| *** Dinesh_Bhor has quit IRC | 02:21 | |
| *** Dinesh_Bhor has joined #openstack-oslo | 06:08 | |
| *** dkehn has quit IRC | 07:03 | |
| *** lpetrut has joined #openstack-oslo | 07:29 | |
| *** lpetrut has quit IRC | 07:30 | |
| *** lpetrut has joined #openstack-oslo | 07:31 | |
| *** rpittau|afk is now known as rpittau | 08:10 | |
| *** tkajinam has quit IRC | 08:12 | |
| *** iurygregory has joined #openstack-oslo | 08:19 | |
| *** tesseract has joined #openstack-oslo | 08:36 | |
| *** ralonsoh has joined #openstack-oslo | 08:41 | |
| *** tosky has joined #openstack-oslo | 08:48 | |
| AJaeger | stephenfin, hberaud, here's a py27 cleanup for openstackdocstheme, could you review this, please? https://review.opendev.org/706082 | 08:55 |
|---|---|---|
| hberaud | AJaeger: o/ ack | 08:55 |
| hberaud | omw | 08:55 |
| AJaeger | thanks! | 09:00 |
| hberaud | AJaeger: done | 09:02 |
| openstackgerrit | Merged openstack/openstackdocstheme master: [ussuri][goal] Drop python 2.7 support and testing https://review.opendev.org/706082 | 09:06 |
| *** obondarev has joined #openstack-oslo | 09:09 | |
| obondarev | harlowja: hi, could you please check https://bugs.launchpad.net/neutron/+bug/1862315/comments/2 ? | 09:10 |
| openstack | Launchpad bug 1862315 in neutron "Sometimes VMs can't get IP when spawned concurrently" [High,New] - Assigned to Oleg Bondarev (obondarev) | 09:10 |
| obondarev | harlowja: it’s a question about read_lock from the fasteners lib | 09:11 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.db master: Ignore releasenote artifacts files. https://review.opendev.org/705486 | 09:12 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.db master: reword releasenote for py27 support dropping https://review.opendev.org/706448 | 09:12 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.db master: reword releasenote for py27 support dropping https://review.opendev.org/706206 | 09:14 |
| *** AJaeger has left #openstack-oslo | 09:16 | |
| hberaud | obondarev: harlowja isn't really active on oslo/openstack nowaday | 09:18 |
| obondarev | hberaud: I see.. thanks. Do you by any chance know how to contact him? | 09:19 |
| hberaud | obondarev: you might try directly on https://github.com/harlowja/fasteners/ ... maybe by chance | 09:20 |
| obondarev | hberaud: got it, thanks! | 09:23 |
| hberaud | obondarev: you're welcome | 09:23 |
| hberaud | obondarev: what's your needs? | 09:26 |
| obondarev | hberaud: I think there might be a bug in the lib, described in https://bugs.launchpad.net/neutron/+bug/1862315/comments/2 | 09:27 |
| openstack | Launchpad bug 1862315 in neutron "Sometimes VMs can't get IP when spawned concurrently" [High,New] - Assigned to Oleg Bondarev (obondarev) | 09:27 |
| hberaud | obondarev: are you using oslo.concurrency? | 09:27 |
| hberaud | obondarev: or directly fasteners? | 09:27 |
| obondarev | hberaud: yes, neutron uses oslo.concurrency | 09:28 |
| hberaud | obondarev: https://github.com/openstack/oslo.concurrency/blob/master/oslo_concurrency/lockutils.py | 09:28 |
| hberaud | obondarev: ok | 09:28 |
| obondarev | hberaud: here https://github.com/openstack/oslo.concurrency/blob/master/oslo_concurrency/lockutils.py#L83 | 09:29 |
| * hberaud look | 09:29 | |
| hberaud | if I understand your case correctly you spawn 2 different process who needs to lock on ressource (the IP)? I'm right? | 09:31 |
| hberaud | obondarev: ^^ | 09:31 |
| obondarev | hberaud: let me try to explain it.. | 09:32 |
| hberaud | obondarev: go wild | 09:32 |
| obondarev | hberaud: neutron dhcp agent has a “resync” thread which takes a write lock from ReaderWriterLock. Also it spawns many threads to process regular events (resource changes) - these threads are taking read locks from the same ReaderWriterLock | 09:33 |
