| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2019-08-16 | |||
| 10:27:51 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Move final bridge commands to privsep. https://review.opendev.org/639580 | |
| 10:27:51 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Cleanup the _execute shim in nova/network. https://review.opendev.org/639581 | |
| 10:27:52 | openstackgerrit | Stephen Finucane proposed openstack/nova master: We no longer need rootwrap. https://review.opendev.org/554438 | |
| 10:27:52 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Cleanup no longer required filters and add a release note. https://review.opendev.org/639826 | |
| 11:03:55 | tbarron | when one live-migrates an instance that has a floating IP does one need to do anything to bring over the FIP ? | |
| 11:04:06 | sean-k-mooney | no | |
| 11:04:18 | tbarron | sean-k-mooney: thanks for confirming | |
| 11:04:33 | tbarron | sean-k-mooney: even with DVR, right? | |
| 11:04:55 | tbarron | not reporting a bug, just checking :) | |
| 11:04:59 | sean-k-mooney | the floating ip translation happens on the networking nodes on in the dvr fip namespeace. that will be set up in all routers that service the neutron router | |
| 11:05:41 | sean-k-mooney | tbarron: this is all handeled in neutron by the way not nova. but for dvr it will jsut work too | |
| 11:06:13 | tbarron | sean-k-mooney: yeah i realize i might ought to be asking in neutron but behaviorally it's a nova migration so | |
| 11:06:28 | tbarron | I figured people here would know the answer | |
| 11:06:45 | tbarron | sean-k-mooney: thanks much | |
| 11:07:24 | sean-k-mooney | worst case senarior your upstream router/switch might send the packet to the old host which will do the translation but when the vm replies it will go direct form the new node and it should learn that the mac/fip has moved port on the TOR | |
| 11:08:52 | tbarron | sean-k-mooney: makes sense | |
| 11:09:25 | sean-k-mooney | stephenfin: that privsep seriese still has a lot of conflit. do you think we should just bit the bullet and try and merge it all n one go | |
| 11:09:40 | stephenfin | sean-k-mooney: I imagine we should, yeah | |
| 11:11:04 | stephenfin | well, it's really just the last patch that causes most of them, looking at it now | |
| 11:11:11 | sean-k-mooney | cool it would be nice if we could merge it today/over teh weekend | |
| 11:11:13 | stephenfin | the first has a lot of conflicts but with older patches | |
| 11:11:42 | stephenfin | efried is our man, I suspect | |
| 11:12:05 | sean-k-mooney | or perhas gibi i saw him around earlier | |
| 11:12:37 | sean-k-mooney | it would be nice to finally be rid of rootwrap in train. then we can start using privsep properly | |
| 11:40:49 | gibi | sean-k-mooney, stephenfin: I will try to look at it but no promises for today | |
| 11:42:20 | sean-k-mooney | gibi: its no urgent its just it would be nice to finally be done with that series | |
| 11:42:39 | sean-k-mooney | so no worries if you dont get to it | |
| 11:46:40 | openstackgerrit | sean mooney proposed openstack/nova master: Libvirt: add support for vPMU configuration. https://review.opendev.org/671338 | |
| 11:47:14 | openstackgerrit | sean mooney proposed openstack/nova master: Libvirt: add support for vPMU configuration. https://review.opendev.org/671338 | |
| 12:04:10 | sean-k-mooney | why do we have some unit test that take 160s to run | |
| 12:05:48 | sean-k-mooney | there has to be something wrong it nova.tests.unit.api.openstack.compute.test_services.ServicesTestV21.test_services_disable | |
| 12:06:44 | openstackgerrit | sean mooney proposed openstack/nova master: Libvirt: add support for vPMU configuration. https://review.opendev.org/671338 | |
| 12:13:31 | gibi | sean-k-mooney: that rings a bell to me | |
| 12:14:01 | gibi | sean-k-mooney: https://review.opendev.org/#/c/675935/ | |
| 12:14:06 | sean-k-mooney | there are several that at least locally are several minutes | |
| 12:14:19 | gibi | sean-k-mooney: I fixed a very same issue in the functional test suite | |
| 12:15:38 | sean-k-mooney | im running them on my workstation where i dont have devstack now to see if that makes a difference | |
| 12:15:57 | sean-k-mooney | well actully i did not have devstack on my laptop... | |
| 12:17:22 | gibi | it seems it takes long on my end too | |
| 12:17:32 | sean-k-mooney | but ya it could be a similar missing stub | |
| 12:17:37 | gibi | let me mock the RPC call to see if that helps | |
| 12:19:28 | gibi | sean-k-mooney: yep the same mock helps | |
| 12:19:40 | gibi | pushing a fix... | |
| 12:19:47 | sean-k-mooney | cool | |
| 12:20:12 | sean-k-mooney | so it was just trying to hit rabbit and timeing out or somehting wehn enableding/disabling rabbit | |
| 12:20:49 | gibi | sean-k-mooney: yes. since https://review.opendev.org/#/q/If32bca070185937ef83f689b7163d965a89ec10a the service disable/enable api call call down to the compute so the compute can update traits in placement | |
