| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-01-31 | |||
| 19:21:08 | dansmith | mriedem: this isn't latent | |
| 19:21:10 | melwitt | mriedem: it's not a latent bug in cells v1. it should have been working fine until we added cells v2 instance listing from all cells | |
| 19:22:59 | openstackgerrit | Matt Riedemann proposed openstack/nova: doc: add upgrade notes to the placement devref https://review.openstack.org/427367 | |
| 19:23:01 | mriedem | cdent: bauzas: ^ | |
| 19:23:11 | mriedem | dansmith: melwitt: ok | |
| 19:24:02 | browne | mriedem: do the RPC versions need to be updated for Ocata? seemed so, so i put up this patch: https://review.openstack.org/#/c/425958/ | |
| 19:24:36 | lyarwood | dansmith: https://review.openstack.org/#/c/427267 should be ready for a review if you have time later today | |
| 19:25:06 | melwitt | artom: I think we've left the test alone since it helps catch our bugs, but technically you're right, a 202 means wait | |
| 19:26:03 | artom | melwitt, well, is the problem that the server state update takes too long, or doesn't happen at all? Because the latter isn't affected by waiting | |
| 19:26:22 | mriedem | it would have to be a race, | |
| 19:26:26 | mriedem | else this would be 100% failure | |
| 19:26:35 | melwitt | artom: the problem is our code is reading instances from the child cells instead of the parent cell, for cells v1 | |
| 19:26:40 | artom | It'll still catch the bugs, except it'll do it by timing out on the wait_for_server_status call | |
| 19:27:10 | melwitt | most of the time, the sync down gets to the child before our read does, and in that case, it will pass the test | |
| 19:27:23 | melwitt | no, it wouldn't catch this bug I don't think | |
| 19:27:45 | melwitt | this is like a split second difference | |
| 19:28:29 | artom | Ah, so the status ends up being synced in all cases | |
| 19:28:36 | melwitt | yes | |
| 19:28:39 | mriedem | lyarwood: i threw the rc-potential tag on that one since it's a regression | |
| 19:29:03 | lyarwood | mriedem: ack thanks | |
| 19:29:23 | artom | melwitt, wait, just to make this clear in my head: the write happens on the parent cell, the read from the child cells, so if you time your read really quickly after the write, you won't get the new state? | |
| 19:29:42 | artom | And the fix you have in mind is to read from the parent cell? | |
| 19:30:05 | melwitt | yes. cells v1 normally reads from the parent cell. it has instances table in the parent and the childs and it syncs them | |
| 19:31:32 | artom | Hrmpf, this is indeed a case of a test being technically wrong but actually useful | |
| 19:31:47 | dansmith | artom: no | |
| 19:31:54 | dansmith | artom: the test case can't tell where it's being read from | |
| 19:32:23 | melwitt | he means the 202 return code implies async so it shouldn't just assume the state change is available after the call returns | |
| 19:32:35 | artom | melwitt, words out of my mouth :) | |
| 19:32:44 | melwitt | we know that it should be, from the code. but the API doesn't tell that | |
| 19:32:51 | dansmith | is it not polling? | |
| 19:33:00 | artom | dansmith, nope | |
| 19:33:00 | melwitt | no, it's just immediately reading after | |
| 19:33:03 | dansmith | ah, okay | |
| 19:33:07 | dansmith | then yeah | |
| 19:33:30 | jaypipes | mriedem, cdent: yes, I'm still working on functional tests for the Ironic patch. Didn't want you to think I forgot or am slacking. | |
| 19:34:08 | mriedem | i'm pretty far removed from that patch by now, | |
| 19:34:12 | mriedem | once john manhandled it | |
| 19:34:44 | mriedem | like, before i start reviewing it again, someone is probably going to have to give me a recap of what happened | |
| 19:35:57 | mriedem | browne: probably... | |
| 19:36:26 | mriedem | i feel like russellb used to have a cheat sheet for this | |
| 19:42:08 | openstackgerrit | Matt Riedemann proposed openstack/nova: doc: add upgrade notes to the placement devref https://review.openstack.org/427367 | |
| 19:45:37 | mordred | mriedem: I've got this error coming up on nodepool patches: http://logs.openstack.org/06/425706/7/check/gate-dsvm-nodepool-src/7004c5b/logs/screen-nodepool.txt.gz#_2017-01-31_19_02_00_177 | |
| 19:45:57 | mordred | mriedem: which is reporting a 500 error from nova and <class 'sqlalchemy.exc.ProgrammingError'> as the error | |
| 19:46:33 | mriedem | nice http://logs.openstack.org/06/425706/7/check/gate-dsvm-nodepool-src/7004c5b/logs/screen-n-api.txt.gz?level=TRACE#_2017-01-31_19_02_00_169 | |
| 19:46:41 | mordred | yah | |
| 19:46:51 | mordred | isn't that pretty? (I was just about to paste that link next :) ) | |
| 19:46:53 | mriedem | Table 'nova_cell0.instance_extra' doesn't exist | |
| 19:47:00 | mriedem | mordred: you don't have your cells v2 stuff setup | |
| 19:47:02 | mriedem | in whatever that thing is | |
| 19:47:18 | mordred | well, it's nodepool's functional tests that run on top of a devstack node | |
