| *** swest has quit IRC | 01:11 | |
| *** swest has joined #zuul | 01:25 | |
| *** bhavikdbavishi has joined #zuul | 02:57 | |
| *** bhavikdbavishi1 has joined #zuul | 02:59 | |
| *** bhavikdbavishi has quit IRC | 03:01 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 03:01 | |
| *** rlandy|bbl is now known as rlandy | 03:30 | |
| *** rlandy has quit IRC | 03:49 | |
| *** sanjayu_ has joined #zuul | 04:12 | |
| *** pcaruana has joined #zuul | 04:55 | |
| *** bolg has quit IRC | 05:15 | |
| *** evrardjp_ has joined #zuul | 05:49 | |
| *** tbarron has quit IRC | 05:54 | |
| *** evrardjp has quit IRC | 05:55 | |
| *** irclogbot_0 has quit IRC | 05:56 | |
| *** irclogbot_3 has joined #zuul | 05:58 | |
| *** avass has joined #zuul | 06:52 | |
| openstackgerrit | Albin Vass proposed zuul/nodepool master: Static provider defaults to ssh connection https://review.opendev.org/688043 | 06:53 |
|---|---|---|
| *** bolg has joined #zuul | 07:15 | |
| *** tobberydberg has quit IRC | 07:20 | |
| *** tobberydberg has joined #zuul | 07:26 | |
| *** jpena|off is now known as jpena | 07:41 | |
| *** hashar has joined #zuul | 07:53 | |
| *** phildawson has joined #zuul | 08:29 | |
| *** avass has quit IRC | 08:31 | |
| *** avass has joined #zuul | 08:44 | |
| *** mhu has joined #zuul | 08:48 | |
| *** bolg has quit IRC | 08:49 | |
| avass | Is it only maintainers that can recheck/regate on opendev? | 08:52 |
| AJaeger | avass: anybody can recheck. Why do you ask? | 09:08 |
| avass | AJaeger: It didn't recheck when I tried. Maybe I was too quick | 09:15 |
| AJaeger | avass: which change? | 09:30 |
| avass | AJeager: https://review.opendev.org/#/c/688043/1 But it worked now | 09:58 |
| AJaeger | avass: you can always check at http://zuul.opendev.org/t/zuul/status whether jobs got started | 10:10 |
| AJaeger | avass: the jobs run for over 30mins - I think you were just impatient and it worked directly... | 10:11 |
| AJaeger | and then often jobs need to wait for free nodes to run the tests on | 10:11 |
| avass | AJeager: I did but I couldn't find it so it didn't look like it started. If they don't have nodes they should still show up as 'queued' right? | 10:13 |
| avass | AJeager: Ah wait, now I see the problem. I was looking here: https://zuul.openstack.org/status | 10:14 |
| avass | Wrong tenant :) | 10:15 |
| *** bhavikdbavishi has quit IRC | 10:31 | |
| AJaeger | avass: yes, different tenant... | 10:57 |
| *** bolg has joined #zuul | 10:58 | |
| *** rfolco has joined #zuul | 11:02 | |
| *** bolg has quit IRC | 11:09 | |
| *** rfolco is now known as rfolco|ruck | 11:09 | |
| *** hashar has quit IRC | 11:14 | |
| *** bolg has joined #zuul | 11:20 | |
| *** jpena is now known as jpena|lunch | 11:33 | |
| *** avass has quit IRC | 11:52 | |
| *** avass has joined #zuul | 11:52 | |
| *** avass has quit IRC | 11:57 | |
| *** hashar has joined #zuul | 12:22 | |
| openstackgerrit | Paul Belanger proposed zuul/zuul master: WIP: Support Ansible 2.9 https://review.opendev.org/674854 | 12:29 |
| *** rlandy has joined #zuul | 12:32 | |
| *** ironfoot has left #zuul | 12:32 | |
| *** jpena|lunch is now known as jpena | 12:35 | |
| *** recheck has quit IRC | 12:44 | |
| *** recheck has joined #zuul | 12:44 | |
| openstackgerrit | Merged zuul/nodepool master: Static provider defaults to ssh connection https://review.opendev.org/688043 | 13:00 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: configure-mirrors: use dnf when needed https://review.opendev.org/688118 | 13:01 |
| *** bhavikdbavishi has joined #zuul | 13:12 | |
| pabelanger | Friday hacking, getting centos-8 jobs working | 13:19 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: configure-mirrors: use dnf when needed https://review.opendev.org/688118 | 13:21 |
| *** gtema_ has joined #zuul | 13:34 | |
| pabelanger | I've sent an email to ML about ansible version changes for Zuul. It would be awesome if people could look and comment | 13:44 |
| pabelanger | zbr:^ | 13:44 |
| mordred | corvus, tristanC: the registry stack has 2x+2 now, but I left off +A in case anyone else wanted to read it | 13:55 |