| hberaud | obondarev: ok got it | 09:34 |
| obondarev | hberaud: “resync” thread is supposed to be of highest priority and block all other threads | 09:34 |
| obondarev | but it seems that many “reader” threads can actually block “writer” thread, because not checking for pending writers | 09:36 |
| hberaud | oh ok | 09:36 |
| hberaud | obondarev: then yes it's fasteners's internal mechanisms here as you already pointed out in your LP, I'll try to keep an eye on this topic too and try to do some local tests | 09:41 |
| obondarev | hberaud: great, thank you! | 09:41 |
| hberaud | obondarev: if I get some info I'll forward to you | 09:41 |
| obondarev | hberaud: cool! | 09:42 |
| *** purplerbot has quit IRC | 10:50 | |
| *** purplerbot has joined #openstack-oslo | 10:51 | |
| *** rpittau is now known as rpittau|bbl | 11:18 | |
| openstackgerrit | Stephen Finucane proposed openstack/oslo.versionedobjects master: [ussuri][goal] Drop python 2.7 support and testing https://review.opendev.org/705725 | 11:45 |
| openstackgerrit | Stephen Finucane proposed openstack/oslo.versionedobjects master: Switch from inspect.getargspec to inspect.getfullargspec https://review.opendev.org/700509 | 11:45 |
| openstackgerrit | Stephen Finucane proposed openstack/oslo.versionedobjects master: Drop use of six https://review.opendev.org/706476 | 11:45 |
| openstackgerrit | Stephen Finucane proposed openstack/oslo.versionedobjects master: Bump to hacking 2.x https://review.opendev.org/706477 | 11:45 |
| openstackgerrit | Merged openstack/oslo.rootwrap master: Remove universal wheel configuration https://review.opendev.org/706336 | 11:48 |
| *** raildo has joined #openstack-oslo | 12:02 | |
| openstackgerrit | Merged openstack/oslo.versionedobjects master: gitignore: Ignore reno artefacts https://review.opendev.org/703969 | 13:05 |
| *** rpittau|bbl is now known as rpittau | 13:14 | |
| *** Luzi has joined #openstack-oslo | 13:16 | |
| *** ansmith has quit IRC | 13:32 | |
| *** ansmith has joined #openstack-oslo | 13:32 | |
| *** gmann_ has joined #openstack-oslo | 13:40 | |
| *** gmann_ is now known as gmann | 13:40 | |
| *** gmann has quit IRC | 13:40 | |
| *** gmann has joined #openstack-oslo | 13:41 | |
| *** yan0s has joined #openstack-oslo | 14:00 | |
| *** lbragstad has quit IRC | 14:17 | |
| *** tesseract-RH has joined #openstack-oslo | 14:19 | |
| *** lbragstad has joined #openstack-oslo | 14:19 | |
| *** tesseract has quit IRC | 14:21 | |
| *** tesseract has joined #openstack-oslo | 14:34 | |
| *** tesseract has quit IRC | 14:43 | |
| *** tesseract-RH has quit IRC | 14:43 | |
| *** tesseract has joined #openstack-oslo | 14:44 | |
| *** Luzi has quit IRC | 14:45 | |
| dansmith | stephenfin: so, have you thought about or considered trying to recreate the getargspec output from the getfullargspec so that our hashes won't change? | 14:49 |
| stephenfin | dansmith: I have, but the issue is that getargspec is apparently buggy and doesn't handle "function annotations and keyword-only parameters." | 14:51 |
| stephenfin | From https://docs.python.org/3/library/inspect.html#inspect.getargspec | 14:51 |
| dansmith | I know that it doesn't | 14:51 |
| dansmith | because it existed before those things | 14:51 |
| dansmith | we don't have those things in nova at least, which would mean we should be able to generate the same fingerprint I would think | 14:52 |
| dansmith | and if we can know whether or not those things are used, we could use the old or new method depending on whether or not we have those, because once someone adds a kw-only param to something the fingerprint would change anyway | 14:53 |
| dansmith | I'm not even sure about the rpc impact of using one, so I'd have to think about that anyway | 14:53 |
