| dmsimard | pep8 all the things | 00:11 |
|---|---|---|
| corvus | mordred: \o/ | 01:56 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Temporary fix for race in quick start job https://review.openstack.org/611476 | 01:58 |
| corvus | mordred, tobiash: ^ that should stabilize the quick-start job until i figure out what to do with the "tenant isn't ready" status code | 01:59 |
| *** bhavikdbavishi has joined #zuul | 03:36 | |
| *** bhavikdbavishi1 has joined #zuul | 03:50 | |
| tobiash | corvus: we could just do a retry loop until the expected data is there | 03:51 |
| *** bhavikdbavishi has quit IRC | 03:52 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 03:52 | |
| corvus | tobiash: yes, but i want to change the expected data :) | 04:05 |
| tobiash | corvus: fine with me :) | 04:05 |
| tobiash | corvus: still awake? | 04:06 |
| corvus | tobiash: yeah, it's still a bit before bed here; i'm preparing for a trip tomorrow | 04:06 |
| tobiash | corvus: I just single core approved your fix so I don't have recheck zuul changes the whole day | 04:07 |
| corvus | tobiash: ++ | 04:07 |
| openstackgerrit | Merged openstack-infra/zuul master: Temporary fix for race in quick start job https://review.openstack.org/611476 | 04:20 |
| openstackgerrit | Merged openstack-infra/zuul master: Fedora docker-compose fixes for selinux https://review.openstack.org/611417 | 04:23 |
| openstackgerrit | Merged openstack-infra/zuul master: Use zuul/nodepool-launcher container for docker-compose https://review.openstack.org/611442 | 04:29 |
| openstackgerrit | Merged openstack-infra/zuul master: Add mysql to quick-start https://review.openstack.org/610697 | 05:10 |
| openstackgerrit | Merged openstack-infra/zuul-base-jobs master: Correct zuul-jobs path https://review.openstack.org/599607 | 05:13 |
| openstackgerrit | Merged openstack-infra/zuul master: Fix periodic job display in builds page https://review.openstack.org/611352 | 05:15 |
| *** bhavikdbavishi has quit IRC | 05:23 | |
| *** bhavikdbavishi has joined #zuul | 05:25 | |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Extract pep8 messages for inline comments https://review.openstack.org/589634 | 05:39 |
| *** rlandy|bbl is now known as rlandy | 05:39 | |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Limit ensure-python to Debian/Ubuntu use https://review.openstack.org/610948 | 06:04 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Fix RST formatting https://review.openstack.org/610198 | 06:11 |
| *** hashar has joined #zuul | 06:15 | |
| *** themroc has joined #zuul | 07:18 | |
| *** persia has quit IRC | 07:19 | |
| *** persia has joined #zuul | 07:21 | |
| *** openstackgerrit has quit IRC | 07:35 | |
| *** jesusaur has quit IRC | 08:38 | |
| *** sshnaidm_ has joined #zuul | 08:59 | |
| AJaeger | Zuul team, openstack-zuul-jobs linter fails now, see http://logs.openstack.org/86/610386/3/gate/openstack-zuul-jobs-linters/a8c857d/job-output.txt.gz#_2018-10-18_07_14_42_431592 - pointing to "zuul-jobs/roles/tox/tasks/main.yaml" | 08:59 |
| AJaeger | Looks like a problem with https://review.openstack.org/589634 | 09:00 |
| AJaeger | mordred, corvus ^ | 09:00 |
| tobiash | AJaeger: do we need a rollback? | 09:08 |
| *** openstackgerrit has joined #zuul | 09:11 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul-jobs master: Revert "Extract pep8 messages for inline comments" https://review.openstack.org/611549 | 09:11 |
| tobiash | AJaeger: in case you need to revert ^ | 09:12 |
| tobiash | but if you can fix forward, that might be preferable | 09:13 |
| tobiash | hint is that zuul is probably not installed during the linter run so it doesn't detect the zuul_return task | 09:13 |
| AJaeger | tobiash: if it's not fixed today - yes. | 09:13 |
| AJaeger | thanks | 09:13 |
| AJaeger | tobiash: sorry, no time right now | 09:13 |
| * tobiash is at lunch now | 09:14 | |
| *** sshnaidm_ has quit IRC | 09:18 | |
| *** goern has quit IRC | 09:22 | |
| *** jesusaur has joined #zuul | 09:31 | |
| gundalow | mordred: nice (inline pep8) Once jlk has GitHubv3 & Checks API we can do the same for GitHub. | 09:41 |
