| openstackgerrit | James E. Blair proposed zuul/zuul master: WIP: Add ExecutorApi https://review.opendev.org/c/zuul/zuul/+/787832 | 00:18 |
|---|---|---|
| openstackgerrit | James E. Blair proposed zuul/zuul master: WIP change zone handling in ExecutorApi https://review.opendev.org/c/zuul/zuul/+/787833 | 00:18 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: WIP ExecutorAPI test rude scheduler build request deletions https://review.opendev.org/c/zuul/zuul/+/788622 | 00:18 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: WIP ExecutorApi add hold test https://review.opendev.org/c/zuul/zuul/+/788623 | 00:18 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: WIP ExecutorApi handle locking nonexistent requests https://review.opendev.org/c/zuul/zuul/+/788624 | 00:18 |
| *** hamalq has quit IRC | 01:34 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul master: getHoldRequest: pad zeros if not full length https://review.opendev.org/c/zuul/zuul/+/788668 | 01:50 |
| *** rlandy|rover|bbl is now known as rlandy|rover | 02:00 | |
| *** rlandy|rover has quit IRC | 02:05 | |
| *** evrardjp has quit IRC | 02:33 | |
| *** evrardjp has joined #zuul | 02:33 | |
| *** sam_wan has joined #zuul | 02:34 | |
| *** bhavikdbavishi has joined #zuul | 03:11 | |
| *** _erlon_ has quit IRC | 03:21 | |
| *** bhavikdbavishi1 has joined #zuul | 03:22 | |
| *** bhavikdbavishi has quit IRC | 03:24 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 03:24 | |
| *** jamesmcarthur_ has quit IRC | 04:04 | |
| *** saneax has joined #zuul | 04:04 | |
| *** jamesmcarthur has joined #zuul | 04:04 | |
| *** jamesmcarthur has quit IRC | 04:09 | |
| *** saneax has quit IRC | 04:16 | |
| *** vishalmanchanda has joined #zuul | 04:24 | |
| *** jamesmcarthur has joined #zuul | 04:27 | |
| *** bhavikdbavishi has quit IRC | 04:39 | |
| *** bhavikdbavishi has joined #zuul | 04:40 | |
| *** jfoufas1 has joined #zuul | 04:44 | |
| *** openstackgerrit has quit IRC | 05:01 | |
| *** jamesmcarthur has quit IRC | 05:12 | |
| *** jamesmcarthur has joined #zuul | 05:13 | |
| *** bhavikdbavishi has quit IRC | 05:14 | |
| *** jamesmcarthur has quit IRC | 05:17 | |
| *** sam_wan has quit IRC | 05:42 | |
| *** jamesmcarthur has joined #zuul | 05:43 | |
| *** jamesmcarthur has quit IRC | 05:55 | |
| *** saneax has joined #zuul | 06:06 | |
| *** jamesmcarthur has joined #zuul | 06:10 | |
| *** avass has quit IRC | 06:13 | |
| *** ajitha has joined #zuul | 06:21 | |
| *** jamesmcarthur has quit IRC | 06:28 | |
| *** bhavikdbavishi has joined #zuul | 06:35 | |
| *** jamesmcarthur has joined #zuul | 06:41 | |
| *** openstackgerrit has joined #zuul | 06:43 | |
| openstackgerrit | Benjamin Schanzel proposed zuul/zuul master: Add tenant name on NodeRequests for Nodepool https://review.opendev.org/c/zuul/zuul/+/788680 | 06:43 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: WIP: Optimize layout recalculations https://review.opendev.org/c/zuul/zuul/+/788681 | 06:45 |
| *** jamesmcarthur has quit IRC | 06:46 | |
| *** reiterative has quit IRC | 06:49 | |
| *** reiterative has joined #zuul | 06:49 | |
| *** avass has joined #zuul | 06:58 | |
| *** jamesmcarthur has joined #zuul | 07:00 | |
| *** hashar has joined #zuul | 07:04 | |
| *** jamesmcarthur has quit IRC | 07:05 | |
| *** nils has joined #zuul | 07:19 | |
| *** jpena|off is now known as jpena | 07:21 | |
| *** rpittau|afk is now known as rpittau | 07:27 | |
| *** jamesmcarthur has joined #zuul | 07:30 | |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: [wip] switch to using devstack's clouds.yaml https://review.opendev.org/c/zuul/nodepool/+/788684 | 07:37 |
| *** tosky has joined #zuul | 07:44 | |
| *** bhavikdbavishi has quit IRC | 07:51 | |
| *** bhavikdbavishi has joined #zuul | 07:53 | |
| *** jcapitao has joined #zuul | 08:00 | |
| *** bhavikdbavishi has quit IRC | 08:11 | |
| *** bhavikdbavishi has joined #zuul | 08:23 | |
