| *** openstack has joined #openstack-oslo | 07:32 | |
| *** ChanServ sets mode: +o openstack | 07:32 | |
| *** rcernin_ has quit IRC | 07:34 | |
| *** tosky has joined #openstack-oslo | 07:37 | |
| *** dtantsur|afk is now known as dtantsur | 07:55 | |
| openstackgerrit | Jorhson Deng proposed openstack/oslo.cache master: Support SASL protocol for memcached https://review.opendev.org/742060 | 08:06 |
|---|---|---|
| openstackgerrit | Jorhson Deng proposed openstack/oslo.cache master: Support SASL protocol for memcached https://review.opendev.org/742060 | 08:15 |
| *** rcernin_ has joined #openstack-oslo | 08:16 | |
| *** rcernin_ has quit IRC | 08:26 | |
| *** moguimar has joined #openstack-oslo | 08:37 | |
| *** rcernin_ has joined #openstack-oslo | 08:47 | |
| *** rcernin_ has quit IRC | 09:07 | |
| *** rcernin_ has joined #openstack-oslo | 09:41 | |
| *** rcernin_ has quit IRC | 09:55 | |
| *** rcernin_ has joined #openstack-oslo | 10:28 | |
| openstackgerrit | Hervé Beraud proposed openstack/automaton master: Add pre-commit https://review.opendev.org/742125 | 11:27 |
| openstackgerrit | Hervé Beraud proposed openstack/debtcollector master: Add pre-commit https://review.opendev.org/742127 | 11:28 |
| openstackgerrit | Hervé Beraud proposed openstack/futurist master: Add pre-commit https://review.opendev.org/742128 | 11:28 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.concurrency master: Add pre-commit https://review.opendev.org/742129 | 11:29 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.config master: Add pre-commit https://review.opendev.org/742130 | 11:30 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.context master: Add pre-commit https://review.opendev.org/742131 | 11:30 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.db master: Add pre-commit https://review.opendev.org/742132 | 11:31 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.i18n master: Add pre-commit https://review.opendev.org/742133 | 11:31 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.limit master: Add pre-commit https://review.opendev.org/742134 | 11:32 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.log master: Add pre-commit https://review.opendev.org/742135 | 11:32 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.messaging master: Add pre-commit https://review.opendev.org/742137 | 11:33 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.middleware master: Add pre-commit https://review.opendev.org/742138 | 11:33 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.policy master: Add pre-commit https://review.opendev.org/742139 | 11:34 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.privsep master: Add pre-commit https://review.opendev.org/742140 | 11:35 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.reports master: Add pre-commit https://review.opendev.org/742141 | 11:35 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.rootwrap master: Add pre-commit https://review.opendev.org/742142 | 11:36 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.serialization master: Add pre-commit https://review.opendev.org/742143 | 11:36 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.service master: Add pre-commit https://review.opendev.org/742144 | 11:37 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.upgradecheck master: Add pre-commit https://review.opendev.org/742147 | 11:37 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.utils master: Add pre-commit https://review.opendev.org/742148 | 11:38 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.versionedobjects master: Add pre-commit https://review.opendev.org/742149 | 11:38 |
| openstackgerrit | Hervé Beraud proposed openstack/oslo.vmware master: Add pre-commit https://review.opendev.org/742150 | 11:39 |
| openstackgerrit | Hervé Beraud proposed openstack/oslotest master: Add pre-commit https://review.opendev.org/742151 | 11:39 |
| openstackgerrit | Hervé Beraud proposed openstack/osprofiler master: Add pre-commit https://review.opendev.org/742152 | 11:40 |
| openstackgerrit | Hervé Beraud proposed openstack/stevedore master: Add pre-commit https://review.opendev.org/742153 | 11:40 |
| openstackgerrit | Hervé Beraud proposed openstack/taskflow master: Add pre-commit https://review.opendev.org/742154 | 11:41 |