| *** bhavikdbavishi has quit IRC | 09:56 | |
| *** sshnaidm has joined #zuul | 10:06 | |
| openstackgerrit | Merged openstack-infra/zuul master: Add more information to build page https://review.openstack.org/610138 | 10:08 |
| *** sshnaidm has quit IRC | 10:12 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Support node caching in the nodeIterator https://review.openstack.org/604648 | 10:38 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Use node cache in quota calculations https://review.openstack.org/604649 | 10:38 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Cache iterations over ready nodes https://review.openstack.org/604650 | 10:38 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Use cache when counting poolnodes https://review.openstack.org/604651 | 10:38 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Use cache when deleting oldest unused nodes https://review.openstack.org/604652 | 10:38 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Cache node iterations in cleanup workers https://review.openstack.org/604691 | 10:38 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Support node caching in the nodeIterator https://review.openstack.org/604648 | 10:52 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Use node cache in quota calculations https://review.openstack.org/604649 | 10:52 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Cache iterations over ready nodes https://review.openstack.org/604650 | 10:52 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Use cache when counting poolnodes https://review.openstack.org/604651 | 10:52 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Use cache when deleting oldest unused nodes https://review.openstack.org/604652 | 10:52 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Cache node iterations in cleanup workers https://review.openstack.org/604691 | 10:52 |
| *** bhavikdbavishi has joined #zuul | 10:54 | |
| *** sshnaidm has joined #zuul | 10:58 | |
| *** hashar has quit IRC | 11:02 | |
| *** themroc has quit IRC | 11:31 | |
| panda | does this seem useful for anyone ? https://review.openstack.org/606062 | 11:36 |
| *** electrofelix has joined #zuul | 11:38 | |
| *** themroc has joined #zuul | 11:38 | |
| *** panda is now known as panda|lunch | 11:39 | |
| *** sshnaidm_ has joined #zuul | 11:50 | |
| *** sshnaidm has quit IRC | 11:53 | |
| *** panda|lunch is now known as panda | 12:13 | |
| *** jesusaur has quit IRC | 12:33 | |
| *** rlandy has joined #zuul | 12:36 | |
| openstackgerrit | Gonéri Le Bouder proposed openstack-infra/zuul master: encrypt_secret: support OpenSSL 1.1.1 https://review.openstack.org/611414 | 13:03 |
| mordred | gundalow: yup! looking forward to that | 13:21 |
| Shrews | mordred: morning! see comments in #infra about reverting your pep8 comment change | 13:25 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Revert "Extract pep8 messages for inline comments" https://review.openstack.org/611549 | 13:31 |
| Shrews | tobiash: do you think a release note for https://review.openstack.org/610993 would be good to inform admins about the bug fix? | 13:40 |
| tobiash | Shrews: yes, makes sense | 13:41 |
| tobiash | Shrews: shall I add it to that or a follow up? | 13:41 |
| Shrews | tobiash: add it to that one and i'll just +3 | 13:41 |
| tobiash | ok | 13:41 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Initialize label statistics to zero https://review.openstack.org/610993 | 13:45 |
| tobiash | Shrews: ^ | 13:45 |
| Shrews | thx | 13:46 |
| tobiash | thanks for review | 13:47 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul-jobs master: Fix zuul_work_dir default for build-reno-releasenotes https://review.openstack.org/611616 | 13:57 |
| *** themroc has quit IRC | 14:04 | |
| openstackgerrit | Merged openstack-infra/nodepool master: Run zuul-quick-start job https://review.openstack.org/610159 | 14:16 |
| *** themroc has joined #zuul | 14:19 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Update ANSIBLE_LIBRARY to use envsitepackagesdir https://review.openstack.org/611622 | 14:23 |
| *** jesusaur has joined #zuul | 15:20 | |
