| -@gerrit:opendev.org- Benjamin Schanzel proposed: [zuul/nodepool] 839514: Default limits for k8s labels and quota support https://review.opendev.org/c/zuul/nodepool/+/839514 | 07:36 | |
| -@gerrit:opendev.org- Dominik Jaeger proposed: [zuul/zuul] 828125: Add support for filesets in job configuration https://review.opendev.org/c/zuul/zuul/+/828125 | 08:33 | |
| -@gerrit:opendev.org- Benjamin Schanzel proposed: [zuul/nodepool] 839514: Default limits for k8s labels and quota support https://review.opendev.org/c/zuul/nodepool/+/839514 | 09:35 | |
| -@gerrit:opendev.org- Dominik Jaeger proposed: [zuul/zuul] 828125: Add support for filesets in job configuration https://review.opendev.org/c/zuul/zuul/+/828125 | 10:32 | |
| @fungicide:matrix.org | > <@mnaser:matrix.org> is there a good way of getting out of a situation where a config-project is not passing tests? | 12:19 |
|---|---|---|
| in opendev, we have a gerrit administrator merge a fix bypassing zuul, but yes it's painful. that's why we adopted a testing policy for nontrivial changes to our base job: https://opendev.org/opendev/base-jobs/src/commit/d821d115975563fbb15f2c874c557303faeaec70/zuul.d/jobs.yaml#L5-L23 | ||
| -@gerrit:opendev.org- Dominik Jaeger proposed: [zuul/zuul] 828125: Add support for filesets in job configuration https://review.opendev.org/c/zuul/zuul/+/828125 | 12:55 | |
| -@gerrit:opendev.org- Joshua Watt proposed: [zuul/zuul] 838924: merger: Handle merges with cherry-pick merge-mode https://review.opendev.org/c/zuul/zuul/+/838924 | 13:13 | |
| -@gerrit:opendev.org- Joshua Watt proposed: [zuul/zuul] 838924: merger: Handle merges with cherry-pick merge-mode https://review.opendev.org/c/zuul/zuul/+/838924 | 13:21 | |
| @mhuin:matrix.org | Hello, do components require 2-way communication with the zookeeper service? What if we'd like to run an executor in a private network with no public IP - in the sense that zk, the scheduler etc can't directly connect to the executor | 13:51 |
| @jpew:matrix.org | I'd like to connect Zuul to MS Teams for reporting project build results; the easiest way I can see to do this is with the SMTP driver emailing the channel e-mail address. However, I'd like each project somehow be able to configure which e-mail address it sends results to. Is this possible somehow? | 13:53 |
| @avass:vassast.org | mhu: I think that should work, but you may need to set up a https://www.zuul-ci.org/docs/zuul/latest/configuration.html#finger-gateway to get live logs working | 13:53 |
| @avass:vassast.org | I may be wrong though :) | 13:54 |
| @avass:vassast.org | jpew: you could create a post/cleanup playbook that reports to MS Teams with the channel/e-mail being configurable with a variable by the project | 13:55 |
| @jpew:matrix.org | I was hoping to do it at the pipeline level. The SMTP driver is about perfect, except that I can't change the address per-project (that I know of) | 13:57 |
| @mhuin:matrix.org | Albin Vass: the finger gw is for streaming logs only? | 13:58 |
| @avass:vassast.org | mhu: I think so, yep | 13:58 |
| @mhuin:matrix.org | Albin Vass: the thing is, the components index in zk is incorrect, the executor node in the private network appears down | 14:00 |
| @mhuin:matrix.org | it's probably a matter of opening the right ports or setting up the right port forwarding, but I can't find out which one | 14:01 |
| @mhuin:matrix.org | it looks like zk is polling the nodes but I can't find doc on that | 14:03 |
| @avass:vassast.org | mhu: zookeeper needs to be accessible on the configured `secureClientPort` by zuul-executor at least | 14:09 |
| @mhuin:matrix.org | so there's no traffic from zk to the zuul-executor? | 14:10 |
| @avass:vassast.org | no zk doesn't open any connection towards the zuul-executor | 14:10 |
| @avass:vassast.org | so that should only need a one way port-forward to let zuul-executor connect to zk | 14:10 |
| @mhuin:matrix.org | okay, interesting ... then being in a private network shouldn't be a problem | 14:11 |
| @mhuin:matrix.org | okay turns out the problem was between the chair and the keyboard 🙂 thanks for your help | 14:52 |
| @fungicide:matrix.org | > <@jpew:matrix.org> I was hoping to do it at the pipeline level. The SMTP driver is about perfect, except that I can't change the address per-project (that I know of) | 15:06 |
| yes, the smtp reporter is per-pipeline. it would probably make more sense to make a specific reporter driver for ms teams, or i guess you could set up a fan-out gateway which processes the received smtp reports and then redelivers delivers them to more specific addresses based on other heuristics like project name (i know how i'd do that trivially in exim, but i have no idea what mta you're relying on) | ||
