| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2019-05-23 | |||
| 21:41:00 | melwitt | nice detective work :) | |
| 21:42:31 | sean-k-mooney | melwitt: what was the console auth service used for with cells v1 again | |
| 21:43:26 | melwitt | all console token auth. for cells v1 the database backend stuff was _not_ implemented | |
| 21:44:07 | sean-k-mooney | ah right so its jsust issueing the tokens to connect to the websockets? | |
| 21:44:27 | mriedem | osa is done https://github.com/openstack/openstack-ansible-os_nova/commit/1df0bd51ee5031039d3852cefbb63162e29b22a5 | |
| 21:45:20 | melwitt | sean-k-mooney: console auth service? yeah it stores the token auths locally and responds yea/nay when a request to connect to console proxy comes in | |
| 21:46:04 | melwitt | and the change we did for non-cells v1 was store the token auths in the cell databases and have console proxies (websockets) run per cell so they can query the database to validate token auths | |
| 21:46:20 | sean-k-mooney | melwitt ok so looking at the patches to remove it in osa and triplo if we just dont deploy it will it work with nova today? | |
| 21:47:17 | sean-k-mooney | i dont see any config cahnge in those patches so it seam like nova just does the right thing now? | |
| 21:47:22 | melwitt | sean-k-mooney: yeah, the default behavior is to use the database as the console token auth storage. so running it is not necessary. unless you're running cells v1 | |
| 21:47:37 | sean-k-mooney | which kolla never supported | |
| 21:48:18 | sean-k-mooney | so i can write a quick patch to delete the container form kolla and remove it form the kolla-ansibel playbook if that is all thats needed | |
| 21:48:53 | sean-k-mooney | well the upgade playbook would also need to stop any running containers | |
| 21:49:09 | sean-k-mooney | but that seams trivial too | |
| 21:50:26 | melwitt | I don't know if this is relevant at this point but during a live, rolling upgrade from queens to rocky, you would need to keep nova-consoleauth running until everything is rocky | |
| 21:50:49 | melwitt | if you were not OK with disrupting existing console sessions | |
| 21:50:49 | sean-k-mooney | looks like https://bugs.launchpad.net/kolla-ansible/+bug/1804474 is open but i am gessing its not been worked on | |
| 21:50:50 | openstack | Launchpad bug 1804474 in kolla-ansible "nova-consoleauth is deprecated" [Undecided,In progress] - Assigned to caoyuan (cao-yuan) | |
| 21:51:23 | sean-k-mooney | melwitt: well if i just remove it in the train branch that should not be an issue right | |
| 21:51:36 | melwitt | yeah should not be | |
| 21:51:37 | sean-k-mooney | e.g. going form stien to train it wont be used anyway | |
| 21:51:55 | sean-k-mooney | and it woudl be safe to backprot technically to rock->stien too | |
| 21:53:04 | sean-k-mooney | ... that said im not ment to be starting new things until i finish some of my curent in flight tasks. | |
| 21:53:29 | melwitt | hah | |
| 22:30:58 | openstackgerrit | Merged openstack/nova master: Do not start nova-network in the notification func test https://review.opendev.org/660975 | |
| 22:34:41 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Link versioned notification talk into docs https://review.opendev.org/661115 | |
| 22:36:45 | openstackgerrit | Matt Riedemann proposed openstack/nova master: WIP: Enable cross-cell resize in the nova-multi-cell job https://review.opendev.org/656656 | |
| 23:57:55 | openstackgerrit | Merged openstack/python-novaclient stable/stein: Set the lower limit of api_version for volume_type https://review.opendev.org/661024 | |
| #openstack-nova - 2019-05-24 | |||
| 02:03:02 | openstackgerrit | melanie witt proposed openstack/nova master: Use instance mappings to count server group members https://review.opendev.org/638324 | |
| 02:03:02 | openstackgerrit | melanie witt proposed openstack/nova master: Add documentation for counting quota usage from placement https://review.opendev.org/653845 | |
