| *** noonedeadpunk has quit IRC | 01:19 | |
| *** noonedeadpunk has joined #openstack-dns | 01:21 | |
| *** hamalq has quit IRC | 01:47 | |
| *** mugsie has quit IRC | 06:37 | |
| *** mugsie has joined #openstack-dns | 11:13 | |
| eandersson | nicolasbock it's valuable for two reasons | 18:21 |
|---|---|---|
| eandersson | 1) It ensures that different services to co-exist | 18:21 |
| eandersson | 2) Not all repositories have all versions of packages available, pip, yum, apt, deb etc | 18:22 |
| eandersson | nicolasbock johnsom https://bugs.launchpad.net/designate/+bug/1915756 | 18:22 |
| openstack | Launchpad bug 1915756 in Designate "Designate-api/central children workers are not getting killed on service stop" [Undecided,New] | 18:23 |
| eandersson | This is fixed, but doubt we ever released any of the fixes. | 18:23 |
| johnsom | eandersson Thanks for the tip. I can track that down. We were just talking about auditing where we are at for stable releases. It's on my todo list | 18:24 |
| nicolasbock | To make sure I understand this: When I `pip install -c lower-constraints.txt -r requirements.txt` I will end up with the version that's in the lower constraints for a given package, right? | 18:25 |
| eandersson | Yea the lowest supported version | 18:25 |
| eandersson | for that service | 18:25 |
| nicolasbock | But in order for different services to co-exist, wouldn't they have to coordinate their lower constraints? | 18:26 |
| johnsom | Yeah, the reason we invested a lot of time in the lower constraint check in Octavia was specifically CentOS/RHEL which uses RPM packaged versions of some python modules. Upstream we test everything normally with the latest available module in pip. I.e. requirements >=x.y.z. However, the distros were packaging an older version based on x.y.z, which we were not testing to still be valid | 18:27 |
| eandersson | nicolasbock co-exist on all platforms is probably more like it. | 18:28 |
| nicolasbock | But how is that coordinated? As far as I can tell every project has their own version of lower constraints | 18:29 |
| eandersson | Yea - it's not properly coordinated | 18:29 |
| johnsom | So, for example, requirements says acme-module>=1.0.0, the zuul jobs would pull acme-module=1.55.0, but CentOS/RHEL has packaged acme-module=1.33.0. OpenStack was claiming 1.33.0 was valid, but there was no test to confirm that. And things went boom when 1.33.0 was missing some new method. | 18:29 |
| nicolasbock | Would we need to fix lower constraints for _all_ dependencies ? | 18:29 |
| eandersson | but "good enough" for most cases | 18:29 |
| nicolasbock | Ah ok, I can see that johnsom | 18:30 |
| eandersson | I actually really like the idea of understanding impact of my code as well | 18:30 |
| eandersson | If I use a new feature of oslo.messaging and that job starts failing | 18:30 |
| nicolasbock | Yes, the general idea of lower constraints I like | 18:30 |
| nicolasbock | I am just not sure how to best and most efficiently make sure they are correct | 18:30 |
| eandersson | I better understand the impact of my change and know that it might cause issues | 18:30 |
| nicolasbock | Are we still happy with dropping lower constraints for the stable branches? | 18:32 |
| nicolasbock | Or did I misunderstand that? | 18:32 |
| eandersson | We don't have a lot of developer time in Designate today | 18:33 |
| johnsom | Yeah, I am fine with dropping it on stable branches. I think the most critical is maintaining it for the current release | 18:33 |
| eandersson | Yea | 18:33 |
| eandersson | I would like to keep them, but there are bigger ticket items we can spend time on. | 18:33 |
| nicolasbock | Ok | 18:33 |
| johnsom | Yeah, agreed, ideal world we would keep them, but we can drop them for now and if it becomes a problem we can re-enable them on the stable branches. | 18:34 |
| nicolasbock | I proposed to drop them in https://review.opendev.org/c/openstack/designate/+/775626, https://review.opendev.org/c/openstack/designate/+/775660, https://review.opendev.org/c/openstack/designate/+/775663, and https://review.opendev.org/c/openstack/designate/+/775664 | 18:34 |
| eandersson | LGTM | 18:34 |
| *** hamalq has joined #openstack-dns | 18:35 | |
| johnsom | There are two of us now working on Designate, but we have been a bit focused on our deployment tooling, etc so far. That should improve however and we should get some more cycles for actual Designate work. | 18:35 |
| johnsom | Ugh, I never got around to getting stable +2 added. Sigh | 18:36 |
| nicolasbock | You mean, 2 in your team johnsom ? | 18:37 |
| johnsom | Thanks for dropping the file too, some projects seem to be forgetting and leaving the misleading lower-constraints.txt file. | 18:37 |