| 19:47:36 | mordred | does devstack-gate need cellsv2 now but not set it up by default? :( | |
| 19:47:47 | mriedem | devstack sets up cells v2 | |
| 19:48:05 | mriedem | https://github.com/openstack-dev/devstack/blob/master/lib/nova#L684 | |
| 19:48:09 | mordred | neat. well, apparently something I'm doing there is not letting it I guess - Ill look fruther | |
| 19:48:10 | mriedem | ^ creates the cell0 db | |
| 19:48:28 | mordred | wow. so it should just do it without any flags it looks like | |
| 19:48:43 | mriedem | mordred: and this stuff at the end https://github.com/openstack-dev/devstack/blob/master/lib/nova#L947 | |
| 19:48:53 | mriedem | mordred: yes, cells v2 is required now in ocata | |
| 19:49:12 | mriedem | all ocata nova deployments will be at least a cell of one | |
| 19:49:24 | mordred | mriedem: is there a cell service I need to enable? | |
| 19:49:41 | mriedem | no | |
| 19:49:49 | mriedem | that's cells v1 | |
| 19:49:55 | mriedem | cells v2 is baked in | |
| 19:50:05 | mriedem | here is the nova_cell0 db create in your run http://logs.openstack.org/06/425706/7/check/gate-dsvm-nodepool-src/7004c5b/logs/devstacklog.txt.gz#_2017-01-31_18_58_29_045 | |
| 19:50:20 | mriedem | map_cell0 http://logs.openstack.org/06/425706/7/check/gate-dsvm-nodepool-src/7004c5b/logs/devstacklog.txt.gz#_2017-01-31_19_01_45_588 | |
| 19:50:31 | mriedem | and simple_cell_setup http://logs.openstack.org/06/425706/7/check/gate-dsvm-nodepool-src/7004c5b/logs/devstacklog.txt.gz#_2017-01-31_19_01_48_335 | |
| 19:50:48 | mordred | WEIRD | |
| 19:51:03 | mriedem | that's an appropriate reaction | |
| 19:51:25 | mriedem | http://docs.openstack.org/developer/nova/cells.html | |
| 19:52:37 | mriedem | so when your thing blew up in n-api, cell0 would be mapped | |
| 19:52:42 | mriedem | so i'm not sure why it's not finding the table | |
| 19:52:43 | mriedem | *db | |
| 19:52:53 | mriedem | or, maybe it found the db but the schema wasn't migrated? | |
| 19:53:05 | mriedem | simple_cell_setup will migrate the db schema on the cell0 db | |
| 19:53:16 | mordred | does it do the migration asynchronously? | |
| 19:53:18 | mriedem | you can see it doing that here http://logs.openstack.org/06/425706/7/check/gate-dsvm-nodepool-src/7004c5b/logs/devstacklog.txt.gz#_2017-01-31_19_02_00_128 | |
| 19:53:20 | mriedem | no | |
| 19:53:29 | diana_clarke | perhaps this? "WARNING: cell0 mapping not found - not syncing cell0." | |
| 19:53:31 | mriedem | Warning: (1831, u"Duplicate index 'block_device_mapping_instance_uuid_virtual_name_device_name_idx' defined on the table 'nova_cell0.block_device_mapping' | |
| 19:53:34 | diana_clarke | http://logs.openstack.org/06/425706/7/check/gate-dsvm-nodepool-src/7004c5b/logs/devstacklog.txt.gz#_2017-01-31_18_58_31_449 | |
| 19:54:06 | mriedem | diana_clarke: that's expected | |
| 19:54:17 | mriedem | we've created the dbs, but haven't mapped cell0 yet, that comeslater | |
| 19:54:19 | mriedem | *comes later | |
| 19:54:46 | mriedem | https://github.com/openstack/nova/blob/master/nova/cmd/manage.py#L638 | |
| 19:54:52 | mriedem | db sync attempts to fan out by default | |
| 19:55:03 | mriedem | i'm not sure that should be a WARNING?!?!??! | |
| 19:55:25 | mriedem | https://review.openstack.org/#/c/420973/ fixes the fanout thing | |
| 19:56:31 | mriedem | we see this https://github.com/openstack/nova/blob/master/nova/cmd/manage.py#L1100 | |
| 19:56:45 | mriedem | and then it migrates the schema for the cell0 db here https://github.com/openstack/nova/blob/master/nova/cmd/manage.py#L1107 | |
| 19:57:53 | diana_clarke | mriedem: but only in the non-cells v1 case, it looks to me like cells v1 is enabled here. (ignore me though, I'm probably wrong) | |
| 19:58:16 | openstackgerrit | Eric Fried proposed openstack/nova: PowerVM Driver: power_on/off and reboot https://review.openstack.org/427380 | |
| 19:58:51 | diana_clarke | (b/c is_service_enabled n-cell returns 1, which I assume means true) | |
| 19:58:52 | mriedem | diana_clarke: in the cells v1 case, cell0 db schema would be migrated here https://github.com/openstack-dev/devstack/blob/master/lib/nova#L956 | |
| 19:59:18 | mriedem | but no, cells v1 isn't enabled here | |
| 19:59:18 | mordred | mriedem: I just pushed this up: remote: https://review.openstack.org/427381 Update nodepool dsvm test to disable services | |
| 19:59:30 | mordred | mriedem: just in case something is wonk with us whitelisting a set of services | |
| 19:59:42 | mordred | mriedem: I don't see any other real differences between job configs | |
| 20:00:07 | mriedem | mordred: there is no extra service for cells so it wouldn't be a problem - but you probably need ^ for placement-api anyway | |
| 20:00:15 | mriedem | since nova requires placement-api in ocata to start computes | |
| 20:00:30 | mordred | shouldn't that default to on in devstack then? | |