joined #openstack-nova | 03:59 | |
| *** annp has joined #openstack-nova | 04:04 | |
| *** praveen has joined #openstack-nova | 04:05 | |
| *** ssurana has quit IRC | 04:09 | |
| mgagne | melwitt: still no idea about why it's failing. Maybe it's a network config within the kernel. For now, if I limit the number of GreenPool workers to 100 (down from default 1000), it works fine. I think implement a Nova config for this would be the solution. (power_sync_greenpool_size? I don't know) | 04:10 |
| mgagne | Even if we find the kernel config to tweak, I don't think we can increase that value forever, especially if you end up with like 1000 hypervisors on the same nova-compute service. No server or load-balancer will gladly accept that amount of connections without some form of tuning. | 04:11 |
| mgagne | sync_power_state_pool_size | 04:15 |
| *** amotoki has quit IRC | 04:16 | |
| *** su_zhang has joined #openstack-nova | 04:17 | |
| *** hamzy has joined #openstack-nova | 04:19 | |
| *** woodster_ has quit IRC | 04:19 | |
| *** harlowja_at_home has joined #openstack-nova | 04:23 | |
| *** artom has quit IRC | 04:26 | |
| *** fragatina has joined #openstack-nova | 04:26 | |
| *** zul has joined #openstack-nova | 04:27 | |
| *** adisky has joined #openstack-nova | 04:27 | |
| *** artom has joined #openstack-nova | 04:27 | |
| *** fragatina has quit IRC | 04:30 | |
| *** akshai has quit IRC | 04:32 | |
| *** rama_y has joined #openstack-nova | 04:38 | |
| *** artom has quit IRC | 04:38 | |
| *** artom has joined #openstack-nova | 04:40 | |
| *** karthiks has joined #openstack-nova | 04:42 | |
| *** prateek has joined #openstack-nova | 04:43 | |
| *** coolsvap has joined #openstack-nova | 04:43 | |
| *** lixiaoy11 has quit IRC | 04:46 | |
| *** rkrum has joined #openstack-nova | 04:47 | |
| openstackgerrit | Mathieu Gagné proposed openstack/nova: Add sync_power_state_pool_size option https://review.openstack.org/367746 | 04:47 |
| *** lixiaoy1 has joined #openstack-nova | 04:52 | |
| *** thorst has quit IRC | 04:55 | |
| *** su_zhang has quit IRC | 04:55 | |
| *** su_zhang has joined #openstack-nova | 04:56 | |
| *** Apoorva has quit IRC | 04:58 | |
| *** Jeffrey4l_ has joined #openstack-nova | 05:00 | |
| *** su_zhang has quit IRC | 05:00 | |
| *** thorst has joined #openstack-nova | 05:03 | |
| *** thorst has quit IRC | 05:03 | |
| *** thorst has joined #openstack-nova | 05:03 | |
| *** Jeffrey4l has quit IRC | 05:04 | |
| *** pbandark has joined #openstack-nova | 05:08 | |
| *** su_zhang has joined #openstack-nova | 05:12 | |
| *** zul has quit IRC | 05:13 | |
| *** Shashi has quit IRC | 05:16 | |
| *** karthiks has quit IRC | 05:16 | |
| *** artom has quit IRC | 05:17 | |
| *** artom has joined #openstack-nova | 05:18 | |
| *** unicell has joined #openstack-nova | 05:18 | |
| *** claudiub|2 has joined #openstack-nova | 05:19 | |
| *** jaypipes has quit IRC | 05:21 | |
| *** jaosorior has joined #openstack-nova | 05:21 | |
| *** claudiub has joined #openstack-nova | 05:21 | |
| *** jaypipes has joined #openstack-nova | 05:23 | |
| *** zul has joined #openstack-nova | 05:23 | |
| *** jfding has joined #openstack-nova | 05:23 | |
| *** rama_y has quit IRC | 05:24 | |
| *** claudiub|2 has quit IRC | 05:25 | |
| *** moshele has joined #openstack-nova | 05:27 | |
| *** burgerk has quit IRC | 05:28 | |
| *** vishwanathj_ is now known as vishwanathj_zzz | 05:28 | |
| *** sridharg has joined #openstack-nova | 05:32 | |
| *** karthiks has joined #openstack-nova | 05:32 | |
| *** harlowja_at_home has quit IRC | 05:33 | |
| *** lpetrut has joined #openstack-nova | 05:35 | |
| *** gcb has joined #openstack-nova | 05:36 | |
| *** zul has quit IRC | 05:38 | |
| *** su_zhang has quit IRC | 05:40 | |
| *** su_zhang has joined #openstack-nova | 05:41 | |
| *** su_zhang_ has joined #openstack-nova | 05:42 | |
| *** amotoki has joined #openstack-nova | 05:43 | |
| *** su_zhang has quit IRC | 05:45 | |
| *** ratailor has joined #openstack-nova | 05:46 | |
| *** nkrinner_afk is now known as nkrinner | 05:55 | |
| *** adisky has quit IRC | 05:58 | |
| *** adisky has joined #openstack-nova | 05:58 | |
| *** praveen has quit IRC | 05:59 | |
| *** tonytan4ever has quit IRC | 06:00 | |
| *** mjura has joined #openstack-nova | 06:01 | |
| *** claudiub has quit IRC | 06:01 | |
| *** paul-carlton has joined #openstack-nova | 06:06 | |
| *** artom has quit IRC | 06:09 | |
| *** yamahata has joined #openstack-nova | 06:09 | |
| *** mjura has quit IRC | 06:09 | |
| *** mjura has joined #openstack-nova | 06:09 | |
| *** artom has joined #openstack-nova | 06:14 | |
| *** praveen has joined #openstack-nova | 06:18 | |
| *** amotoki has quit IRC | 06:19 | |
| *** moshele has quit IRC | 06:21 | |
| *** praveen has quit IRC | 06:23 | |
| *** pcaruana has joined #openstack-nova | 06:23 | |
| *** praveen has joined #openstack-nova | 06:24 | |
| *** amotoki has joined #openstack-nova | 06:28 | |
| *** harlowja_at_home has joined #openstack-nova | 06:29 | |
| *** avolkov` has joined #openstack-nova | 06:30 | |
| *** phuongnh has quit IRC | 06:31 | |
| *** phuongnh has joined #openstack-nova | 06:32 | |
| *** praveen has quit IRC | 06:32 | |
| *** praveen has joined #openstack-nova | 06:33 | |
| *** lpetrut has quit IRC | 06:36 | |
| *** ttrumm has joined #openstack-nova | 06:37 | |
| *** hferenc has joined #openstack-nova | 06:38 | |
| *** ttrumm has quit IRC | 06:39 | |
| *** ttrumm has joined #openstack-nova | 06:39 | |
| *** ttrumm has quit IRC | 06:40 | |
| *** lpetrut has joined #openstack-nova | 06:42 | |
| *** sridharg has quit IRC | 06:43 | |
| *** ttrumm has joined #openstack-nova | 06:44 | |
| *** su_zhang_ has quit IRC | 06:45 | |
| *** su_zhang has joined #openstack-nova | 06:46 | |
| *** andreas_s has joined #openstack-nova | 06:46 | |
| *** harlowja_at_home has quit IRC | 06:47 | |
| *** Dinesh_Bhor has joined #openstack-nova | 06:47 | |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Don't pass argument sqlite_db in method set_defaults https://review.openstack.org/350947 | 06:47 |
| slaweq_ | mdbooth: kashyap: hello, do You know who from core team I could ask to check patches https://review.openstack.org/#/c/363926/ and https://review.openstack.org/#/c/365756/ ? | 06:47 |
| *** harlowja_at_home has joined #openstack-nova | 06:48 | |
| *** amotoki has quit IRC | 06:48 | |
| *** su_zhang has quit IRC | 06:50 | |
| *** amotoki has joined #openstack-nova | 06:52 | |
| openstackgerrit | garyk proposed openstack/nova: Compute: ensure that InvalidDiskFormat is handled correctly https://review.openstack.org/367772 | 06:56 |
| *** harlowja_at_home has quit IRC | 06:57 | |
| *** sridharg has joined #openstack-nova | 07:00 | |
| *** sahid has joined #openstack-nova | 07:02 | |
| *** jlanoux has joined #openstack-nova | 07:02 | |
| *** amotoki has quit IRC | 07:02 | |
| *** praveen has quit IRC | 07:03 | |
| *** jamielennox|away is now known as jamielennox | 07:03 | |
| *** ttrumm has quit IRC | 07:04 | |
| openstackgerrit | garyk proposed openstack/nova: Compute: ensure that InvalidDiskFormat is handled correctly https://review.openstack.org/367772 | 07:05 |
| openstackgerrit | Prateek Arora proposed openstack/nova: Delete traces of in-progress snapshot on VM being deleted https://review.openstack.org/294513 | 07:05 |
| *** gcb has quit IRC | 07:05 | |
| *** praveen has joined #openstack-nova | 07:06 | |
| *** tesseract- has joined #openstack-nova | 07:07 | |
| *** jlanoux_ has joined #openstack-nova | 07:10 | |
| *** jlanoux has quit IRC | 07:10 | |
| *** jsheeren has joined #openstack-nova | 07:11 | |
| *** rkrum has quit IRC | 07:11 | |
| *** ttrumm has joined #openstack-nova | 07:11 | |
| *** matrohon has joined #openstack-nova | 07:12 | |
| *** ttrumm has quit IRC | 07:12 | |
| *** claudiub has joined #openstack-nova | 07:13 | |
| *** jichen has joined #openstack-nova | 07:13 | |
| *** EinstCrazy has quit IRC | 07:14 | |
| *** amotoki has joined #openstack-nova | 07:14 | |
| *** praveen has quit IRC | 07:17 | |
| *** oanson has joined #openstack-nova | 07:19 | |
| *** ihrachys has joined #openstack-nova | 07:20 | |
| *** ttrumm has joined #openstack-nova | 07:20 | |
| *** ttrumm has quit IRC | 07:20 | |
| *** lpetrut has quit IRC | 07:22 | |
| *** mgoddard has joined #openstack-nova | 07:23 | |
| *** EinstCrazy has joined #openstack-nova | 07:24 | |
| *** tuanla_ has quit IRC | 07:27 | |
| *** annp has quit IRC | 07:27 | |
| *** hieulq has quit IRC | 07:27 | |
| *** annp has joined #openstack-nova | 07:27 | |
| *** tuanla_ has joined #openstack-nova | 07:27 | |
| *** hieulq has joined #openstack-nova | 07:28 | |
| *** ttrumm has joined #openstack-nova | 07:28 | |
| *** phuongnh has quit IRC | 07:28 | |
| *** phuongnh has joined #openstack-nova | 07:29 | |
| *** ttrumm_ has joined #openstack-nova | 07:30 | |
| *** ttrumm has quit IRC | 07:33 | |
| *** code-R_ has quit IRC | 07:34 | |
| *** paul-carlton has quit IRC | 07:38 | |
| *** praveen has joined #openstack-nova | 07:39 | |
| *** gouthamr has quit IRC | 07:40 | |
| *** jlanoux_ has quit IRC | 07:40 | |
| *** Oku_OS-away is now known as Oku_OS | 07:42 | |
| *** vishwana_ has joined #openstack-nova | 07:42 | |
| *** vishwana_ has quit IRC | 07:42 | |
| *** vishwana_ has joined #openstack-nova | 07:43 | |
| *** vishwanathj_zzz has quit IRC | 07:44 | |
| *** shangxdy has joined #openstack-nova | 07:45 | |
| *** tonytan4ever has joined #openstack-nova | 07:46 | |
| *** jfding has quit IRC | 07:48 | |
| *** zenoway has joined #openstack-nova | 07:50 | |
| *** paul-carlton has joined #openstack-nova | 07:51 | |
| *** tonytan4ever has quit IRC | 07:52 | |
| *** sridharg has quit IRC | 07:54 | |
| *** EinstCrazy has quit IRC | 07:55 | |
| *** EinstCrazy has joined #openstack-nova | 07:55 | |
| *** twaldrop has quit IRC | 07:56 | |
| *** sridharg has joined #openstack-nova | 07:56 | |
| *** zz_dimtruck is now known as dimtruck | 07:57 | |
| *** thorst has quit IRC | 07:57 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** sridhargaddam has joined #openstack-nova | 08:01 | |
| *** EinstCrazy has quit IRC | 08:01 | |
| *** thorst has joined #openstack-nova | 08:01 | |
| *** sridharg has quit IRC | 08:01 | |
| *** EinstCrazy has joined #openstack-nova | 08:02 | |
| *** zzzeek has joined #openstack-nova | 08:02 | |
| *** derekh has joined #openstack-nova | 08:05 | |
| *** yassine has joined #openstack-nova | 08:05 | |
| *** yassine is now known as Guest47093 | 08:05 | |
| *** thorst has quit IRC | 08:06 | |
| *** Guest47093 is now known as yassou | 08:06 | |
| *** paul-carlton has quit IRC | 08:07 | |
| *** tpatzig has joined #openstack-nova | 08:08 | |
| *** tpatzig_ has quit IRC | 08:09 | |
| *** thorst has joined #openstack-nova | 08:09 | |
| *** mkoderer has quit IRC | 08:10 | |
| *** SS-SP has quit IRC | 08:10 | |
| *** liuyulong has quit IRC | 08:10 | |
| *** liuyulong has joined #openstack-nova | 08:10 | |
| *** mkoderer has joined #openstack-nova | 08:11 | |
| kashyap | slaweq_: I think most core folks might be busy with some other priority stuff, let's give it a day or two. | 08:13 |
| slaweq_ | kashyap: ok, I will wait more | 08:17 |
| *** mingyu has joined #openstack-nova | 08:17 | |
| *** thorst has quit IRC | 08:17 | |
| *** mingyu has quit IRC | 08:20 | |
| openstackgerrit | Mikhail Feoktistov proposed openstack/nova: Check hypervisor version while getting bus type https://review.openstack.org/282398 | 08:21 |
| *** hoangcx has quit IRC | 08:22 | |
| *** markus_z has joined #openstack-nova | 08:24 | |
| *** vishwana_ has quit IRC | 08:25 | |
| *** vishwanathj_zzz has joined #openstack-nova | 08:25 | |
| *** EinstCrazy has quit IRC | 08:26 | |
| *** hoangcx has joined #openstack-nova | 08:26 | |
| *** thorst has joined #openstack-nova | 08:27 | |
| *** EinstCra_ has joined #openstack-nova | 08:29 | |
| *** EinstCra_ has quit IRC | 08:30 | |
| *** salv-orlando has joined #openstack-nova | 08:30 | |
| *** mingyu has joined #openstack-nova | 08:31 | |
| *** salv-orl_ has quit IRC | 08:32 | |
| *** EinstCrazy has joined #openstack-nova | 08:33 | |
| *** EinstCrazy has quit IRC | 08:35 | |
| *** EinstCrazy has joined #openstack-nova | 08:35 | |
| *** jaosorior is now known as jaosorior_lunch | 08:40 | |
| *** Jeffrey4l_ has quit IRC | 08:42 | |
| *** jlanoux has joined #openstack-nova | 08:45 | |
| *** zarkone has joined #openstack-nova | 08:53 | |
| zarkone | Hello! I've faced some troubles with using 3rd party scheduler driver | 08:54 |
| zarkone | seems like in Newton scheduler uses entry points mechanism to find driver module | 08:54 |
| zarkone | but entry points are creating at build\setup phase | 08:55 |
| zarkone | that means that I could use only built-in drivers... | 08:56 |
| *** gszasz has joined #openstack-nova | 08:56 | |
| *** lixiaoy1 has quit IRC | 08:57 | |
| *** efoley has joined #openstack-nova | 08:57 | |
| slaweq_ | zarkone: can't You add own entry point for new driver? | 08:58 |
| zarkone | slaweq_: thanks, I just thought that entry points could only be added at build phase.. could you please point me where to find info about adding entry point for installed module? | 08:59 |
| *** avolkov` has quit IRC | 09:00 | |
| zarkone | I've done it by editing etry_points.txt in .egg-info | 09:00 |
| *** zenoway has quit IRC | 09:00 | |
| zarkone | how to o that properly? | 09:00 |
| zarkone | *do | 09:00 |
| *** bauzas is now known as bauwser | 09:00 | |
| *** zenoway has joined #openstack-nova | 09:01 | |