| @fungicide:matrix.org | > <@mhuin:matrix.org> so there's no traffic from zk to the zuul-executor? | 15:09 |
| for reference, while the diagram doesn't have directional indicators, the prose here is intended to describe which components are establishing connections to which other components, directionally: https://zuul-ci.org/docs/zuul/latest/components.html | ||
| @fungicide:matrix.org | the idea is thatyou should be able to put each of those components on different networks and have sufficient information from that document to be able to write firewall rules filtering the traffic between them | 15:15 |
| @clarkb:matrix.org | corvus: fungi any idea if that netaddr issue is a problem for ansible v5? I wonder if we should encourage them to add it as a dep to v5 if so. IPv6 isn't really optional functionality if IPv4 just works | 15:22 |
| @jim:acmegating.com | Clark: i do not know (i don't believe we exercised v5 during the time it was broken); one could probably run a test playbook in a venv to test it. | 15:24 |
| @jpew:matrix.org | fungi: Ah, I can run an intermediate exim server... would you happen to have an example on how to do the fan-out? | 16:14 |
| @fungicide:matrix.org | jpew: http://www.exim.org/exim-html-4.90/doc/html/spec_html/filter_ch-exim_filter_files.html | 16:24 |
| @fungicide:matrix.org | that's the way i'm most familiar with, but there are a few different ways you could go about it with exim specifically, or you could do something similar i'm sure with postfix, procmail... | 16:25 |
| @fungicide:matrix.org | * jpew: http://www.exim.org/exim-html-current/doc/html/spec_html/filter_ch-exim_filter_files.html | 16:27 |
| @clarkb:matrix.org | corvus: re the stack at https://review.opendev.org/c/zuul/zuul/+/839672 is that the sort of thing you'd prefer to land after v6 is tagged? Or should I go ahead and rereview and probably approve things nowish? | 18:16 |
| @clarkb:matrix.org | its the no_log handling in the callback for json stuff | 18:16 |
| @jim:acmegating.com | Clark: i think we can merge that now | 18:46 |
| @clarkb:matrix.org | corvus: cool I ended up +2'ing the stack but not +A'ing did you want to +A or shoudl I go ahead and do that? | 18:46 |
| @jim:acmegating.com | Clark: feel free to +A, i'm about to step away for a bit | 18:46 |
| @clarkb:matrix.org | done | 18:47 |
| @jim:acmegating.com | i think if there are still no fires by the end of the day, we can probably tag what opendev is running as 6.0, and that will just be in a later release; if there are fires, then that will be in 6.0 :) | 18:47 |
| @clarkb:matrix.org | sounds good to me | 18:47 |
| @clarkb:matrix.org | corvus: when you get back you may want ot look at https://review.opendev.org/c/zuul/zuul/+/839784 it does seem like quickstart's test2 has been forgotten? I'm not entirely sure what its purpose was/is | 19:00 |
| @jim:acmegating.com | Clark: it's there so users have other repos to experiment with. including cross-repo-deps. | 19:23 |
| @fungicide:matrix.org | i suppose one solution would be to add a depends-on exercise to the quickstart | 19:40 |
| @fungicide:matrix.org | then it's no longer unused | 19:40 |
| @fungicide:matrix.org | > <@jim:acmegating.com> i think if there are still no fires by the end of the day, we can probably tag what opendev is running as 6.0, and that will just be in a later release; if there are fires, then that will be in 6.0 :) | 19:50 |
| no fires yet at least! today's not a particularly heavy activity day for us, but there's still enough going on i expect we'd have spotted significant issues by now | ||
| @clarkb:matrix.org | https://review.opendev.org/c/zuul/zuul/+/840136 is another docs update that has me confused | 20:08 |
| @clarkb:matrix.org | fungi: I'm trying to determine if netaddr needs to be installed for ansible v5. Do you know what tasks were failing under 2.9 without netaddr? | 20:28 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 20:34 | |
| - [zuul/nodepool] 840241: Measure rate limits from start times https://review.opendev.org/c/zuul/nodepool/+/840241 | ||
| - [zuul/nodepool] 840242: Fix object refresh in the AWS adapter https://review.opendev.org/c/zuul/nodepool/+/840242 | ||
| @jim:acmegating.com | Clark: i think i linked to a build either here or in #_oftc_#opendev:matrix.org | 20:35 |
| @clarkb:matrix.org | ya grepping around I think it is the | ipv6 filter | 20:35 |
| @clarkb:matrix.org | the |ipv4 filter fails similarly and they both fail under ansible 5 | 20:36 |
