| opendevreview | Julia Kreger proposed openstack/ironic master: WIP: Set the backend to threading. https://review.opendev.org/c/openstack/ironic/+/953683 | 00:20 |
|---|---|---|
| opendevreview | Julia Kreger proposed openstack/ironic master: WIP: Set the backend to threading. https://review.opendev.org/c/openstack/ironic/+/953683 | 01:30 |
| TheJulia | so, looks like some of the tests which invoke the service are actually failing/going sideways test wise because aspects like startup checks blow up. | 01:31 |
| rpittau | good morning ironic! o/ | 05:41 |
| rpittau | uhmmmm why is https://review.opendev.org/c/openstack/bifrost/+/948245 not merged? and what "ready to submit" means?!? :/ | 07:02 |
| janders | good morning Ironic o/ | 08:13 |
| janders | quick Q for my testing worflow: is bifrost generally reboot-proof? (relevant to some of the testing I am about to attempt) | 08:14 |
| dtantsur | TheJulia: I got the impression that we get a session per thread because oslo.db has some sort of a thread-local cache for session. I might be horribly wrong. | 08:30 |
| dtantsur | Mmm, does anyone know how to increase backlog in irccloud? It's tiny, I miss some responses from the day before apparently | 08:55 |
| opendevreview | Abhishek Bongale proposed openstack/ironic-tempest-plugin master: Add Tempest tests for inspection rules in Ironic https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/951761 | 08:57 |
| frickler | rpittau: it needs a rebase, the gerrit status is somewhere between misleading and wrong. priteau asked about that the other day already. I just guess that in order to submit the rebase you'll also need to add a DCO footer now | 09:13 |
| rpittau | frickler: ok, thanks | 09:18 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: Default ansible to version 10.x https://review.opendev.org/c/openstack/bifrost/+/948245 | 09:19 |
| iurygregory | good morning ironic | 12:22 |
| TheJulia | dtantsur: some of the tests... I was able to identify it is because of self._start_service running and fixing issues due to conductor code firing up | 12:48 |
| dtantsur | TheJulia: are you talking about the database still or about the later issue with tests? | 12:49 |
| TheJulia | still seems to present as db issues, ultimately I suspect its going to require some evaluation of failing tests to determine if the issue is actually that the test has some unexpected behavior occuring which is presenting as database failures | 12:51 |
| * TheJulia proceeds to try and wake up for meeting in a half hour | 12:52 | |
| dtantsur | janders: missed your question: we expect bifrost to be reboot-proof, but its testenv setup is not. | 12:52 |
| dtantsur | specifically, we kill libvirt's dnsmasq during installation, but it will start back on reboot | 12:53 |
| dtantsur | TheJulia: I wonder if we're doing something unsafe in tests themselves (dunno, passing sessions around) | 12:54 |
| TheJulia | dtantsur: that is what I'm thinking, but I want to get a long lived conductor up to see how it behaves | 12:56 |
| TheJulia | hmm, no errors | 12:56 |
| TheJulia | still seems to be sitting there happily power syncing | 12:57 |
| dtantsur | happy power syncing \o/ | 12:58 |
| TheJulia | We've got a weird issue with the "none" transport option, if you could take a look at it, that should get us a little close. I need to take a closer look at the single process stuff today | 12:59 |
| TheJulia | I... at least now that I've kind of fixed some tests which I could get reliably erroring with just a single test run which now seem rock solid, it feels like we're on an improving track | 13:00 |
| opendevreview | Syed Haseeb Ahmed proposed openstack/ironic master: Add interface name to inventory for inspection hooks https://review.opendev.org/c/openstack/ironic/+/953414 | 13:32 |
| opendevreview | Mahnoor Asghar proposed openstack/ironic master: Fix names of steps in comments as per function https://review.opendev.org/c/openstack/ironic/+/953959 | 14:14 |
| opendevreview | Mahnoor Asghar proposed openstack/ironic master: Fix names of steps in comments as per function https://review.opendev.org/c/openstack/ironic/+/953960 | 14:16 |
| TheJulia | masghar: did you mean to do two separate changes? | 14:18 |
| TheJulia | at a glance, it looks like the same change | 14:18 |
| iurygregory | TheJulia, tks for the review yesterday o/ | 14:29 |
| masghar | TheJulia, no, the first one had two change ids so I abandoned it | 14:52 |
| masghar | (Its the same change yes) | 14:53 |
| TheJulia | ack | 14:59 |
| TheJulia | when that happens, generally it is tracking on the last change ID, you yeah, sort of makes sense to just abandon then | 15:00 |
| opendevreview | Julia Kreger proposed openstack/ironic master: WIP: Set the backend to threading. https://review.opendev.org/c/openstack/ironic/+/953683 | 15:50 |
| opendevreview | Iury Gregory Melo Ferreira proposed openstack/sushy master: Don't fail to eject vmedia in case there is no vmedia https://review.opendev.org/c/openstack/sushy/+/952730 | 17:26 |
| iurygregory | dtantsur, should be fixed now ^ | 17:26 |
| opendevreview | Verification of a change to openstack/ironic master failed: Fix Sensor Data values https://review.opendev.org/c/openstack/ironic/+/953395 | 18:12 |
| cardoe | So I think… https://review.opendev.org/c/openstack/ironic/+/952168 is more correct. But still speaking with the neutron folks about the fact that in the L2 case the port should have some kind of identification to the segment it’s bound on. | 18:35 |
| cardoe | I’ll be honest I haven’t run that patch set on actual metal yet. I plan to. | 18:36 |
| TheJulia | I'll try to take a look soon, but might not be until next week | 18:55 |
| TheJulia | I'm very heads down on eventlet stuffs | 18:55 |
| TheJulia | On a plus side, real threads with ironic :) | 18:55 |
| TheJulia | dtantsur: an update, a ton of our issues are largely we're just trying to do threads on some actions because of test style and invoking executors, and just a lot of substrate complexity tests don't always need, peeling some of it off or limiting it to a syncrhonous executor seems to be the way to go. Just working through each item one by one as I find them | 18:57 |
| TheJulia | also, stuff like undo a thing which is supposed to also be undone previously but .... yeah. I think eventlet was hiding a lot of stuff which is now surfacing | 18:58 |
| opendevreview | Julia Kreger proposed openstack/ironic master: WIP: Set the backend to threading. https://review.opendev.org/c/openstack/ironic/+/953683 | 20:24 |
| TheJulia | dtantsur: tomorrow lests chat regarding threads/databases and our tests | 20:41 |
| TheJulia | cid: you'll likely want to join the discussion as well | 20:47 |
| TheJulia | I'll just try and get us on a call tomorrow and we'll go from there | 20:48 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!