| *** jamesmcarthur has quit IRC | 08:23 | |
| *** jamesmcarthur has joined #zuul | 08:39 | |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Fix broken jobs overriding checkout to an unprotected branch https://review.opendev.org/c/zuul/zuul/+/788694 | 08:43 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Fix broken jobs overriding checkout to an unprotected branch https://review.opendev.org/c/zuul/zuul/+/788694 | 08:47 |
| *** nhicher has quit IRC | 09:37 | |
| *** jpena is now known as jpena|off | 09:45 | |
| *** tobberydberg has quit IRC | 09:47 | |
| *** hashar has quit IRC | 09:48 | |
| *** tobberydberg has joined #zuul | 09:52 | |
| *** jpena|off has quit IRC | 09:54 | |
| *** fbo has joined #zuul | 10:09 | |
| *** nhicher has joined #zuul | 10:10 | |
| *** sam_wan has joined #zuul | 10:39 | |
| *** ajitha has quit IRC | 10:40 | |
| *** jpena has joined #zuul | 10:46 | |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: WIP: Optimize layout recalculations https://review.opendev.org/c/zuul/zuul/+/788681 | 10:46 |
| *** iurygregory has quit IRC | 10:54 | |
| *** iurygregory has joined #zuul | 10:58 | |
| *** jcapitao is now known as jcapitao_lunch | 11:10 | |
| *** jpena is now known as jpena|lunch | 11:31 | |
| *** rlandy has joined #zuul | 11:35 | |
| *** Shrews has joined #zuul | 11:46 | |
| *** ajitha has joined #zuul | 11:54 | |
| *** hashar has joined #zuul | 12:09 | |
| *** jcapitao_lunch is now known as jcapitao | 12:13 | |
| *** bhavikdbavishi has quit IRC | 12:18 | |
| *** bhavikdbavishi has joined #zuul | 12:19 | |
| *** bhavikdbavishi has quit IRC | 12:25 | |
| *** jpena|lunch is now known as jpena | 12:27 | |
| *** jamesmcarthur has quit IRC | 12:38 | |
| *** jamesmcarthur has joined #zuul | 12:38 | |
| *** bhavikdbavishi has joined #zuul | 12:42 | |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: Optimize layout re-calculation after re-enqueue https://review.opendev.org/c/zuul/zuul/+/788681 | 12:45 |
| *** bhavikdbavishi1 has joined #zuul | 12:47 | |
| *** bhavikdbavishi has quit IRC | 12:49 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 12:49 | |
| *** sam_wan has quit IRC | 13:16 | |
| *** nhicher has quit IRC | 13:27 | |
| *** fbo has quit IRC | 13:27 | |
| *** nhicher has joined #zuul | 13:28 | |
| *** ssmashnuk has joined #zuul | 13:30 | |
| *** vishalmanchanda has quit IRC | 13:30 | |
| *** bhavikdbavishi has quit IRC | 13:53 | |
| *** nhicher has quit IRC | 14:05 | |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: Optimize layout re-calculation after re-enqueue https://review.opendev.org/c/zuul/zuul/+/788681 | 14:08 |
| *** nhicher has joined #zuul | 14:09 | |
| *** rlandy is now known as rlandy|mtg | 14:21 | |
| *** hashar has quit IRC | 14:26 | |
| openstackgerrit | Ade Lee proposed zuul/zuul-jobs master: Add role to enable FIPS on a node https://review.opendev.org/c/zuul/zuul-jobs/+/788778 | 14:49 |
| *** rlandy|mtg is now known as rlandy | 14:50 | |
| *** Shrews is now known as Shrews_ | 14:51 | |
| *** Shrews_ is now known as Shrews | 14:51 | |
| openstackgerrit | Ade Lee proposed zuul/zuul-jobs master: Add role to enable FIPS on a node https://review.opendev.org/c/zuul/zuul-jobs/+/788778 | 14:59 |
| *** bhavikdbavishi has joined #zuul | 15:01 | |
| *** bhavikdbavishi1 has joined #zuul | 15:04 | |
| *** bhavikdbavishi has quit IRC | 15:05 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 15:05 | |
| *** jfoufas1 has quit IRC | 15:10 | |
| *** bhavikdbavishi has quit IRC | 15:36 | |
| *** jamesmcarthur has quit IRC | 16:07 | |
| *** jamesmcarthur has joined #zuul | 16:14 | |
| *** jpena|off has joined #zuul | 16:18 | |
| *** hamalq has joined #zuul | 16:22 | |
| *** hamalq has quit IRC | 16:23 | |
| *** hamalq has joined #zuul | 16:24 | |
| *** jpena has quit IRC | 16:25 | |
| *** jpena|off has quit IRC | 16:32 | |
| *** jamesmcarthur has quit IRC | 16:34 | |