| tristanC | mordred: thanks, though i don't have time to review it today | 13:57 |
| mordred | tristanC: cool. we can probably just land it and deal with issues as followups too | 14:00 |
| *** fdegir has quit IRC | 14:06 | |
| *** fdegir has joined #zuul | 14:06 | |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: configure-mirrors: use dnf when needed https://review.opendev.org/688118 | 14:18 |
| *** gtema_ has quit IRC | 14:18 | |
| mordred | corvus: morning! I was looking at timburke's comment about the ksa exceptions vs the openstacksdk exceptions in registry. there are several places where you use rest calls directly on a session instead of doing so on the service adapter - and instead of using an sdk call. are those specifically on purpose or just how it made sense as you were working on it? | 14:18 |
| *** hashar has quit IRC | 14:19 | |
| corvus | mordred: some of them stem from places where i didn't know how to do the same in sdk (which may be a doc bug) -- others might be places where i copy-pastad from those | 14:20 |
| corvus | mordred: main thing is, i wasn't trying to avoid it on purpose (the opposite), but i also didn't root-cause each speedbump | 14:22 |
| mordred | corvus: cool | 14:23 |
| mordred | corvus: yeah - there's some places where it's straightforward to swap out sdk - some where it's not, which is a good example of needing to improve sdk | 14:24 |
| mordred | (stream object is a good example - you need size *and* the streaming content - sdk only gives you streaming content) | 14:24 |
| corvus | mordred: ack -- though i think i originally wrote it dumber than that and only got the streaming content; i think using both was timburke's improvement | 14:25 |
| corvus | mordred: that may have been a case where it was like "i know how to stream with requests" | 14:26 |
| corvus | mordred, timburke: do you know if openstacksdk can talk to the swift in the saio container image that swift is building? (does sdk need a keystone? does saio have one?) | 14:26 |
| tdasilva | corvus: saio does not have keystone | 14:27 |
| mordred | sdk doesn't need keystone - we can put in an endpoint_override | 14:27 |
| corvus | mordred: that's a clouds.yaml thing? | 14:27 |
| mordred | that said - sdk doesn't support non-keystone *for auth* - so if the saio swift needs auth we'd still be sol | 14:27 |
| mordred | corvus: yah | 14:28 |
| corvus | tdasilva: ^ know the answer to the auth question? | 14:28 |
| tdasilva | :( | 14:28 |
| corvus | :( | 14:28 |
| tdasilva | yeah, we use tempauth | 14:28 |
| corvus | for the test job i'd like to do, i don't care about auth, but i do just need to pass a clouds.yaml in and get an sdk object out | 14:29 |
| tdasilva | what does a cloud.yaml file looks like? where can i find where sdk does auth for object storage? | 14:30 |
| clarkb | you can devstack with only keystone and swift enabled | 14:30 |
| clarkb | and that will give you a working clouds.yaml | 14:30 |
| mordred | tdasilva: sdk uses keystoneauth for auth | 14:30 |
| mordred | so the auth layer is down in the guts of the system - that said - it *is* pluggable at the ksa layer - so it's possible we could write a ksa auth plugin that implements swift auth | 14:31 |
| tdasilva | fwiw, removing auth from a saio container should be pretty simple...in case sdk can work with no auth at all | 14:32 |
| mordred | yes - sdk can definitely handle noauth | 14:32 |
| tdasilva | just remove tempauth from the pipeline in /etc/swift/proxy-server.conf | 14:32 |
| corvus | i'd like to test the part of the code that makes the container though :/ | 14:32 |
| corvus | oh that sounds like something we could do in setup | 14:32 |
| mordred | let me find a clouds.yaml ironic noauth example that should work | 14:33 |
| tdasilva | https://github.com/openstack/swift/blob/master/docker/rootfs/etc/swift/proxy-server.conf#L12 | 14:33 |
| corvus | maybe we can sed that out in a run command or something when we start the container | 14:34 |
| corvus | (or bind-mount in a local copy) | 14:34 |
| mordred | corvus: http://paste.openstack.org/show/783012/ | 14:35 |