| @jim:acmegating.com | Clark: https://matrix.to/#/!GXiijUJAGqDLBuZqwV:matrix.org/$Txm1-EY85nujDaMdOuM-dKqDB7rPwDATIlvy9SBdNM0?via=matrix.org&via=acmegating.com&via=inaugust.com | 20:36 |
| @clarkb:matrix.org | https://docs.ansible.com/ansible/latest/user_guide/playbooks_filters_ipaddr.html is the docs for that. I personally find it odd that they wouldn't depend on that for simple useful filters like this | 20:37 |
| @clarkb:matrix.org | Looks like 8MB difference between my venv without netaddr installed and with netaddr installed | 20:39 |
| @clarkb:matrix.org | https://github.com/ansible-collections/ansible.netcommon/blob/main/requirements.txt it does list its requirements so now it is extra weird to me that the combo package called ansible on pypi wouldn't do something similar | 20:47 |
| @jim:acmegating.com | Clark: maybe time to ask dmsimard ? | 20:48 |
| @jim:acmegating.com | (maybe not including the collections' requirements is an oversight?) | 20:48 |
| @jim:acmegating.com | * (maybe not including the superset of collections' requirements is an oversight?) | 20:49 |
| @clarkb:matrix.org | dmsimard: hey do you know why netaddr isn't installed when you `pip install 'ansible<6'`, but it is listed as a dependency of `ansible.netcommon` in https://github.com/ansible-collections/ansible.netcommon/blob/main/requirements.txt ? Are the various requirements of the subrepos collected together into `ansible` not also merged? does it make sense to maybe do that? | 20:49 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 20:57 | |
| - [zuul/zuul] 840243: Fix node request failures in dependency cycles https://review.opendev.org/c/zuul/zuul/+/840243 | ||
| - [zuul/zuul] 840244: Deduplicate jobs in dependency cycles https://review.opendev.org/c/zuul/zuul/+/840244 | ||
| -@gerrit:opendev.org- lotorev vitaly proposed wip: [zuul/nodepool] 833515: Log static node info on when failed to sync node https://review.opendev.org/c/zuul/nodepool/+/833515 | 21:22 | |
| -@gerrit:opendev.org- lotorev vitaly proposed wip: [zuul/nodepool] 833515: Log node info when failed to sync static node https://review.opendev.org/c/zuul/nodepool/+/833515 | 21:24 | |
| -@gerrit:opendev.org- lotorev vitaly proposed wip: [zuul/nodepool] 833515: Log node info when failing to sync static node https://review.opendev.org/c/zuul/nodepool/+/833515 | 21:28 | |
| -@gerrit:opendev.org- lotorev vitaly marked as active: [zuul/nodepool] 833515: Log node info when failing to sync static node https://review.opendev.org/c/zuul/nodepool/+/833515 | 21:34 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 840246: Add a spec for global semaphores https://review.opendev.org/c/zuul/zuul/+/840246 | 21:34 | |
| @clarkb:matrix.org | > <@clarkb:matrix.org> dmsimard: hey do you know why netaddr isn't installed when you `pip install 'ansible<6'`, but it is listed as a dependency of `ansible.netcommon` in https://github.com/ansible-collections/ansible.netcommon/blob/main/requirements.txt ? Are the various requirements of the subrepos collected together into `ansible` not also merged? does it make sense to maybe do that? | 22:16 |
| hrm the ipv4 and ipv6 filters may actually live in this collection: https://github.com/ansible-collections/ansible.utils/blob/main/requirements.txt but it lists netaddr too | ||
| @jim:acmegating.com | zuul-maint: how do these look for releases: | 22:34 |
| zuul: commit a89ce345c0fb5fb304713b0b0aa8097312be59f5 (HEAD -> master, tag: 6.0.0, origin/master, gerrit/master, refs/changes/55/840055/1) | ||
| nodepool: commit a2e5e640ad13b5bf3e7322eb3b62005484e21765 (HEAD -> master, tag: 6.0.0, origin/master, gerrit/master) | ||
| (if those look good, i'll cut them during my morning tomorrow) | ||
| @clarkb:matrix.org | corvus: did nodepool get restarted on that in opendev? If not do we think that is necessary? | 22:35 |
| @clarkb:matrix.org | corvus: the zuul commit lgtm though | 22:36 |
| @clarkb:matrix.org | corvus: looks like opendev launchers last restarted about 8 weeks ago so I don't think they are running that nodepool commit which is more recent | 22:37 |
| @jim:acmegating.com | Clark: oh, i thought they were getting restarted for the jammy work | 22:42 |
| @clarkb:matrix.org | corvus: only the builders | 22:42 |
| @jim:acmegating.com | mm hmm. | 22:42 |
| @clarkb:matrix.org | its possible that the builders are the only things that got updates too (the builders automatically restart but not the launchers) | 22:42 |
| @jim:acmegating.com | well, i guess i'll restart them now. | 22:42 |
| @clarkb:matrix.org | wfm | 22:42 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!