| rpittau | happy Friday ironic! o/ | 06:23 |
|---|---|---|
| mmalchuk | morning Ironic o/ | 06:43 |
| mmalchuk | rpittau CI unhappy in stable brances? needs backport of your patch? | 06:44 |
| rpittau | mmalchuk: we've decided to apply the workaround only on master | 06:45 |
| mmalchuk | asking becouse of this | 06:45 |
| rpittau | I expect the new pkg to be available next week | 06:45 |
| mmalchuk | stay calm and stay for a week))) | 06:46 |
| *** dmellado819 is now known as dmellado81 | 06:55 | |
| dtantsur | yeah. if we backport all patches, we may need to pretty quickly revert them once the package is there. | 07:29 |
| dtantsur | TheJulia, JayF, we still see pathological locking behavior in metal3, now with Retrying None in 18.05774155513265 seconds as it raised OperationalError: (sqlite3.OperationalError) database is locked | 07:53 |
| dtantsur | logs at https://jenkins.nordix.org/job/metal3_bmo_main_integration_test_ubuntu/291/ I'll try to make some sense out of them later | 07:53 |
| dtantsur | the logs just before the 1st (I think) retry: https://paste.opendev.org/show/bD6FBjpx9cz2kL3iOlKS/ | 07:55 |
| adam-metal3 | Hi Ironic | 08:26 |
| adam-metal3 | rpittau, TheJulia, I have created the bug ticket based on yesterday's discussion https://bugs.launchpad.net/ironic-python-agent/+bug/2031092 | 08:27 |
| adam-metal3 | dtantsur, thanks for reporting the locking issue in IRC (I planned to mention it also), do you still see this locking in opendev ci or now it is limited to metal3 ci? | 08:28 |
| rpittau | thanks adam-metal3, I've added a comment there :) | 08:58 |
| iurygregory | good morning ironic | 12:02 |
| rpittau | JayF, TheJulia, did we move also ironic-python-agent-builder to launchpad? cc masghar | 12:10 |
| dtantsur | adam-metal3: I was checking why my BMO patch failed | 12:30 |
| TheJulia | dtantsur: at least it looks like looks like nothing is actually raising an exception, although the length of some of that is worrisome | 13:24 |
| TheJulia | but... wow | 13:29 |
| TheJulia | 992 seconds to complete a node_update | 13:29 |
| TheJulia | looks like the logging needs another option set | 13:31 |
| * TheJulia starts to wonder if we've got a lock hanging in plain sight | 13:43 | |
| dtantsur | quite possibly.. | 13:54 |
| TheJulia | If nothing is exploding once I clean up my "get_serivce_steps" to the agent patch from yesterday's security investigation, I'll try to divert some cycles, around tasks themselves. Things always seem to start right after ipmi commands execute which can also throw things sideways a bit | 14:02 |
| rpittau | bye everyone, have a great weekend, see you next wednesday! o/ | 16:01 |
| TheJulia | o/ | 16:03 |
| mmalchuk | \o | 16:17 |
| JayF | Please let me know if you think there have been contributors in the community with no commits but that should still be considered an active contributor. | 18:05 |
| JayF | As PTL I have until 8/17 to submit them as contributors | 18:05 |
| iurygregory | is this a new thing a PTL should do? O.o | 18:07 |
| iurygregory | asking because no one ever asked me to do such thing during the time I was PTL | 18:08 |
| TheJulia | no, it has been a thing for a long time | 18:09 |
| iurygregory | oh wow | 18:09 |
| TheJulia | the person asking thing I think is the challenge | 18:09 |
| JayF | I have been an extra AC in the past for Ironic | 18:11 |
| JayF | when I was not committing code but still putting aside time to review | 18:11 |
| JayF | it's usually somethign where if someone cares, they'll ask for it | 18:12 |
| JayF | the only "real" benefit is being able to vote for and hold governance offices like ptl/tc | 18:12 |
| JayF | TheJulia: dansmith: one of you able to spare a moment to help w/oslo messaging bits? | 19:22 |
| JayF | I'm getting rapidly convinced it may not be reasonable to fix with an allowlist without adding a dep on oslo.context | 19:22 |
| dansmith | JayF: ask the question before I decide | 19:22 |
| JayF | Right now, oslo.messaging does not dep on oslo.context | 19:23 |
| JayF | to make a proper oslo context object, I would have to dep on it, and create a fresh one | 19:23 |
| JayF | new dep = not backportable AIUI | 19:23 |
| dansmith | there's no reason for a dependency | 19:23 |
| JayF | plus I know oslo<->oslo deps are avoided to prevent cycles | 19:23 |
| dansmith | you just need to construct a fake analog fixture for the tests right? | 19:23 |
| JayF | No | 19:24 |
| JayF | I discovered a real world breakage of turning it into dicts | 19:24 |
| JayF | services can provide their own serializer | 19:24 |
| JayF | neutron uses a serializer that calls ctxt.to_dict() | 19:24 |
| dansmith | right but they provide the context and the serializer yeah? | 19:24 |
| JayF | let me reframe my question | 19:25 |
| JayF | https://review.opendev.org/c/openstack/oslo.messaging/+/891096/8/oslo_messaging/notify/notifier.py#206 | 19:25 |
| JayF | This is obviously wrong. | 19:25 |
| JayF | safe_ctxt must be the same type as ctxt for $reasons discovered above | 19:25 |
| JayF | to make it a requestcontext, the only way I know to get there is to add a dep on oslo.context and instantiate a new RequestContext | 19:25 |
| dansmith | safe_ctxt = ctxt.__class__() ? | 19:26 |
| dansmith | but also, doesn't requestcontext have a clone operation? | 19:26 |