| mordred | corvus: that's what the clouds.yaml should look like | 14:36 |
| corvus | cool, maybe we have enough to try a swift-backed gate test for this | 14:36 |
| tdasilva | corvus: if you drop a little script here https://github.com/openstack/swift/tree/master/docker/rootfs/etc/cont-init.d it would fix on start up | 14:37 |
| corvus | that's handy | 14:37 |
| tdasilva | i've used crudini for those kind of changes. Example: https://github.com/swiftstack/file-connector/blob/master/configs/rootfs/etc/cont-init.d/50_set_env | 14:38 |
| corvus | that looks like just what we need | 14:39 |
| mordred | tdasilva: how do accounts work in swift aio if there's no auth? | 14:39 |
| mordred | because normally endpoints are like https://swift.example.com/v1/AUTH_{project_id} - I'd assune without keystone projects there's still a swift account that would go in the url to be where containers go? | 14:40 |
| mordred | or would it just be https://swift.example.com/v1/{container}/{object} without auth? | 14:40 |
| tdasilva | mordred: we have auto_create enabled: https://github.com/openstack/swift/blob/master/docker/rootfs/etc/swift/proxy-server.conf#L94 | 14:41 |
| tdasilva | mordred: nope, it would still be AUTH_whatever you pass in | 14:41 |
| *** michael-beaver has joined #zuul | 14:41 | |
| mordred | gotcha - but we'd just make up something to put after AUTH and swift would create that? | 14:41 |
| tdasilva | correct | 14:41 |
| mordred | cool | 14:41 |
| mordred | corvus: in the exmaple clouds.yaml I gave you - we'll need to append the AUTH_whatever to that object_store_endpoint_override | 14:42 |
| corvus | whatever sounds good to me | 14:45 |
| corvus | on another note -- https://gerrit-review.googlesource.com/c/plugins/checks/+/231201 has our first interaction with the gerrit checks api (click the 'checks' tab) | 14:52 |
| corvus | the good news is that the "unevaluated" means "zuul has looked at this and decided that it doesn't have anything to do". for some reason, that's how they display "not relevant" in the ui | 14:53 |
| mordred | corvus: that's super awesome | 14:53 |
| corvus | the bad news is that part of the reason it decided that was that it couldn't actually clone the repo, because some of the characters in our password are confusing git's url parser | 14:54 |
| corvus | so we may be missing some escaping somewher | 14:54 |
| *** bolg has quit IRC | 15:05 | |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: Enable zuul-jobs-test-base-roles-centos-8 as nv https://review.opendev.org/688146 | 15:18 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: configure-mirrors: use dnf when needed https://review.opendev.org/688118 | 15:21 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: configure-mirrors: use dnf when needed https://review.opendev.org/688118 | 15:22 |
| *** jamesmcarthur has joined #zuul | 15:39 | |
| corvus | yeah, we need to url-encode the password | 15:40 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Use gzip to compress files uploaded to swift https://review.opendev.org/688154 | 15:50 |
| *** jangutter has quit IRC | 15:50 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: URL quote username/password in gerrit https://review.opendev.org/688155 | 15:53 |
| *** sanjayu_ has quit IRC | 16:01 | |
| openstackgerrit | Clark Boylan proposed zuul/nodepool master: Handle case where nova server is in DELETED state https://review.opendev.org/688157 | 16:01 |
| pabelanger | Yay, centos-8 job working for tox: https://dashboard.zuul.ansible.com/t/ansible/build/930bc196797648fbaec408da8658da33 | 16:04 |
| AJaeger | congrats! | 16:06 |
| *** mhu has quit IRC | 16:06 | |
| pabelanger | AJaeger: thanks goes to ianw! It just worked when I added it to nodepool | 16:07 |
| AJaeger | lucky you ;) | 16:07 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: configure-mirrors: use dnf when needed https://review.opendev.org/688118 | 16:07 |
| *** jamesmcarthur has quit IRC | 16:19 | |
| *** jamesmcarthur has joined #zuul | 16:23 | |
| *** bhavikdbavishi has quit IRC | 16:30 | |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: configure-mirrors: use dnf when needed https://review.opendev.org/688118 | 16:43 |