| *** jamesmcarthur has joined #zuul | 16:36 | |
| *** bhavikdbavishi has joined #zuul | 16:39 | |
| *** jcapitao has quit IRC | 16:45 | |
| *** jamesmcarthur has quit IRC | 16:55 | |
| *** jamesmcarthur has joined #zuul | 16:56 | |
| *** iurygregory has quit IRC | 16:58 | |
| *** nils has quit IRC | 16:59 | |
| *** rpittau is now known as rpittau|afk | 16:59 | |
| *** rlandy is now known as rlandy|rover | 16:59 | |
| *** saneax has quit IRC | 17:03 | |
| *** iurygregory has joined #zuul | 17:31 | |
| *** iurygregory has quit IRC | 17:31 | |
| *** josefwells has joined #zuul | 17:37 | |
| *** iurygregory has joined #zuul | 17:38 | |
| *** bhavikdbavishi has quit IRC | 17:53 | |
| *** bhavikdbavishi has joined #zuul | 17:54 | |
| *** rlandy|rover is now known as rlandy|mtg | 18:00 | |
| *** bhavikdbavishi has quit IRC | 18:00 | |
| *** bhavikdbavishi has joined #zuul | 18:00 | |
| *** ssmashnuk has quit IRC | 18:01 | |
| openstackgerrit | Merged zuul/zuul master: Switch to ZooKeeper backed build result events https://review.opendev.org/c/zuul/zuul/+/782939 | 18:06 |
| *** bhavikdbavishi has quit IRC | 18:32 | |
| *** nils has joined #zuul | 18:39 | |
| josefwells | hello zuulers. I'm a bit confused about the working directory for stuff. I've scanned the old IRC logs but that has left me more confused. I see "prepare_workspace" role dumps stuff in "zuul_workspace_root", but "prepare_workspace_git" and pretty much everything else, drops data in "ansible_user_dir". The "ansible_user_dir" seems like a | 18:51 |
| josefwells | built in ansible thing (but I can't confirm that looking at ansible docs). | 18:51 |
| clarkb | josefwells: yes ansible_user_dir is the homedir for the user that ansible used to ssh | 18:52 |
| clarkb | I think zuul_workspace_root may be executor side | 18:52 |
| josefwells | oh, I didn't catch that distinction. Is it possible to override "ansible_user_dir"? | 18:53 |
| clarkb | josefwells: you could change the user's homedir or change the user. I believe it also arrives as a fact and you may be able to override facts (but I would need to read ansible var precedence docs to know) | 18:58 |
| *** rlandy|mtg is now known as rlandy|rover | 19:01 | |
| mordred | yah - but ... I wouldn't do so lightly - many of the jobs in zuul-jobs make some assumptions about src dir locations | 19:06 |
| mordred | look for instances of zuul_work_dir and zuul.project.src_dir - as well as the src_dir attribute in each item of the zuul.projects[] list | 19:06 |
| *** hashar has joined #zuul | 19:13 | |
| *** ssmashnuk has joined #zuul | 19:22 | |
| josefwells | clarkb: changing the homedir sounds like the best solution.. I can build that into the zuul container config so that would at least all be together (vs. symlinks in the users actual home-dir) | 19:25 |
| *** jamesmcarthur has quit IRC | 19:26 | |
| josefwells | Also need to figure out how to have a separate workdir-per-job, since I'm not able to containerize all the work and instead am using static nodes with a big ugly NFS disk | 19:26 |
| ssmashnuk | First off, thanks for everyone's help getting WinRM working. I took the advice and abandoned WinRM for SSH and I have connectivity to my windows nodes now. Build key setup is working, but I'm now running into a issues with 'prepare-workspace'. How far should I expect to be getting with the out of the box Zuul modules on Windows nodes? Not | 19:30 |
| ssmashnuk | sure if the current failures are my lack of experience or if there is extremely limited Windows node support? | 19:30 |
| *** ikhan has quit IRC | 19:31 | |
| clarkb | ssmashnuk: prepare-workspace requires rsync (it uses the synchronize module). Not sure if that is expected to work with windows or not | 19:33 |
| avass | ssmashnuk: I think there's a synchronize-repos for windows but it may not be merged. let me check | 19:35 |
