| *** odyssey4me has quit IRC | 00:11 | |
| *** odyssey4me has joined #zuul | 00:11 | |
| *** graffatcolmingov is now known as sigmavirus24 | 00:26 | |
| *** sigmavirus24 has joined #zuul | 00:26 | |
| *** JasonCL has joined #zuul | 00:41 | |
| *** JasonCL has quit IRC | 00:42 | |
| *** JasonCL has joined #zuul | 00:42 | |
| *** JasonCL has quit IRC | 00:48 | |
| *** harlowja has quit IRC | 01:21 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool master: WIP: Add fedora-28 to nodepool dsvm https://review.openstack.org/559211 | 02:01 |
|---|---|---|
| *** JasonCL has joined #zuul | 02:28 | |
| *** JasonCL has quit IRC | 02:33 | |
| *** JasonCL has joined #zuul | 02:43 | |
| *** patriciadomin has quit IRC | 02:45 | |
| *** JasonCL has quit IRC | 02:48 | |
| *** patriciadomin has joined #zuul | 02:50 | |
| *** pwhalen has joined #zuul | 02:51 | |
| *** pwhalen has joined #zuul | 02:51 | |
| *** gouthamr has quit IRC | 03:05 | |
| *** harlowja has joined #zuul | 03:53 | |
| *** snapiri has joined #zuul | 04:44 | |
| *** harlowja has quit IRC | 05:13 | |
| *** threestrands has quit IRC | 07:36 | |
| *** electrofelix has joined #zuul | 07:49 | |
| *** jpena|off is now known as jpena | 07:54 | |
| *** ssbarnea_ has joined #zuul | 09:18 | |
| *** jesusaur has quit IRC | 09:26 | |
| *** jesusaur has joined #zuul | 09:48 | |
| *** bhavik1 has joined #zuul | 10:30 | |
| *** bhavik1 has quit IRC | 10:38 | |
| *** JasonCL has joined #zuul | 10:58 | |
| *** JasonCL has quit IRC | 11:01 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack-infra/zuul master: Add branch to Zuul dashboard https://review.openstack.org/559313 | 11:18 |
| *** jpena is now known as jpena|lunch | 11:38 | |
| *** gouthamr has joined #zuul | 11:45 | |
| *** ssbarnea_ has quit IRC | 12:03 | |
| *** ssbarnea_ has joined #zuul | 12:05 | |
| *** maeca has joined #zuul | 12:13 | |
| *** ssbarnea_ has quit IRC | 12:22 | |
| mrhillsman | any thoughts on why 'nc -v -w2 ip 9000' works and i can see traffic but payloads from github fail to be delivered? | 12:30 |
| mrhillsman | i did upgrade via pip | 12:31 |
| mordred | mrhillsman: did you catch the REST API url changes (including the change in the path for github webhooks?) | 12:37 |
| mrhillsman | nope | 12:38 |
| mordred | mrhillsman: it's now going to have an /api at the front of the path | 12:38 |
| mrhillsman | thx | 12:38 |
| mordred | mrhillsman: fwiw, we also added some docs about different web deployment options ... | 12:38 |
| mrhillsman | looking at the docs now | 12:39 |
| mordred | mrhillsman: https://zuul-ci.org/docs/zuul/admin/installation.html#web-deployment-options | 12:39 |
| mordred | cool | 12:39 |
| mrhillsman | glad i tried this actually | 12:39 |
| mrhillsman | not sure folks would have picked up on this change without some pain | 12:39 |
| mrhillsman | we update last friday of every month | 12:40 |
| mordred | mrhillsman: yah - in the future we will not make breaking changes like that | 12:40 |
| *** jpena|lunch is now known as jpena | 12:41 | |
| mordred | mrhillsman: although it's good to know you're going a monthly update cadence - that means that if we do a "add new thing, deprecate old thing, later remove old thing" - we probably want to wait at least a month before removing old thing | 12:42 |
| mrhillsman | i think the changelog is a good addition as well | 12:43 |
| mrhillsman | even though it does not have this change | 12:43 |
| mordred | yah - we added it as the first thing right after the 3.0 release | 12:43 |
| mrhillsman | we can at least check there initially before we do anything | 12:43 |
| mordred | ++ | 12:43 |
| mordred | we're hoping to use that to be able to communicate changes more clearly to people | 12:44 |