| *** paladox has quit IRC | 16:49 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: URL quote username/password in gerrit https://review.opendev.org/688155 | 16:54 |
| corvus | tobiash, mordred: ^ added a unit test -- we don't exercise that path in the more functional tests because we override it to do local clones. so that at least runs the code to make sure i didn't typo "urllib" :) | 16:54 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Use gzip to compress files uploaded to swift https://review.opendev.org/688154 | 16:55 |
| *** paladox has joined #zuul | 16:57 | |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Use gzip to compress files uploaded to swift https://review.opendev.org/688154 | 16:58 |
| corvus | clarkb: to test this, should we make a copy of the role, update opendev base-test to use it, then parent something to base-test? | 16:59 |
| clarkb | corvus: it probably wouldn't hurt | 16:59 |
| clarkb | corvus: as an alternative maybe we can just run it by hand against rax? | 16:59 |
| corvus | clarkb: yeah, i think that would be sufficient | 17:00 |
| clarkb | hrm though that won't check the ansible python version concerns | 17:00 |
| corvus | oh right | 17:00 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Use gzip to compress files uploaded to swift https://review.opendev.org/688154 | 17:04 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Test role for upload-logs-swift https://review.opendev.org/688177 | 17:04 |
| clarkb | corvus: ^ ok that adds the test role first | 17:04 |
| clarkb | we should be able to safely merge that, then update base-test, then recheck a change proposal against base-test | 17:04 |
| clarkb | and do it a few times until we hit rax | 17:04 |
| corvus | clarkb: i single-core-approved that since the real review is happening in the other change | 17:05 |
| clarkb | sgtm | 17:05 |
| *** jpena is now known as jpena|off | 17:05 | |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: configure-mirrors: use dnf when needed https://review.opendev.org/688118 | 17:18 |
| corvus | clarkb: zuul-jobs change is sad | 17:21 |
| clarkb | corvus: hrm I wonder if the copied over test cases make too manyassumptions /me looks | 17:21 |
| corvus | one of the things is going to be no docs | 17:22 |
| clarkb | oh it isn't happy under python27 which was expected | 17:22 |
| corvus | (ie, it needs to be linked into the docs somehow) | 17:22 |
| corvus | clarkb: the py27 thing is an interesting dilemma -- maybe some kind of a conditional test skip for that? | 17:23 |
| clarkb | corvus: I think I can also convert the vendored code to be python2 compat | 17:25 |
| clarkb | it uses the gzip interface that exists in 2 and 3 | 17:25 |
| clarkb | and io.BytesIO is also in 2 and 3 | 17:25 |
| clarkb | maybe start there? | 17:25 |
| corvus | clarkb: yeah, that sounds better | 17:27 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Use gzip to compress files uploaded to swift https://review.opendev.org/688154 | 17:29 |
| clarkb | oh derp | 17:29 |
| clarkb | I didn't rebase properly. One moment | 17:29 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Test role for upload-logs-swift https://review.opendev.org/688177 | 17:34 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Use gzip to compress files uploaded to swift https://review.opendev.org/688154 | 17:34 |
| clarkb | https://review.opendev.org/688154 is unmodified as I want to avoid making updates to it that will just be noise until we nail down that this works iwth the test changes | 17:34 |
| *** rlandy is now known as rlandy|brb | 17:36 | |
| *** noorul has joined #zuul | 17:37 | |
| noorul | ofosos: hi | 17:37 |
| clarkb | arg I still got the rebase wrong | 18:00 |
| openstackgerrit | Clark Boylan proposed zuul/nodepool master: Handle case where nova server is in DELETED state https://review.opendev.org/688157 | 18:01 |
| clarkb | mordred: ^ now with testing. corvus Shrews review of ^ would be appreciated. Possibly fixes an issue we are seeing in the wild | 18:01 |