| JayF | I don't want a clone, I want an empty one | 19:26 |
| JayF | > safe_ctxt = ctxt.__class__() | 19:26 |
| JayF | is what I need I think | 19:26 |
| dansmith | a clone is perfectly fine | 19:26 |
| JayF | I knew there had to be a way | 19:26 |
| dansmith | yeah, so I'd do this actually: | 19:27 |
| dansmith | safe_ctxt = ctxt.__class__.from_dict(ctxt.to_dict()) | 19:27 |
| dansmith | that will get you a copy and then you can delattr the things you don't want (or set them to None) | 19:27 |
| JayF | so like, then call safe_ctxt.getmembers() and del any not in the list? | 19:28 |
| dansmith | (i.e. del safe_ctxt.auth_token) | 19:29 |
| * TheJulia reads | 19:30 | |
| mmalchuk | JayF which period count for an active contribution? if I cmommit one change per year am I active?) | 19:31 |
| TheJulia | I'd honestly do the same, or an "eliminate any keys not in my safe key list" | 19:31 |
| JayF | mmalchuk: single commit per release | 19:31 |
| dansmith | yeah, you could nuke them from the to_dict() result before from_dict() | 19:31 |
| JayF | mmalchuk: basically you are automatically a contributor in our automation if you have a commit in anything openstack for that release | 19:31 |
| * TheJulia burries brain back into db/sqlalchemy/api.py | 19:31 | |
| JayF | What about building my dict, *then* calling from_dict() on it | 19:31 |
| JayF | e.g. | 19:31 |
| JayF | for key in safe_keys: | 19:31 |
| JayF | if getattr(ctxt, key, None) is not None: | 19:31 |
| JayF | new_ctxt[key] = getattr(ctxt, key) | 19:32 |
| JayF | safe_ctxt = ctxt.__class__.from_dict(new_ctxt) | 19:32 |
| dansmith | sure | 19:32 |
| mmalchuk | JayF I'm about your question. not about me | 19:32 |
| dansmith | JayF: I was more just showing how to get a copy of the same type of object, but optimize like that for sure | 19:32 |
| mmalchuk | whp is active contributor without commits? | 19:32 |
| mmalchuk | who* | 19:32 |
| JayF | mmalchuk: think about it this way: there are two paths to being considered an active contributor to a (project,release) pair. Commits in the project during the release, or having the PTL or TC recognize you have non-code contributions. | 19:33 |
| JayF | mmalchuk: For instance, in the past, when working at a non-openstack job, I still made a point to review all changes in ironic-python-agent so I was named a contributor and given voting rights anyway. | 19:33 |
| mmalchuk | oh... yep... forgot reviews) | 19:33 |
| mmalchuk | I think there are more people who do only reviews | 19:34 |
| JayF | Please have them reach out to me if they care about getting the ability to vote in PTL or TC elections (or to run for PTL/TC) | 19:35 |
| JayF | I'll note AC has been given in the past before to people who worked on marketing/outreach as well. Basically just it's an effort to ensure that all forms of contribution get to vote on the future of the project; not just those making easy-to-measure contributions. | 19:37 |
| clarkb | historically people translating strings have also been included | 19:43 |
| clarkb | I think a long time ago when docs were less in the projects themselves the docs writers may have been added this way too | 19:43 |
| JayF | I think in practice, especially now, it's a small list. | 19:45 |
| JayF | In philosophy, it's important to do so that folks get recognzied for all kinds of contributions. | 19:46 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Add service steps call to agent logic https://review.opendev.org/c/openstack/ironic/+/891229 | 19:47 |
| TheJulia | JayF: philosophy wise, agree 1000% | 19:48 |
| TheJulia | dtantsur: starting to get an idea, maybe. I think a fresh brain on monday might help | 20:13 |
| JayF | those things are dangerous, be careful ;) | 20:16 |
| TheJulia | lol | 20:18 |
| mmalchuk | JayF is there a place where list of contributors published instead of repo? wiki? | 20:36 |
| JayF | I'm not certain | 20:37 |
| JayF | If I were checking for myself or a collegue, I'd check stackalytics.io and see if it attributed any commits to me in the Bobcat cycle | 20:37 |
| TheJulia | it gets published with each release | 20:58 |
| TheJulia | the acutal "how, dunno" | 20:59 |
| clarkb | TheJulia: the list for all of openstack? its generated via a script fungi runs that queries gerrit | 21:00 |
| clarkb | and I think you are supposed to put people outside of git commits into an extra contributors file eitheri n your repo or the openstack governance repo | 21:00 |
| TheJulia | I'm indeed | 21:00 |
| TheJulia | I, personally, just didn't remember that it was a script somewhere | 21:01 |
| fungi | yes i simply run the script which generates the tc electorate, except i give it a date range from rc1 of the previous cycle to rc1 of the current cycle | 21:01 |
| fungi | so same tools the election officials run in https://opendev.org/openstack/election | 21:02 |
| JayF | https://review.opendev.org/c/openstack/oslo.messaging/+/891096 tested and fixes the bug on Ironic \o/ | 21:23 |
| TheJulia | \o/ | 21:25 |
| JayF | also IDK if anyone is crazy enough to have done this | 21:26 |
| JayF | but that test was on my json-rpc ironic | 21:26 |
| JayF | with notifications pointed at rmq | 21:26 |
| JayF | and everything worked as it should | 21:26 |
| JayF | I shouldn't be surprised but yet somehow I am :D | 21:27 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!