| 02:42:38 | openstackgerrit | Merged openstack/nova stable/stein: Disable limit if affinity(anti)/same(different)host is requested https://review.opendev.org/659239 | |
| 03:23:19 | yaawang | bauzas: Hi, in spec expose-auto-converge-post-copy(https://review.opendev.org/#/c/651681/), what do you mean by "preffering"? | |
| 04:58:25 | openstackgerrit | Sundar Nadathur proposed openstack/nova-specs master: Nova Cyborg interaction specification. https://review.opendev.org/603955 | |
| 05:10:07 | openstackgerrit | Sundar Nadathur proposed openstack/nova master: ksa auth conf and client for cyborg access https://review.opendev.org/631242 | |
| 05:10:07 | openstackgerrit | Sundar Nadathur proposed openstack/nova master: WIP: Add Cyborg device profile groups to request spec. https://review.opendev.org/631243 | |
| 05:10:08 | openstackgerrit | Sundar Nadathur proposed openstack/nova master: WIP: Create and bind Cyborg ARQs. https://review.opendev.org/631244 | |
| 05:10:09 | openstackgerrit | Sundar Nadathur proposed openstack/nova master: WIP: Get resolved Cyborg ARQs and add PCI BDFs to VM's domain XML. https://review.opendev.org/631245 | |
| 05:49:11 | eandersson | What is the general recommended value for heal_instance_info_cache_interval on a large deployment (~1k nodes)? | |
| 05:50:01 | eandersson | With the default we couldn't scale neutron fast enough to actually handle the load. | |
| 06:18:08 | openstackgerrit | Boxiang Zhu proposed openstack/nova master: Add compute_nodes_uuid field to Destination object https://review.opendev.org/661188 | |
| 06:31:39 | mnaser | melwitt, efried: wrt nova-consoleauth, we've killed it in production (and we've killed it in OSA and looks like Matt proposed a change to remove the cleanup code) -- https://review.opendev.org/#/c/661126/ | |
| 06:42:30 | openstackgerrit | Leehom Li proposed openstack/nova master: RT keep report usage in case CPUPinning conflict https://review.opendev.org/661208 | |
| 06:44:40 | openstackgerrit | Leehom Li proposed openstack/nova master: RT keep report usage in case CPUPinning conflict https://review.opendev.org/661208 | |
| 07:24:38 | jangutter | If only this comic came out at the start of the month: https://xkcd.com/2153/ | |
| 07:39:44 | openstackgerrit | Yongli He proposed openstack/nova master: clean up orphan instances https://review.opendev.org/627765 | |
| 08:25:55 | lyarwood | ganso: LGTM but as tonyb said, as this is stable-only we typically get full Nova Core members to also sign off on things | |
| 08:26:55 | lyarwood | dansmith / efried ; https://review.opendev.org/#/c/659338 - as above, as this is stable-only would you mind giving this a once over? | |
| 08:28:16 | lyarwood | mdbooth: https://review.opendev.org/#/c/658903/ - tgif, am I missing the point here? | |
| 08:28:54 | mdbooth | lyarwood: looking | |
| 08:31:33 | lyarwood | ah nvm I think I see it now | |
| 08:31:56 | lyarwood | they want to pull directly from rbd when using a file based backend | |
| 08:34:17 | mdbooth | lyarwood: Right | |
| 08:35:41 | mdbooth | lyarwood: I hesitate to say it, but I think we want an os-glance library for this | |
| 08:36:07 | mdbooth | Otherwise we're embedding details of glance backend drivers in nova | |
| 08:36:20 | mdbooth | I suspect cinder would also use it | |
| 08:36:25 | lyarwood | we already have rbd_utils.py | |
| 08:36:38 | mdbooth | That's generic | |
| 08:36:48 | mdbooth | And glance presumably has other backend drivers | |
| 08:36:51 | lyarwood | as is this stuff tbh | |
| 08:37:00 | mdbooth | which aren't ceph | |
| 08:37:07 | lyarwood | there's nothing glance specific you'd need here | |
| 08:37:13 | lyarwood | you already have the location | |
| 08:37:28 | lyarwood | if anything rbd_utils etc could all live in os-brick | |
| 08:38:50 | mdbooth | lyarwood: See: file, http, rbd, swift, sheepdog, cinder, vmware | |
| 08:42:07 | openstackgerrit | Boxiang Zhu proposed openstack/nova master: Add valid compute nodes with host and/or node in api layer https://review.opendev.org/661237 | |