| dansmith | I guess I will try to put up a parallel patch and see how it goes | 14:57 |
| openstackgerrit | Cyril Roelandt proposed openstack/oslo.log master: Switch to hacking 2.x https://review.opendev.org/706538 | 15:00 |
| *** rpittau is now known as rpittau|brb | 15:04 | |
| *** lpetrut has quit IRC | 15:08 | |
| *** rpittau|brb is now known as rpittau | 15:20 | |
| openstackgerrit | Dan Smith proposed openstack/oslo.versionedobjects master: Provide stable getargspec() behavior for method fingerprinting https://review.opendev.org/706548 | 15:43 |
| dansmith | stephenfin: ^ so that won't pass our own fast8 test because that runs on py2 and I can't define a method in the test to test the newer spec format without failing pep8 | 15:43 |
| dansmith | also that made me think that maybe just "try: getargspec() except ValueError: getfullargspec()" would be plenty of compat without my namedtuple that I add there | 15:44 |
| dansmith | it would generate the deprecation warning, which i assume is like the worst possible thing you can imagine | 15:45 |
| dansmith | but Ieft it in there from my debugging because I can't run pep8 anyway | 15:45 |
| dansmith | but, point is, generating a compatible spec is possible | 15:46 |
| stephenfin | dansmith: I don't care about the DeprecationWarning (except for it blowing up subunit :/), so much as what it indicates | 15:46 |
| stephenfin | Just replied but tl;dr: my reasoning was take the hit now to avoid further pain down the line | 15:47 |
| dansmith | that we're a mature product and don't break users unless there's a major reason to? :) | 15:47 |
| stephenfin | "This feature is being entirely removed from Python" seems pretty major :) | 15:48 |
| stephenfin | but lemme check that solution. If it works, why not | 15:48 |
| dansmith | when that happens, then sure, but your reply just now sounds like actually removing it was not very popular | 15:49 |
| dansmith | take away the first try..except for debugging and it generates the same spec for methods that don't need the full one, even without calling the deprecated method | 15:50 |
| stephenfin | removing getfullargspec (not getargspec) was not, because it provides a mostly compatible bridge for Py2-Py3 compatible code | 15:50 |
| stephenfin | as for the pep8 issues, I suspect you want https://review.opendev.org/#/c/706477/ | 15:50 |
| dansmith | oh I see | 15:50 |
| dansmith | ah yeah | 15:50 |
| dansmith | well, I can reorder that whole series if you want, with the pep8 thing under my fix, cleaned up and ready | 15:51 |
| dansmith | if you're happy with that soluton | 15:51 |
| stephenfin | Yeah, looks pretty good to me. Just so I'm clear, there's no reason to keep the try-except if we don't have to support Python 2, right? | 15:54 |
| stephenfin | i.e. if it's rebased onto the "remove py27" patch | 15:54 |
| dansmith | there's no reason to keep it at all in that code | 15:54 |
| dansmith | it was just for debugging, you see spec is never used | 15:54 |
| stephenfin | Ah, I understand now | 15:54 |
| stephenfin | Sweet. Yeah, wfm | 15:55 |
| dansmith | okay, will reorder that whole stack and clean this up | 15:55 |
| stephenfin | Sweet. Thanks for looking at that | 15:55 |
| *** bnemec has quit IRC | 15:55 | |
| *** bnemec has joined #openstack-oslo | 15:56 | |
| dansmith | stephenfin: er, that hacking patch doesn't seem to make pep8 run on python3 | 15:59 |
| dansmith | if you can point me at a patch to do that for some other project I can add it in here | 16:00 |
| dansmith | er, actually maybe I need to rebuild my pep8 venv | 16:00 |
| smcginnis | Is basepython=python3 actually supposed to be under [testenv] instead of [tox]? | 16:00 |
| smcginnis | Otherwise it looks like it should work. | 16:01 |
