| *** Goneri has quit IRC | 00:09 | |
| *** rlandy has quit IRC | 00:42 | |
| *** Goneri has joined #zuul | 01:01 | |
| *** vblando has quit IRC | 01:11 | |
| *** ChrisShort has quit IRC | 01:13 | |
| *** samccann has quit IRC | 01:13 | |
| *** swest has quit IRC | 01:14 | |
| *** vblando has joined #zuul | 01:19 | |
| *** mnasiadka has quit IRC | 01:19 | |
| *** ChrisShort has joined #zuul | 01:22 | |
| *** samccann has joined #zuul | 01:22 | |
| *** mnasiadka has joined #zuul | 01:22 | |
| *** swest has joined #zuul | 01:28 | |
| *** Goneri has quit IRC | 02:55 | |
| *** bhavikdbavishi has joined #zuul | 03:35 | |
| *** evrardjp has quit IRC | 04:33 | |
| *** evrardjp has joined #zuul | 04:33 | |
| *** wuchunyang has joined #zuul | 04:36 | |
| *** decimuscorvinus has quit IRC | 04:38 | |
| *** bhavikdbavishi has quit IRC | 04:47 | |
| *** bhavikdbavishi has joined #zuul | 04:48 | |
| *** decimuscorvinus has joined #zuul | 04:48 | |
| *** wuchunyang has quit IRC | 05:15 | |
| *** sgw has quit IRC | 05:44 | |
| *** bhavikdbavishi1 has joined #zuul | 06:01 | |
| *** dustinc has quit IRC | 06:02 | |
| *** bhavikdbavishi has quit IRC | 06:03 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 06:03 | |
| *** bhavikdbavishi has quit IRC | 06:46 | |
| *** yolanda has quit IRC | 07:31 | |
| *** bhavikdbavishi has joined #zuul | 07:36 | |
| *** threestrands has quit IRC | 07:41 | |
| *** hashar has joined #zuul | 07:56 | |
| *** iurygregory has joined #zuul | 07:59 | |
| *** iurygregory has quit IRC | 08:02 | |
| *** tosky has joined #zuul | 08:04 | |
| *** bhavikdbavishi has quit IRC | 08:24 | |
| *** bhavikdbavishi has joined #zuul | 08:33 | |
| *** ianychoi__ has quit IRC | 08:38 | |
| *** bhavikdbavishi has quit IRC | 08:48 | |
| *** harrymichal has joined #zuul | 08:57 | |
| *** leoluk has joined #zuul | 09:04 | |
| *** bhavikdbavishi has joined #zuul | 09:34 | |
| *** donnyd has joined #zuul | 09:37 | |
| openstackgerrit | Albin Vass proposed zuul/nodepool master: WIP: aws: add support for uploading diskimages https://review.opendev.org/735217 | 10:14 |
|---|---|---|
| openstackgerrit | Albin Vass proposed zuul/nodepool master: aws: add support for attaching instance profiles https://review.opendev.org/734774 | 10:23 |
| *** wuchunyang has joined #zuul | 10:54 | |
| *** wuchunyang has quit IRC | 10:58 | |
| *** bhavikdbavishi has quit IRC | 11:12 | |
| *** wuchunyang has joined #zuul | 11:26 | |
| *** hashar has quit IRC | 11:45 | |
| *** bhavikdbavishi has joined #zuul | 11:47 | |
| *** rfolco|rover has joined #zuul | 11:51 | |
| *** bhavikdbavishi has quit IRC | 11:53 | |
| *** wuchunyang has quit IRC | 12:01 | |
| *** wuchunyang has joined #zuul | 12:33 | |
| openstackgerrit | Albin Vass proposed zuul/nodepool master: WIP: aws: add support for uploading diskimages https://review.opendev.org/735217 | 12:59 |
| *** wuchunyang has quit IRC | 13:02 | |
| *** bhavikdbavishi has joined #zuul | 13:08 | |
| *** rfolco|rover has quit IRC | 13:11 | |
| *** bhavikdbavishi has quit IRC | 13:36 | |
| avass | corvus: thanks for catching the arn/name exlusivity for instance profiles yesterday, the change should be ready now: https://review.opendev.org/#/c/734774/ :) | 13:57 |
| avass | corvus: tested it by launching an instance with boto3 against a vpc I set up :) | 13:57 |
| openstackgerrit | Guillaume Chauvel proposed zuul/zuul-jobs master: ensure-tox: Avoid Reinstalling tox venv if present https://review.opendev.org/735339 | 13:58 |
| openstackgerrit | Guillaume Chauvel proposed zuul/zuul-jobs master: ensure-twine: Avoid Reinstalling twine if present https://review.opendev.org/735340 | 13:58 |