| 08:53:00 | lyarwood | mdbooth: right, I just don't think we need yet another lib for this | |
| 10:08:34 | openstackgerrit | Silvan Kaiser proposed openstack/nova stable/stein: Added mount fstype based validation of Quobyte mounts https://review.opendev.org/660706 | |
| 10:23:32 | openstackgerrit | Merged openstack/nova stable/stein: Fix target used in nova.policy.check_is_admin https://review.opendev.org/660330 | |
| 10:23:45 | openstackgerrit | Merged openstack/nova stable/queens: Define irrelevant-files for tempest-full-py3 job https://review.opendev.org/650460 | |
| 10:23:52 | openstackgerrit | Merged openstack/nova stable/queens: Error out migration when confirm_resize fails https://review.opendev.org/652150 | |
| 10:32:37 | openstackgerrit | Merged openstack/nova stable/queens: Teardown networking when rolling back live migration even if shared disk https://review.opendev.org/658149 | |
| 10:32:47 | openstackgerrit | Merged openstack/nova master: Link versioned notification talk into docs https://review.opendev.org/661115 | |
| 10:57:51 | openstackgerrit | Adam Spiers proposed openstack/nova master: Move patch_exists() to nova.test.TestCase for reuse https://review.opendev.org/660500 | |
| 11:07:42 | openstackgerrit | Adam Spiers proposed openstack/nova master: Move patch_exists() to nova.test.TestCase for reuse https://review.opendev.org/660500 | |
| 11:39:15 | lyarwood | has anyone ever seen LM try to claim resources on the source at the start of a migration? | |
| 11:39:36 | lyarwood | IHAC suggesting that's happening but that seems totally wrong | |
| 12:03:59 | openstackgerrit | Adam Spiers proposed openstack/nova master: Move patch_exists() to nova.test.TestCase for reuse https://review.opendev.org/660500 | |
| 12:09:02 | efried | stephenfin: in case you didn't see http://eavesdrop.openstack.org/irclogs/%23openstack-nova/%23openstack-nova.2019-05-24.log.html#t2019-05-24T06:31:39 | |
| 12:09:41 | openstackgerrit | Adam Spiers proposed openstack/nova master: Move patch_exists() to nova.test.TestCase for reuse https://review.opendev.org/660500 | |
| 12:15:18 | aspiers | god sometimes Python really pisses me off | |
| 12:16:45 | aspiers | Python 2/3 unicode stuff is a never-ending nightmare | |
| 12:17:05 | aspiers | I'm literally seeing heisenbugs with it | |
| 12:17:18 | aspiers | one minute everything's fine, the next it magically breaks | |
| 12:17:23 | cdent | i've never met a language that does it in a way that makes me happy. python is one of many that make me want to flip tables | |
| 12:17:59 | cdent | aspiers: bugs on your own machine or remote nodes? if the latter there are so many variables involved (system encoding, python build, blah blah) | |
| 12:18:08 | aspiers | local | |
| 12:18:32 | aspiers | I'm on the verge | |
| 12:18:54 | aspiers | I just want to slurp in the contents of __file__ (sounds weird I know, but there is a good reason) and then do a substring match on it | |
| 12:19:28 | aspiers | if I try to decode('utf-8') on each line I get UnicodeDecodeError: 'utf8' codec can't decode byte 0xf3 in position 1: unexpected end of data | |
| 12:19:54 | aspiers | despite it being an ASCII file (nova/nova/tests/unit/test_test.py) | |
| 12:20:34 | aspiers | ascii is no better: UnicodeDecodeError: 'ascii' codec can't decode byte 0xf3 in position 1: ordinal not in range(128) | |
| 12:24:34 | aspiers | WTF!! I literally ran the test twice in a row and it only failed the second time!! | |
| 12:27:35 | aspiers | OK, I've spotted a pattern | |
| 12:27:41 | aspiers | It always passes on py36 | |
| 12:28:02 | aspiers | It *only* passes on py27 if I run it after editing and saving the file | |
| 12:28:11 | aspiers | If I re-run it, subsequent runs fail | |
| 12:28:39 | aspiers | WT actual F | |
| 12:31:31 | cdent | if __file__ always the .py file and never the .pyc file? | |
| 12:31:38 | aspiers | ohhhhhhh | |
| 12:31:39 | cdent | s/if/is/ | |