| *** themroc has quit IRC | 15:21 | |
| *** goern has joined #zuul | 15:40 | |
| jlk | I'll get time to work on that again, soonish I think | 15:59 |
| pabelanger | \o/ | 15:59 |
| pabelanger | also happy to help too | 15:59 |
| openstackgerrit | Merged openstack-infra/nodepool master: Initialize label statistics to zero https://review.openstack.org/610993 | 16:10 |
| *** sshnaidm_ is now known as sshnaidm | 16:14 | |
| pabelanger | corvus: tobiash: SpamapS: we just had an issue in ansible-network and deleting project branch, it seems zuul didn't see the event properly from github and 'branch' configuration was still on zuul-merger. Which lead to merge conflict issues when we try to propose new PRs to the same project. I am going to try and get logs from SF.io, but wanted to see if you've also seen this | 16:42 |
| pabelanger | mhu: tristanC: ^ | 16:42 |
| clarkb | pabelanger: I think we've had to fix similar bugs with the gerrit driver when it didn't update its cached config (or invalidate it) when a branch is deleted | 16:42 |
| clarkb | I want to say the fix in gerrit was to discard the entire cached config when that happens? | 16:43 |
| pabelanger | clarkb: yah, we did SIGHUP to scheduler, which fixed the issue. But required a human to do it. Fixing in gerrit does ring a bell | 16:44 |
| clarkb | I wonder if this is more difficutl with github because github encourages branching | 16:44 |
| clarkb | would be potentially expensive to reload config on every github branch delete? | 16:45 |
| pabelanger | Great question, I am un sure | 16:45 |
| pabelanger | in our case, it was an old 'master' branch from when the project move to 'devel' branch. Usually, our workflow in ansible-network won't be to delete many branches | 16:46 |
| pabelanger | actually, I do not know if SIGHUP fixed it, I believe scheduler was stopped / started. Will confirm in logs | 16:47 |
| *** caphrim007 has joined #zuul | 16:57 | |
| *** panda is now known as panda|off | 17:30 | |
| *** dkehn has quit IRC | 17:30 | |
| *** rfolco is now known as rfolco|rucker | 17:31 | |
| *** dkehn has joined #zuul | 17:37 | |
| tobiash | pabelanger: hrm, actually this should work | 17:47 |
| tobiash | pabelanger: maybe this fixes your issue: https://review.openstack.org/600088 | 17:47 |
| pabelanger | tobiash: ack, will look shortly! | 17:48 |
| *** electrofelix has quit IRC | 17:48 | |
| pabelanger | tobiash: thanks, +2. maybe corvus or mordred can also review | 17:53 |
| ssbarnea | hi! does anyone know how to change the current url when you do a search in zuul? so it would endup being "http://zuul.openstack.org/status#66666" so you can easily share it with others? | 17:56 |
| ssbarnea | the url works, is only the search action that does not update the current page url | 17:56 |
| ssbarnea | i hope it would work | 17:56 |
| mordred | ssbarnea: the builds page does it - so it's definitely possible | 18:05 |
| ssbarnea | mordred: glad to hear that, my JS experience sucks, otherwise i would have raised a CR | 18:08 |
| mordred | tristanC: if you get a sec, could you point me in the right direction with that ^^? | 18:10 |
| dmsimard | mordred: unless mistaken tristanC is on PTO for a week or two | 18:14 |
| mordred | dmsimard: NOOOOOO | 18:14 |
| dmsimard | mordred: I'm told it's one week -- so he'll be back sometime next week :p | 18:24 |
| *** panda|off has quit IRC | 18:59 | |
| *** panda has joined #zuul | 19:00 | |
| *** bhavikdbavishi has quit IRC | 19:22 | |
| *** ssbarnea_ has joined #zuul | 20:02 | |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool master: WIP: Cleanup down ports https://review.openstack.org/609829 | 20:08 |
| SpamapS | pabelanger: since I actively avoid project branches, no, I haven't seen the delete branch thing you mentioned. | 20:13 |
| mordred | Shrews: should we put cleanupLeakedPorts into sdk like delete_unattached_floating_ips? or I suppose it's likely normal of a thing to need? | 20:21 |
| *** openstackgerrit has quit IRC | 20:36 | |
| Shrews | mordred: i have *no* idea | 20:39 |
| mordred | Shrews: me either! | 20:40 |