| *** zenoway has quit IRC | 09:01 | |
| *** zenoway has joined #openstack-nova | 09:01 | |
| *** sdake has joined #openstack-nova | 09:02 | |
| *** amotoki has quit IRC | 09:04 | |
| *** thorst has quit IRC | 09:04 | |
| *** danpb has joined #openstack-nova | 09:05 | |
| *** thorst has joined #openstack-nova | 09:05 | |
| *** ttrumm_ has quit IRC | 09:05 | |
| *** ttrumm has joined #openstack-nova | 09:05 | |
| *** mfedosin has joined #openstack-nova | 09:06 | |
| *** EinstCrazy has quit IRC | 09:06 | |
| *** lpetrut has joined #openstack-nova | 09:08 | |
| *** EinstCrazy has joined #openstack-nova | 09:08 | |
| *** Jeffrey4l has joined #openstack-nova | 09:08 | |
| *** dimtruck is now known as zz_dimtruck | 09:10 | |
| *** thorst has quit IRC | 09:10 | |
| *** cdent has joined #openstack-nova | 09:11 | |
| *** code-R has joined #openstack-nova | 09:12 | |
| *** code-R_ has joined #openstack-nova | 09:13 | |
| johnthetubaguy | gmann: alex_xu: +1 to tidying the rest of that up next cycle | 09:16 |
| *** sdake_ has joined #openstack-nova | 09:16 | |
| alex_xu | johnthetubaguy: yea | 09:16 |
| *** code-R has quit IRC | 09:17 | |
| *** mingyu has quit IRC | 09:17 | |
| *** mingyu has joined #openstack-nova | 09:17 | |
| *** sdake has quit IRC | 09:18 | |
| *** mingyu_ has joined #openstack-nova | 09:19 | |
| *** mingyu_ has quit IRC | 09:19 | |
| *** mingyu has quit IRC | 09:19 | |
| *** aloga has quit IRC | 09:23 | |
| *** aloga has joined #openstack-nova | 09:24 | |
| *** artom has quit IRC | 09:25 | |
| gmann | johnthetubaguy: yup. Thanks | 09:25 |
| *** mingyu has joined #openstack-nova | 09:26 | |
| *** thorst has joined #openstack-nova | 09:26 | |
| *** artom has joined #openstack-nova | 09:27 | |
| *** mingyu has quit IRC | 09:27 | |
| *** moshele has joined #openstack-nova | 09:27 | |
| *** kevinz has quit IRC | 09:29 | |
| prateek | johnthetubaguy, can you please review https://review.openstack.org/#/c/294513/, i have addressed your comments | 09:31 |
| *** thorst has quit IRC | 09:31 | |
| *** jinquan has quit IRC | 09:34 | |
| *** oanson has quit IRC | 09:35 | |
| *** takedakn has joined #openstack-nova | 09:43 | |
| slaweq_ | zarkone: You can add it manually in entry_point.txt file where python package is installed or add to setup.cfg file in Your project before install it | 09:44 |
| zarkone | yes, but I want to add entry point to scheduler | 09:45 |
| *** rwmjones has quit IRC | 09:45 | |
| zarkone | without rebuild | 09:46 |
| *** rwmjones has joined #openstack-nova | 09:46 | |
| *** takedakn has quit IRC | 09:47 | |
| openstackgerrit | Jianghua Wang proposed openstack/nova: XenAPI: resolve the fetch_bandwidth failure https://review.openstack.org/365838 | 09:49 |
| *** artom has quit IRC | 09:50 | |
| *** artom has joined #openstack-nova | 09:51 | |
| *** PaulMurray has joined #openstack-nova | 09:52 | |
| PaulMurray | mdbooth, hi - did you make progress with your DVR setup ? | 09:53 |
| *** tojuvone_ is now known as tojuvone | 09:55 | |
| *** Guest52040 has joined #openstack-nova | 09:56 | |
| *** sdake has joined #openstack-nova | 09:57 | |
| *** praveen has quit IRC | 09:57 | |
| *** yamahata has quit IRC | 09:59 | |
| *** hoangcx has quit IRC | 09:59 | |
| *** sdake_ has quit IRC | 10:00 | |
| openstackgerrit | sahid proposed openstack/nova: libvirt: add ps2mouse in choice for pointer_model https://review.openstack.org/367909 | 10:02 |
| *** annp has quit IRC | 10:05 | |
| *** hieulq has quit IRC | 10:06 | |
| *** shangxdy has quit IRC | 10:07 | |
| *** cdent has quit IRC | 10:09 | |
| *** phuongnh has quit IRC | 10:10 | |
| *** EinstCrazy has quit IRC | 10:10 | |
| *** tuanla_ has quit IRC | 10:11 | |
| *** thorst has joined #openstack-nova | 10:11 | |
| prateek | alex_xu, can you please look into https://review.openstack.org/#/c/327706/ | 10:19 |
| *** riccardo_ has joined #openstack-nova | 10:20 | |
| *** moshele has quit IRC | 10:23 | |
| *** artom has quit IRC | 10:26 | |
| *** karthiks has quit IRC | 10:26 | |
| slaweq_ | zarkone: You can add entry point in Your project and scheduler should find it if it will be in same "namespace" | 10:29 |
| *** artom has joined #openstack-nova | 10:31 | |
| *** sdake_ has joined #openstack-nova | 10:31 | |
| *** moshele has joined #openstack-nova | 10:32 | |
| *** sdake has quit IRC | 10:33 | |
| *** mfedosin has quit IRC | 10:34 | |
| *** mfedosin has joined #openstack-nova | 10:36 | |
| *** cdent has joined #openstack-nova | 10:36 | |
| *** tbachman has quit IRC | 10:37 | |
| *** jaosorior_lunch is now known as jaosorior | 10:41 | |
| *** ttrumm_ has joined #openstack-nova | 10:41 | |
| *** ttrumm_ has quit IRC | 10:42 | |
| *** mingyu has joined #openstack-nova | 10:43 | |
| *** ttrumm has quit IRC | 10:45 | |
| *** artom has quit IRC | 10:50 | |
| *** artom has joined #openstack-nova | 10:53 | |
| zarkone | slaweq_: ah! thanks for this! | 10:53 |
| *** takedakn has joined #openstack-nova | 10:53 | |
| slaweq_ | zarkone: np, I hope I helped You :) | 10:54 |
| zarkone | yes, for sure | 10:54 |
| slaweq_ | :) | 10:54 |
| zarkone | I had this idea, but haven't tried, thought that it shouldn't work for some reason | 10:54 |
| zarkone | need to improve my python + openstack libs knowledge, may be I don't get some basics | 10:55 |
| slaweq_ | zarkone: I should work | 10:56 |
| slaweq_ | I'm doing such things for my own projects | 10:56 |
| slaweq_ | to extend mostly neutron but should be the same | 10:57 |
| *** cdelatte has joined #openstack-nova | 10:58 | |
| *** zarkone has quit IRC | 11:00 | |
| *** artom has quit IRC | 11:01 | |
| *** artom has joined #openstack-nova | 11:01 | |
| *** rtheis has joined #openstack-nova | 11:03 | |
| *** mingyu has quit IRC | 11:04 | |
| *** praveen has joined #openstack-nova | 11:04 | |
| *** mfedosin has quit IRC | 11:05 | |
| PaulMurray | mdbooth, you there ? | 11:07 |
| mdbooth | PaulMurray: I am | 11:07 |
| PaulMurray | mdbooth, I produced the network outage you were looking for | 11:07 |
| mdbooth | PaulMurray: Incidentally, I've found a friendly neutroner at Red Hat who's hopefully going to help with my reproducer shortly | 11:08 |
| mdbooth | PaulMurray: Excellent! What was the setup? | 11:08 |
| PaulMurray | Its a DVR setup - actually a standard install of Helion - so its not on head | 11:08 |
| PaulMurray | similar to devstack | 11:09 |
| PaulMurray | running in vagrant vms | 11:09 |
| mdbooth | I tried DVR in devstack briefly, but couldn't get networking working at all | 11:09 |
| PaulMurray | (vagrant managed - that's irrelevant I guess) | 11:09 |
| mdbooth | PaulMurray: But crucially was your test the same as mine? | 11:09 |
| PaulMurray | I did what you described in your email | 11:09 |
| mdbooth | Namely 2 instances on the same tenant network pinging each other? | 11:09 |
| PaulMurray | I put a 30 second pause in post_live_migration_at_destination | 11:09 |
| mdbooth | Cool | 11:10 |
| PaulMurray | I was pinging them from the desktop | 11:10 |
| mdbooth | Could you try again, pinging between them if you still have it running? | 11:10 |
| PaulMurray | When the switchover happens there is a short pause and some packets lost before pings restart | 11:10 |
| *** mingyu has joined #openstack-nova | 11:10 | |
| mdbooth | It's just that it doesn't require me to get external routing working :) | 11:10 |
| *** akshai has joined #openstack-nova | 11:11 | |
| mdbooth | However, if it's required to validate then I'll do it | 11:11 |
| PaulMurray | with a 30 second sleep in the method there is a 30s pause | 11:11 |
| PaulMurray | When running with DVR it shouldn't make any difference (I think) | 11:11 |
| mdbooth | So, when my friendly red hatter gets back from lunch I'm going to have a go with devstack, just because it's less activation energy for me :) | 11:12 |
| mdbooth | However, if we're bogging down, could I get steps to do what you did? | 11:12 |
| PaulMurray | of course - actually I have two desktops, one always has devstack running the other always has hellion, but the devstack one is not DVR, so I wanted to quickly prove that this actually happens with DVR first | 11:13 |
| PaulMurray | I could have a go at doing the setup on devstack too, but I have the same problem you do - my friendly neutron people are in US | 11:13 |
| mdbooth | Mine's in spain, so we've been talking already | 11:14 |
| mdbooth | Should be back shortly | 11:14 |
| mdbooth | It's ajo, in fact | 11:14 |
| *** sahid has quit IRC | 11:15 | |
| *** stewie925 has quit IRC | 11:16 | |
| *** syjulian has joined #openstack-nova | 11:17 | |
| *** sdague has joined #openstack-nova | 11:18 | |
| *** stewie925 has joined #openstack-nova | 11:21 | |
| *** baoli has joined #openstack-nova | 11:24 | |
| *** baoli_ has joined #openstack-nova | 11:27 | |
| *** jschlueter has joined #openstack-nova | 11:28 | |
| *** panbalag has quit IRC | 11:29 | |
| *** panbalag has joined #openstack-nova | 11:30 | |
| *** dikonoor has quit IRC | 11:30 | |
| *** baoli has quit IRC | 11:30 | |
| *** GB21 has joined #openstack-nova | 11:31 | |
| *** gcb has joined #openstack-nova | 11:32 | |
| *** moshele has quit IRC | 11:33 | |
| *** takedakn has quit IRC | 11:33 | |
| *** baoli_ has quit IRC | 11:34 | |
| *** dikonoor has joined #openstack-nova | 11:34 | |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: inject files when config drive is not requested https://review.openstack.org/335676 | 11:35 |
| *** baoli has joined #openstack-nova | 11:38 | |
| *** baoli has quit IRC | 11:44 | |
| cdent | sdague are you a cookies or cake kind of guy? | 11:46 |
| cdent | I asked because I think you deserve whichever is your preference | 11:46 |
| sdague | either is good :) | 11:46 |
| sdague | and thanks | 11:46 |
| cdent | I think, although, I'm not certain, that people weren't understanding the point of the logging you are doing, until your recent comment to say "we're trying to make sure the server is behaving" | 11:47 |
| cdent | instead they were thinking of it is as confirming action (rather than _correct_ action) | 11:48 |
| cdent | in any case | 11:48 |
| * cdent cookies sdague | 11:48 | |
| *** tonytan4ever has joined #openstack-nova | 11:48 | |
| sdague | yeh, maybe, it feels like there is confusion in that review | 11:49 |
| prateek | cdent, sdague, i am feeling hungry by reading all this :), cdent, can you please review https://review.openstack.org/#/c/327706/ whenever you are free ? | 11:49 |
| cdent | prateek: yeah, gonna get more coffee and then will have a look at that and a couple other things | 11:50 |
| *** artom has quit IRC | 11:50 | |
| prateek | cdent, ok thanks, | 11:50 |
| *** praveen has quit IRC | 11:51 | |
| *** artom has joined #openstack-nova | 11:51 | |
| sdague | prateek: what are we doing today with that? | 11:51 |
| *** mgoddard has quit IRC | 11:52 | |
| *** wangqun has quit IRC | 11:52 | |
| prateek | sdague, there was one method that was the an issue of discussion in that patch, alex_xu gave a suggestion to make it fine | 11:53 |
| prateek | sdague, i am sorry if i misunderstood that question :/ | 11:53 |
| *** tonytan4ever has quit IRC | 11:53 | |
| sdague | prateek: no, I mean what is current behavior | 11:53 |
| prateek | it throws a 500 | 11:53 |
| *** GB21 has quit IRC | 11:53 | |
| prateek | sdague, it throws a 500 | 11:53 |
| *** lucasagomes is now known as lucas-hungry | 11:55 | |
| openstackgerrit | Mikhail Feoktistov proposed openstack/nova: Check hypervisor version while getting bus type https://review.openstack.org/282398 | 11:55 |
| sdague | I guess my concern is wrapping another thing in here - https://review.openstack.org/#/c/327706/14/nova/api/openstack/extensions.py | 11:55 |
| *** jed56 has quit IRC | 11:55 | |
| sdague | and that Neutron Connect failure looks different than Cinder / Glance connect failure | 11:56 |
| *** baoli has joined #openstack-nova | 11:59 | |
| *** code-R_ has quit IRC | 12:00 | |
| *** code-R has joined #openstack-nova | 12:00 | |
| *** gongysh has quit IRC | 12:00 | |
| prateek | sdague, when i came across the bug ,i thought the solution to overcome this problem would be writing a client wrapper and then catching the exception | 12:01 |
| prateek | sdague, also there was already a patch to write the client wrapper and i am extending that to catch this exception, | 12:02 |
| prateek | from what i see glance also had similar mechanism | 12:02 |
| sdague | prateek: so this really only has an impact on the proxy api calls that are deprecated, right? | 12:04 |
| sdague | although it changes things for the compute -> neutron path as well? | 12:05 |
| prateek | sdague, in nova/volume/cinder.py, they have the similar thing which they have done as decorator | 12:05 |
| *** kaisers__ has joined #openstack-nova | 12:06 | |
| *** code-R has quit IRC | 12:06 | |
| prateek | sdague, i am not sure about the statement "impact on the proxy api calls that are deprecated, right?" | 12:06 |
| *** raj_singh_ has joined #openstack-nova | 12:06 | |
| sdague | the only time a user is going to see the different return code is if they make a nova api call, which calls neutron on the spot | 12:07 |
| sdague | but, most nova api calls don't do that | 12:07 |
| sdague | except the direct proxy interfaces, which are all deprecated | 12:07 |
| *** cdent has quit IRC | 12:08 | |
| sdague | http://developer.openstack.org/api-ref/compute/#networks-os-networks-deprecated | 12:08 |
| *** kaisers_ has quit IRC | 12:08 | |
