Earlier  
Posted Nick Remark
#openstack-nova - 2022-10-27
09:46:40 bauzas correct, so a lazy import
09:46:43 sean-k-mooney its the in tree ovs plugin
09:47:01 bauzas hence me wondering how we could generate the config without importing it directly
09:47:02 sean-k-mooney ya but when we generate the config refence we need to list the namespaces to generate
09:47:05 sean-k-mooney and we dont
09:47:06 bauzas without not*
09:48:27 gibi auniyal_: left +2s on the first and left some feedback in https://review.opendev.org/c/openstack/nova/+/862603
09:49:34 sean-k-mooney amorin: anyway back to yoru question its just a case of we never got aroudn to it
09:50:04 sean-k-mooney so patches are welcome i was just loking for where to chagne it but i dont rememebr off the top of my head
09:50:28 sean-k-mooney i think you need to add it here https://github.com/openstack/nova/blob/master/etc/nova/nova-config-generator.conf#L18
09:50:30 bauzas https://github.com/openstack/nova/blob/master/etc/nova/nova-config-generator.conf
09:50:36 sean-k-mooney ya
09:50:37 amorin ack, perfect
09:50:37 bauzas damn, again, beated
09:50:47 amorin I will try to propose a change then
09:51:01 bauzas amorin: test it locally
09:51:11 bauzas amorin: you have a tox target
09:51:13 sean-k-mooney can you add all the in tree plugsin so also linux bridge
09:51:17 bauzas tox -egenconfig
09:51:33 amorin ok
09:51:45 sean-k-mooney the noop pluging does not have config so you can ignore it
09:53:06 amorin so it's os-vif + linux-bridge
09:53:18 amorin is the list somewhere?
09:53:53 sean-k-mooney https://github.com/openstack/os-vif
09:54:04 sean-k-mooney there are 3 plugins the vif_plug diretories
09:54:14 sean-k-mooney and the ovs and linux bridge ones have config options
09:55:51 sean-k-mooney amorin: this is how the cofnig is regestered https://github.com/openstack/os-vif/blob/master/os_vif/plugin.py#L71-L89
09:57:28 sean-k-mooney the linux bridge name is https://github.com/openstack/os-vif/blob/b0298b5350f21f858f2bf72c5d18e3019a696514/vif_plug_linux_bridge/constants.py#L13
09:57:46 sean-k-mooney os os_vif_linux_bridge
09:58:24 sean-k-mooney that i think will be the namespace you need to list but i have not tried that to see if that is correct or if the namespace shoudl just be os_vif or something like that
10:09:27 darkhorse sean-k-mooney: thank you for responding to the issue that I have reported to launchpad. https://bugs.launchpad.net/nova/+bug/1993736
10:10:27 darkhorse I think I did not make clear description there. I have updated the description and put a link to a patchset that I have published.
10:11:03 darkhorse Will you take a look at it when you find a moment?
10:12:39 amorin ack, thanks sean-k-mooney
10:12:44 amorin and bauzas
10:51:10 auniyal_ thanks gibi
11:28:50 opendevreview Sahid Orentino Ferdjaoui proposed openstack/nova-specs master: spec: allowing target state for evacuate https://review.opendev.org/c/openstack/nova-specs/+/857838
11:44:19 opendevreview Sahid Orentino Ferdjaoui proposed openstack/nova-specs master: spec: allowing target state for evacuate https://review.opendev.org/c/openstack/nova-specs/+/857838
11:51:20 songwenping_ sean-k-mooney: hi, in our Rocky version env, placement integrate with nova, placement-api service comes across this error https://bugs.launchpad.net/nova/+bug/1314554/, and return 500 when call its api. do you have any idea?
11:53:12 sean-k-mooney i dont think we support nosetests in rocky by the way but that not really the issue
11:55:39 sean-k-mooney well it might be related
11:56:25 sean-k-mooney songwenping_: do you get the same issue if you use stestr or tox to run the tests
12:02:29 songwenping_ our tests cannot run, we have give up the test in our product env.
12:03:29 sean-k-mooney we will be end of lifing the rocky branch shortly with no other release planned
12:03:57 sean-k-mooney if this is broken its likely not a recent breakage
12:04:24 sean-k-mooney the last commit to stable rocky was the 22nd of april this year
12:04:37 sean-k-mooney so your product tests have been broken since then or longer
12:06:39 songwenping_ yes we have broken for long time
12:06:42 sean-k-mooney im running the test locally now to confirm but this could be due to a downstream change you have in your product
12:14:06 sean-k-mooney songwenping_: so the test pass correctly with tox using stestr
12:14:31 sean-k-mooney nosetest was never supported in the rocky release so this might be just down to using the incorrect test runner
12:14:59 sean-k-mooney so this sound like its either a downstream bug or a bug in how you are runnign the tests
12:21:55 sean-k-mooney sahid: comments left on https://review.opendev.org/c/openstack/nova-specs/+/857838
12:23:24 songwenping_ sean-k-mooney: i'm finding the lastest changid for our code.
