| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-02-07 | |||
| 20:43:49 | openstackgerrit | Maciej Szankin proposed openstack/nova master: conf: mark glance.api_servers opt as required https://review.openstack.org/428241 | |
| 21:03:15 | smcginnis | johnthetubaguy: Start here: https://review.openstack.org/#/c/297140/ | |
| 21:03:28 | smcginnis | johnthetubaguy: Then here: https://review.openstack.org/#/c/346438/ | |
| 21:04:28 | smcginnis | johnthetubaguy: Actually, the second one isn't the one I was looking for. This one might be more interesting: https://review.openstack.org/#/c/346470/ | |
| 21:07:27 | openstackgerrit | Diana Clarke proposed openstack/nova master: Reduce the number of COMMIT & SELECT 1 statements https://review.openstack.org/430390 | |
| 21:11:28 | johnthetubaguy | smcginnis: cool, thank you, something for the morning :) | |
| 21:12:40 | smcginnis | ;) | |
| 21:15:34 | cdent | diana_clarke: I want to subscribe to your newsletter | |
| 21:19:14 | diana_clarke | cdent: Nah... it's just a low-hanging fruit, migration 216 only, thing. Nothing much to see there. | |
| 21:20:06 | cdent | diana_clarke I don't know, there's just something fascinating about the collected works of your commits. Not sure what it is. | |
| 21:23:00 | diana_clarke | cdent: perhaps it's because I have no clue what openstack is, so my commits always dance around that, lol. | |
| 21:23:36 | cdent | oh you still worry about that? yeah, apparently you're supposed to stop worrying about that and just accept it | |
| 21:24:28 | artom | diana_clarke, don't tell anyone, openstack is everything to everyone | |
| 21:26:29 | mriedem | diana_clarke: any numbers on how much faster that makes that migration? | |
| 21:26:54 | mriedem | i came across an open nova bug last week from tripleo about how the nova db migrations were timing them out | |
| 21:26:55 | artom | Telcos want to orchestrate some NFV stuff, Rackspace wants a webhosting cloud, you want a job, and I want something I can use to forget the loneliness when alcohol isn't doing it | |
| 21:27:17 | mriedem | artom: there is always opium | |
| 21:27:54 | artom | mriedem, that's next on the list: 1. Alcohol 2. Openstack 3. Opium 4. ??? 5. Profit! | |
| 21:28:05 | cdent | artom, that's the best description of openstack I've ever read | |
| 21:28:08 | artom | I'll proll'y never reach 5 though | |
| 21:28:34 | mriedem | diana_clarke: this one https://bugs.launchpad.net/nova/+bug/1661396 | |
| 21:28:34 | openstack | Launchpad bug 1661396 in OpenStack Compute (nova) "undercloud install fails (nova-db-sync timeout) on VM on an SATA disk hypervisor" [Undecided,Incomplete] | |
| 21:28:45 | artom | cdent, :) | |
| 21:28:55 | dansmith | mriedem: that's ours and it's something with our setup as far as we can tell | |
| 21:29:23 | mriedem | "This seems to be related to the switch to innodb_file_per_table to ON which has doubled the amount of time the db sync takes on this class of hardware." | |
| 21:29:54 | diana_clarke | mriedem: It doesn't make 216 all that much faster, just a few seconds. Here are the nova db migrations, sorted slowest to fastest: http://paste.openstack.org/show/598003/ | |
| 21:29:57 | dansmith | it's way more than double, it's like 100x or something | |
| 21:30:52 | mriedem | heh, 267 is mine | |
| 21:30:56 | mriedem | \o/ | |
| 21:31:13 | mriedem | but that's also because it's scanning all of the tables | |
| 21:31:35 | cdent | https://www.percona.com/blog/2015/02/24/mysqls-innodb_file_per_table-slowing/ | |
| 21:31:46 | cdent | "Many of you in the audience are responsible for running automated tests on your codebase before deploying to production" | |
| 21:31:53 | diana_clarke | mriedem: I started looking at them b/c of that tripleo bug, but honestly I don't think this will help much. Something else is up (like dansmith already said, I see now). | |
| 21:32:07 | dansmith | yep | |
| 21:32:38 | dansmith | but if it makes unit tests faster and doesn't alter the schema at all, then that's cool, although I'm skeptical since we cache the schema the first time we create it in the fixture | |
| 21:33:19 | diana_clarke | dansmith: before I take the WIP off my patch, I'll do a DB dump before and after and diff them. It's already on my TODO list. | |
| 21:33:40 | dansmith | diana_clarke: excellent, thanks | |
| 21:34:09 | diana_clarke | mriedem: And here are the api database migration times, if you're curious. This time sorted by migration number: http://paste.openstack.org/show/598005/ | |
| 21:36:06 | mriedem | diana_clarke: fyi https://github.com/openstack/nova/blob/master/tools/db/schema_diff.py | |
| 21:36:19 | mriedem | dprince wrote that when he was collapsing migrations way back when | |
| 21:36:32 | diana_clarke | mriedem: ah, thanks! | |
| 22:05:31 | pumaranikar | sdague, can you please check https://review.openstack.org/#/c/395194/. Placement json formatter change. | |
| 22:10:44 | openstackgerrit | Diana Clarke proposed openstack/nova master: Reduce the number of COMMIT & SELECT 1 statements https://review.openstack.org/430390 | |