| sdague | personally, I'd rather see this look the same as the cinder / glance approaches, because the more different ways we have to do this in the code, the harder it is to keep track of | 12:09 |
| *** kaisers__ has quit IRC | 12:10 | |
| *** raj_singh_ has quit IRC | 12:11 | |
| prateek | sdague, the approach to that is already there and different from glance | 12:11 |
| *** artom has quit IRC | 12:12 | |
| *** mfedosin has joined #openstack-nova | 12:12 | |
| *** artom has joined #openstack-nova | 12:13 | |
| prateek | sdague, the calls were already wrapped in a decorator and the excpetions catched http://paste.openstack.org/show/570186/ | 12:13 |
| sdague | prateek: right that's fine | 12:14 |
| prateek | sdague, the neutron wrapper is already present and is a recently and already merged patch, so i guess we would have to have different approaches | 12:14 |
| prateek | sdague, its just that it was missing some cases and i am handling that | 12:14 |
| sdague | here is the thing, the 504 change that you have there, will only be activated when people make API calls that we tell them not to, and that will be deleted in the medium future | 12:15 |
| *** sahid has joined #openstack-nova | 12:15 | |
| sdague | it feels like the bulk of the patch is for getting that part correct | 12:15 |
| sdague | but that part is about to all be deleted | 12:15 |
| prateek | sdague, also i feel that since it connects to neutron also , the part which is not being deprecated is also hit | 12:16 |
| PaulMurray | mdbooth, FYI - repeated the test pinging from another VM. Same result - I only have two compute hosts so the vm doing the pings is on one of the hosts involved in the migration. Migrated both ways so one way its collated with the source, the other way with the destination. Same result both times | 12:16 |
| sdague | prateek: which part? | 12:16 |
| prateek | for connection to neutron also and not only nova-net | 12:17 |
| sdague | prateek: ok, but there a 2 ways we talk to neutron | 12:17 |
| sdague | directly on behalf of a user - when doing os-networks or something | 12:17 |
| sdague | that's all deprecated and slated for deletion | 12:17 |
| sdague | and behind the scenes, from the nova-compute daemon as a service user | 12:17 |
| sdague | which is not during an API call | 12:18 |
| sdague | it is a behind the scenes async operation | 12:18 |
| sdague | so will never return the neutron failure to the user | 12:18 |
| sdague | it will instead either be quietly dropped, or put the compute instance into ERROR state | 12:18 |
| *** hoangcx has joined #openstack-nova | 12:18 | |
| mdbooth | PaulMurray: Thanks | 12:19 |
| sdague | we really really really try to avoid synchronously calling another API system in the middle of complex logic for our APIs, like in server create | 12:19 |
| mdbooth | Currently trying to reproduce without dvr with ajo | 12:19 |
| prateek | sdague, maybe we can drop the code when we drop the whole, but would anyways help users using old versions of openstack , maybe in backporting a patch for a particular problem | 12:20 |
| sdague | prateek: I really don't want to backport things like the 504 change | 12:21 |
| prateek | sdague, ok | 12:22 |
| mdbooth | PaulMurray: Incidentally, are you testing with the recent change to setup_networks_..... ? | 12:22 |
| sdague | prateek: because then we have to document that as an expection | 12:22 |
| sdague | but when it's expected, is massively vague | 12:22 |
| sdague | because it's not tied to a microversion | 12:23 |
| prateek | sdague, ok | 12:25 |
| *** mriedem has joined #openstack-nova | 12:26 | |
| *** ratailor has quit IRC | 12:26 | |
| ftersin | sdague: hi. do you now, who can help me with live migration. i was wandering if it is possible to migrate a booted from volume instance with config drive | 12:27 |
| *** pradk has joined #openstack-nova | 12:28 | |
| *** hoangcx2 has joined #openstack-nova | 12:29 | |
| prateek | sdague, can you suggest a way on how i can take the patch forward ? | 12:30 |
| *** mingyu has quit IRC | 12:31 | |
| PaulMurray | mdbooth, no, mine is on mitaka | 12:31 |
| *** hoangcx has quit IRC | 12:31 | |
| sdague | prateek: how about just remove the exception translation | 12:31 |
| sdague | keep the logging | 12:32 |
| riccardo_ | Hi guys, I a Python file I have the following line: fairness_hosts = self.servicegroup_api.get_all("fairness") and it makes a call to get_all of nova/servicegroup/api.py. But this method is not existent anymore. How can I get a list of all hosts that run the service "fairness"? | 12:32 |
| *** rfolco has joined #openstack-nova | 12:33 | |
| sdague | prateek: I'm looking at the bug in detail now, so this was about security groups on list instances | 12:35 |
| prateek | sdague, yes | 12:35 |
| sdague | which is one of the few places it exposes on non proxy api calls | 12:35 |
| *** mfedosin has quit IRC | 12:36 | |
| *** sridhargaddam has quit IRC | 12:36 | |
| sdague | so, honestly, I think this should just get solved with the proper approach of caching the security groups, I think papering around which error response happens to the user here isn't really making things substantial better for folks | 12:37 |
| prateek | sdague, yes, basically , if the neutron server goes down and it tries to list instances, the API throws 500 | 12:37 |
| sdague | right, and in your patch it will still not work, but throw a 504 | 12:37 |
| prateek | yes | 12:37 |
| *** sridhargaddam has joined #openstack-nova | 12:37 | |
| prateek | but you would know what actually happened | 12:37 |
| sdague | right, that seems not much better | 12:37 |
| sdague | only if you have that patch | 12:37 |
| prateek | rather than getting a 500 | 12:37 |
| sdague | which isn't behind a microversion, so you have no idea if you are supposed to get that error or not | 12:38 |
| johnthetubaguy | so the network_info object doesn't have security groups already? | 12:38 |
| johnthetubaguy | I say object, I mean mega blob | 12:38 |
| prateek | no, it doesn't cache | 12:39 |
| johnthetubaguy | seems like a mistake to only cache half the data | 12:39 |
| johnthetubaguy | the thing about report errors, is they need to be useful to the person who gets them | 12:40 |
| *** wznoinsk has quit IRC | 12:40 | |
| johnthetubaguy | if the admin can see it was clearly a neutron API availability issue, I think we are good enough for now | 12:40 |
| prateek | yup, i included the error message along with NeutronConnectionFailed for that | 12:40 |
| johnthetubaguy | i.e. +1 what I think sdague is saying | 12:40 |
| johnthetubaguy | so its not the API response consumer | 12:41 |
| sdague | johnthetubaguy: right, I'm fine with the more explicit log message | 12:41 |
| johnthetubaguy | its the person reading the API logs that can fix things | 12:41 |
| *** tblakes has joined #openstack-nova | 12:41 | |
| johnthetubaguy | yeah, +1 better logs for this case | 12:41 |
| sdague | but changing the API returns here seems to have little gain | 12:41 |
| johnthetubaguy | +1 | 12:41 |
| johnthetubaguy | its not a contract we want to keep long term | 12:41 |
| sdague | and it seems like there has been 5 months working on just that instead of the security_group cache which would solve the problem for the user | 12:41 |
| sdague | johnthetubaguy: agreed | 12:41 |
| prateek | even if we cache the security groups the neutron server is still down, | 12:44 |
| prateek | the user would get an error somewhere | 12:44 |
| *** mfedosin has joined #openstack-nova | 12:44 | |
| sdague | prateek: where? | 12:44 |
| johnthetubaguy | we wouldn't know or care at that point, when listing instances, sure boot instance would then fail | 12:44 |
| sdague | but boot instance fails after the API disconnects, so it doesn't return to the user | 12:45 |
| prateek | sdague, wherever they would need neutron server, yeah like booting a server | 12:45 |
| sdague | prateek: the user wouldn't get that error | 12:45 |
| mriedem | sahid: i had a question in https://review.openstack.org/#/c/367909/ about making ps2mouse the default | 12:45 |
| *** cdent has joined #openstack-nova | 12:45 | |
| openstackgerrit | Ukesh Kumar proposed openstack/nova: Wrong description for "name" field of rebuild API https://review.openstack.org/367989 | 12:45 |
| mriedem | sahid: actually i wonder if we should make usbtablet the default, | 12:45 |
| mriedem | since use_usb_tablet defaults to True | 12:45 |
| *** salv-orlando has quit IRC | 12:46 | |
| prateek | sdague, agreed the user might not get an error ther | 12:46 |
| sdague | prateek: not might not, definitely won't | 12:46 |
| prateek | ok :) | 12:47 |
| johnthetubaguy | and if we ever gave them that error, there is nothing much they could do with it | 12:47 |
| *** salv-orlando has joined #openstack-nova | 12:47 | |
| *** wznoinsk has joined #openstack-nova | 12:47 | |
| *** sdake_ is now known as sdake | 12:47 | |
| sdague | mriedem: I responded to your comments here - https://review.openstack.org/#/c/366186/ - the expensive bits are already flagged behind being at debug log level | 12:47 |
| *** salv-orlando has quit IRC | 12:48 | |
| mriedem | sdague: ah totally missed that | 12:48 |
| *** salv-orlando has joined #openstack-nova | 12:48 | |
| mriedem | on the sec group caching thing you guys are talking about above, we've had a bug for that for awhile | 12:49 |
| sdague | cdent provided that feedback in the first iterations of that patch, which was good feedback | 12:49 |
| prateek | johnthetubaguy, sdague, so the better alternative here is to catch the exception that is coming here and inform the user about it (withot having a new exception) and log the error message properly | 12:49 |
| prateek | is it ? | 12:49 |
| johnthetubaguy | were saying inform the operator by some good logging | 12:49 |
| prateek | or a different approach altogether is needed ? | 12:49 |
| sdague | prateek: yeh, just add the extra log message for the operator | 12:49 |
| mriedem | prateek: sdague: johnthetubaguy: this bug https://bugs.launchpad.net/nova/+bug/1567655 | 12:49 |
| openstack | Launchpad bug 1567655 in OpenStack Compute (nova) "500 error when trying to list instances and neutron-server is down" [Medium,In progress] - Assigned to Prateek Arora (parora) | 12:49 |
| sdague | mriedem: I know | 12:49 |
| *** tonytan4ever has joined #openstack-nova | 12:49 | |
| mriedem | ok, there have been several duplicate bugs like this | 12:50 |
| sdague | there is a patch to just change the error reporting, which I don't think is worth it | 12:50 |
| sdague | we need to actually fix the thing | 12:50 |
| mriedem | ^ is the one where dansmith suggested we cache secgroups | 12:50 |
| sdague | and stop jumping around the edges | 12:50 |
| sdague | mriedem: yep | 12:50 |
| sdague | we should do that | 12:50 |
| sdague | but there has been no work on that from what I can tell, the focus was on the suggestion to just change the error code | 12:50 |
| mdbooth | PaulMurray: So, just had a pretty intense session with ajo trying to reproduce that with 'regular' ovs (there must be a proper name for that), and we think it just works in that case. | 12:51 |
| mriedem | alright, well the caching thing is ocata anyway | 12:51 |
| mdbooth | PaulMurray: So, looks like I'm going to have to go down the DVR route. | 12:51 |
| sdague | mriedem: agreed | 12:51 |
| *** lucas-hungry is now known as lucasagomes | 12:51 | |
| prateek | sdague, John341 mriedem , i can take that up | 12:51 |
| sdague | mriedem: it just doesn't seem worth changing the API, having to document a new error that's not used anywhere else, when our goal is to delete all paths that need it | 12:52 |
| *** jaosorior has quit IRC | 12:52 | |
| *** salv-orlando has quit IRC | 12:52 | |
| *** jaosorior has joined #openstack-nova | 12:52 | |
| *** gcb has quit IRC | 12:53 | |
| prateek | sdague, mriedem , johnthetubaguy , it would be good if you can point me to something that is having any (or none) degree of similarity to the caching we are talking about here otherwise i would start from scratch | 12:53 |
| *** salv-orlando has joined #openstack-nova | 12:53 | |
| sdague | mriedem: anyway, I'm going to be offline for a good chunk of the day. I feel like there is a lot of confusion on https://review.openstack.org/#/c/366186/. I'm going to try to update the comment message to explain where I'm coming from and why I think it's kind of critical. | 12:53 |
| *** tonytan4ever has quit IRC | 12:54 | |
| johnthetubaguy | prateek: was just looking for a link, have a peak in here: https://github.com/openstack/nova/blob/master/nova/network/neutronv2/api.py#L1267 | 12:54 |
| mriedem | sdague: i just replied in that one | 12:55 |
| mriedem | prateek: there is a comment in the bug report about adding a new column to the instance_info_caches table for serialized security groups | 12:55 |
| mriedem | we'd basically be serializing and deserializing SecurityGroupList for a given instance.security_groups into that table | 12:56 |
| mriedem | on a GET if it's not there we'd have to pull the security groups from neutron, and then store it in the cache, | 12:56 |
| mriedem | and there is code to 'heal' that cache, | 12:56 |
| mriedem | and there is a 'refresh_cache' decorator in the network API code for updating that whenever we change network resources tied to an instance | 12:57 |
| mriedem | it's definitely not trivial | 12:57 |
| mriedem | especially because the network_api and security_group_api are in different interfaces | 12:57 |