| corvus | tobiash, clarkb, mordred: can i get some +3 on https://review.opendev.org/688155 ? it had 2x+2 before i added a unit test... | 18:01 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Test role for upload-logs-swift https://review.opendev.org/688177 | 18:04 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Use gzip to compress files uploaded to swift https://review.opendev.org/688154 | 18:04 |
| clarkb | maybe this time I got the rebase correct | 18:04 |
| clarkb | corvus: looking | 18:04 |
| *** rlandy|brb is now known as rlandy | 18:08 | |
| mordred | clarkb: +2 | 18:11 |
| mordred | corvus: +A | 18:11 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Test role for upload-logs-swift https://review.opendev.org/688177 | 18:15 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Use gzip to compress files uploaded to swift https://review.opendev.org/688154 | 18:15 |
| clarkb | *, somearg isn't valid argument listing in python2 but is in 3 | 18:16 |
| clarkb | tests pass locally now | 18:16 |
| mordred | clarkb: wow - you can put named arg after * in 3? | 18:17 |
| mordred | I | 18:17 |
| mordred | TIL | 18:17 |
| fungi | yep | 18:17 |
| clarkb | yup | 18:17 |
| fungi | has to be passed by name exclusively though, for obvious reasons | 18:17 |
| corvus | mordred: i did not know that :) | 18:17 |
| mordred | oh - right - that's actually the thing to indicate "anything after this is keyword only" | 18:17 |
| mordred | def foo(bar, *, bang): means bang must be keyword only | 18:18 |
| mordred | and then there's something in 3.8 for disallowing things from being keyword args too | 18:18 |
| fungi | python 2.8 also gives you a way to have positional-only args (before the *) | 18:18 |
| mordred | yeah | 18:18 |
| fungi | er, 3.8 yeah | 18:18 |
| fungi | it's in the 3.8 release notes new features section | 18:19 |
| mordred | def f(a, b, /, c, d, *, e, f): | 18:19 |
| *** hashar has joined #zuul | 18:19 | |
| mordred | a and b are positional only, c and d are normal, e and f are keyword only | 18:19 |
| corvus | mordred: you're pulling my leg. that's totally perl5. | 18:20 |
| fungi | /, is an interesting choice syntactically | 18:20 |
| mordred | https://docs.python.org/3.9/whatsnew/3.8.html#positional-only-parameters | 18:20 |
| mordred | corvus: https://docs.python.org/3.9/whatsnew/3.8.html#assignment-expressions is the thing that will really blow your mind | 18:20 |
| fungi | it's not entirely clear to me why you'd want positional-only args over normal nameable positional args, but i guess there must be some use case | 18:21 |
| mordred | fungi: they make a couple of arguments in favor of positional-only use cases | 18:21 |
| mordred | fungi: for instance, len(obj='hello') - obj doesn't make it more readable, even though that's the argument name | 18:22 |
| clarkb | in this case it was being used to mask out args they didn't care about but otherwise fit to an existing api | 18:22 |
| clarkb | because we don't care about them for this use case I simply removed the * | 18:22 |
| mordred | also "Since the parameters to the left of / are not exposed as possible keywords, the parameters names remain available for use in **kwargs:" | 18:22 |
| clarkb | hrm though now that I've said that I think there is a bug in my change | 18:23 |
| fungi | mordred: yeah, rereading now, i see where emulating wrapped c functions might be convenient, though the other arguments they make seem a bit more specious to me | 18:24 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Test role for upload-logs-swift https://review.opendev.org/688177 | 18:25 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Use gzip to compress files uploaded to swift https://review.opendev.org/688154 | 18:25 |
| clarkb | latest patchsets set the compression level to a default of 9 | 18:26 |
| mordred | fungi: I agree | 18:26 |
| fungi | it does allow you to prevent callers from relying on the arbitrary parameter names you've chosen, but there are also plenty of ways to tell authors of consuming code to not do it or that you don't support it... it lets you reuse the same names for positional-only arguments and kwargs, but you could probably also just write the function less ambiguously... it keeps people who use your function from unnecessarily | 18:28 |