| ssmashnuk | clarkb: Thanks. The other issue was zuul_console. It appears to try and run the python code for this via powershell. How important is the zuul_console step? | 19:37 |
| avass | smyers: I think we're using this pretty much as is: https://review.opendev.org/c/zuul/zuul-jobs/+/740005 | 19:37 |
| smyers | ssmashnuk ^ :D | 19:37 |
| avass | ssmashnuk: zuul_console is not yet supported on windows :) | 19:37 |
| avass | samccann: oh we probably added this: https://review.opendev.org/c/zuul/zuul-jobs/+/747640/2 I can check tomorrow if there's anything else we've changed with that role | 19:39 |
| ssmashnuk | Thanks everyone. Lots of good leads here. | 19:39 |
| avass | clarkb: I don't know if there's an rsync for windows. but git with ssh works :) | 19:40 |
| *** ssmashnuk has quit IRC | 19:42 | |
| corvus | josefwells: if you have multiple jobs running on a static node, i would recommend setting up multiple users and registering each as a distinct static node rather than using the same user and trying to have different workspaces; as mordred mentioned, there's a lot that assumes ansible_user_dir in zuul-jobs. | 19:47 |
| avass | Interesting comment from daniel: https://review.opendev.org/c/zuul/zuul-jobs/+/780662/16/roles/add-build-sshkey/tasks/create-key-and-replace.yaml#31 | 19:52 |
| clarkb | avass: that likely does affect linux nodes. I suspect the controlpersist setting is related to that | 19:53 |
| avass | since linux jobs never reset the connection there is that ping still using the inital session? | 19:53 |
| avass | clarkb: so there should probably be a meta: reset_connection before the ping tasks in that case | 19:54 |
| corvus | avass: likely so and i think your solution is correct | 19:54 |
| corvus | it is entirely possible that was written before CP was working correctly :) | 19:55 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Reset connection before testing build ssh-keys https://review.opendev.org/c/zuul/zuul-jobs/+/788826 | 19:56 |
| avass | samccann: you might also want this if you're using static nodes: https://review.opendev.org/c/zuul/zuul-jobs/+/780662 | 20:02 |
| corvus | avass: i think you mean ssmashnuk, who left a bit ago | 20:03 |
| avass | oops | 20:03 |
| *** samccann has left #zuul | 20:05 | |
| josefwells | corvus:I can't easily make new users on these static nodes. But this is all within the zuul-jobs stuff right? If I build my own jobs/playbooks/roles I should be ok? Just manual and have to keep an eye on zuul-dev to make sure you don't change anything that requires a corresponding change to the jobs? | 20:24 |
| avass | josefwells: if you write your own playbooks to set that up it should be fine. we have similar problems with static windows nodes | 20:27 |
| corvus | josefwells: yep, as avass said. hopefully the prepare-workspace roles can cope with an alternate work dir? | 20:28 |
| corvus | yeah, there's a zuul_workspace_root variable that prepare-workspace uses | 20:30 |
| corvus | i think the issue is that nothing else in zuul-jobs uses that | 20:30 |
| avass | looks like preparere-workspace-git uses "{{ ansible_user_dir }}/{{ zuul.project.src_dir }}" everywhere | 20:30 |
| avass | corvus: maybe they should, and default it to that ^ ? | 20:31 |
| *** jamesmcarthur has joined #zuul | 20:31 | |
| corvus | avass: probably so... guillaumec: that might be a better approach to take in https://review.opendev.org/735980 ? | 20:32 |
| avass | or rather default zuul_workspace_root to ansible_user_dir | 20:32 |
| corvus | avass: right that. definitely not the second thing (that's zuul_work_dir) | 20:32 |
| avass | corvus: I don't like that it changes to zuul.build by default since the workdir can be used as a git cache | 20:33 |
| corvus | right. default should be current behavior, but if zuul-jobs used zuul_workspace_root globally, then a user could set zuul_workspace_root to {{zuul.build}} | 20:34 |