| mrhillsman | and it looks like it is just a simple change of the webhook in the app so no need to bother users to approve anything | 12:45 |
| mordred | yah. everything else should be good | 12:45 |
| *** odyssey4me has quit IRC | 12:47 | |
| *** odyssey4me has joined #zuul | 12:47 | |
| *** pwhalen has quit IRC | 12:54 | |
| mrhillsman | i do not remember where i saw this but mordred there was a patch regarding zuul not getting all the data from github | 13:00 |
| mrhillsman | github3.exceptions.IncompleteResponse: None The library was expecting more data in the response (KeyError('patch',)). Either GitHub modified it's response body, or your token is not properly scoped to retrieve this information. | 13:01 |
| mrhillsman | do you happen to know the patch that fixes this? | 13:01 |
| mrhillsman | or is it related to github3 | 13:01 |
| *** JasonCL has joined #zuul | 13:05 | |
| *** JasonCL has quit IRC | 13:06 | |
| *** JasonCL has joined #zuul | 13:07 | |
| *** pwhalen has joined #zuul | 13:07 | |
| *** pwhalen has joined #zuul | 13:07 | |
| Shrews | corvus: i have created a test to reproduce the error: 'utf-8' codec can't decode byte 0xe2 in position 1023: unexpected end of data | 13:07 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/zuul master: WIP: Fix streaming decoding boundaries https://review.openstack.org/559326 | 13:09 |
| Shrews | corvus: ^^^ | 13:09 |
| *** JasonCL_ has joined #zuul | 13:10 | |
| *** patriciadomin has quit IRC | 13:10 | |
| Shrews | now to figure out the fix | 13:11 |
| *** JasonCL has quit IRC | 13:13 | |
| tobiash | mrhillsman: that is fixed in github3.py but wairing for a release | 13:16 |
| mrhillsman | i jumped back down from 1.0.2 | 13:16 |
| tobiash | ping jlk for that | 13:16 |
| mrhillsman | based on - http://git.zuul-ci.org/cgit/zuul/commit/?h=3.0.0&id=66aae2f42eacb473af2e2ca4a40ec2f12d174e68 | 13:17 |
| mrhillsman | that resolved it | 13:18 |
| tobiash | this is the pr that fixes this: https://github.com/sigmavirus24/github3.py/pull/817 | 13:20 |
| * SpamapS fixing reject label/status .. ahh.. zuul dev so much simpler than openstack deploying.. :-P | 13:29 | |
| SpamapS | tobiash: thanks for that fix btw.. I have it pulled into our env. :) | 13:30 |
| *** elyezer has joined #zuul | 13:35 | |
| *** JasonCL_ has quit IRC | 13:36 | |
| *** JasonCL has joined #zuul | 13:37 | |
| *** JasonCL has quit IRC | 13:37 | |
| *** JasonCL has joined #zuul | 13:37 | |
| *** JasonCL has quit IRC | 13:46 | |
| *** evrardjp has joined #zuul | 14:05 | |
| evrardjp | hey | 14:05 |
| evrardjp | tristanC: or mordred could you have a look at this https://review.openstack.org/#/c/559313/ ? | 14:06 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul master: Make label/status mirrors in reject for github https://review.openstack.org/559340 | 14:06 |
| evrardjp | it would make the mornings less painful. | 14:06 |
| *** JasonCL has joined #zuul | 14:10 | |
| *** JasonCL has quit IRC | 14:13 | |
| dmsimard | tobiash: do you use max-cores/max-ram in nodepool yet ? | 14:18 |
| *** JasonCL has joined #zuul | 14:21 | |
| *** JasonCL has quit IRC | 14:23 | |
| tobiash | dmsimard: no, I just use all | 14:23 |
| dmsimard | tobiash: all ? | 14:24 |
| tobiash | All nodepool can get ;) | 14:28 |
| dmsimard | tobiash: what does that mean ? max-server instead ? | 14:28 |
| dmsimard | or is there somehow a config that lets nodepool just use whatever it can (without failing on quota errors?) | 14:29 |
| pabelanger | max-server is still a cap, but optional now with quota. So it if was missing, nodepool would use everything in the tenant | 14:30 |
| pabelanger | removing max-server should do that, what errors are you seeing? | 14:31 |
| *** JasonCL has joined #zuul | 14:35 | |
| tobiash | Yes, just don't specify a limit | 14:37 |