| avass | corvus: uh, I guess I could build nodepool from that and test it as well to be extra sure that everything works :) | 14:00 |
| corvus | avass: if you decide to do that, what i like to do is just set up a single launcher with a min-ready of 1 for a label, then watch it create a server. if i need to iterate, i just 'nodepool delete' that server and let it launch again. then min-ready:0 and delete a final time to clean up. | 14:05 |
| corvus | avass: also, yay, code review works :) it's a bummer about moto not having that yet. | 14:05 |
| avass | corvus: I also created an issue to hopefully get someone working on it: https://github.com/spulec/moto/pull/3064#issuecomment-642423395 but I'll have to update that test case :) | 14:07 |
| avass | corvus: yeah, that sounds like a good idea, I'm planning on testing nodepool-builder with aws later anyway, so I'll get back to you when I've tested it :) | 14:08 |
| avass | going to the gym for now, I'll check in later | 14:09 |
| *** sgw has joined #zuul | 14:38 | |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Add test-build-container-image-release job https://review.opendev.org/735249 | 14:44 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Add test-build-container-image-release job https://review.opendev.org/735249 | 15:41 |
| *** harrymichal has quit IRC | 16:01 | |
| *** harrymichal has joined #zuul | 16:18 | |
| *** bhavikdbavishi has joined #zuul | 16:51 | |
| *** irclogbot_2 has quit IRC | 18:29 | |
| *** irclogbot_0 has joined #zuul | 18:34 | |
| *** rfolco|rover has joined #zuul | 18:39 | |
| *** ysandeep|away is now known as ysandeep | 18:39 | |
| openstackgerrit | Sagi Shnaidman proposed zuul/zuul-jobs master: Add jobs for testing ensure-ansible https://review.opendev.org/734584 | 18:39 |
| *** rfolco|rover has quit IRC | 18:49 | |
| *** ysandeep is now known as ysandeep|away | 19:06 | |
| avass | corvus: 734774 works :) | 19:16 |
| openstackgerrit | Matthieu Huin proposed zuul/zuul master: web UI: user login with OpenID Connect https://review.opendev.org/734082 | 19:19 |
| avass | mordred: 734774 is working, I set up nodepool-launcher from that change as launched a node to make sure it starts with the instance profile | 19:20 |
| corvus | mordred, avass, clarkb, fungi: could we land https://review.opendev.org/735249 soonish so i can re-run the tag jobs to try to get tagged docker images? | 19:55 |
| fungi | sure, looking now | 19:55 |
| clarkb | corvus: does change_url exist if change doesn't? looking at https://review.opendev.org/#/c/735249/4/roles/build-docker-image/tasks/build.yaml | 19:56 |
| corvus | clarkb: yes it does. it's poorly named. | 19:56 |
| fungi | there are some alternative values stored in it depending on what the trigger provides | 19:56 |
| corvus | points to the tag in gitea in our case | 19:56 |
| *** bhavikdbavishi has quit IRC | 20:02 | |
| avass | corvus: looking | 20:03 |
| fungi | avass: i had approved it, but i've undone my workflow +1 just now, feel free to approve it yourself if you're comfortable with it | 20:05 |
| avass | corvus: just a small thing, it looks like it supports both docker and the container roles (podman) but there's only a test job for for docker | 20:14 |
| avass | corvus: I'm guessing it's to make it easier to re-user it later | 20:14 |
| corvus | avass: yes -- most of that is copied from existing test jobs. we only need the docker support to get our zuul images published. i want to prioritize that, but it would be more effort to remove the podman stuff from the tests, so i left it in for future expansion. | 20:16 |