| fungi | including a pointless argument name forcing their code to be more readable? i don't even get that one | 18:28 |
| fungi | they seem to be advertising it as a mechanism for taking away flexibility | 18:29 |
| corvus | clarkb: comment on 177 | 18:31 |
| corvus | at least, that's where we put test-mirror-workspace-git-repos | 18:31 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Test role for upload-logs-swift https://review.opendev.org/688177 | 18:32 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Use gzip to compress files uploaded to swift https://review.opendev.org/688154 | 18:32 |
| clarkb | makes sense | 18:32 |
| openstackgerrit | Monty Taylor proposed zuul/zuul-registry master: Catch openstack exceptions instead of keystoneauth https://review.opendev.org/688185 | 18:37 |
| *** rlandy has quit IRC | 18:38 | |
| *** rlandy has joined #zuul | 18:39 | |
| *** pcaruana has quit IRC | 18:55 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Test role for upload-logs-swift https://review.opendev.org/688177 | 18:56 |
| openstackgerrit | Merged zuul/zuul master: URL quote username/password in gerrit https://review.opendev.org/688155 | 19:00 |
| clarkb | I've rechecked https://review.opendev.org/#/c/680178/2 to test the gzipped swift logs | 19:11 |
| *** jamesmcarthur has quit IRC | 19:26 | |
| *** jamesmcarthur has joined #zuul | 19:40 | |
| *** jamesmcarthur has quit IRC | 19:52 | |
| *** noorul has quit IRC | 20:23 | |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul master: DNM: Test changes to base-test job https://review.opendev.org/688199 | 20:31 |
| *** VitaliyL has joined #zuul | 21:01 | |
| VitaliyL | Hi, I have a loggin file logging.conf (see below). This file is used as log_config parameter for all zuul services (scheduler, merger, executor, web). | 21:04 |
| VitaliyL | Hi, I have a loggin file logging.conf (see below). This file is used as log_config parameter for all zuul services (scheduler, merger, executor, web). The issue is all messages from scheduler are double printed like this^ | 21:05 |
| VitaliyL | Hi, I have a loggin file logging.conf (see below). This file is used as log_config parameter for all zuul services (scheduler, merger, executor, web). The issue is all messages from scheduler are double printed like this: scheduler_1 | 2019-10-11 20:54:23,974 INFO zuul.TenantParser: _cacheTenantYAML: configloader: | 21:05 |
| VitaliyL | /usr/local/lib/python3.7/site-packages/zuul/configloader.py: MainProcess: 11: 1 | 21:05 |
| VitaliyL | 39676539156224: Loading configuration from zuul/zuul-jobs/zuul.d/python-jobs.yaml@master | 21:05 |
| VitaliyL | 2019-10-11 20:54:24,088 INFO zuul.Scheduler: _doReconfigureEvent: scheduler: /usr/local/lib/python3.7/site-packages/zuul/scheduler.py: MainProcess: 11: 139676539156224: Full reconfiguration complete (duration: 58.264 seconds) | 21:05 |
| clarkb | VitaliyL: I think that means you have two handlers matching the scheduler logger? | 21:06 |
| clarkb | so you get it emitted twice once for each handler | 21:06 |
| VitaliyL | Sorry, I'm also struggling with webchat console on freenode. My logging.conf file: | 21:07 |
| VitaliyL | [loggers] | 21:07 |
| VitaliyL | %(module)s: %(pathname)s: %(processName)s: %(process)d: %(thread)d: %(message)s | 21:07 |
| clarkb | VitaliyL: you can use a paste tool like paste.openstack.org or github's gist to paste large amounts of text | 21:07 |
| clarkb | that is typically better than pasting it verbatim here | 21:07 |
| VitaliyL | clarkb, thanks for an advice, here is my logging.conf: http://paste.openstack.org/show/783028/, it's based on https://opendev.org/zuul/zuul/src/branch/master/etc/logging.conf-sample, also I removed extra handlers to keep it simple. | 21:12 |
| clarkb | I wonder if root and zuul are both matching since root should cover everything and zuul will cover the scheduler | 21:15 |
| clarkb | I would expect it to simply apply the info leve to zuul and below though | 21:15 |