| dansmith | yeah, still fails after rebuild | 16:01 |
| stephenfin | dansmith: it should. Wanna push it up anyway and I'll take a look? | 16:01 |
| dansmith | moving to testenv makes it work | 16:02 |
| smcginnis | That hacking patch's pep8 job ran with python3.6. | 16:02 |
| dansmith | smcginnis: when I run it here I get all the py2.7 deprecation messages and failure to install o.vo because of the py3 requirement | 16:03 |
| smcginnis | Huh, odd. | 16:03 |
| dansmith | stephenfin: I'm not sure what's up with that sys import removal either | 16:04 |
| smcginnis | dansmith: pep8 for your patch works for me. | 16:04 |
| smcginnis | ./oslo_versionedobjects/fixture.py:221:9: F841 local variable 'spec' is assigned to but never used | 16:05 |
| dansmith | smcginnis: weird, I don't get it | 16:05 |
| dansmith | if I move it to testenv it works fine, move it back, it runs in py2 | 16:05 |
| openstackgerrit | Dan Smith proposed openstack/oslo.versionedobjects master: [ussuri][goal] Drop python 2.7 support and testing https://review.opendev.org/705725 | 16:06 |
| openstackgerrit | Dan Smith proposed openstack/oslo.versionedobjects master: Drop use of six https://review.opendev.org/706476 | 16:06 |
| openstackgerrit | Dan Smith proposed openstack/oslo.versionedobjects master: Bump to hacking 2.x https://review.opendev.org/706477 | 16:06 |
| smcginnis | Not sure what the difference would be in environment, but I've always set that under testenv, so that may be the "more correct" way to do it. | 16:06 |
| openstackgerrit | Dan Smith proposed openstack/oslo.versionedobjects master: Provide stable getargspec() behavior for method fingerprinting https://review.opendev.org/706548 | 16:06 |
| dansmith | well, here it is ^ | 16:06 |
| dansmith | stephenfin also randomly removed a sys import from exception in his getargspec patch, but I'm not sure why that wasn't failing before, so I figure it probably belongs in some other patch -- maybe the flake8 one | 16:07 |
| *** tesseract has quit IRC | 16:38 | |
| *** yan0s has quit IRC | 16:43 | |
| *** tesseract has joined #openstack-oslo | 16:57 | |
| *** rpittau is now known as rpittau|afk | 17:02 | |
| openstackgerrit | Stephen Finucane proposed openstack/oslo.versionedobjects master: Drop use of six https://review.opendev.org/706476 | 17:07 |
| openstackgerrit | Stephen Finucane proposed openstack/oslo.versionedobjects master: Bump to hacking 2.x https://review.opendev.org/706477 | 17:07 |
| openstackgerrit | Stephen Finucane proposed openstack/oslo.versionedobjects master: Provide stable getargspec() behavior for method fingerprinting https://review.opendev.org/706548 | 17:07 |
| *** tesseract has quit IRC | 17:36 | |
| *** tosky has quit IRC | 17:40 | |
| *** obondarev has quit IRC | 17:40 | |
| openstackgerrit | Dan Smith proposed openstack/oslo.versionedobjects master: Drop use of six https://review.opendev.org/706476 | 18:37 |
| openstackgerrit | Dan Smith proposed openstack/oslo.versionedobjects master: Bump to hacking 2.x https://review.opendev.org/706477 | 18:37 |
| openstackgerrit | Dan Smith proposed openstack/oslo.versionedobjects master: Provide stable getargspec() behavior for method fingerprinting https://review.opendev.org/706548 | 18:37 |
| *** mmethot has joined #openstack-oslo | 20:07 | |
| *** mmethot_ has quit IRC | 20:09 | |
| *** lbragstad has quit IRC | 20:48 | |
| *** raildo_ has joined #openstack-oslo | 21:34 | |
| *** raildo has quit IRC | 21:37 | |
| *** raildo_ has quit IRC | 21:53 | |
| *** iurygregory has quit IRC | 22:29 | |
| *** tosky has joined #openstack-oslo | 22:31 | |
| *** camelCaser has quit IRC | 22:57 | |
| *** camelCaser has joined #openstack-oslo | 22:58 | |
| *** ralonsoh has quit IRC | 23:56 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!