| avass | alright, I think it looks alright then, just trying to understand why it looks like that :) | 20:17 |
| avass | but that's something that could be a bit confusing later, since it looks like there should be something that is testing it with the 'container' roles but there isn't | 20:19 |
| openstackgerrit | Merged zuul/zuul-jobs master: Add test-build-container-image-release job https://review.opendev.org/735249 | 20:28 |
| -openstackstatus- NOTICE: The Gerrit service on review.opendev.org is going offline momentarily at 21:00 UTC for project rename maintenance, but should return within a few minutes: http://lists.opendev.org/pipermail/service-announce/2020-June/000004.html | 20:58 | |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Fix upload-docker-image when in release pipeline https://review.opendev.org/735393 | 20:58 |
| *** paladox has quit IRC | 21:28 | |
| *** paladox has joined #zuul | 21:33 | |
| corvus | mordred: can you +3 https://review.opendev.org/735393 when you have a sec? | 21:43 |
| mordred | corvus: I can! | 21:44 |
| mordred | corvus: done. also - exciting | 21:44 |
| *** y2kenny has joined #zuul | 21:44 | |
| corvus | mordred: yeah, we got really close to uploading :) | 21:44 |
| mordred | corvus: I think close is good enough | 21:45 |
| mordred | corvus: it's the thought that counts, right? | 21:45 |
| mordred | corvus: I almost landed a patch earlier today in cliff - but didnt' because a library it subclasses has a new version that has a docstring on a method in a class that sphinx in cliff finds and complains about | 21:46 |
| mordred | corvus: I *definitely* think it's awesome that sphinx on cliff complains about the content in a parent class that isn't in the source tree, don't you? | 21:46 |
| clarkb | nova's gate is broken beacuse xetex can't emit a pdf doc file due to an input file being too large? | 21:46 |
| mordred | clarkb: that doesn't seem awesome. | 21:46 |
| corvus | clarkb: right channel? | 21:47 |
| avass | corvus, mordred: that doesn't look too hard to set up tests for if we just add a flag to use a self hosted repository to login to | 21:47 |
| mordred | clarkb: where is that being discussed? | 21:47 |
| corvus | avass: yeah, we could probably do that; we do similar for the registry roles | 21:47 |
| clarkb | corvus: I just wanted to commiserate the sphinx problems | 21:47 |
| corvus | clarkb: oh, well, i'm confused because that problem has been solved for nova before | 21:48 |
| y2kenny | For nodepool, if I see CleanupWorker stuck looping between "Cleaning up held nodes" and "Cleaning up empty nodes", what might that indicate? (I am pretty sure it's my custom driver causing problem but I am not sure where to start to debug this.) | 21:48 |
| corvus | y2kenny: that might be normal | 21:49 |
| corvus | y2kenny: it's a daemon thread that wakes up every minute or so and does that loop | 21:49 |
| y2kenny | corvus: um... interesting... ok then my test must be timing out for other reasons. This is good to know. Thanks :) | 21:50 |
| y2kenny | corvus: actually... what if those message comes up like 4 times a seconds... that's too much right? | 21:53 |
| corvus | y2kenny: in a unit test? maybe not, we reduce the delay in that case to make the tests run faster | 21:55 |
| y2kenny | corvus: oh ok... yes it is a unit test | 21:55 |
| openstackgerrit | Merged zuul/zuul-jobs master: Fix upload-docker-image when in release pipeline https://review.opendev.org/735393 | 21:56 |
| -openstackstatus- NOTICE: gerrit is being taken offline for emergency cleanup, will return to service again shortly | 22:00 | |