| *** rfolco|ruck has quit IRC | 21:17 | |
| VitaliyL | root logger is configured to have WARNING level, while zuul logger is configured for INFO one. I have INFO messages doubled, which means they come from zuul logger. | 21:19 |
| VitaliyL | Here is output from 'docker-compose logs scheduler': http://paste.openstack.org/show/783030/ | 21:22 |
| clarkb | corvus: do we accept both the logging config but then still write to stdout by default with the quickstart setup? | 21:25 |
| clarkb | that might explain this behavior too (if we aren't overriding the default with provided config) | 21:25 |
| VitaliyL | Actually only I need is to switch between DEBUG/WARNING levels, is there a way to achive that simpler rather than using logging.conf? | 21:35 |
| clarkb | reading the code I don't think we are misapplying the config like I theorized may happen | 21:42 |
| fungi | there was a discussion not long ago about adding a command-line option or rpc command to set the stdout log level | 21:44 |
| fungi | the python logging module config is not a very user-friendly mechanism for managing such things, even if it's more flexible than what we would probably implement | 21:45 |
| clarkb | VitaliyL: the goal is to log zuul logs at info level and everything else at warning? | 21:46 |
| clarkb | VitaliyL: can you share your docker-compose edits too? | 21:46 |
| VitaliyL | > the goal is to log zuul logs at info level and everything else at warning? | 21:49 |
| VitaliyL | The goal is to be able to switch zuul logs between debug/warning level and not interested in other logs (like gearman). | 21:51 |
| clarkb | oh if this is going through the reload path then that may explain it too | 21:52 |
| clarkb | VitaliyL: does it double log prior to loading the new config? | 21:52 |
| VitaliyL | clarkb, could you clarify that? | 21:54 |
| clarkb | VitaliyL: does the logging do what you expect prior to the switching that you want to do | 21:54 |
| *** rlandy has quit IRC | 21:57 | |
| VitaliyL | I still don't understand the question. zuul.conf always refers to logging.conf. After changing debug level in logging.conf I do 'docker-compose restart scheduler' and watch logs 'docker-compose logs -f scheduler'. | 21:58 |
| VitaliyL | Debug level specified in logging.conf for zuul logger always matches that in stdout. Whatever logging level is zuul.* message are always doubled. | 21:59 |
| clarkb | VitaliyL: are you using the quickstart docker-compose file? If so that does not use a logging.conf but instead relies on a default behavior of zuul for logging iirc | 21:59 |
| openstackgerrit | Merged zuul/zuul-jobs master: Use gzip to compress files uploaded to swift https://review.opendev.org/688154 | 22:02 |
| VitaliyL | Yes, I started with quick start docker-compose file. I've already changed etc/zuul.conf so that {scheduler,merger,web,executor}.log_config=/etc/zuul/logging.conf | 22:02 |
| clarkb | can you share the zuul.conf and the logging.conf (both versions) | 22:03 |
| VitaliyL | clarkb, here is a paste with both files: http://paste.openstack.org/show/783031/ | 22:08 |
| clarkb | hrm ya I don't know that all looks fine to me | 22:12 |
| VitaliyL | I'll try to reproduce with vanilla zuul quick-start + logging.conf added to zuul.conf | 22:31 |
| *** hashar has quit IRC | 22:34 | |
| *** sgw has quit IRC | 22:44 | |
| *** jamesmcarthur has joined #zuul | 22:45 | |
| VitaliyL | clarkb, I reproduced with vanilla zuul quick-start. Here is a complete instruction: http://paste.openstack.org/show/783108/ | 23:06 |
| *** sgw has joined #zuul | 23:12 | |
| *** sgw has quit IRC | 23:20 | |
| *** VitaliyL has quit IRC | 23:20 | |
| *** sgw has joined #zuul | 23:21 | |
| clarkb | Vitaliyl is gone btu it having a paste of the actual doubled log output would be good too | 23:35 |
| clarkb | I think irc truncated the previous example | 23:35 |
| *** michael-beaver has quit IRC | 23:38 | |
| *** armstrongs has joined #zuul | 23:46 | |
| *** jamesmcarthur has quit IRC | 23:53 | |
| *** jamesmcarthur has joined #zuul | 23:57 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!