| corvus | clarkb: responded on https://review.openstack.org/558643 tobiash, mordred: if you have a moment to review that, i'd appreciate it. i consider it a pretty important bugfix that we should get out soon to reduce the chances of relying on erroneous behavior. | 14:37 |
| corvus | dmsimard: nodepool now understands several important openstack limits and can avoid quota errors automatically. it's designed to work with heterogenous instance sizes too. we're not taking advantage of either of those things yet in openstack-infra. | 14:40 |
| corvus | (though i would very much like to) | 14:41 |
| *** JasonCL has quit IRC | 14:41 | |
| dmsimard | corvus: right, I remember discussing that with tobiash back in Denver I think -- it's a great feature but I guess my question was.. is the nodepool current default behavior (if lacking max-server/cores/ram config) to just maximize the quotas provided by openstack ? | 14:42 |
| dmsimard | pabelanger: not seeing any errors, inquiring about the feature :) | 14:42 |
| tobiash | dmsimard: yes | 14:43 |
| dmsimard | tobiash: wow nice | 14:44 |
| dmsimard | ++ | 14:44 |
| tobiash | corvus: I'm currently on a city trip but can look probably in a few hours | 14:44 |
| mordred | corvus: it looks great to me | 14:44 |
| corvus | tobiash: thx | 14:44 |
| corvus | mordred: thx | 14:45 |
| corvus | my plan is to get that merged, work on writing the containers spec, then get back to the review backlog next week | 14:46 |
| pabelanger | I think we need to update nodepool for volume quota too. Noticed that in vexxhost with boot-from-volume flag | 14:47 |
| *** JasonCL has joined #zuul | 14:48 | |
| *** JasonCL has quit IRC | 14:49 | |
| *** JasonCL has joined #zuul | 14:49 | |
| openstackgerrit | Stephen Finucane proposed openstack-infra/zuul-jobs master: Default warning-is-error to True for non-legacy Sphinx projects https://review.openstack.org/559348 | 14:51 |
| *** JasonCL has quit IRC | 14:54 | |
| openstackgerrit | Stephen Finucane proposed openstack-infra/zuul-jobs master: Default warning-is-error to True for non-legacy Sphinx projects https://review.openstack.org/559348 | 15:11 |
| *** JasonCL has joined #zuul | 15:19 | |
| dmsimard | Shrews, corvus: is it possible to update the reason for a held node ? | 15:19 |
| dmsimard | (already held) | 15:19 |
| *** JasonCL has quit IRC | 15:20 | |
| *** yolanda_ is now known as yolanda | 15:20 | |
| *** JasonCL has joined #zuul | 15:21 | |
| Shrews | dmsimard: no | 15:21 |
| *** JasonCL has quit IRC | 15:21 | |
| *** JasonCL has joined #zuul | 15:22 | |
| *** JasonCL has quit IRC | 15:23 | |
| *** JasonCL has joined #zuul | 15:24 | |
| *** JasonCL has quit IRC | 15:24 | |
| *** JasonCL has joined #zuul | 15:25 | |
| mrhillsman | man i have to run but going to come back to this, dropping it here for some thoughts | 15:31 |
| mrhillsman | question first | 15:31 |
| mrhillsman | for the noop | 15:31 |
| mrhillsman | does that go through the executor | 15:31 |
| mrhillsman | or scheduler handles it | 15:31 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/zuul master: Fix streaming decoding boundaries https://review.openstack.org/559326 | 15:33 |
| *** openstackgerrit has quit IRC | 15:33 | |
| Shrews | ianw: corvus: ^^ should fix the logging bug | 15:34 |
| evrardjp | tristanC: I am not sure what you meant in your review https://review.openstack.org/#/c/559313/, does that prevent a +w ? | 15:35 |
| evrardjp | oh you meant for ease of use | 15:35 |
| corvus | mrhillsman: it stays within the scheduler | 15:46 |
| clarkb | corvus: I've +2'd role path change. Left one tiny little nit comment but not worth changing | 15:55 |
| clarkb | I didn't approve it as again I think maybe tobias would be a good reviewer to have on it? | 15:55 |
| corvus | clarkb: yeah, tobiash said he should be able to get to it later today | 15:58 |