| corvus | or {{ansible_user_dir}}/{{zuul.build}}. or /tmp/{{zuul.build}} or whatever :) | 20:34 |
| josefwells | Coming from jenkins, so forgive my thinking, but you get a working directory per-build, so I'm trying to emulate that since I need a workspace per-project-per-prospective-build, but I don't grok the zuul variables to desribe that | 20:35 |
| corvus | josefwells: we all came from jenkins, no apology necessary :) | 20:35 |
| josefwells | just most of our workloads are not yet container-able, and for now I have to use LSF to get to any real compute. Enough of my sad story | 20:36 |
| corvus | josefwells: to get highly technical here: zuul itself doesn't care; and it does in fact create a workspace per build on its own executors. it's just that the "standard library" roles in zuul-jobs mostly assume that they own the remote node they will be run on. | 20:36 |
| corvus | josefwells: so that's why you won't have any probs if you don't use stuff from zuul-jobs. but we really want you to at least use the git prep roles (the ones that copy the git repos from the zuul executor to the remote node) because those are a pain to get right. | 20:37 |
| josefwells | ah, hence the prepare_workspace does an rsync of an already ready-to-go filesystem | 20:37 |
| josefwells | I noticed that the git jobs were quite detailed, which is why I started looking again at some of my assumptions about variables | 20:38 |
| corvus | josefwells: yep. and it looks like you can use prepare-workspace today, but it's kind of the old-and-busted way of doing things; prepare-workspace-git is nicer and can take advantage of long-lived caches on the remote node to speed things up. but it looks like it may not be ready for that use case. | 20:38 |
| josefwells | corvus: yes, and I even have (readonly) git caches on my static nodes | 20:39 |
| josefwells | corvus: but if I could just have zuul create its own internal-workspace in NFS space, I could skip the middle-man and get down to work | 20:39 |
| avass | josefwells, corvus: it could also work if nodepool had a way to keep track of which "id" or something it gave zuul when using max-parallel-jobs | 20:42 |
| josefwells | corvus: if I add a symlink in the ansible_user_dir I guess I'm still left with only one workspace | 20:42 |
| avass | so if you set max-parallel-job: 4 you also get a workspace from 1 to 4. that could be used to work the same way jenkins does | 20:43 |
| corvus | avass: i don't see how without another user on the host | 20:43 |
| corvus | avass: it's true that nodepool could supply, say, an integer, but that's no different than zuul.uuid; there's nothing magic about it like there is for a user's homedir | 20:44 |
| corvus | sorry zuul.build | 20:44 |
| avass | corvus: could nodepool be made to split a node into multiple nodes internally when given a max-parallel-jobs? | 20:44 |
| corvus | avass: i don't get it, why? | 20:45 |
| avass | so it can keep track of which does it has delegated to zuul or not and set a nodepool.<attribute> accordingly that can be consumed by the job | 20:45 |
| corvus | avass: but how is that different than just using zuul.build? | 20:46 |
| avass | corvus: it wouldn't be unique | 20:46 |
| corvus | or any other unique thing? | 20:46 |
| corvus | you don't want it to be unique? | 20:46 |
| avass | i mean unique as in unique per worker, not unique each time the node is delegated | 20:46 |
| corvus | so you want the exact behavior you would get by having multiple users but without actually issuing an "adduser" command? | 20:47 |
| avass | yeah | 20:47 |
| avass | if for some reason it's not possible to do that | 20:47 |
| corvus | i'm sure nodepool could be modified to accomplish that, but it's not set up for that today | 20:48 |
| corvus | today nodepool identifies unique nodes via (hostname, username, port) | 20:49 |
| corvus | so if you can vary any one of those, you get a new node in nodepool | 20:49 |