| mriedem | the network api would just have a handle to the secgroup api in order to refresh that cache i suppose | 12:57 |
| mriedem | anyway, it's probably more than a bug, it's probably a bp/spec given the various considerations for getting this right | 12:58 |
| prateek | mriedem, I can try doing it if you think it's worth it , also "As a partial fix we can backport, we should gracefully handle the connection error when trying to talk to neutron API and return a more appropriate response code (503 maybe?). ," i was taking the shorter route the second part of the comment ;) | 12:58 |
| mriedem | prateek: sounds like that's been rejected as an idea | 12:58 |
| mriedem | i'm ok with not doing a hacky workaround for now since this is a latent bug | 12:58 |
| prateek | if the idea is rejected , then shouldn't we just go ahead with what we have right now ? | 12:59 |
| prateek | we would have something rather than nothing | 12:59 |
| prateek | mriedem, ^ | 13:00 |
| *** Matias has joined #openstack-nova | 13:00 | |
| *** Matias has left #openstack-nova | 13:00 | |
| *** Matias has joined #openstack-nova | 13:01 | |
| sahid | mriedem: ah yes.. let me look at your comment on the review | 13:01 |
| *** thorst has quit IRC | 13:02 | |
| *** andreas_s has quit IRC | 13:02 | |
| mriedem | prateek: i don't know what patch you're talking about | 13:03 |
| prateek | https://review.openstack.org/#/c/327706/14 | 13:04 |
| openstackgerrit | Sean Dague proposed openstack/nova: Additional logging for placement API https://review.openstack.org/366186 | 13:04 |
| sdague | mriedem: ok, I updated the commit message | 13:04 |
| *** woodster_ has joined #openstack-nova | 13:04 | |
| prateek | mriedem, https://review.openstack.org/#/c/327706, i was talking about this | 13:05 |
| *** mingyu has joined #openstack-nova | 13:05 | |
| *** sdake has quit IRC | 13:05 | |
| *** mingyu has quit IRC | 13:05 | |
| mriedem | sdague: can you make it TRACE logging rather than DEBUG? | 13:05 |
| sdague | basically, there is almost no other usage of the debug log level during this release, so the concern about not using DEBUG for this seems odd to me | 13:05 |
| sdague | mriedem: I thought you said trace is not in newton | 13:05 |
| mriedem | we can't configure it w/o those oslo.log patches | 13:05 |
| mriedem | but we could if we needed to debug the gate via a depends-on | 13:06 |
| mriedem | if needed | 13:06 |
| mriedem | since i thought that was your concern | 13:06 |
| sdague | I am fine making it trace in ocata | 13:06 |
| mriedem | honestly we're already hurting the scheduler performance in the gate because of https://review.openstack.org/#/c/366283/ | 13:06 |
| sdague | I guess, that's the part I don't understand. | 13:06 |
| sdague | mriedem: that's pretty different | 13:06 |
| mriedem | i'm suggesting TRACE since it's an audit block that queries the DB for sanity checking and debugging, | 13:07 |
| mriedem | you want something in for debugging this in newton, | 13:07 |
| mriedem | i'd be fine with getting it in under TRACE, | 13:07 |
| sdague | mriedem: right, but you said you can't configure it in newton | 13:07 |
| mriedem | yes we don't have a config option to toggle that in newton w/o the oslo.log patch, but we could use that if necessary via depends-on until it lands in ocata if needed | 13:07 |
| mriedem | i thought your main reason for getting this in was gate debugging? | 13:08 |
| sdague | mriedem: it's any debugging | 13:08 |
| sahid | mriedem: I replied to your comment I don't really have a strong opinion on that point | 13:08 |
| sdague | mriedem: can you just go through the commit message and read what I tried to write there? | 13:09 |
| *** mdrabe has joined #openstack-nova | 13:09 | |
| mriedem | sdague: so if i'm a newton operator, and the placement stuff isn't being used in the scheduler for placement decisions, why would i care to debug this? | 13:09 |
| sdague | mriedem: then don't run placement under debug | 13:09 |
| mriedem | it seems like we want this debugging in place as devs to sort this out, but in reality it's not actually being used | 13:09 |
| sdague | and problem solved | 13:09 |
| sahid | danpb: can you have a look at https://review.openstack.org/#/c/367909/ | 13:09 |
| *** claudiub has quit IRC | 13:10 | |
| *** smatzek has joined #openstack-nova | 13:10 | |
| mriedem | fair point | 13:10 |
| johnthetubaguy | so no one should run placement in production at this point anyways, I am assuming? | 13:10 |
| sdague | there are 3 messages at debug level in newton | 13:10 |
| *** riccardo_ has quit IRC | 13:10 | |
| johnthetubaguy | or at least, it doesn't help them yet | 13:10 |
| sdague | 2 of which are added in this patch | 13:11 |
| mriedem | so one plan is land this debug patch in newton for our sake, change to trace in ocata once the oslo.log change is released, | 13:11 |
| sdague | the 3rd is the api start | 13:11 |
| sdague | mriedem: yes, I would be happy with that plan | 13:11 |
| openstackgerrit | Gábor Antal proposed openstack/nova: Throw exception if numa_nodes is not set to integer greater than 0 https://review.openstack.org/337380 | 13:11 |
| mriedem | and for newton, ops should just not run placement service in debug mode | 13:11 |
| *** gongysh has joined #openstack-nova | 13:11 | |
| mriedem | if they even run it at all in newton | 13:11 |
| johnthetubaguy | mriedem: I thought they shouldn't run placement at all yet? | 13:11 |
| sdague | mriedem: the should only run in debug if they want to debug things | 13:11 |
| mriedem | johnthetubaguy: they can, but it's optional, so no idea if they will | 13:12 |
| *** ptm__ has joined #openstack-nova | 13:12 | |
| sdague | mriedem: which means they mostly won't | 13:12 |
| mriedem | johnthetubaguy: like, i don't know if i'd suggest to our cloud to run it right now | 13:12 |
| johnthetubaguy | so its just extra CPU usage for zero gain right now, it doesn't seem worth it yet | 13:12 |
| sdague | the whole purpose of having it was so that we would have working newton computes that could talk to it correctly | 13:12 |
| sdague | and report correctly | 13:13 |
| mriedem | well, the gain i see is you get a jump on tracking RPs and usage/inventory | 13:13 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: WIP: Update BuildRequest if instance currently being scheduled https://review.openstack.org/365740 | 13:13 |
| sdague | so that for ocata, when we want to use it, you can with old computes | 13:13 |
| sdague | and we know it works | 13:13 |
| *** mjura has quit IRC | 13:13 | |
| *** PaulMurray has quit IRC | 13:13 | |
| sdague | because we'd have nothing to test the compute clients against otherwise | 13:13 |
| *** edleafe is now known as figleaf | 13:14 | |
| mriedem | alternatively in ocata we remove this from the placement api and move it to the RT | 13:15 |
| *** su_zhang has joined #openstack-nova | 13:16 | |
| *** prateek has quit IRC | 13:16 | |
| *** lpetrut has quit IRC | 13:16 | |
| *** andreas_s has joined #openstack-nova | 13:19 | |
| mriedem | i have to get my daughter on the bus and then head into the office, i'll check back in on this after morning meetings | 13:19 |
| mriedem | danpb: can you check this fix and the one below https://review.openstack.org/#/c/335676/ | 13:20 |
| sdague | mriedem: possibly, though if it's not in the same request as the allocation changes, corolating events id going to be super hard | 13:20 |
| *** mriedem has quit IRC | 13:20 | |
| sdague | which, I also have a paragraph about in the commit message | 13:21 |
| *** mriedem has joined #openstack-nova | 13:22 | |
| alex_xu | cdent: sdague mriedem, one fix for placement https://review.openstack.org/367307, but probably it isn't critical one. | 13:23 |
| *** andreas_s has quit IRC | 13:23 | |
| cdent | thanks alex_xu, will have a look | 13:24 |
| alex_xu | cdent: thanks | 13:24 |
| *** hoangcx has joined #openstack-nova | 13:24 | |
| *** hoangcx2 has quit IRC | 13:25 | |
| *** xyang1 has joined #openstack-nova | 13:25 | |
| *** mriedem has quit IRC | 13:26 | |
| alex_xu | sdague: thanks | 13:27 |
| figleaf | alex_xu: looks good | 13:28 |
| alex_xu | figleaf: thanks :) | 13:28 |
| *** BjoernT has joined #openstack-nova | 13:28 | |
| *** BjoernT is now known as Bjoern_zZzZzZzZ | 13:28 | |
| *** sdake has joined #openstack-nova | 13:30 | |
| *** gszasz has quit IRC | 13:31 | |
| *** nkrinner is now known as nkrinner_afk | 13:31 | |
| *** pwnall1 has joined #openstack-nova | 13:31 | |
| *** pradk has quit IRC | 13:32 | |
| *** pradk has joined #openstack-nova | 13:32 | |
| *** smatzek has quit IRC | 13:34 | |
| *** jichen has quit IRC | 13:34 | |
| *** tonytan4ever has joined #openstack-nova | 13:35 | |
| *** Guest52040 has quit IRC | 13:40 | |
| *** sridhargaddam has quit IRC | 13:41 | |
| *** Bjoern_zZzZzZzZ is now known as BjoernT | 13:41 | |
| *** jschlueter has quit IRC | 13:41 | |
| *** lpetrut has joined #openstack-nova | 13:43 | |
| *** jed56 has joined #openstack-nova | 13:45 | |
| *** mingyu has joined #openstack-nova | 13:47 | |
| *** sdague has quit IRC | 13:50 | |
| *** dansmith is now known as superdan | 13:50 | |
| *** brian_price has quit IRC | 13:51 | |
| *** ociuhandu has joined #openstack-nova | 13:51 | |
| *** markus_z has quit IRC | 13:53 | |
| *** smatzek has joined #openstack-nova | 13:53 | |
| *** alaski is now known as lascii | 13:54 | |
| openstackgerrit | garyk proposed openstack/nova: Compute: ensure that InvalidDiskFormat is handled correctly https://review.openstack.org/367772 | 13:55 |
| *** kaisers_ has joined #openstack-nova | 13:56 | |
| *** AlexeyAbashkin has joined #openstack-nova | 13:57 | |
| *** kaisers_ has quit IRC | 13:57 | |
| *** marst has quit IRC | 13:58 | |
| *** mriedem has joined #openstack-nova | 13:59 | |
| *** vladikr has joined #openstack-nova | 13:59 | |
| *** anush has joined #openstack-nova | 14:00 | |
| *** fragatina has joined #openstack-nova | 14:01 | |
| *** thorst has joined #openstack-nova | 14:02 | |
| *** mfedosin has quit IRC | 14:02 | |
| *** mriedem1 has joined #openstack-nova | 14:03 | |
| *** Dinesh_Bhor has quit IRC | 14:03 | |
| *** mriedem has quit IRC | 14:04 | |
| *** mfedosin has joined #openstack-nova | 14:04 | |
| *** rlrossit has joined #openstack-nova | 14:04 | |
| *** fragatina has quit IRC | 14:06 | |
| mriedem1 | huh http://logs.openstack.org/07/367307/1/check/gate-tempest-dsvm-neutron-full-ubuntu-xenial/cf54759/logs/screen-n-cpu.txt.gz?level=TRACE#_2016-09-08_12_05_26_139 | 14:06 |
| *** mriedem1 is now known as mriedem | 14:06 | |
| mriedem | we shouldn't stacktrace on a port not found when unbinding ports | 14:06 |
| *** thorst has quit IRC | 14:07 | |
| *** marst has joined #openstack-nova | 14:07 | |
| *** TravT has joined #openstack-nova | 14:07 | |
| *** akshai_ has joined #openstack-nova | 14:08 | |
| *** lpetrut has quit IRC | 14:08 | |
| mriedem | https://bugs.launchpad.net/nova/+bug/1621883 | 14:08 |
| openstack | Launchpad bug 1621883 in OpenStack Compute (nova) "PortNotFoundClient stacktrace in n-cpu when unbinding ports" [Undecided,New] | 14:08 |
| *** TravT has quit IRC | 14:08 | |
| *** brian_price has joined #openstack-nova | 14:10 | |
| *** artom has quit IRC | 14:10 | |
| *** TravT has joined #openstack-nova | 14:10 | |
| *** mingyu has quit IRC | 14:11 | |
| *** akshai has quit IRC | 14:11 | |
| mriedem | alex_xu: do you have an example of this warning in a ci run? https://bugs.launchpad.net/nova/+bug/1621437 | 14:11 |
| openstack | Launchpad bug 1621437 in OpenStack Compute (nova) "Add Allocations record failed in the startup of compute node" [Medium,In progress] - Assigned to Alex Xu (xuhj) | 14:11 |
| mriedem | i'm not finding that in logstash | 14:11 |
| *** Jeffrey4l has quit IRC | 14:11 | |
| mriedem | cdent: something for ocata, we need to mark the error messages from the placement-api for translation | 14:12 |
| *** thorst has joined #openstack-nova | 14:12 | |
| mriedem | maybe not even ocata if you can do that today | 14:12 |
| *** Jeffrey4l has joined #openstack-nova | 14:12 | |
| *** hoangcx has quit IRC | 14:13 | |
| mriedem | i'll open a bug | 14:13 |
| *** kaisers_ has joined #openstack-nova | 14:13 | |
| *** lpetrut has joined #openstack-nova | 14:14 | |
| cdent | mriedem: thanks. when you say "error" and "from the placement-api" do you mean: everything LOG.error and/or the http error responses; everything produced both in the api handlers and the object methods. | 14:14 |
| *** lpetrut has quit IRC | 14:14 | |
| mriedem | well anything logged above debug should be marked with the _LI, _LW, _LE markers | 14:15 |
| mriedem | i think we have pep8 for that | 14:15 |
| mriedem | i'm talking about errors in the http responses | 14:15 |
| *** edmondsw has joined #openstack-nova | 14:15 | |
| mriedem | like here https://github.com/openstack/nova/blob/master/nova/api/openstack/placement/handlers/allocation.py#L101 | 14:15 |
| superdan | jaypipes: did you ever write the patch from yesterday morning? | 14:16 |
| superdan | I don't see it anywhere | 14:16 |
| *** zul has joined #openstack-nova | 14:16 | |
| *** Jeffrey4l has quit IRC | 14:17 | |
| mriedem | cdent: https://bugs.launchpad.net/nova/+bug/1621888 | 14:17 |
| openstack | Launchpad bug 1621888 in OpenStack Compute (nova) "placement-api http responses are not marked for translation" [Undecided,New] | 14:17 |
| *** ptm__ is now known as PaulMurray | 14:17 | |
| *** artom has joined #openstack-nova | 14:17 | |
| cdent | thanks mriedem I'll look into that | 14:17 |
| *** avolkov` has joined #openstack-nova | 14:19 | |