| openstackgerrit | Hervé Beraud proposed openstack/tooz master: Add pre-commit https://review.opendev.org/742155 | 11:42 |
| openstackgerrit | Hervé Beraud proposed openstack/openstack-doc-tools master: Add pre-commit https://review.opendev.org/742157 | 11:43 |
| openstackgerrit | Hervé Beraud proposed openstack/openstackdocstheme master: Add pre-commit https://review.opendev.org/742158 | 11:43 |
| openstackgerrit | Hervé Beraud proposed openstack/os-api-ref master: Add pre-commit https://review.opendev.org/742159 | 11:44 |
| openstackgerrit | Hervé Beraud proposed openstack/pbr master: Add pre-commit https://review.opendev.org/742160 | 11:44 |
| openstackgerrit | Hervé Beraud proposed openstack/sphinx-feature-classification master: Add pre-commit https://review.opendev.org/742161 | 11:45 |
| openstackgerrit | Merged openstack/oslo.service master: Do not import "oslo.log" in the main module https://review.opendev.org/741972 | 11:47 |
| *** sean-k-mooney has quit IRC | 11:48 | |
| *** iurygregory has quit IRC | 11:49 | |
| dtantsur | hberaud: I believe you folks have discussed it already, but I do like to point out that all this ^^^ opens up remote code execution on all developer machines | 11:50 |
| dtantsur | can you at least pull the commit hooks from a trusted location rather than a random dude's github? | 11:51 |
| openstackgerrit | Jorhson Deng proposed openstack/oslo.cache master: Support SASL protocol for memcached https://review.opendev.org/742060 | 11:51 |
| * dtantsur disables hooks on his machine until this is resolved | 11:51 | |
| hberaud | dtantsur: sure | 11:53 |
| openstackgerrit | Jorhson Deng proposed openstack/oslo.cache master: Support SASL protocol for memcached https://review.opendev.org/742060 | 11:53 |
| dtantsur | hberaud: what I mean is something like https://opendev.org/openstack/oslo-commit-hooks | 11:54 |
| dtantsur | which I'm very cool with trusting to execute code on my machine | 11:54 |
| hberaud | dtantsur: sorry I was not aware of about this discussion | 11:54 |
| dtantsur | I dunno if it has been discussed or not, just pointing out a potential problem | 11:54 |
| hberaud | dtantsur: I agree I was thinking about a centralized official repo pulled on all your pre-commit config to avoid to patch ~40 repos all the time | 11:55 |
| hberaud | dtantsur: thanks for the heads up | 11:56 |
| dtantsur | yep. maybe even skip "oslo" and just start collecting useful hooks | 11:56 |
| *** tkajinam has quit IRC | 11:56 | |
| dtantsur | in openstack/commit-hooks | 11:56 |
| hberaud | yep | 11:56 |
| hberaud | dtantsur: FYI, long short story, all my patches are based on => https://review.opendev.org/#/c/720771/ | 11:58 |
| *** iurygregory has joined #openstack-oslo | 11:59 | |
| *** raildo has joined #openstack-oslo | 12:01 | |
| openstackgerrit | Merged openstack/tooz master: hashring: allow choosing hash function https://review.opendev.org/739725 | 12:12 |
| *** dave-mccowan has joined #openstack-oslo | 12:28 | |
| dtantsur | unrelated, but could you folks release tooz with ^^ please? | 12:40 |
| moguimar | dtantsur, I was also a bit unsure about random dude commit hook | 12:42 |
| moguimar | but what about https://github.com/pre-commit/pre-commit-hooks | 12:42 |
| moguimar | and https://gitlab.com/pycqa/flake8 | 12:42 |
| dtantsur | pycqa is an authority I know, the pre-commit project - dunno | 12:42 |
| moguimar | I don't think dropping https://github.com/Lucas-C/pre-commit-hooks would hurt that much | 12:43 |
| dtantsur | even like that, I'm still worried that one unfortunate commit (not necessary malicious) can make local development of oslo impossible | 12:43 |
| dtantsur | (unless I misunderstand how it works) | 12:43 |
| moguimar | for now it is not gating | 12:43 |
| moguimar | and not automaticly installed | 12:44 |
| moguimar | you can run with tox -e pre-commit | 12:44 |
| moguimar | just to experiment | 12:44 |
| moguimar | but you can also `pre-commit install` | 12:44 |
| moguimar | and have the hooks on | 12:44 |
| moguimar | the idea is to get it gating in pep8 testenv | 12:45 |
| moguimar | otherwhise we wouldn't be able to enforce the hooks style | 12:45 |
| dtantsur | while it's opt-in it's fine. as soon as it gets run automatically with `tox -epep8`, it can become a problem | 12:51 |