| 12:21:23 | gibi | in these tests there is no compute so we are waiting for a timeout that is then catched and droped by the caller | |
| 12:21:34 | gibi | s/catched/caught/ | |
| 12:22:15 | sean-k-mooney | im surpised no one completed about long running ci jobs | |
| 12:22:28 | sean-k-mooney | although i guess tempest will always be longer | |
| 12:27:16 | gibi | I saw the slow functional tests locally but somehow I did not found the same for the unit test, I guess functional takes longer so I always waited for that | |
| 12:28:39 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Avoid timeout from service update api unit tests https://review.opendev.org/676923 | |
| 12:29:06 | sean-k-mooney | i just noticed because i went to grab my phone form my bedroom and then got a glass of water and noticed it was still running and it looked like it was hung. | |
| 12:29:14 | sean-k-mooney | then it updated with a long test | |
| 12:30:30 | sean-k-mooney | im seeing some other errors too. i just checkout out the latest master and ran them and im getting "TypeError: missing a required argument: 'project_id'" | |
| 12:31:28 | sean-k-mooney | i wonder if i could have some stale pyc files? | |
| 12:32:24 | efried | stephenfin: Thanks for catching up the privsep series. Hopefully I'll get a chance to +A those suckers today. | |
| 12:36:36 | gibi | sean-k-mooney: the rest of the py27 run was clean for me top of master | |
| 12:37:29 | sean-k-mooney | gibi: im seeing this on py36 http://paste.openstack.org/show/757964/ | |
| 12:37:42 | sean-k-mooney | im cloneing a fresh repo to see if it makes any difference | |
| 12:37:59 | sean-k-mooney | there are 3 other failure all in the sam test class | |
| 12:40:31 | sean-k-mooney | the project id is not set in the context object | |
| 12:40:36 | sean-k-mooney | oh it is | |
| 12:40:47 | sean-k-mooney | its uuids.my_tenant | |
| 12:44:28 | gibi | sean-k-mooney: py36 was clean for me too | |
| 12:45:06 | sean-k-mooney | ill know in a second if the clean repo is clean | |
| 12:45:37 | sean-k-mooney | i delete the pyc files and it still happened so i dont know what it could be in the other one | |
| 12:46:10 | sean-k-mooney | oh ya i have to wait for those tests without the rpc stub... | |
| 12:47:54 | sean-k-mooney | nope same 4 failtures on a clean clone | |
| 12:48:01 | gibi | if it still fails for you then it might be the virtualenv. I did not recreate that | |
| 12:49:15 | sean-k-mooney | http://paste.openstack.org/show/757965/ | |
| 12:50:20 | openstackgerrit | Eric Fried proposed openstack/nova master: DNM: Test sdk Connection(strict_proxies) https://review.opendev.org/676926 | |
| 12:50:28 | efried | mordred: There's the sniffer ^ | |
| 12:50:52 | mordred | efried: fingers crossed | |
| 12:56:54 | sean-k-mooney | gibi it looks like an issue with " mocked_client = mock.create_autospec(client.Client)" | |
| 12:57:23 | sean-k-mooney | i progess passted that but when the client is used we get teh exception | |
| 12:57:24 | efried | what? Autospec causing problems? Never! | |
| 12:57:48 | gibi | sean-k-mooney: I still cannot reproduce it with a recreated virtualenv. Im using Python 3.6.9 | |
| 12:58:22 | sean-k-mooney | Python 3.6.8 | |
| 12:58:30 | sean-k-mooney | that is on rhel8 | |
| 12:58:37 | sean-k-mooney | ill try it on a ubunut system | |
| 13:00:58 | gibi | sean-k-mooney: I'm on 16.04 | |
| 13:01:29 | gibi | sean-k-mooney: with ppa:deadsnakes/ppa as the python source | |
| 13:01:46 | sean-k-mooney | it passed for my on python 3.6.8 on popos which is based ubuntu 18.04 | |
| 13:01:53 | sean-k-mooney | so this might be a rhel8 issue | |
| 13:02:33 | sean-k-mooney | i keep finding rhel8 issues that make me wonder why im trying to using... | |
| 13:02:59 | sean-k-mooney | i am so tempted to just install ubunut 18.04 but not quite yet... | |
| 13:06:05 | sean-k-mooney | gibi: thanks for checking :) | |
| 13:21:57 | KeithMnemonic | Can someone please see if they can WF+1 this change now https://review.opendev.org/#/c/674916/ ? | |
| 13:25:58 | efried | KeithMnemonic: You'll need a stable. mriedem is on vacation. dansmith or melwitt should be able to help once they arrive on the scene. (They're US Pacific, so it might be a bit.) | |
| 13:27:58 | KeithMnemonic | efried: thanks! | |
| 13:33:50 | dansmith | efried: 'scuse me? | |
| 13:34:39 | efried | dansmith: KeithMnemonic is looking for a second stable to push https://review.opendev.org/#/c/674916/ | |
| 13:34:50 | dansmith | efried: I know, | |
| 13:34:52 | efried | and yes, I know you're an early bird | |
| 13:34:58 | dansmith | mkay. | |
| 13:35:02 | efried | I was giving you an excuse to finish your coffee | |
| 13:35:09 | dansmith | hrmf. | |
| 13:35:13 | efried | you're welcome. | |
| 13:43:29 | KeithMnemonic | dansmith: Thanks! | |