| clarkb | Shrews: comment on https://review.openstack.org/#/c/559326/2 worried, the test could have a race and be flaky depending on timing (suggested a fix though). | 16:03 |
| *** maeca has quit IRC | 16:03 | |
| Shrews | clarkb: the code you're worried about isn't the code we are testing | 16:06 |
| Shrews | it's within ZuulWeb itself | 16:06 |
| Shrews | but it does make me wonder about the possibility within there | 16:07 |
| clarkb | Shrews: I think even then it could buffer n bytes < 1025 | 16:07 |
| Shrews | clarkb: but i think that's ok? it should still pass either way (so wouldn't be flakey) | 16:08 |
| Shrews | i'm not sure how to guarantee that code is ALWAYS reading the max buffer size though | 16:09 |
| clarkb | Shrews: the last assert requires that the full file have been read all the way through | 16:09 |
| clarkb | Shrews: in which case it would fail if you read less than the 1025 bytes I think | 16:09 |
| clarkb | as for guaranteeing that I think you just have a loop that checks for the buffer size in the client to go higher than 1024 | 16:10 |
| clarkb | before ending the job | 16:10 |
| Shrews | wait... you just confuse dme | 16:10 |
| clarkb | self.assertEqual(file_contents, self.ws_client_results) file_contents will be 1025 bytes (maybe 1026 with a newline) | 16:11 |
| Shrews | i think maybe we're talking about different things | 16:11 |
| clarkb | but self.ws_client_results will be some number of bytes between 0 and log_file_size | 16:11 |
| *** acozine1 has joined #zuul | 16:11 | |
| clarkb | oh actually where does self.ws_client_results get that data passed to it? its defined as '' initially but I'm not seeing where it gets bigger | 16:13 |
| Shrews | clarkb: that is filled by the thread. that thread will not end until the entire file contents have been read. | 16:15 |
| clarkb | aha its in self.runWSClient | 16:15 |
| *** patriciadomin has joined #zuul | 16:15 | |
| clarkb | Shrews: right but isn't the entire file content determed by how long the job remains running? because if the job stops early it stops streaming | 16:16 |
| Shrews | clarkb: what job are we stopping early? | 16:16 |
| clarkb | Shrews: the one in the test, we stop it before we join the ws thread | 16:16 |
| clarkb | Shrews: which means depending on cpu contention and io speed I think you could potentially write say only 1024 bytes or 0 bytes (maybe somewhere in between) then stop the job and no more bytes will be written to the client | 16:17 |
| Shrews | clarkb: sorry, i do not see where we stop a job | 16:17 |
| Shrews | it's allowed to run to completion | 16:18 |
| clarkb | Shrews: yes, running to completion stops the job | 16:18 |
| clarkb | we don't forcefully stop it, we let it stop on its own | 16:18 |
| clarkb | and iirc once the job stops running no more bytes are written to the streamer | 16:18 |
| clarkb | so depending on whether or not we've read all the bytes at that point the last assert could fail | 16:19 |
| Shrews | clarkb: streamer reads from the ansible log file | 16:19 |
| mrhillsman | thx corvus | 16:21 |
| Shrews | fwiw, this isn't new code. it's an exact copy of the test below it, and it has never been flakey | 16:21 |
| clarkb | Shrews: through the logging stream handler so its something like ws_client -> ZuulWeb <- LoggingStreamHandler. The loggingstreamhandler goes away when the job stops | 16:22 |
| mrhillsman | grabbing some food then I think the rest of my day will be occupied with trying to get this damn executor to execute :) | 16:22 |
| clarkb | if zuulweb hasn't read all the bytes at athat point I think you fail | 16:22 |
| *** rbergero1 is now known as rbergeron | 16:22 | |
| *** rbergeron has quit IRC | 16:22 | |
| *** rbergeron has joined #zuul | 16:22 | |
| jlk | tobiash: I'll try to get a release out today | 16:24 |