| *** Jeffrey4l has joined #openstack-nova | 14:19 | |
| *** lpetrut has joined #openstack-nova | 14:20 | |
| *** adisky has quit IRC | 14:21 | |
| *** zz_dimtruck is now known as dimtruck | 14:23 | |
| *** mingyu has joined #openstack-nova | 14:24 | |
| jaypipes | superdan: https://review.openstack.org/368035 still trying to fix one bug in functional tests. | 14:24 |
| superdan | jaypipes: ah, okay | 14:24 |
| *** lpetrut has quit IRC | 14:25 | |
| *** mingyu has quit IRC | 14:26 | |
| *** mingyu has joined #openstack-nova | 14:26 | |
| *** hoangcx has joined #openstack-nova | 14:28 | |
| *** dikonoor has quit IRC | 14:29 | |
| *** pbandark has quit IRC | 14:30 | |
| *** andreas_s has joined #openstack-nova | 14:30 | |
| *** mlavalle has joined #openstack-nova | 14:30 | |
| *** burgerk has joined #openstack-nova | 14:31 | |
| *** burgerk_ has joined #openstack-nova | 14:33 | |
| burgerk_ | mriedem: Thanks for the help getting the MediumText change merged | 14:36 |
| *** burgerk has quit IRC | 14:37 | |
| *** su_zhang has quit IRC | 14:38 | |
| *** su_zhang has joined #openstack-nova | 14:38 | |
| *** jaypipes is now known as leakypipes | 14:39 | |
| *** jheroux has joined #openstack-nova | 14:39 | |
| *** su_zhang_ has joined #openstack-nova | 14:40 | |
| *** hoangcx has quit IRC | 14:41 | |
| *** ducttape_ has joined #openstack-nova | 14:41 | |
| *** andymaier has joined #openstack-nova | 14:42 | |
| cdent | superdan, jaypipes: I've added jay's patch to the etherpad for safe keeping | 14:42 |
| superdan | cdent: ah thanks | 14:42 |
| *** andymaier has quit IRC | 14:42 | |
| *** su_zhang has quit IRC | 14:42 | |
| *** Oku_OS is now known as Oku_OS-away | 14:44 | |
| *** READ10 has joined #openstack-nova | 14:45 | |
| *** READ10 has quit IRC | 14:45 | |
| *** READ10 has joined #openstack-nova | 14:45 | |
| mriedem | burgerk_: np | 14:48 |
| *** mingyu has quit IRC | 14:49 | |
| *** sudipto has quit IRC | 14:49 | |
| *** sudipto_ has quit IRC | 14:49 | |
| *** kaisers_ has quit IRC | 14:50 | |
| alex_xu | mriedem: I didn't see any example from the CI, I can see that from my local. | 14:50 |
| *** zul has quit IRC | 14:51 | |
| alex_xu | mriedem: it happened everytime in my local after clearup the db. Very strange to why that isn't in the CI. | 14:53 |
| leakypipes | cdent: cheers | 14:53 |
| openstackgerrit | Andrey Volkov proposed openstack/nova: Add schedule_and_build_instances conductor method https://review.openstack.org/367557 | 14:54 |
| openstackgerrit | Andrey Volkov proposed openstack/nova: WIP Move instance creation to conductor https://review.openstack.org/319379 | 14:54 |
| *** tanvirt has joined #openstack-nova | 14:57 | |
| *** burt has joined #openstack-nova | 14:58 | |
| *** annegentle has joined #openstack-nova | 14:58 | |
| *** raj_singh_ has joined #openstack-nova | 14:59 | |
| *** matrohon has quit IRC | 14:59 | |
| *** mdrabe has quit IRC | 15:00 | |
| *** armax has joined #openstack-nova | 15:02 | |
| *** su_zhang_ has quit IRC | 15:04 | |
| *** su_zhang has joined #openstack-nova | 15:04 | |
| *** mdrabe has joined #openstack-nova | 15:05 | |
| *** siva_krishnan has left #openstack-nova | 15:06 | |
| *** tbachman has joined #openstack-nova | 15:06 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Add {get_,set_}aggregates to objects.ResourceProvider https://review.openstack.org/355263 | 15:07 |
| openstackgerrit | Chris Dent proposed openstack/nova: Implement get and set aggregates in the placement API https://review.openstack.org/357993 | 15:07 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] add a placement_aggregates table to api_db https://review.openstack.org/362863 | 15:07 |
| *** jaosorior has quit IRC | 15:07 | |
| *** pcaruana has quit IRC | 15:08 | |
| *** ihrachys has quit IRC | 15:08 | |
| *** su_zhang has quit IRC | 15:09 | |
| alex_xu | mriedem: I see now, in my local env, the compute node already have some instances. When node startup, RT will create Allocation record. In the CI, the node startup without any instance on it. | 15:10 |
| *** wasmum has quit IRC | 15:10 | |
| *** mingyu has joined #openstack-nova | 15:11 | |
| *** rrecio_ has joined #openstack-nova | 15:13 | |
| *** gongysh has quit IRC | 15:13 | |
| mriedem | alex_xu: hmm, yeah...i guess we'd have instances in a grenade job, but we don't run the placement-api in a grenade job yet | 15:13 |
| mriedem | superdan: ^ right? i can't remember if we ever got a grenade job with placement yet | 15:14 |
| mriedem | but i don't think so | 15:14 |
| *** prateek has joined #openstack-nova | 15:14 | |
| *** gouthamr has joined #openstack-nova | 15:15 | |
| *** tjones has joined #openstack-nova | 15:16 | |
| *** gouthamr_ has joined #openstack-nova | 15:16 | |
| mriedem | we should get a grenade job in the experimental queue with the placement api, i can work on that today | 15:17 |
| mriedem | single node | 15:17 |
| mriedem | leakypipes: you're probably better to look at this https://review.openstack.org/#/c/367307/ | 15:19 |
| superdan | mriedem: yeah, I don't think so | 15:19 |
| superdan | mriedem: and yeah, I said last week that we really need to be validating that path at some point | 15:19 |
| *** pwnall2 has joined #openstack-nova | 15:20 | |
| mriedem | i can work on a job for that today | 15:20 |
| *** gouthamr has quit IRC | 15:20 | |
| *** pwnall1 has quit IRC | 15:21 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: neutron: don't trace on port not found when unbinding ports https://review.openstack.org/368079 | 15:22 |
| superdan | mriedem: okay cool | 15:22 |
| bauwser | mriedem: superdan: yeah, we don't have grenade testing against the placement api | 15:24 |
| *** mingyu has quit IRC | 15:25 | |
| bauwser | mriedem: superdan: but what could you guys think we should test ? | 15:25 |
| mriedem | bauwser: just that it doesn't shit it's patns | 15:25 |
| mriedem | *pants | 15:25 |
| *** pwnall2 has quit IRC | 15:25 | |
| superdan | non-pants-shitting would be good | 15:25 |
| bauwser | I appreciate the analogy :) | 15:25 |
| mriedem | experimental job is free | 15:25 |
| bauwser | mriedem: so, like asking grenade to run the placement api ? | 15:25 |
| mriedem | well, | 15:26 |
| *** pwnall1 has joined #openstack-nova | 15:26 | |
| mriedem | we just enable the placement-api service in the job, | 15:26 |
| mriedem | so when it upgrades from mitaka to neutron, devstack sets up the placement api on the neutron side | 15:26 |
| mriedem | and we have some instances in the db | 15:26 |
| bauwser | on the full grenade job or in the aio+compute ? | 15:26 |
| *** jsheeren has quit IRC | 15:26 | |
| mriedem | single node | 15:26 |
| mriedem | full upgrade | 15:26 |
| bauwser | k | 15:26 |
| *** vishwanathj_zzz is now known as vishwanath | 15:27 | |
| *** vishwanath is now known as vishwanathj | 15:27 | |
| *** browne has joined #openstack-nova | 15:27 | |
| leakypipes | mriedem: reviewing now. | 15:27 |
| bauwser | mriedem: and I guess s/neutron/newton/ :p | 15:27 |
| *** mingyu has joined #openstack-nova | 15:27 | |
| mriedem | bauwser: yeah | 15:27 |
| *** amotoki has joined #openstack-nova | 15:27 | |
| mriedem | did i say neutron? | 15:27 |
| mriedem | yeah | 15:27 |
| mriedem | but both | 15:27 |
| mriedem | so i'll work on that and you guys can marvel over my beautiful neutron bug fix above | 15:28 |
| bauwser | mriedem: that's where IRC is unfortunate when you want to lie :p | 15:28 |
| *** yassou has quit IRC | 15:28 | |
| *** su_zhang has joined #openstack-nova | 15:29 | |
| *** gongysh has joined #openstack-nova | 15:30 | |
| *** ZZelle has quit IRC | 15:34 | |
| *** mdrabe has quit IRC | 15:34 | |
| *** mdrabe has joined #openstack-nova | 15:35 | |
| *** jistr is now known as jistr|biab | 15:35 | |
| *** ZZelle has joined #openstack-nova | 15:36 | |
| *** FrankZhang has joined #openstack-nova | 15:37 | |
| *** pwnall1 has quit IRC | 15:40 | |
| *** pwnall1 has joined #openstack-nova | 15:40 | |
| FrankZhang | Hi, everyone. This is Frank from Rackspace. Very glad to be here. Actually I need some help to review one Nova DB patch on the upstream. I'll be really grateful someone can take a look at it: https://review.openstack.org/#/c/329231/ | 15:41 |
| mriedem | bauwser: superdan: https://review.openstack.org/#/c/368087/ | 15:41 |
| bauwser | roger. | 15:41 |
| *** jdurgin has quit IRC | 15:42 | |
| mriedem | heh, looks like rax private cloud is really hurting for that stable/liberty change | 15:42 |
| FrankZhang | Yeah, that's true. I couldn't denny there's too much buggy things existing in Liberty | 15:44 |
| *** Jeffrey4l has quit IRC | 15:47 | |
| *** hshiina has joined #openstack-nova | 15:48 | |
| *** paul-carlton2 has quit IRC | 15:48 | |
| *** openstackgerrit has quit IRC | 15:49 | |
| *** tesseract- has quit IRC | 15:49 | |
| *** openstackgerrit has joined #openstack-nova | 15:49 | |
| *** mingyu has quit IRC | 15:50 | |
| *** andreas_s has quit IRC | 15:51 | |
| *** zul has joined #openstack-nova | 15:52 | |
| -openstackstatus- NOTICE: New setuptools release appears to have a circular import which is breaking many jobs - check for ImportError: cannot import name monkey. | 15:53 | |
| *** ChanServ changes topic to "New setuptools release appears to have a circular import which is breaking many jobs - check for ImportError: cannot import name monkey." | 15:53 | |
| mriedem | FrankZhang: is setting unlimited quota a feature for rpc? | 15:54 |
| *** dimtruck is now known as zz_dimtruck | 15:57 | |
| FrankZhang | mriedem IMO it's more like a guarantees in case user miss-set project quota | 15:59 |
| *** Jeffrey4l has joined #openstack-nova | 15:59 | |
| *** zenoway has quit IRC | 16:01 | |
| *** whenry has joined #openstack-nova | 16:01 | |
| *** amotoki_ has joined #openstack-nova | 16:02 | |
| *** zz_dimtruck is now known as dimtruck | 16:02 | |
| *** jdurgin has joined #openstack-nova | 16:02 | |
| *** gongysh has quit IRC | 16:03 | |
| *** amotoki_ has quit IRC | 16:03 | |
| *** rlrossit has quit IRC | 16:04 | |
| FrankZhang | mriedem really appreciate the help | 16:04 |
| mriedem | i will accept payment in ham sandwiches | 16:05 |
| *** amotoki has quit IRC | 16:05 | |
| *** yamahata has joined #openstack-nova | 16:06 | |
| FrankZhang | You deserve much more than this, haha | 16:06 |
| *** jdurgin has quit IRC | 16:06 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/368122 | 16:09 |
| *** thorst has quit IRC | 16:11 | |
| *** raj_singh_ has quit IRC | 16:11 | |
| *** vladikr has quit IRC | 16:11 | |
| *** akshai_ has quit IRC | 16:12 | |
| *** thorst has joined #openstack-nova | 16:12 | |
| *** pbandark has joined #openstack-nova | 16:13 | |
| *** thorst_ has joined #openstack-nova | 16:15 | |
| *** thorst_ has quit IRC | 16:16 | |
| *** thorst has quit IRC | 16:16 | |
| *** thorst has joined #openstack-nova | 16:19 | |
| *** sudipto has joined #openstack-nova | 16:20 | |
| *** sudipto_ has joined #openstack-nova | 16:20 | |
| *** jdurgin has joined #openstack-nova | 16:21 | |
| *** mkoderer has quit IRC | 16:21 | |
| *** tpatzig has quit IRC | 16:21 | |
| *** pbandark has quit IRC | 16:21 | |
| *** jlanoux has quit IRC | 16:23 | |
| *** thorst_ has joined #openstack-nova | 16:24 | |
| *** thorst has quit IRC | 16:24 | |
| *** akshai has joined #openstack-nova | 16:25 | |
| *** jistr|biab is now known as jistr | 16:25 | |
| *** jdurgin has quit IRC | 16:27 | |
| *** sdake has quit IRC | 16:28 | |
| *** thorst_ has quit IRC | 16:29 | |
| *** vladikr has joined #openstack-nova | 16:29 | |
| *** jdurgin has joined #openstack-nova | 16:29 | |
| *** su_zhang has quit IRC | 16:29 | |
| *** su_zhang has joined #openstack-nova | 16:30 | |
| *** thorst has joined #openstack-nova | 16:31 | |
| *** hshiina has quit IRC | 16:31 | |
| *** fragatina has joined #openstack-nova | 16:32 | |
| *** su_zhang_ has joined #openstack-nova | 16:32 | |
| *** READ10 has quit IRC | 16:32 | |
| *** su_zhang has quit IRC | 16:32 | |
| *** thorst has quit IRC | 16:32 | |
| *** dimtruck is now known as zz_dimtruck | 16:33 | |
| mgagne | melwitt: today I'm getting 140k less log entries after I reduced the number of Greenpool workers for power state sync periodic task. And no failure (BadStatusLine exception) :D | 16:33 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: hardware: Rework docstrings https://review.openstack.org/367468 | 16:34 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: hardware: Standardize '_get_XXX_constraints' functions https://review.openstack.org/367469 | 16:34 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: hardware: Flatten functions https://review.openstack.org/367470 | 16:34 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: hardware: Split out 'numa' and 'compat' modules https://review.openstack.org/367467 | 16:34 |
| melwitt | mgagne: interesting | 16:34 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: hardware: Group functions by broader function https://review.openstack.org/368144 | 16:34 |
| leakypipes | cdent: any chance you can help me diagnose the issue with the one functional test failure from http://logs.openstack.org/35/368035/1/check/gate-nova-tox-db-functional-ubuntu-xenial/98b7e2d/testr_results.html.gz? I cannot understand the output of the console in http://logs.openstack.org/35/368035/1/check/gate-nova-tox-db-functional-ubuntu-xenial/98b7e2d/console.html because something (testr?) is | 16:34 |
| leakypipes | making it all garbled :( | 16:34 |
| *** zz_dimtruck is now known as dimtruck | 16:34 | |