| 22:24:09 | openstackgerrit | Alex Szarka proposed openstack/nova master: Refactor init_instance:resume_guests_state https://review.openstack.org/349582 | |
| 22:25:16 | openstackgerrit | Alex Szarka proposed openstack/nova master: Changed maximum code complexity to 25 https://review.openstack.org/351176 | |
| 22:29:35 | mordred | mriedem: so - I've got this question about boot from volume | |
| 22:30:51 | mordred | mriedem: http://developer.openstack.org/api-ref/compute/?expanded=create-server-detail#create-server says that block_device_mapping_v2.boot_index takes a string and also that None should be used for volumes not used as boot volumes | |
| 22:32:00 | mriedem | was just packing up... | |
| 22:32:20 | mordred | mriedem: but we just got an error "Invalid input for field/attribute boot_index. Value: None. None is not of type 'integer" | |
| 22:32:24 | mriedem | mordred: http://docs.openstack.org/developer/nova/block_device_mapping.html might be helpful | |
| 22:32:33 | mordred | mriedem: yup. it also says None is valid | |
| 22:34:01 | mriedem | i feel like something was just recently changed in the json schema there, is this master? | |
| 22:34:18 | mordred | dhellmann: you were doing master devstack? | |
| 22:34:30 | mordred | mriedem: I think so yeah? | |
| 22:34:43 | dhellmann | mordred : yes, master devstack (from a few hours ago) | |
| 22:34:51 | mriedem | https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/schemas/block_device_mapping.py#L56 | |
| 22:35:28 | mordred | fantastic | |
| 22:36:06 | mriedem | what's the image and dest type? | |
| 22:36:19 | mriedem | source_type and destination_type i mean | |
| 22:37:01 | mriedem | because https://github.com/openstack/nova/blob/master/nova/block_device.py#L198 | |
| 22:37:02 | dhellmann | it's a qcow2 ubuntu image | |
| 22:37:16 | mordred | destination_type would be 'volume' | |
| 22:37:22 | mriedem | and source_type is? | |
| 22:37:23 | mriedem | image? | |
| 22:37:25 | dhellmann | I want to boot it ephemeral and attach a volume | |
| 22:37:47 | mordred | yah- source_type also volume | |
| 22:38:01 | mriedem | have you tried omitting boot_index | |
| 22:38:02 | mriedem | ? | |
| 22:38:05 | mordred | yes. that works | |
| 22:38:12 | mordred | as does providing -1 or '-1' | |
| 22:38:46 | mordred | however, None worked in the past and is still referneced in the docs ... so maybe there's a bug somewhere, I'm just not sure what to file/write up | |
| 22:38:59 | mordred | (we can make shade dtrt by omitting or using -1 | |
| 22:39:14 | mriedem | oomichi: Kevin_Zheng: gmann: alex_xu: do you remember anything recently related to this? ^ | |
| 22:39:40 | mriedem | https://review.openstack.org/#/c/410006/ | |
| 22:40:50 | mriedem | mordred: i wouldn't put a ton of faith in the api-ref docs on this one | |
| 22:40:57 | mordred | mriedem: ok | |
| 22:41:01 | openstackgerrit | Pushkar Umaranikar proposed openstack/nova master: Placement api: set custom json_error_formatter in resource_class https://review.openstack.org/430463 | |
| 22:41:04 | mriedem | i'd be interested in knowing if it's different on like a liberty install | |
| 22:41:29 | mordred | mriedem: None worked against dreamhost - whatever version is running there | |
| 22:41:43 | mriedem | mordred: because my guess is the wording in the api-ref was pulled from http://docs.openstack.org/developer/nova/block_device_mapping.html and that's more about the actual object code in nova rather than the API | |
| 22:42:05 | mordred | ah - k. so it's likely just a docs bug | |
| 22:42:06 | dhellmann | dreamhost is either mitaka or newton | |
| 22:42:15 | mriedem | i'd file a bug either way so we can investigate | |
| 22:42:27 | dhellmann | I can do that | |
| 22:42:37 | mriedem | plus, if the api is ok but the docs are wrong, we can use the bug to fix the docs | |
| 22:42:51 | mordred | cool. | |
| 22:42:52 | mordred | thanks dhellmann ! | |
| 22:43:05 | openstackgerrit | Alex Szarka proposed openstack/nova master: Changed maximum code complexity to 30 https://review.openstack.org/351176 | |
| 22:43:10 | dhellmann | thanks, mordred & mriedem | |
| 22:45:00 | mriedem | also, | |
| 22:45:07 | mriedem | i wonder if you were hitting a nova v2 vs v2.1 endpoint before... | |
| 22:45:11 | mriedem | v2 didn't have schema validation | |
| 22:45:22 | mriedem | mordred: dhellmann: ^ | |
| 22:45:39 | mordred | oh - it's entirely possible | |
| 22:46:06 | mordred | shade is (currently) just piggybacking novaclient which grabs latest microversion aiui | |
| 22:46:12 | mordred | I haven't gotten to doing explicit microversion code myself | |
| 22:46:29 | mordred | (that'll be next month - expect plenty of questions from me) | |
| 22:47:02 | mriedem | there is where the legacy v2 api bdm v2 stuff happened https://github.com/openstack/nova/blob/liberty-eol/nova/api/openstack/compute/legacy_v2/servers.py#L424 | |
| 22:47:54 | mriedem | which only checked boot_index here https://github.com/openstack/nova/blob/liberty-eol/nova/block_device.py#L201 | |
| 22:48:00 | mriedem | if source_type == 'image' and destination_type == 'local': | |