| jlk | although Friday releases are generally bad. So maybe Monday | 16:25 |
| Shrews | clarkb: LogStreamingHandler does not go away when the job does. It goes away when it finishes reading the entire log file under the log file disappears (which is why we run a job that intentionally halts in the middle... so we can get the file contents | 16:29 |
| Shrews | s/under/until/ | 16:29 |
| clarkb | Shrews: reading more its the thing logstreamhandler reads from that goes away the actual finger server | 16:30 |
| clarkb | Shrews: because when the job stops running the finger server can't read it off of disk anymore as its been cleaned up aiui | 16:31 |
| clarkb | but maybe thats only if we forcefully stop the job which is why you were asking about that earlier? | 16:31 |
| clarkb | (and would explain my confusion) | 16:31 |
| Shrews | clarkb: right, which is why we run the playbook that halts for us | 16:31 |
| clarkb | got it | 16:31 |
| Shrews | clarkb: http://git.openstack.org/cgit/openstack-infra/zuul/tree/tests/fixtures/config/streamer/git/common-config/playbooks/python27.yaml | 16:32 |
| *** bhavik1 has joined #zuul | 16:37 | |
| tobiash | jlk: yes, monday is probably better | 16:41 |
| tobiash | corvus: lgtm | 16:41 |
| *** bhavik1 has quit IRC | 16:43 | |
| *** electrofelix has quit IRC | 16:50 | |
| *** Wei_Liu has quit IRC | 16:57 | |
| *** Wei_Liu1 has joined #zuul | 16:57 | |
| *** Wei_Liu1 is now known as Wei_Liu | 17:00 | |
| *** jpena is now known as jpena|off | 17:01 | |
| *** openstackgerrit has joined #zuul | 17:04 | |
| openstackgerrit | Merged openstack-infra/zuul master: Add branch to Zuul dashboard https://review.openstack.org/559313 | 17:04 |
| *** gouthamr has quit IRC | 17:17 | |
| openstackgerrit | Merged openstack-infra/zuul master: Allow some plugins to read from playbook dir https://review.openstack.org/559132 | 17:20 |
| *** gouthamr has joined #zuul | 18:13 | |
| *** gouthamr has quit IRC | 18:13 | |
| *** gouthamr has joined #zuul | 18:14 | |
| openstackgerrit | Merged openstack-infra/zuul master: Check out more appropriate branches of role and playbook repos https://review.openstack.org/558643 | 18:22 |
| *** logan- has quit IRC | 18:56 | |
| *** logan- has joined #zuul | 18:56 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul master: Increase unit testing of host / group vars https://review.openstack.org/559405 | 19:20 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul master: Inventory groups should be under children key https://review.openstack.org/559406 | 19:20 |
| pabelanger | mordred: ^noticed that while trying to add groups of groups for zuul inventory files | 19:22 |
| pabelanger | having a little issue getting configloader.py working for nested groups, but hopefully figure something out | 19:23 |
| *** maeca has joined #zuul | 19:42 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul master: Inventory groups should be under children key https://review.openstack.org/559406 | 19:59 |
| evrardjp | ohoh it sems I broke stuff | 20:56 |
| evrardjp | my bad. hard refresh worked. Have a nice week-end folks! | 20:57 |
| *** elyezer has quit IRC | 21:10 | |
| *** elyezer has joined #zuul | 21:16 | |
| *** gouthamr has quit IRC | 21:22 | |
| *** pwhalen has quit IRC | 21:28 | |
| *** pwhalen has joined #zuul | 21:28 | |
| *** pwhalen has joined #zuul | 21:28 | |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul master: Make label/status mirrors in reject for github https://review.openstack.org/559340 | 21:49 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul master: Make reject a mirror of require for github https://review.openstack.org/559340 | 21:51 |
| *** nhicher has quit IRC | 21:55 | |
| *** nhicher has joined #zuul | 21:55 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!