| dtantsur | and (if airplanes are still a thing), please think of people running tests on an airplane ;) | 12:52 |
| *** rcernin_ has quit IRC | 12:54 | |
| hberaud | moguimar: I agree we could easily drop this one https://github.com/Lucas-C/pre-commit-hooks without much issues | 13:17 |
| hberaud | dtantsur: concerning the pycqa hooks I thing we can trust him, it will only call flake8 (https://github.com/pycqa/flake8/blob/master/.pre-commit-hooks.yaml) so if the hooks is broken then flake8 will be broken in the same manner even if we don't use pre-commit | 13:20 |
| hberaud | while the requirements version of flake8 stay sync with the hook's version | 13:21 |
| openstackgerrit | Jorhson Deng proposed openstack/oslo.cache master: Support SASL protocol for memcached https://review.opendev.org/742060 | 13:25 |
| hberaud | I mean while these versions stay synced then if an issue exist in flake8 we will facing this one whatever if we use tox or pre-commit | 13:25 |
| hberaud | moguimar, dtantsur: concerning https://github.com/pre-commit/pre-commit-hooks I think that while the version is pinned then we have no much chances to seen unexpected behaviors and if I'm right pre-commit force us to pin repo versions, and this is already the case on our config => https://review.opendev.org/#/c/720771/6/.pre-commit-config.yaml | 13:29 |
| hberaud | thoughts? | 13:30 |
| dtantsur | hberaud: tags are replaceable (although this is a problem only in case of malicious modifications) | 13:34 |
| dtantsur | maybe use a git hash? these are hard to reproduce | 13:34 |
| hberaud | dtantsur: right | 13:34 |
| *** kevko has joined #openstack-oslo | 13:34 | |
| hberaud | dtantsur: indeed git hash could be more secure | 13:34 |
| kevko | hi guys, can someone discuss with me oslo.cache , i found a big bug | 13:34 |
| hberaud | kevko: which one? | 13:34 |
| hberaud | kevko: go wild | 13:35 |
| kevko | hberaud: https://bugs.launchpad.net/oslo.cache/+bug/1888394 | 13:35 |
| openstack | Launchpad bug 1888394 in oslo.cache "Oslo.cache exponencially raising up connection to memcached" [Undecided,New] | 13:35 |
| kevko | hberaud: we met this bug when upgrading to latest oslo.cache in stein on thursday , big deployment, big traffic to memcache | 13:36 |
| hberaud | kevko: got it | 13:36 |
| hberaud | kevko: I seen you've a patch to propose, right? | 13:37 |
| hberaud | kevko: how do you plan to solve it? | 13:37 |
| *** bnemec has joined #openstack-oslo | 13:38 | |
| hberaud | kevko: I mean the flush is something important to manage stale datas | 13:38 |
| hberaud | kevko: but I understood your use case | 13:38 |
| hberaud | moguimar: FYI ^^^ (oslo.cache) | 13:39 |
| kevko | hberaud: wait a minute .. just pushing to review | 13:41 |
| hberaud | kevko: yep could be worth to continue the discussion on it | 13:41 |
| openstackgerrit | Michal Arbet proposed openstack/oslo.cache master: Do not hardcode flush_on_reconnect, move to oslo.cache config https://review.opendev.org/742193 | 13:42 |
| kevko | hberaud: just pushed to review - > https://review.opendev.org/742193 , idea of patch is that it is not hardcondig param in code ..but use from oslo.cache config backend_argument if specified .. moreover ..this option makes sense mostly in keystone ..so all services can have turned off | 13:44 |
| hberaud | kevko: yep I seen thanks, I already left a comment on it | 13:44 |
| moguimar | I added myself to review too | 13:46 |
| kevko | thank you guys :) | 13:46 |
| kevko | i hate release notes :D | 13:47 |
| hberaud | kevko: why, reno is something really well designed, based on git branches etc... and really useful for the releases team :) | 13:48 |
| * hberaud don't want to trigger a debate :) | 13:49 | |
| kevko | hberaud: i left a reply to your comment hberaud :) | 13:52 |
| hberaud | kevko: thanks | 13:53 |
| kevko | hberaud: oh, thank you ..well, then i will rework as you said :) | 13:59 |
| hberaud | kevko: you're welcome :) | 13:59 |
| *** Luzi has quit IRC | 14:34 | |
| *** Luzi has joined #openstack-oslo | 14:48 | |
| *** Luzi has quit IRC | 14:50 | |
| *** rcernin_ has joined #openstack-oslo | 14:52 | |
| *** rcernin_ has quit IRC | 14:57 | |