| Shrews | mordred: my gut says "no" because a sdk should be just that | 20:40 |
| Shrews | mordred: not a general "clean up after all the bugs for me" development kit | 20:41 |
| Shrews | it's hard enough keeping up with the various apis, imagine keeping up with the bugs too! | 20:42 |
| * Shrews shuts up | 20:42 | |
| clarkb | thats kind of why shade exists :P | 20:42 |
| clarkb | mordred: if you put it into the shade portion of the sdk I think it makes sense there | 20:43 |
| Shrews | shade is "hide the cloud differences" mainly | 20:43 |
| clarkb | and workaround the bugs | 20:43 |
| *** pcaruana has quit IRC | 20:44 | |
| Shrews | bugs in api, yes. bugs in services themselves? i disagree with trying to do that, but i'm a stubborn mule | 20:45 |
| clarkb | that is why the floating ip stuff exists aiui | 20:45 |
| Shrews | back when nodepool was the main/only user, yep. | 20:46 |
| clarkb | well because the code comes from nodepool :P | 20:46 |
| clarkb | I don't think that makes it wrong | 20:47 |
| clarkb | nodepool produced a useful thing because it worked around these issues for users | 20:47 |
| Shrews | if the "main" openstack api works around the bugs for the services, where is the incentive for the services to fix their bugs? | 20:48 |
| clarkb | I think what we have learned is that users don't choose the version of openstack they have to live with and no amount of incentivizing say neutron fix a bug will ensure that the end user gets that bug fix in the next N years | 20:49 |
| clarkb | I do think project teams have taken a much better approach at this type of thing since shade existed though. The incentive seems to be "look how bad it is we made an official workaround" | 20:50 |
| mordred | also - to be fair - we still haven't added support for get-me-a-network to shade after complaining it didn't exist for ages | 20:51 |
| clarkb | mordred: I thought the point with get me a network was that it was automagic | 20:51 |
| clarkb | shade shouldn't have to do anything with it, you boot an instance and it gets you a network transparently | 20:51 |
| mordred | clarkb: almost, but not quite | 21:05 |
| mordred | clarkb: it's a flag you pass to a server creation | 21:05 |
| clarkb | ah | 21:06 |
| dmsimard | mordred: the famous "get me a network".. from rax ? | 21:07 |
| mordred | clarkb: networks: auto, starting from microversion 2.37 | 21:07 |
| clarkb | dmsimard: no it was a nova + neutron feature to fix the problem of not having networking by default with openstack and needing to make a router, network and subnet in the correct order and tie them all together | 21:07 |
| clarkb | its about a 15 step process to do it manually | 21:07 |
| mordred | dmsimard: the famout "get me a network" from me making very loud complaining noises that in a new cloud account before I could boot a vm I had to create a bunch of network objects | 21:08 |
| mordred | yah | 21:08 |
| *** panda has quit IRC | 21:08 | |
| clarkb | Horizon was the best way to do it prior to this as a new user because it asked you all of the info before booting an instance and basically did get me a network for you in horizon | 21:09 |
| clarkb | but that required using a web browser | 21:09 |
| mordred | clarkb: my thinking is that in sdk create_server we'll default network to auto if the remote side supports it and if you didn't request anything explicit | 21:10 |
| mordred | I haven't done it yet because if the remote side _doesn't_ support it, I think we should simulate it client-side and make all the network objects for you | 21:10 |
| clarkb | mordred: with that flag it will only create the network if one isn't already available right? if so then I think that is the right behavior | 21:10 |
| mordred | and that seems like a lot of work | 21:10 |
| mordred | clarkb: yes | 21:11 |
| *** panda has joined #zuul | 21:12 | |
| clarkb | mordred: the downside to not doing that being people with older clouds will have different experience than people with newer clouds? | 21:16 |
| clarkb | seems like enabling it for newer clouds to start is a reasonable compromise? | 21:16 |