| josefwells | corvus: port? like ssh to something besides 22, like unprivileged ports and it looks like a new node? I can do that. | 20:50 |
| corvus | if it's in your power to set hostnames, you could make a bunch of hostnames like "worker1".."workerN" all with the same ip, register each as a node in nodepool but all with the same username and you'd get that effect. | 20:50 |
| corvus | josefwells: or port, yes. | 20:50 |
| avass | it's something we've wanted to take a look at but haven't had time to do so we've sort of accepted slow checkouts to unique workspaces instead. | 20:51 |
| josefwells | corvus: I can request aliases, that is crazy but doable. Our repository is not small.. so we can't really live with slow there | 20:51 |
| avass | and that we will eventually get rid of the need to use max-parallel-jobs | 20:52 |
| josefwells | corvus: even if we have separate nodes (via port/hostname magic) how does that give us separate work-dirs? | 20:53 |
| avass | josefwells: you could map the host name to a specific workdir | 20:53 |
| corvus | josefwells: it doesn't; it only solves avass's problem of wanting some kind of persistent id to use in a separet workdir | 20:53 |
| corvus | so you'd still have the "nothing uses zuul_workspace_root except prepare-workspace" problem, but for prepare-workspace and your own playbooks, you could tell it to use the nodepool hostname as a semi-permanent/semi-unique workspace root. | 20:54 |
| *** jamesmcarthur has quit IRC | 21:14 | |
| *** Shrews has quit IRC | 21:20 | |
| *** jamesmcarthur has joined #zuul | 21:22 | |
| *** sduthil has quit IRC | 21:24 | |
| josefwells | corvus: seems like I can put a symlink on the zuul-executor for /var/lib/zuul/builds -> /NFS/wherever | 21:26 |
| josefwells | "additional level of indirection..." | 21:26 |
| josefwells | that will not work when we move this to k8s using your fancy new operator, since our k8s is not in NFS space. | 21:33 |
| fungi | the cluster host can't expose nfs mounts into the containers, i guess? | 21:33 |
| josefwells | not without lots of work, or using other protocols. mostly I try to keep our container world from mounting anything and get data in more modern ways (artifactory, git, etc) | 21:37 |
| *** ajitha has quit IRC | 21:40 | |
| fungi | 'heh, we even map afs mounts into containers | 21:44 |
| *** nils has quit IRC | 21:48 | |
| openstackgerrit | Matthieu Huin proposed zuul/zuul-client master: Add output formatters when relevant https://review.opendev.org/c/zuul/zuul-client/+/788832 | 21:53 |
| *** hashar has quit IRC | 22:06 | |
| josefwells | Ok, unrelated to my previous queries about workdirs, I'm running a modified-for-github docker quickstart, and my web instance and on my Build pages, I get an error about not being able to load pages from the servier:8000 (log server), but I can load those separately. Any idea what could be going wrong there? | 22:07 |
| clarkb | CORS maybe? | 22:09 |
| fungi | the web instance is serving javascript which asks your browser to load the manifest from the log archive location. could the domains be different and your browser is unhappy? | 22:09 |
| clarkb | I think those errors will show up in the dev console | 22:09 |
| josefwells | yeah, I thought cross-site scripting but it is all the same host and the error is a zuul error, not a browser error | 22:10 |
| josefwells | I don't see errors in the web or log docker logs | 22:11 |
| clarkb | is it a 500 or 404 or something else? I'd probably styart with identifying the type of error then use that to determine where to look next | 22:11 |
| fungi | zuul will embed an error message in the page if your browser fails to load the manifest | 22:11 |
| josefwells | ok, slight lie, but I think unrelated : "GET /1/1/0ba6a23db9f688bb383259ba52a5232ccce5155c/check/q6job/11b5e1c/job-output.json.gz HTTP/1.1" 404 196 | 22:11 |
| josefwells | th rest of the logs show 200 or 206 response | 22:12 |