| dtantsur | bnemec: morning! would it be possible to release tooz with the hashring change that merged today? | 15:16 |
| bnemec | dtantsur: https://review.opendev.org/742226 | 15:36 |
| bnemec | Had some issues with the latest version of the release tooling. :-/ | 15:36 |
| dtantsur | thank you! | 15:39 |
| openstackgerrit | Michal Arbet proposed openstack/oslo.cache master: Do not hardcode flush_on_reconnect, move to oslo.cache config https://review.opendev.org/742193 | 15:45 |
| kevko | hberaud: hope that reno is OK :P | 15:46 |
| hberaud | kevko: nice changes, kudos | 15:48 |
| hberaud | kevko: I'll review on PS2 directly | 15:48 |
| hberaud | thanks | 15:48 |
| dhellmann | bnemec : zzzeek is looking at adopting stevedore in dogpile.cache but stevedore requires python >=3.6 and dogpile still supports 3.5. Do you know why that restriction was made on stevedore? | 15:52 |
| hberaud | dhellmann: I suppose it's related to https://governance.openstack.org/tc/reference/runtimes/victoria.html#python-runtimes-for-train | 15:55 |
| hberaud | (officially supported runtimes) | 15:56 |
| openstackgerrit | Doug Hellmann proposed openstack/stevedore master: restore support for python 3.5 https://review.opendev.org/742234 | 15:56 |
| dhellmann | hberaud : saying 'at a minimum' there doesn't mean no longer supporting earlier versions. it means the version specified must be among the support list | 15:56 |
| hberaud | dhellmann: AFAIK all the reasons are there => https://github.com/openstack/stevedore/blob/master/releasenotes/notes/drop-python2-support-3f0f717570cad8cb.yaml https://governance.openstack.org/tc/goals/selected/ussuri/drop-py27.html | 16:02 |
| bnemec | dhellmann: hberaud: I think part of the motivation for dropping official py35 support was that our default test nodes were not going to have that version installed anymore. | 16:24 |
| bnemec | We have kept py35 support in a few specific projects like pbr, so I think we _could_ reinstate that if it would be helpful. | 16:24 |
| hberaud | bnemec: yep you are right about the nodes, I missed that point | 16:25 |
| kevko | hberaud: i'm just curious , what is PS2 ? :) | 16:27 |
| *** hamalq has joined #openstack-oslo | 16:27 | |
| hberaud | bnemec: I think it's possible to reintroduce official support but after few discussions about this with the concerned people to reach a consensus | 16:27 |
| hberaud | kevko: Patch Set 2 | 16:28 |
| *** hamalq has quit IRC | 16:28 | |
| *** hamalq has joined #openstack-oslo | 16:29 | |
| *** ralonsoh_ has joined #openstack-oslo | 16:41 | |
| *** ralonsoh has quit IRC | 16:43 | |
| dhellmann | bnemec , hberaud : it would be nice if we could keep support. if there's no way to do that on opendev, which I can understand, I wouldn't mind moving stevedore back to a github repo and using travis for tests. that wouldn't be my first choice, of course | 16:52 |
| dhellmann | I'm also OK with saying the differences between 3.5 and 3.6 aren't going to affect how the library works, so 3.5 is just not tested | 16:52 |
| smcginnis | I think we can still test on 3.5 with xenial nodes. We would just need to defined the job to use that. | 16:53 |
| hberaud | smcginnis: it could be a good approach | 16:53 |
| dhellmann | is there a job definition for that already? | 17:09 |
| *** dtantsur is now known as dtantsur|afk | 17:15 | |
| *** hamalq has quit IRC | 17:16 | |
| *** hamalq has joined #openstack-oslo | 17:16 | |
| smcginnis | dhellmann: I think we can just add "openstack-python35-jobs" template to .zuul.yaml. | 17:17 |
| dhellmann | ok, let me try that... | 17:17 |
| openstackgerrit | Doug Hellmann proposed openstack/stevedore master: restore support for python 3.5 https://review.opendev.org/742234 | 17:18 |
| dhellmann | I'm still surprised at the difference between dropping 3.5 from the trove classifiers and setting the required version in the package metadata. | 17:19 |
| smcginnis | Yeah, not sure where that came in. | 17:19 |
| smcginnis | There were a few folks that were pushing that. | 17:19 |
| dhellmann | I could see doing it for the services, I guess? But even there it feels like overkill. | 17:19 |
| dhellmann | anyway, let's see how things fare with the new job(s) in place | 17:20 |
| smcginnis | ++ | 17:20 |