| johnsom | Yeah, Michael Chapman is on my team. We are both focused on Designate now. | 18:38 |
| nicolasbock | Nice! | 18:38 |
| johnsom | Sadly, I have been doing a lot of planning and tripleo work. lol | 18:38 |
| nicolasbock | Haha, that's how it goes | 18:38 |
| johnsom | We also have a QA/QE person working on the tempest coverage | 18:39 |
| nicolasbock | Wow, that is nice. I haven't had any time so far to look at those tests | 18:39 |
| johnsom | https://review.opendev.org/c/openstack/designate-tempest-plugin/+/773477 | 18:39 |
| nicolasbock | Thanks for the reviews on the lower constraints johnsom and eandersson ! | 18:44 |
| *** openstackgerrit has joined #openstack-dns | 19:39 | |
| *** ChanServ sets mode: +v openstackgerrit | 19:39 | |
| openstackgerrit | Merged openstack/designate stable/victoria: Drop lower-constraints test https://review.opendev.org/c/openstack/designate/+/775626 | 19:39 |
| openstackgerrit | Merged openstack/designate stable/train: Drop lower-constraints test https://review.opendev.org/c/openstack/designate/+/775663 | 21:06 |
| nicolasbock | The Stein gate keeps failing, espcially around the pdns4 tests. We might have to backport frickler 's fix (https://review.opendev.org/c/openstack/designate/+/754604 ). I'll try to get it in for Ussuri and then backport down the branches (https://review.opendev.org/c/openstack/designate/+/755138). | 22:04 |
| openstackgerrit | Merged openstack/designate stable/ussuri: Drop lower-constraints test https://review.opendev.org/c/openstack/designate/+/775660 | 22:19 |
| openstackgerrit | Nicolas Bock proposed openstack/designate stable/ussuri: Fix pdns4 devstack plugin https://review.opendev.org/c/openstack/designate/+/755138 | 22:22 |
| openstackgerrit | Nicolas Bock proposed openstack/designate stable/train: Fix pdns4 devstack plugin https://review.opendev.org/c/openstack/designate/+/775902 | 23:05 |
| openstackgerrit | Nicolas Bock proposed openstack/designate stable/stein: Fix pdns4 devstack plugin https://review.opendev.org/c/openstack/designate/+/775903 | 23:07 |
| eandersson | nicolasbock you might need to bundle those two changes for Stein | 23:11 |
| eandersson | Since both probably need to be in place to get the CI to pass | 23:12 |
| nicolasbock | Ah, that's a good point eandersson | 23:24 |
| nicolasbock | Does anyone know where the source of https://releases.openstack.org/teams/designate.html lives? | 23:36 |
| johnsom | It is auto generated by the release zuul jobs | 23:37 |
| nicolasbock | Ah. There are some spelling issues on that page though | 23:43 |
| nicolasbock | "designated-tempest-plugin" | 23:43 |
| nicolasbock | For example | 23:43 |
| nicolasbock | I was wondering where that's being pulled in from | 23:43 |
| johnsom | Which release? I'm not seeing that | 23:44 |
| nicolasbock | Wallaby | 23:46 |
| johnsom | Oh, that is google translate doing that. I think. | 23:47 |
| nicolasbock | Oh. That's funny :) | 23:47 |
| nicolasbock | You mean just on my client? | 23:47 |
| nicolasbock | Or when the page is being autogenerated? | 23:48 |
| johnsom | That content should be built by the "deliverables" files in the releases repo. | 23:48 |
| nicolasbock | Those files all look good | 23:48 |
| nicolasbock | I mean in terms of spelling | 23:49 |
| johnsom | Yeah, on the chrome browser. If I open it in firefox it's right, if chrome doesn't auto-translate it's fine. It's just when chrome thinks it is italian and translates it that things go wrong | 23:49 |
| nicolasbock | Wow, that's sneaky | 23:49 |
| nicolasbock | Ok, sorry for the false alarm then | 23:49 |
| nicolasbock | I didn't realize that my browser is autotranslating english to english | 23:50 |
| nicolasbock | Now I see it | 23:50 |
| johnsom | Chrome does a lot of sneak things... lol | 23:50 |
| johnsom | Though I keep using it, so I guess bad on me | 23:50 |
| nicolasbock | When I click on "show me the page in it's original, i.e. Italian" it's all correct | 23:50 |
| nicolasbock | I hadn't noticed that little Google translate symbol in the URI bar | 23:51 |
| nicolasbock | How does it figure tha page is in Italian? | 23:51 |
| johnsom | I have no idea. I doubt the OpenStack site is inserting a language hint in the page that is wrong. But could be. | 23:52 |
| johnsom | After digging for hours I suspect I would find that some oddball google AI is seeing a word and deciding it's Italian. | 23:53 |
| nicolasbock | Well, the page actually goes out of its way to declare it in English: `<html lang="en" xml:lang="en" xmlns="http://www.w3.org/1999/xhtml">` | 23:53 |
| johnsom | Interesting enough, the Octavia page doesn't come up as Italian | 23:53 |
| nicolasbock | Yes, that is curious | 23:54 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!