| clarkb | for a 404 I would go check the docroot and see what is present | 22:12 |
| clarkb | since it is basically saying that path isn't valid | 22:12 |
| josefwells | yeah, it sure looks like cross-site: Access to XMLHttpRequest at 'http://docker-host:8000/1/1/0ba6a23db9f688bb383259ba52a5232ccce5155c/check/q6job/11b5e1c/zuul-manifest.json' from origin 'http://docker-host:9000' has been blocked by CORS policy: No 'Access-Control-Allow-Origin' header is present on the requested resource. | 22:14 |
| josefwells | Which is odd, because I see a lot of 'Access-Control-Allow-Origin' in the source-code of zuul/web | 22:16 |
| clarkb | The log server isn't zuul-web though I don't think | 22:17 |
| clarkb | ya the logs container is apache | 22:18 |
| corvus | we might need to update the quickstart apache config? | 22:20 |
| clarkb | Header set Access-Control-Allow-Origin "*" is set in the httpd.confg | 22:20 |
| clarkb | is it possible josefwells has an old image build without that? | 22:21 |
| corvus | true, but that only merged in october... josefwells any chance your quickstart httpd config is from before then? | 22:21 |
| corvus | yeah that :) | 22:21 |
| corvus | josefwells: https://review.opendev.org/755864 is the change if you want to check for that | 22:22 |
| josefwells | OMG totally possible, sorry to be a pain, but moving gerrit->github in the quickstart was quite an operation | 22:23 |
| josefwells | so I am not up on my pulls | 22:24 |
| clarkb | josefwells: I think this is a lcoal build fwiw | 22:25 |
| corvus | yeah, so apply that change and 'docker-compose build' i think? | 22:27 |
| josefwells | clarkb corvus: yep, all works now. again, thanks and sorry for being behind | 22:33 |
| josefwells | clarkb corvus: I'll see if I can clean up my changes so you can have a github-quickstart | 22:34 |
| corvus | josefwells: thanks -- note though that we deliberately chose to focus on just one system for the quickstart so that people could most easily and reliably get a zuul up and running. as you noted, it's quite complicated to use github, and it's not something we can automate. i don't want to be discouraging, but i think it's not straightforward and i don't want you to do a bunch of work that we don't end up | 22:43 |
| corvus | merging. if we were to do something like that, i could maybe see doing it like this -- there are some patches that turn the quick-start into a series of tutorials. i'd very much like to merge them, and then maybe we could do an add-on tutorial for adding a github connection? | 22:43 |
| corvus | forgot to link to tutorial patch series: https://review.opendev.org/732067 | 22:50 |
| josefwells | That would make sense, then just an alternate docker-compose when you reach step X | 22:50 |
| mhu | Is setting the list of possible builds/buildsets results in constants still desirable? | 23:18 |
| mhu | see https://review.opendev.org/c/zuul/zuul/+/738632 | 23:18 |
| ianw | tobiash: could i get you to loop back on https://review.opendev.org/c/zuul/nodepool/+/785821 which stops a pretty consistent error we're seeing on nl03 | 23:35 |
| ianw | note there's now a refactor change underneath it from your comments just to keep things simple | 23:36 |
| *** tosky has quit IRC | 23:40 | |
| ianw | zuul-maint: this is a small one that fixes not trying to copy logs if docker fails like it did the other day : https://review.opendev.org/c/zuul/zuul-jobs/+/787019 | 23:42 |
| openstackgerrit | Matthieu Huin proposed zuul/zuul-client master: Add search filters when relevant https://review.opendev.org/c/zuul/zuul-client/+/788847 | 23:52 |
| mhu | zuul-maint: we can search builds through zuul-client, let's add buildsets, and single build/buildset info to the mix! https://review.opendev.org/q/topic:%2522builds_query%2522+(status:open) :) | 23:57 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!