| smcginnis | I think another option, if we wanted to just define something locally, would be to make sure nodeset is set to xenial. | 17:20 |
| smcginnis | https://opendev.org/openstack/openstack-zuul-jobs/src/branch/master/zuul.d/jobs.yaml#L113 | 17:20 |
| smcginnis | But that template appears to just pull in this anyway, so it's probably good. | 17:20 |
| dansmith | bnemec: yo, buddy | 17:32 |
| bnemec | dansmith: o/ | 17:33 |
| *** ralonsoh_ has quit IRC | 17:33 | |
| smcginnis | dhellmann: Hmm, failing. | 17:33 |
| dansmith | bnemec: hey, so, oslo_concurrency.processutils.. looks to me like it tries to spawn argv[0] (i.e. the current python interpreter) with -m oslo_concurrency.prlimit when it overrides subprocess calls | 17:34 |
| dansmith | bnemec: does that sound legit? | 17:34 |
| smcginnis | dhellmann: Ah! We need different upper-constraints since openstack/requirements now only has 3.6 and up. | 17:34 |
| dansmith | bnemec: problem is, if you're glance and you subprocess exec things from your api, you get /usr/bin/uwsgi-core as your python interpreter :) | 17:34 |
| smcginnis | So need to either point that specific env at ussuri, or keep a local copy in-repo. | 17:34 |
| dhellmann | smcginnis :-( | 17:35 |
| dhellmann | my time and energy for this are pretty limited right now, unfortunately | 17:36 |
| bnemec | dansmith: It should only do that if you set prlimits, right? | 17:37 |
| dansmith | bnemec: it looks like it from the code, but I'm assuming glance is doing that | 17:38 |
| bnemec | If they need that then I think they'll have to provide python_exec to explicitly set the Python interpreter. | 17:38 |
| dansmith | bnemec: to the exec call? | 17:38 |
| bnemec | dansmith: Yeah: https://github.com/openstack/oslo.concurrency/blob/master/oslo_concurrency/processutils.py#L264 | 17:39 |
| dansmith | okay | 17:39 |
| dansmith | I wonder if I can get that from uwsgi somehow | 17:39 |
| dansmith | like "yo dawg, what python are you actually using?" | 17:40 |
| smcginnis | This seems very familiar. :/ | 17:40 |
| dansmith | smcginnis: you talking about me? | 17:40 |
| smcginnis | dansmith: Yeah. I swear I've run into this before somewhere else, but I think it was just long enough ago that I can't recall anything more than having seen it. | 17:41 |
| *** ralonsoh has joined #openstack-oslo | 17:41 | |
| dansmith | smcginnis: ack, I expect it's not common for most api servers to exec things, so probably not common to hit this | 17:41 |
| smcginnis | Could very well have been some of the past attempts with glance api. | 17:42 |
| openstackgerrit | Sean McGinnis proposed openstack/stevedore master: restore support for python 3.5 https://review.opendev.org/742234 | 17:45 |
| *** ralonsoh has quit IRC | 18:08 | |
| dansmith | bnemec: sorry I got distracted, but.. where is oslo getting the interpreter from? | 18:54 |
| bnemec | dansmith: By default it uses sys.executable. | 19:05 |
| dansmith | okay that doesn't match with what sys.executable looks like to me, | 19:05 |
| dansmith | but that too is not really python | 19:05 |
| dansmith | meaning, the path it's logging isn't the same as sys.executable | 19:05 |
| dansmith | anyway, I've been examining the environment we have for uwsgi and I don't see it give us a pointer to the right python anywhere, | 19:05 |
| dansmith | so I'm guessing we'll just have to pass it in so we can pass it to exec() | 19:06 |
| *** adriant has quit IRC | 20:39 | |
| *** adriant has joined #openstack-oslo | 20:40 | |
| *** kevko has quit IRC | 21:01 | |
| *** raildo has quit IRC | 21:08 | |
| *** Dmitrii-Sh has quit IRC | 21:09 | |
| *** Dmitrii-Sh has joined #openstack-oslo | 21:10 | |
| *** kgiusti has left #openstack-oslo | 21:43 | |
| *** kevko has joined #openstack-oslo | 21:46 | |
| *** moguimar has quit IRC | 21:59 | |
| *** rcernin_ has joined #openstack-oslo | 22:22 | |
| *** rcernin_ has quit IRC | 22:22 | |
| *** rcernin has joined #openstack-oslo | 22:34 | |
| *** bnemec has quit IRC | 22:47 | |
| *** tkajinam has joined #openstack-oslo | 22:55 | |
| *** Dmitrii-Sh has quit IRC | 23:05 | |
| *** Dmitrii-Sh has joined #openstack-oslo | 23:06 | |
| *** iurygregory has quit IRC | 23:42 | |
| *** kevko has quit IRC | 23:44 | |
| *** tosky has quit IRC | 23:47 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!