| cdent | leakypipes: yeah, one sec | 16:34 |
| mgagne | melwitt: applied the patch at ~midnight: http://imgur.com/a/VTjSx | 16:35 |
| *** sudipto has quit IRC | 16:35 | |
| *** sudipto_ has quit IRC | 16:35 | |
| mgagne | limiting to 100, instead of 1000 | 16:36 |
| melwitt | wow | 16:36 |
| leakypipes | cdent: nm. reproduced the logs locally. | 16:36 |
| cdent | leakypipes: yeah, was just doing that myself | 16:36 |
| cdent | tox -efunctional placement_api_inventory_post_inventory_bad_resource_class | 16:37 |
| leakypipes | cdent: yeah, found it. | 16:37 |
| *** su_zhang_ has quit IRC | 16:37 | |
| openstackgerrit | Jay Pipes proposed openstack/nova: [placement] make PUT inventory consistent with GET https://review.openstack.org/368035 | 16:37 |
| melwitt | mgagne: and by log entries, do you mean a log statement from inside the power state sync function? | 16:37 |
| mgagne | melwitt: a line in a log file =) | 16:38 |
| *** su_zhang has joined #openstack-nova | 16:38 | |
| mgagne | melwitt: I don't think our system supports multiline well =) | 16:38 |
| cdent | leakypipes: this use to be a little easier to trace because there were less lines of logs. the additional lines are great for some things. | 16:38 |
| mgagne | I can create a graph based on this specific log entry I guess | 16:38 |
| melwitt | mgagne: okay. I was just curious what got logged less as a result | 16:38 |
| mgagne | but the node I'm testing just doesn't have the entry anymore today, checking in file directly | 16:38 |
| mgagne | let me copy paste the error | 16:39 |
| mgagne | http://paste.openstack.org/show/570274/ | 16:39 |
| cdent | leakypipes: if it is giving key error on resource class there doesn't that mean the json schema is not validating? | 16:41 |
| melwitt | mgagne: okay so you mean the reduction is just the lack of a bunch of error logging for the BadStatusLine | 16:42 |
| *** su_zhang has quit IRC | 16:42 | |
| mgagne | melwitt: well, power state sync is now successful and therefore no error generated | 16:42 |
| melwitt | right | 16:42 |
| mgagne | context is nova-compute with Ironic driver power syncing with ironic-api | 16:42 |
| melwitt | yeah | 16:43 |
| cdent | leakypipes: nevermind, I see it is less complex than that | 16:43 |
| *** tpatzig has joined #openstack-nova | 16:44 | |
| *** lucasagomes is now known as lucas-afk | 16:44 | |
| *** mkoderer has joined #openstack-nova | 16:44 | |
| *** efoley has quit IRC | 16:45 | |
| cdent | leakypipes: I have a fix and an improved test if you've been called away | 16:48 |
| *** pilgrimstack has quit IRC | 16:48 | |
| *** READ10 has joined #openstack-nova | 16:49 | |
| *** prateek has quit IRC | 16:50 | |
| *** prateek has joined #openstack-nova | 16:50 | |
| *** markd_ has quit IRC | 16:53 | |
| *** gouthamr_ is now known as gouthamr | 16:53 | |
| *** derekh has quit IRC | 16:54 | |
| cdent | leakypipes: you have squared it away, i see | 16:54 |
| * cdent goes back to sleep | 16:54 | |
| *** Swami has joined #openstack-nova | 16:55 | |
| cfriesen | bauzas: leakypipes: Can you double-check something? I think we've got a race between revert-resize and resource audit on the source that we're reverting back to. In ComputeManger.revert_resize() we setmigration.status = 'reverted' and save the migration. As soon as we do that, MigrationList.get_in_progress_by_host_and_node() will no longer show that migration even though we haven't actually called finish_revert_resize() | 16:55 |
| cfriesen | yet. | 16:55 |
| *** ChanServ changes topic to "This channel is for Nova development. For support of Nova deployments, please use #openstack. Please see: https://wiki.openstack.org/wiki/Nova/Newton_Release_Schedule" | 16:57 | |
| cfriesen | bauwser: ^ | 16:57 |
| -openstackstatus- NOTICE: setuptools 27.1.2 addresses the circular import | 16:57 | |
| melwitt | mgagne: based on a quick look around the code, it seems reasonable the sync power state greenpool size should be configurable too, like the wsgi one is. probably the reason it isn't is because back in the day, there were no rest api calls possible as part of the sync routine | 16:58 |
| mgagne | yea. I'm not sure how vmware and others handle that part | 16:59 |
| melwitt | assuming a rest api call being more heavyweight causes some issue with the pool size | 16:59 |
| mgagne | but I guess they can handle much more load? or in fact, there is no use case where there would be 1000 instances on the same "hypervisor" | 16:59 |
| *** anush has quit IRC | 17:00 | |
| *** sahid has quit IRC | 17:02 | |
| melwitt | yeah, I'm not sure. I thought vcenter was a thing where a cluster would show up as one hypervisor to nova, but I really don't know much about how vmware et al work | 17:02 |
| *** rodrigods has quit IRC | 17:03 | |
| *** rodrigods has joined #openstack-nova | 17:03 | |
| *** tongli has joined #openstack-nova | 17:03 | |
| *** tbachman has quit IRC | 17:03 | |
| leakypipes | cdent: hmm, I *did* rebase on master... | 17:05 |
| cdent | oh dear | 17:05 |
| *** markd_ has joined #openstack-nova | 17:05 | |
| cdent | something is wonky | 17:05 |
| *** yamahata has quit IRC | 17:06 | |
| cdent | leakypipes: I pulled down patchset 2 and it doesn't have those problematic lines | 17:07 |
| *** mnestratov has quit IRC | 17:08 | |
| leakypipes | cdent: ps2 is the latest patchset, though.. and gerrit is showing those lines :( | 17:08 |
| cdent | s/2/1/ | 17:08 |
| *** rlrossit has joined #openstack-nova | 17:08 | |
| leakypipes | cdent: all I did was rebase to master :( | 17:08 |
| cdent | leakypipes: was your master out of date? | 17:08 |
| leakypipes | no | 17:08 |
| cdent | leakypipes: weird. and your master doesn't have those lines? | 17:09 |
| openstackgerrit | Jay Pipes proposed openstack/nova: [placement] make PUT inventory consistent with GET https://review.openstack.org/368035 | 17:10 |
| leakypipes | cdent: lol, the rebase didn't work... I needed to git prune apparently. fixed now. ^ | 17:10 |
| * cdent waves chicken, hopes for best | 17:11 | |
| *** unicell has quit IRC | 17:13 | |
| *** zul has quit IRC | 17:15 | |
| *** zul has joined #openstack-nova | 17:16 | |
| *** tonytan4ever has quit IRC | 17:16 | |
| cdent | leakypipes: I'm seeing no difference between ps2 and ps3 | 17:18 |
| cdent | leakypipes: the offending lines are still there, and my vote is still there because the files haven't changed | 17:18 |
| *** tbachman has joined #openstack-nova | 17:19 | |
| *** su_zhang has joined #openstack-nova | 17:24 | |
| *** rrecio_ has quit IRC | 17:26 | |
| *** bnemec is now known as beekneemech | 17:28 | |
| *** devananda|dinner is now known as devananda | 17:32 | |
| *** thorst_ has joined #openstack-nova | 17:33 | |
| *** ftersin has left #openstack-nova | 17:35 | |
| *** tonytan4ever has joined #openstack-nova | 17:37 | |
| *** thorst_ has quit IRC | 17:38 | |
| *** tonytan_brb has joined #openstack-nova | 17:38 | |
| *** fragatina has quit IRC | 17:39 | |
| *** tjones has left #openstack-nova | 17:39 | |
| *** unicell has joined #openstack-nova | 17:40 | |
| *** tonytan4ever has quit IRC | 17:42 | |
| *** marst has quit IRC | 17:43 | |
| *** marst has joined #openstack-nova | 17:43 | |
| *** fragatina has joined #openstack-nova | 17:43 | |
| *** yamahata has joined #openstack-nova | 17:44 | |
| johnthetubaguy | mriedem: looks like you dropped -2s on some of the config patches, I could go hit those soon if thats what you want there? | 17:44 |
| mriedem | johnthetubaguy: up to you | 17:46 |
| mriedem | i'm not actively reviewing those | 17:47 |
| superdan | config changes need not go in before the release right? | 17:47 |
| mriedem | i'm trying to sort out https://review.openstack.org/#/c/355415 but because i'm not a master of how the image cache and the various image backends work, and there are no details in the commit message, i'm having a hard time with that one | 17:47 |
| mriedem | mikal: you should look at ^ | 17:47 |
| mriedem | superdan: those are just the cleanups | 17:47 |
| mriedem | docs | 17:47 |
| leakypipes | cdent: ok, will look | 17:47 |
| *** fragatina has quit IRC | 17:48 | |
| superdan | mriedem: not moving things you mean? | 17:48 |
| mriedem | but a bunch of ops people have said that https://review.openstack.org/#/c/355415 fixes a bad issue for them so i want to get it in | 17:48 |
| mriedem | i'm not understanding the question | 17:48 |
| mriedem | the centralization part? | 17:48 |
| mriedem | i'd say we shouldn't do those right now | 17:49 |
| mriedem | but the help cleanup stuff is ok i think | 17:49 |
| *** salv-orlando has quit IRC | 17:49 | |
| superdan | right, okay | 17:49 |
| superdan | mriedem: I reviewed that imagecache one a while ago but never voted I guess | 17:49 |
| *** salv-orlando has joined #openstack-nova | 17:49 | |
| mriedem | with all of the image backends and callback methods in that caching code, | 17:50 |
| mriedem | i'm having a real hard time justifying a +1 let alone +2 from me on that fix | 17:51 |
| mriedem | it seems harmless, and people have said it fixes their issue | 17:51 |
| mriedem | but gdi i'd like some details in the commit message about why it's a bug | 17:51 |
| *** mfedosin has quit IRC | 17:51 | |
| mriedem | but when i ask i'm an immediate asshole | 17:51 |
| superdan | mriedem: what do you mean "why it's a bug" ? | 17:51 |
| mriedem | the commit message and code says 'under certain conditions' | 17:52 |
| mriedem | what are those conditions that expose those? | 17:52 |
| mriedem | like, the create_image methods in the image backends all look like they create an image to me, so is it one of those, or is it one of the fetch functions passed into the image backend create_image method? | 17:52 |
| superdan | mriedem: well, you can race with other things that may be culling unused base images right? | 17:53 |
| mriedem | no idea | 17:54 |
| johnthetubaguy | mriedem: superdan: I feel I should spend effort on the bugs instead right now, we can land those as soon as master opens up again, just curious on intention there | 17:54 |
| melwitt | mriedem: I agree it's weird the mysterious "under certain conditions" with no example. I'm reading the lp bug now | 17:55 |
| mriedem | the image cache manager with the partial fetch function callbacks and various backends is just a thing i don't have in my brain | 17:55 |
| mriedem | so someone has to draw me a map to explain the bug | 17:56 |
| superdan | mriedem: so actually I may have found a bug in there | 17:57 |
| superdan | trying to explain it to you | 17:57 |
| superdan | *while trying | 17:59 |
| superdan | mriedem: but anyway, for things like flat, there is an exists check that prevents the actual creation of the image, which I assume we can race with, especially with shared imagecaches, and I assume that is the "some circumstances" | 18:00 |
| *** burgerk_ has quit IRC | 18:01 | |
| *** burgerk has joined #openstack-nova | 18:02 | |
| mriedem | so at t1 exists says it exists, then at t2 the image cache manager thinks it's not used and removes it, so by t3 when we need it, it no longer exists and we fail | 18:02 |
| superdan | yeah | 18:02 |
| mriedem | and create_image() would not create it if it's called between t1 and t2 | 18:03 |
| mriedem | i don't think the bug is for a race, but ok | 18:03 |
| *** shz has joined #openstack-nova | 18:03 | |
| leakypipes | cdent: ok, so fuck it... not sure how those lines snuck back in there. which ones should I remove again? | 18:04 |
| leakypipes | cdent: or maybe I should just start over and rebase on the latest changes you've put up. | 18:04 |
| leakypipes | that merged just recently. | 18:04 |
| leakypipes | god I am so burned out right now. :( | 18:05 |
| cdent | leakypipes: the comment on patchset 1 identifies them I think. basically the stuff about NewCapacityExceended I think | 18:05 |
| cdent | I can just fix it with another patch if you like | 18:05 |
| cdent | or I can push my local change which didn't have it but fixed the other problem | 18:05 |
| cdent | s/other/original | 18:05 |
| cdent | in fact lemme do that | 18:05 |
| *** akshai has quit IRC | 18:05 | |
| leakypipes | cdent: ok | 18:05 |
| leakypipes | cdent: ty | 18:05 |
| *** shihanzhang has quit IRC | 18:06 | |
| mriedem | superdan: looking at mdbooth's test, it says "The Flat backend doesn't write to the image cache when creating a non-image backend." so i'm wondering if this is the issue https://github.com/openstack/nova/blob/master/nova/virt/libvirt/imagebackend.py#L495-L497 | 18:07 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] cleanup some incorrect comments https://review.openstack.org/368171 | 18:09 |
| mriedem | and you only get into ^ if ephemeral or swap | 18:09 |
| *** ssurana has joined #openstack-nova | 18:10 | |
| *** READ10 has quit IRC | 18:10 | |
| * mdbooth is still here (briefly) btw | 18:10 | |
| mdbooth | mriedem: Yes, that's the code which means that flat doesn't write to the image cache when creating ephemeral or swap | 18:11 |
| mdbooth | mriedem: Incidentally I firmly believe that *nothing* should write to the image cache when creating ephemeral or swap | 18:12 |
| mdbooth | That's a job for another day, though | 18:12 |
| *** burgerk has quit IRC | 18:12 | |
| *** akshai has joined #openstack-nova | 18:12 | |
| *** PaulMurray has quit IRC | 18:12 | |
| melwitt | mdbooth: are you looking at the review https://review.openstack.org/#/c/355415/ there's a question from superdan there on whether that call should be self.exists() | 18:13 |