12:32:39 darkhorse Hi team Can I get someone's attention on https://bugs.launchpad.net/nova/+bug/1993736?
12:39:10 sean-k-mooney darkhorse: ah you got the test passign
12:39:19 darkhorse yes
12:39:49 darkhorse sean-k-mooney: yes
12:41:15 sean-k-mooney so just lookign at it quickly
12:41:21 sean-k-mooney it woudl be nice to add a functional test
12:41:32 sean-k-mooney and assert teh instance action log now has the deelted action
12:41:46 sean-k-mooney you just fixed the existing test but did not add any to test the change you made
12:42:15 sean-k-mooney as in there is no test to actully ensure the instnace action log now has the deleted actions
12:47:03 darkhorse sean-k-mooney: will you give me some more details/examples? i think there are already functional tests for deleting shelve offloaded instances.
12:48:09 darkhorse excuse me this is my first time committing to upstream
13:06:43 sean-k-mooney darkhorse: yes there shoudl be an existing one that you can just add a rest call to get the list of instance action events and check that the deleted on is in the list
13:06:50 sean-k-mooney ill see if i can fidn one quickly
13:07:40 darkhorse sean-k-mooney: thank you!
13:08:26 sean-k-mooney https://github.com/openstack/nova/blob/master/nova/tests/functional/test_servers.py#L2578-L2598
13:08:40 sean-k-mooney thats a simple shelve unshleve test with delete at teh end
13:09:15 sean-k-mooney so you can copy that skip the unshelve step and then add a call to the event list api
13:09:21 sean-k-mooney and assert delete is there
13:09:57 sean-k-mooney let me see if there is an existing server event list test you could copy form
13:12:12 sean-k-mooney darkhorse: yes ok so there are example here https://github.com/openstack/nova/blob/master/nova/tests/functional/test_instance_actions.py
13:13:05 sean-k-mooney so you coudl add a new test case by modifying https://github.com/openstack/nova/blob/master/nova/tests/functional/test_instance_actions.py#L96
13:14:20 darkhorse sean-k-mooney: thank you! I will take a look and add functional test for shelved instance deletion.
13:14:22 sean-k-mooney so you really just need this https://github.com/openstack/nova/blob/master/nova/tests/functional/test_instance_actions.py#L229-L234
13:14:50 sahid sean-k-mooney: ack thank you
13:15:27 sean-k-mooney darkhorse: i would put this test in the other file test_servers
13:15:48 sean-k-mooney and just copy the shelve_unshelve test as i said and add in this check
13:15:53 sean-k-mooney then it would be good
13:16:46 darkhorse sean-k-mooney: ok thank you
13:21:32 opendevreview Elod Illes proposed openstack/nova stable/wallaby: Retry image download if it's corrupted https://review.opendev.org/c/openstack/nova/+/819180
13:23:12 opendevreview Elod Illes proposed openstack/nova stable/victoria: Retry image download if it's corrupted https://review.opendev.org/c/openstack/nova/+/819181
15:59:12 melwitt sean-k-mooney[m]: this is what I came up with for that unit test loading guestfs issue, also put some repro steps in the referenced lp bug if you want to see https://review.opendev.org/c/openstack/nova/+/862769
16:00:31 sean-k-mooney[m] cool ill take a look in a while
16:03:45 sean-k-mooney[m] ok ya that looks interesting
16:04:16 sean-k-mooney[m] i set rp+1 on it so ill come back to it in a bit. deliveray arriving shortly
16:04:34 sean-k-mooney[m] so need to step away for a few minutes
16:06:52 melwitt k. no rush of course, just wanted to show you it
16:07:43 sean-k-mooney[m] so its a slightly difffernt approch then our other poison fixtures but its also a slight different usecase
16:08:13 melwitt if you have suggestions lmk. that's just the best I could think of at the time :P
16:09:50 sean-k-mooney[m] the only thing i was thinking about really is this only works if we use the oslo import utils funciton
16:09:56 sean-k-mooney[m] and not a import statement
16:10:20 melwitt yes. I tried to make __import__ work but failed. if you know how, lmk
16:11:28 melwitt I didn't find a way to monkey patch it. I also haven't really monkey patched built ins before that I can remember
16:11:49 sean-k-mooney[m] i feel like i have look it up before but i dont rememebr how to do it currelty
16:12:44 sean-k-mooney[m] we can always make it more robost later so we can likely start with this
16:14:18 sean-k-mooney[m] https://stackoverflow.com/a/63320902
16:16:16 sean-k-mooney[m] we could try doing that ^

Earlier   Later