| y2kenny | corvus: one more question... is the idea of node for nodepool supposed to exist before the associated resources is created? Is that the reason why the static driver is tricky? there is a coupling between the node and the server? | 22:05 |
| corvus | y2kenny: exactly | 22:06 |
| y2kenny | I have been comparing the k8s driver and the static driver. The creation of node to be stored in zk seems to be the main difference. For k8s driver, the node is always created at launch | 22:06 |
| corvus | y2kenny: we create a node when we decide to create a remote resource, then we fill in the remote info when it exists | 22:06 |
| y2kenny | but for static, the node is created when the static server is registered | 22:06 |
| y2kenny | ok... so what nodepool does in terms of scheduling is really just about the label... | 22:07 |
| y2kenny | and for this label I can pick from this pool | 22:08 |
| y2kenny | but then it's not clear to me what listNode really does... | 22:08 |
| y2kenny | so in general there's really no constraint to nodepool scheduling? | 22:09 |
| y2kenny | I feel like I am missing something | 22:09 |
| y2kenny | corvus: is min-ready/max-ready the only constraint that nodepool care about? | 22:11 |
| clarkb | y2kenny: some of the drivers look at quotas too | 22:12 |
| clarkb | y2kenny: openstack does for sure and it will stop booting things if the cloud reports it will be oever quota | 22:13 |
| y2kenny | clarkb: how does quota information get communicated to nodepool? | 22:13 |
| y2kenny | clarkb: what I am really confuse about is how listNode doesn't take in pool as argument | 22:14 |
| SpamapS | AWS has some things to prevent going over limits. | 22:14 |
| clarkb | it periodically asks the cloud for that information, and uses its local records to estimate values from there | 22:14 |
| clarkb | it also does a forced recalculation if the cloud reports ndoepool is over the limit but nodepool thinks it isn't | 22:14 |
| y2kenny | clarkb: is the report done via listNodes or another interface? | 22:15 |
| clarkb | its via the give me my usage api ( I forget the exact name ) | 22:15 |
| y2kenny | hasRemainingQuota? | 22:16 |
| y2kenny | hasProviderQuota? | 22:16 |
| SpamapS | (just max-servers though) | 22:17 |
| clarkb | ya hasRemainingQuota is meant to provide a true false based on what nodepool knows | 22:17 |
| y2kenny | and node_types == labels? | 22:17 |
| clarkb | and there is a quota cache invalidation that happens if we get an exception due to hitting quota from the cloud | 22:18 |
| y2kenny | node_types/ntype | 22:18 |
| clarkb | yes I think so | 22:20 |
| y2kenny | clarkb: can you describe the usage for listNodes()? I read the devguide description but I don't quite understand its use. Does it just list all resources available by a provider across all pool? | 22:21 |
| y2kenny | oh wait... I think I may get it now... it's the static driver that's throwing me off | 22:21 |
| y2kenny | "existing resource" == anything that launched by nodepool... I probably confused by the languge because for static, existing resources are configured resources | 22:22 |
| *** harrymichal has quit IRC | 22:31 | |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Add tests for upload-docker-image https://review.opendev.org/735402 | 22:42 |
| avass | mordred, corvus: I'm thinking something like that ^ :) | 22:42 |
| avass | feel free to continue on that. it's getting late so I'm quitting for tonight :) | 22:42 |
| -openstackstatus- NOTICE: The Gerrit service on review.opendev.org is available again | 22:50 | |
| *** tosky has quit IRC | 23:25 | |
| *** armstrongs has joined #zuul | 23:38 | |
| *** armstrongs has quit IRC | 23:47 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!