| mriedem | mdbooth: can you read my comment in the commit message (just posted) trying to trace this through and let me know if that's correct? | 18:15 |
| mdbooth | melwitt: Just replying directly. | 18:15 |
| melwitt | cool | 18:15 |
| mdbooth | mriedem: Done | 18:21 |
| superdan | I don't understand how that happens based on mdbooth's explanation above | 18:21 |
| superdan | i.e. how create_image doesn't create a base image for those | 18:21 |
| mdbooth | superdan: So, when you call cache(), it passes a bunch of kwargs down to the specific create_image() of the backend | 18:22 |
| superdan | ah, self.exists in create and os.path.exists in cache | 18:22 |
| superdan | maybe that makes more sense | 18:22 |
| mdbooth | Both Flat and Lvm look in those kwargs for (different) arguments, to decide whether or not they're creating an ephemeral/swap or an image | 18:23 |
| mdbooth | In the latter case, they write directly to target, not to base | 18:23 |
| mdbooth | So they don't populate the image cache | 18:23 |
| superdan | gah | 18:23 |
| mdbooth | It's all fantastically convoluted :) | 18:23 |
| mriedem | hate this code | 18:23 |
| superdan | thinking they were all the same exists, I figured it had to be timing related | 18:24 |
| mdbooth | Seriously, somebody should rewrite that stuff | 18:24 |
| mriedem | heh | 18:24 |
| superdan | which made sense to me, but that's what led me to the os.path.exists thing | 18:24 |
| mriedem | ok, so i'd really appreciate a better explanantion in the commit message or comment in the code, but not sure i want to do it myself (i'd screw it up), or wait until next week for mdbooth to do it, realizing it's friday night for him | 18:24 |
| mriedem | and the ops guys that were on it all hate me by now already | 18:25 |
| mdbooth | mriedem: Ah, this is the one I wrote the test for? | 18:25 |
| mriedem | yes | 18:25 |
| mdbooth | Ok | 18:25 |
| *** pwnall1 has quit IRC | 18:25 | |
| mriedem | which thank you for | 18:25 |
| mdbooth | Yeah, they seemed a bit annoyed | 18:25 |
| mriedem | the annoyance is a 2 way street | 18:26 |
| mdbooth | Indeed | 18:26 |
| mriedem | superdan: are you good with this? | 18:26 |
| superdan | mriedem: sure, but especially since I thought it was a different reason we should prolly fix the commit message | 18:26 |
| *** hemanthm is now known as hemanthm|lunch | 18:26 | |
| superdan | mriedem: you wanna do that since you had the right idea and I can +2? | 18:26 |
| mriedem | superdan: do you want to update the commit message? | 18:26 |
| superdan | I want _you_ to :P | 18:27 |
| mriedem | ha | 18:27 |
| mriedem | mdbooth: you want to take 2 minutes to update the commit message? | 18:27 |
| mdbooth | mriedem: I'm hopefully minutes away from reproducing that live-migration bug... | 18:27 |
| mdbooth | Gimme a few | 18:27 |
| mriedem | o | 18:27 |
| mriedem | ok | 18:27 |
| *** su_zhang has quit IRC | 18:28 | |
| *** jichen has joined #openstack-nova | 18:29 | |
| mdbooth | Reproduced :) | 18:29 |
| mdbooth | It's only taken me about 3 days | 18:29 |
| *** su_zhang has joined #openstack-nova | 18:31 | |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] make PUT inventory consistent with GET https://review.openstack.org/368035 | 18:31 |
| *** rrecio has joined #openstack-nova | 18:32 | |
| *** fragatina has joined #openstack-nova | 18:34 | |
| *** fragatina has quit IRC | 18:34 | |
| *** fragatina has joined #openstack-nova | 18:34 | |
| *** jed56 has quit IRC | 18:35 | |
| *** BjoernT has quit IRC | 18:36 | |
| mriedem | mdbooth: but now you get bragging rights | 18:36 |
| mdbooth | mriedem superdan: Is that commit message ok? | 18:36 |
| cdent | superdan: I think that ^ fixes the issues with that stuff. If you have a mo to gander | 18:36 |
| mdbooth | It's not up my normal essay standard. | 18:36 |
| superdan | mdbooth: I see no change | 18:38 |
| openstackgerrit | Matthew Booth proposed openstack/nova: Fix resizing in imagebackend.cache() https://review.openstack.org/355415 | 18:38 |
| mriedem | +1 | 18:38 |
| mriedem | there we go | 18:38 |
| mriedem | UK internets | 18:38 |
| mdbooth | Now about now? | 18:38 |
| *** jdurgin has quit IRC | 18:38 | |
| superdan | cdent: will try to look when I get back but have to go do a thing in a few minutes | 18:38 |
| mdbooth | Nah, I've never used the gerrit editor before | 18:38 |
| mdbooth | I failed to publish it | 18:38 |
| cdent | superdan: I don't reckon there's a huge rush | 18:39 |
| cdent | superdan: just wanted you to be aware it was in a state of readiness | 18:39 |
| mriedem | +2 | 18:39 |
| mriedem | mdbooth: thanks | 18:39 |
| superdan | cdent: thanks | 18:39 |
| *** pwnall1 has joined #openstack-nova | 18:40 | |
| *** TravT has quit IRC | 18:41 | |
| *** lpetrut has joined #openstack-nova | 18:41 | |
| *** zul has quit IRC | 18:44 | |
| vladikr | anyone has a bandwidth to look at a numa related bug fix - been there for a while: https://review.openstack.org/#/c/342709/7 - there is a relevant functional test on top | 18:45 |
| *** rlrossit has quit IRC | 18:46 | |
| *** pwnall1 has quit IRC | 18:47 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Add {get_,set_}aggregates to objects.ResourceProvider https://review.openstack.org/355263 | 18:47 |
| openstackgerrit | Chris Dent proposed openstack/nova: Implement get and set aggregates in the placement API https://review.openstack.org/357993 | 18:47 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] add a placement_aggregates table to api_db https://review.openstack.org/362863 | 18:47 |
| *** pwnall1 has joined #openstack-nova | 18:47 | |
| *** Sukhdev has joined #openstack-nova | 18:48 | |
| *** cdent has quit IRC | 18:49 | |
| *** leakypipes has quit IRC | 18:52 | |
| marst | Hello! Does anyone have a guide/examples how nova boot --hint flag actually works? Also, there used to be force_host option, allowing to skip schedulers completely, but not anymore?! | 18:55 |
| *** jdurgin has joined #openstack-nova | 18:56 | |
| *** su_zhang has quit IRC | 18:56 | |
| *** sdague has joined #openstack-nova | 18:57 | |
| sdague | superdan: for this, the issue with TRACE is that it's not actually possible to turn on TRACE in newton | 18:58 |
| sdague | https://review.openstack.org/#/c/366186/ | 18:58 |
| sdague | can we land it as exists in newton, then downgrade to TRACE as soon as we get the ability to allow logging that in ocata? | 18:58 |
| *** pwnall1 has quit IRC | 19:00 | |
| *** zul has joined #openstack-nova | 19:00 | |
| *** pwnall1 has joined #openstack-nova | 19:00 | |
| *** unicell has quit IRC | 19:03 | |
| *** vladikr has quit IRC | 19:03 | |
| *** pwnall1 has quit IRC | 19:06 | |
| *** hoangcx has joined #openstack-nova | 19:07 | |
| *** unicell has joined #openstack-nova | 19:07 | |
| *** unicell has quit IRC | 19:08 | |
| *** danpb has quit IRC | 19:08 | |
| *** unicell has joined #openstack-nova | 19:08 | |
| *** hoangcx has quit IRC | 19:14 | |
| openstackgerrit | ianeta hutchinson proposed openstack/nova: Removed extra characters https://review.openstack.org/368185 | 19:16 |
| *** pwnall1 has joined #openstack-nova | 19:16 | |
| *** isq has quit IRC | 19:17 | |
| *** tbachman has quit IRC | 19:21 | |
| *** tbachman has joined #openstack-nova | 19:22 | |
| *** mkoderer has quit IRC | 19:23 | |
| *** hoangcx has joined #openstack-nova | 19:23 | |
| *** tpatzig has quit IRC | 19:23 | |
| openstackgerrit | ianeta hutchinson proposed openstack/nova: Removed extra characters https://review.openstack.org/368185 | 19:25 |
| *** rlrossit has joined #openstack-nova | 19:26 | |
| *** jheroux has quit IRC | 19:26 | |
| *** lamt has joined #openstack-nova | 19:26 | |
| openstackgerrit | Sean Dague proposed openstack/nova: Additional logging for placement API https://review.openstack.org/366186 | 19:30 |
| openstackgerrit | Mathieu Gagné proposed openstack/nova: Add sync_power_state_pool_size option https://review.openstack.org/367746 | 19:33 |
| *** isq has joined #openstack-nova | 19:35 | |
| openstackgerrit | Sean Dague proposed openstack/nova: Additional logging for placement API https://review.openstack.org/366186 | 19:36 |
| *** artom has quit IRC | 19:36 | |
| *** artom has joined #openstack-nova | 19:37 | |
| *** hoangcx has quit IRC | 19:37 | |
| *** artom has quit IRC | 19:37 | |
| *** sdake has joined #openstack-nova | 19:38 | |
| *** artom has joined #openstack-nova | 19:38 | |
| *** dimtruck is now known as zz_dimtruck | 19:39 | |
| *** mnestratov has joined #openstack-nova | 19:39 | |
| *** hemanthm|lunch is now known as hemanthm | 19:41 | |
| *** sdake has quit IRC | 19:42 | |
| *** BjoernT has joined #openstack-nova | 19:43 | |
| *** sdake has joined #openstack-nova | 19:44 | |
| *** lpetrut has quit IRC | 19:46 | |
| *** zz_dimtruck is now known as dimtruck | 19:47 | |
| *** jaugustine_ has joined #openstack-nova | 19:50 | |
| *** jaugustine_ is now known as jaugustine | 19:50 | |
| *** cdelatte has quit IRC | 19:56 | |
| *** nic has joined #openstack-nova | 19:59 | |
| *** marst has quit IRC | 20:00 | |
| superdan | sdague: I really really don't think we should have it in there. it's MHO.. I just commented again and invited other people who don't agree to override me | 20:01 |
| sdague | superdan: ok, I guess I didn't understand your previous concern on DEBUG being used by other things though | 20:02 |
| *** ducttape_ has quit IRC | 20:03 | |
| superdan | sdague: and you still don't or do now? | 20:03 |
| *** tanvirt has quit IRC | 20:03 | |
| *** tongli has quit IRC | 20:03 | |
| superdan | the shared config thing is another issue I hadn't thought of | 20:04 |
| sdague | assuming placement api ends up with a dedicated copy of a nova.conf, given that it's configured radically differently, I'm not sure I get the concern | 20:04 |
| superdan | the people that run with debug=True all the time and use unified controllers will not be able to deploy placement without doing the extra work | 20:04 |
| sdague | the shared file, sure | 20:04 |
| superdan | my point however, was that if you want the actual debug output from placement, you have to turn on the extra queries in order to get that, making it painful to get that level of output from it | 20:05 |
| mriedem | from a recent example here, we just disabled debug on our controllers because of the scheduler debug logging being a terrible impact to select_destinations() | 20:05 |
| mriedem | we could separate the configs on the controller but that was going to be a pain in the ass | 20:05 |
| sdague | mriedem: sure | 20:05 |
| mriedem | we could also run nova-scheduler with --nodebug | 20:05 |
| mriedem | and override the config | 20:05 |
| sdague | mriedem: you could also actually use python logging to tweak that service | 20:06 |
| superdan | yeah, and that's just extra logs being emitted, not N extra DB queries for every allocation | 20:06 |
| sdague | superdan: ok, so TRACE doesn't seem viable for newton, because you can't enable it | 20:06 |
| sdague | an extra config flag? | 20:06 |
| mriedem | to be clear, i'm not sure if you can set nova=TRACE in CONF.default_log_levels and make it work | 20:07 |
| mriedem | i haven't tried it, but assumed that only worked for the logging from the libs that nova is using | 20:07 |
| mriedem | like oslo and glanceclient | 20:07 |
| superdan | yeah, and as I said, if that means we can't have this in newton then I'm fine with that, since I don't think it should be there anyway.. if you want a workaround config to _enable_ it then I guess that's the same compromise | 20:07 |
| superdan | mriedem: based on what Iknow of normal logging it should work, but maybe there is some gotcha there | 20:07 |
| mriedem | i didn't think about the workaround config | 20:07 |
| superdan | I actually thought mriedem's trace=True patch was a nova patch, | 20:08 |
| mriedem | we could do that for newton, release as deprecated, and then drop it in ocata when the CONF.trace thing comes up | 20:08 |
| superdan | so I figured workaround would be the same | 20:08 |
| superdan | but it's not, so.. | 20:08 |
| sdague | superdan: right, I think that's where you disagree, maybe because we were debugging different issues this week | 20:08 |
| mriedem | CONF.workarounds.placement_trace | 20:08 |
| sdague | I'm not actually convinced that the allocations logic is right | 20:08 |
| mriedem | *placement_audit | 20:08 |
| mriedem | something like that | 20:08 |
| superdan | sdague: I know, and I definitely disagree | 20:08 |
| superdan | but trying to be compromisey :) | 20:09 |
| *** jdurgin has quit IRC | 20:09 | |
| sdague | superdan: ok, so feel pretty confident now that allocations are right? | 20:10 |
| *** zul has quit IRC | 20:10 | |
| superdan | sdague: I think this audit was useful for a very narrow issue, and won't be for most other things _i've_ already debugged and fixed | 20:10 |
| *** zul has joined #openstack-nova | 20:10 | |
| superdan | sdague: and I've been beating on this pretty hard for the last three weeks, so yeah, I'm pretty comfortable both with what is there now but also with how we debug things going forward if we have an issue | 20:11 |
| sdague | ok, except this was all wrong for the first 2 weeks you were testing it and not noticed | 20:12 |
| superdan | sdague: another thing I haven't mentioned is that since you're not doing this in a transaction, you could very easily be emitting wrong audit information, which is worse than nothing, IMHO | 20:13 |