| mordred | yah - and now that we've landed full discovery support, adding it for people with newer clouds is pretty easy | 21:17 |
| dmsimard | mordred: well, really my understanding of it is that there is just a layer between the request and neutron which gets an actual network id based on availability etc | 21:21 |
| dmsimard | Get me a network was a uuid like 00000-0000-0000 (and probably more zeroes) | 21:23 |
| jlk | I got a nice history lesson on Gerrit, including the "correct pronunciation" from Guido yesterday. He was visiting GitHub HQ and wanted to chat with Python devs at the company. | 21:24 |
| dmsimard | We tried to do something similar at inap :) | 21:24 |
| dmsimard | jlk: no way that's cool | 21:24 |
| dmsimard | What is he doing nowadays? Last I hear he was at Dropbox ? | 21:25 |
| jlk | mypy stuff | 21:25 |
| jlk | type checking for Python. They rolled it out across a bunch of dropbox code | 21:25 |
| mordred | jlk: did he teach you how to do a proper dutch-g ? | 21:25 |
| jlk | he did not include the linguistics lesson :D | 21:25 |
| mordred | :) | 21:26 |
| jlk | but I know I"d be laughed at if I tried to pronounce it that way in OpenStack circles | 21:26 |
| mordred | now I want to start pronouncing it that way | 21:26 |
| mordred | just to be that guy | 21:26 |
| jlk | but it was fun to learn that Gerrit came from a thing that came from a thing he wrote when he joined Google. | 21:26 |
| mordred | ++ | 21:27 |
| jlk | He wrote a thing for source review, as there was none. Then it was rewritten for app engine, and then re-written by Android folks in Java as Gerrit | 21:27 |
| jlk | IT USED TO BE PYTHON... | 21:27 |
| *** rlandy is now known as rlandy|bbl | 21:27 | |
| jlk | oh, what could have been... | 21:27 |
| mordred | jlk: fwiw - we actually looked _briefly_ at resurrecting the python version | 21:27 |
| dmsimard | jlk: interesting | 21:27 |
| mordred | back in the very early days of gerrit at openstack | 21:27 |
| mordred | I think Shrews is the one who dug in to that :) | 21:28 |
| clarkb | and I want to say the thing we decided was you needed an app engine? | 21:28 |
| mordred | yeah. it was too much work to get it back up to feature parity with the state of the java one at that time | 21:29 |
| jlk | that sounds right, it apparently was very tied to app engine | 21:29 |
| clarkb | dmsimard: aiui you have to provision subnet pools in neutron that specify valid subnet ip ranges to create subnets per tenant out of | 21:31 |
| clarkb | dmsimard: and you have to set a default external network, but the rest of the provisioning is dynamic | 21:31 |
| dmsimard | clarkb: yeah, the challenge is you run out of vlans in the same layer 2 when you're handing 2 vlans per customer | 21:34 |
| dmsimard | And then vxlan becomes complicated | 21:34 |
| dmsimard | So having this kind of layer to make something land in the right place is nice :) | 21:34 |
| clarkb | dmsimard: the resource limits don't change, a tenant can't have working networking without it. either they create the resources themselves or you do it for them | 21:35 |
| dmsimard | Indeed | 21:36 |
| mordred | dmsimard: of course, I always argue in favor of the vexxhost model - let me boot vms directly on the public networks and default to that - but then also let me create a private network if I want one | 21:47 |
| mordred | this model seems to bother people though | 21:47 |
| jlk | isn't that kind of what Rackspace did? | 21:47 |
| jlk | it bothers people who have to pay for public ipv4 addresses :D | 21:47 |
| clarkb | jlk: I don't think rackspace ever allowed you to go private only? | 21:48 |
| clarkb | but ya rackspace does similar for most users | 21:48 |
| mordred | jlk: well, rackspace just gave you a public and a private -although the private was on a shared private net | 21:48 |
| mordred | of course, rackspace dislikes dhcp | 21:48 |
| jlk | yeah, RAX had "funny" networking | 21:48 |
| mordred | so, there's that | 21:48 |
| *** ssbarnea_ has quit IRC | 21:53 | |
| *** caphrim007 has quit IRC | 22:52 | |
| *** rlandy|bbl is now known as rlandy | 23:19 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!