| superdan | which is my point about doing this the way it's being done | 20:13 |
| sdague | that's where I get concerned, allocations only were right after https://review.openstack.org/#/c/366245/ merged 48 hours ago | 20:13 |
| superdan | sdague: yeah, good one, I should be ashamed.. anyway, this is MHO and I've invited people to override me, and offered another compromise | 20:14 |
| superdan | not sure what else I can do other than not have an opinion | 20:14 |
| mriedem | i'd be +2 on a workaround option that is deprecated in favor of the trace option that gets into oslo.log, and back this behind that option | 20:16 |
| mriedem | that is the bridge to the trace option for newton | 20:16 |
| mriedem | *ocata | 20:16 |
| mriedem | defaults to False naturally | 20:17 |
| sdague | superdan: ok, the transaction point is a good one | 20:17 |
| superdan | mriedem: I'm not sure why we kept discussing after that | 20:17 |
| *** tonytan_brb has quit IRC | 20:18 | |
| sdague | superdan: I guess I didn't think you said "yes, I'd support that" on the workaround approach | 20:18 |
| *** pwnall1 has quit IRC | 20:18 | |
| superdan | sdague: [13:07:30] <superdan>yeah, and as I said, if that means we can't have this in newton then I'm fine with that, since I don't think it should be there anyway.. if you want a workaround config to _enable_ it then I guess that's the same compromise | 20:18 |
| superdan | sdague: same compromise as trace | 20:18 |
| *** jdurgin has joined #openstack-nova | 20:19 | |
| Matias | hey guys, is a core reviewer available to see if this can be merged? https://review.openstack.org/#/c/358599/ | 20:19 |
| sdague | ok. so if folks are feeling confident on the allocations logic side, I'll drop out the usage bits. I wasn't, as I think it's still really lightly tested | 20:20 |
| *** thorst_ has joined #openstack-nova | 20:20 | |
| sdague | I'll propose back as trace once we have that facility, as it's not like we'll be actively developing the newton code post branch | 20:20 |
| superdan | sdague: you'll still keep the logging of the raw allocation dict through right? | 20:21 |
| sdague | yeh | 20:21 |
| sdague | I'll just delete the audit_usage method | 20:21 |
| sdague | that works for you, right? | 20:21 |
| superdan | yep, wasn't sure if it got lost that I _do_ want the actual allocation info logged | 20:22 |
| mriedem | Matias: i've got that starred, just needed to get back to it | 20:22 |
| sdague | superdan: ok, one last thing | 20:23 |
| sdague | probably worth putting logging on delete as well for the allocations, just to see them in both directions | 20:23 |
| sdague | can I do that after allocations.delete_all() | 20:23 |
| superdan | sdague: sure | 20:23 |
| sdague | or will that have invalidated that | 20:24 |
| sdague | and I need to do it just before | 20:24 |
| Matias | mriedem: thanks | 20:25 |
| superdan | sdague: I'm 1.5h into my "afternoon off" and not doing very well at it, so I'm going to run off now | 20:25 |
| sdague | ok | 20:25 |
| *** smatzek has quit IRC | 20:26 | |
| *** pwnall1 has joined #openstack-nova | 20:26 | |
| sdague | superdan: sorry, didn't realize it was afternoon off | 20:27 |
| superdan | sdague: not your fault | 20:27 |
| *** superdan is now known as dansmith | 20:27 | |
| openstackgerrit | Sean Dague proposed openstack/nova: Additional logging for placement API https://review.openstack.org/366186 | 20:29 |
| *** salv-orl_ has joined #openstack-nova | 20:30 | |
| *** rama_y has joined #openstack-nova | 20:32 | |
| openstackgerrit | Sean Dague proposed openstack/nova: Additional logging for placement API https://review.openstack.org/366186 | 20:32 |
| *** salv-orlando has quit IRC | 20:32 | |
| *** baoli has quit IRC | 20:33 | |
| *** salv-orl_ has quit IRC | 20:33 | |
| *** baoli has joined #openstack-nova | 20:33 | |
| *** salv-orlando has joined #openstack-nova | 20:34 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Additional logging for placement API https://review.openstack.org/366186 | 20:36 |
| *** rama_y has quit IRC | 20:36 | |
| *** mkoderer has joined #openstack-nova | 20:38 | |
| *** tpatzig has joined #openstack-nova | 20:38 | |
| *** rmcall has quit IRC | 20:39 | |
| *** rmcall has joined #openstack-nova | 20:40 | |
| *** su_zhang has joined #openstack-nova | 20:41 | |
| *** rfolco_ has joined #openstack-nova | 20:45 | |
| *** lpetrut has joined #openstack-nova | 20:45 | |
| *** rfolco has quit IRC | 20:46 | |
| *** pwnall2 has joined #openstack-nova | 20:48 | |
| *** tonytan4ever has joined #openstack-nova | 20:50 | |
| *** pwnall1 has quit IRC | 20:52 | |
| mriedem | Matias: reviewed, comments inline | 20:54 |
| Matias | mriedem: thanks | 20:55 |
| *** ssurana has quit IRC | 20:57 | |
| *** openstackstatus has quit IRC | 20:57 | |
| mriedem | mgagne: fyi based on earlier discussions with melwitt https://review.openstack.org/#/c/242006/ | 20:58 |
| *** jaugustine has quit IRC | 20:59 | |
| *** openstackstatus has joined #openstack-nova | 20:59 | |
| *** ChanServ sets mode: +v openstackstatus | 20:59 | |
| mgagne | mriedem: so get_all_power_states would be the call to implement for Ironic? | 20:59 |
| mgagne | mriedem: I'm not sure how Ironic will allow you to fetch nodes filtered by a LONG? list of instance UUIDs | 21:00 |
| mgagne | maybe this is something that would need to be implemented in Ironic API if not already done (I didn't check) | 21:01 |
| mriedem | oh not sure about that, just sounded like a related thing | 21:01 |
| mriedem | there is a lot of overhead with going out to vcenter for thousands of instances one by one is the point | 21:01 |
| *** baoli has quit IRC | 21:02 | |
| mgagne | looks like it's not supported | 21:02 |
| *** baoli has joined #openstack-nova | 21:02 | |
| mgagne | yes, a coworker suggested a similar solution yesterday | 21:02 |
| mgagne | I'm glad that Nova will support the ability to fetch in batch | 21:03 |
| *** rlrossit has quit IRC | 21:03 | |
| *** jichen has quit IRC | 21:04 | |
| *** lascii is now known as alaski | 21:04 | |
| mriedem | not merged yet :) | 21:04 |
| mriedem | needs some love | 21:04 |
| *** lpetrut has quit IRC | 21:07 | |
| *** baoli has quit IRC | 21:07 | |
| mgagne | =) | 21:08 |
| mgagne | so for now, I'm more or less looking for a fix that *could* be easily backported | 21:08 |
| mgagne | and wouldn't depend on a new feature/api in Ironic | 21:08 |
| openstackgerrit | Paulo Matias proposed openstack/nova: Fill destination check data with VNC/SPICE listen addresses https://review.openstack.org/358599 | 21:09 |
| mriedem | ooo a fun NFS + imagecache bug https://bugs.launchpad.net/nova/+bug/1621818 | 21:10 |
| openstack | Launchpad bug 1621818 in OpenStack Compute (nova) "nova-compute unexpected input/output errors on starting instances (NFS + image-cache)" [Undecided,Confirmed] | 21:10 |
| mriedem | mdbooth: your favorite | 21:11 |
| mriedem | browne: need to get the vmware ci running on https://review.openstack.org/#/c/281134/ | 21:12 |
| mriedem | it's been over a day | 21:12 |
| mriedem | well, nearly a day | 21:12 |
| *** haplo37__ has joined #openstack-nova | 21:12 | |
| *** su_zhang has quit IRC | 21:13 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Add sync_power_state_pool_size option https://review.openstack.org/367746 | 21:15 |
| openstackgerrit | Pushkar Umaranikar proposed openstack/nova: Remove internal_id attribute from instances table https://review.openstack.org/360666 | 21:17 |
| mgagne | mriedem: tyvm for the review | 21:18 |
| mriedem | np | 21:18 |
| browne | mriedem: ok will do | 21:20 |
| *** lamt has quit IRC | 21:25 | |
| *** thorst_ has quit IRC | 21:28 | |
| *** thomasem has joined #openstack-nova | 21:29 | |
| *** thomasem has left #openstack-nova | 21:29 | |
| *** prateek has quit IRC | 21:30 | |
| *** haplo37__ has quit IRC | 21:32 | |
| *** akshai has quit IRC | 21:32 | |
| *** annegentle has quit IRC | 21:34 | |
| *** mdrabe has quit IRC | 21:35 | |
| *** mdrabe has joined #openstack-nova | 21:35 | |
| mriedem | do we support attaching 2 ports on different networks with the same mac to the same instance? https://bugs.launchpad.net/nova/+bug/1621076 | 21:35 |
| openstack | Launchpad bug 1621076 in OpenStack Compute (nova) "Can't detach interface from VM (if VM has two interface with same mac addresses)" [Undecided,New] | 21:35 |
| *** pwnall2 has quit IRC | 21:35 | |
| mriedem | armax: ^ ever heard of that? | 21:36 |
| armax | mriedem: that’s not a limitation in neutron per se | 21:37 |
| *** rfolco_ has quit IRC | 21:38 | |
| armax | in that the unique constraint for the mac is applied on a per network basis | 21:38 |
| * armax looks | 21:38 | |
| *** yassou has joined #openstack-nova | 21:38 | |
| armax | looks like the error is coming from libvirt | 21:38 |
| *** wasmum has joined #openstack-nova | 21:39 | |
| armax | it’s impressive this went undetected for so long :) | 21:39 |
| armax | mriedem: what’s weird though is that the interfaces could be attached successfully | 21:40 |
| mriedem | armax: yeah i thought the same | 21:40 |
| mriedem | looks like libvirt complains b/c it can't find which interface to detach | 21:41 |
| mriedem | since it's looking up by mac | 21:41 |
| *** pwnall1 has joined #openstack-nova | 21:41 | |
| armax | mriedem: yeah, makes sense | 21:42 |
| armax | “sense" | 21:42 |
| *** rtheis has quit IRC | 21:43 | |
| mriedem | oh well, maybe danpb can sort it out next week | 21:44 |
| melwitt | I wondered about that bug too. I had thought mac addresses were supposed to be unique | 21:48 |
| clarkb | they are, I think thats reasonable behavior by libvirt | 21:50 |
| mriedem | i assume there is an awesome NFV HA use case lurking here | 21:51 |
| mriedem | carl_baldwin: you might have input on this https://bugs.launchpad.net/nova/+bug/1617409 | 21:53 |
| openstack | Launchpad bug 1617409 in OpenStack Compute (nova) "fixed ip gets assigned when a port is created with no fixed ip" [Undecided,New] - Assigned to Manikantha Srinivas Tadi (manikantha-tadi) | 21:53 |
| mtreinish | melwitt: not always, for example if you use bonding | 21:53 |
| mtreinish | it should be unique on the network though, otherwise you'll have weird issues | 21:53 |
| *** dimtruck is now known as zz_dimtruck | 21:54 | |
| melwitt | I see | 21:54 |
| mtreinish | but it's a weird edge case. I'm sure mriedem is right and it's something nfv why it's coming up | 21:54 |
| clarkb | mtreinish: in that case though you still only have one logical interface andyou can still uniquely identify the physical interfaces iirc | 21:55 |
| *** tblakes has quit IRC | 21:55 | |
| mtreinish | clarkb: right it's one logical interface. But the kernel changes the hwaddr to be the same on both of the physical interfaces | 21:56 |
| mtreinish | clarkb: http://paste.openstack.org/show/570327/ | 21:56 |
| mtreinish | from my server | 21:56 |
| mtreinish | heh, I had to fight this the other day because switching to use systemd to bring it up swapped the macs | 21:57 |
| clarkb | ha | 21:57 |
| clarkb | of course it did | 21:57 |
| mriedem | diana_clarke: fyi https://bugs.launchpad.net/nova/+bug/1613199 | 21:57 |
| openstack | Launchpad bug 1613199 in OpenStack Compute (nova) "nova should accept pre-pended comments in public keys" [Undecided,New] | 21:57 |
| carl_baldwin | mriedem: ack | 22:00 |
| *** hemanthm is now known as hemanthm|away | 22:04 | |
| *** tonytan_brb has joined #openstack-nova | 22:06 | |
| *** artom has quit IRC | 22:06 | |
| *** gjayavelu has joined #openstack-nova | 22:06 | |
| *** tonytan4ever has quit IRC | 22:07 | |
| *** artom has joined #openstack-nova | 22:08 | |
| *** mriedem has quit IRC | 22:11 | |
| *** liangy has quit IRC | 22:18 | |
| *** xinliang has quit IRC | 22:18 | |
| *** pwnall1 has quit IRC | 22:22 | |
| *** yamahata has quit IRC | 22:22 | |
| *** mnestratov has quit IRC | 22:26 | |
| *** mdrabe has quit IRC | 22:27 | |
| *** thorst has joined #openstack-nova | 22:28 | |
| *** xinliang has joined #openstack-nova | 22:31 | |
| *** thorst has quit IRC | 22:33 | |
| *** xyang1 has quit IRC | 22:36 | |
| *** yamahata has joined #openstack-nova | 22:37 | |
| *** sdague has quit IRC | 22:38 | |
| auggy | that bug was kind of weird one | 22:39 |
| auggy | the use case provided in the bug wasn't really valid | 22:40 |
| auggy | but maybe i misunderstood | 22:40 |
| *** sdake has quit IRC | 22:40 | |
| *** raj_singh_ has joined #openstack-nova | 22:43 | |
| *** mingyu has joined #openstack-nova | 22:51 | |
| *** raj_singh_ has quit IRC | 22:54 | |
| *** mingyu has quit IRC | 22:55 | |
| *** ociuhandu has quit IRC | 22:56 | |
| *** figleaf is now known as edleafe | 22:59 | |
| *** rrecio has quit IRC | 23:07 | |
| *** rrecio has joined #openstack-nova | 23:09 | |
| *** rrecio has quit IRC | 23:13 | |
| mgagne | is "XenProject CI" working well? it failed multiple times on what I consider a trivial change. | 23:15 |
| mgagne | in fact, the change is the equivalent of a noop | 23:16 |
| *** armax has quit IRC | 23:20 | |
| *** yassou has quit IRC | 23:20 | |
| *** sdake has joined #openstack-nova | 23:22 | |
| *** pradk has quit IRC | 23:27 | |
| *** BjoernT has quit IRC | 23:30 | |
| *** artom has quit IRC | 23:32 | |
| *** artom has joined #openstack-nova | 23:33 | |
| *** zul has quit IRC | 23:33 | |
| *** Swami has quit IRC | 23:39 | |
| *** armax has joined #openstack-nova | 23:43 | |
| *** nic has quit IRC | 23:44 | |
| *** vishwanathj has quit IRC | 23:52 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!