14:34:13 #startmeeting Pulp Triage 2019-11-01 14:34:13 #info fao89 has joined triage 14:34:13 !start 14:34:13 Meeting started Fri Nov 1 14:34:13 2019 UTC. The chair is fao89. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:34:13 Useful Commands: #action #agreed #help #info #idea #link #topic. 14:34:13 The meeting name has been set to 'pulp_triage_2019-11-01' 14:34:13 fao89: fao89 has joined triage 14:34:21 #info mikedep333 has joined triage 14:34:21 !here 14:34:21 mikedep333: mikedep333 has joined triage 14:34:24 #info bmbouter has joined triage 14:34:24 !here 14:34:24 bmbouter: bmbouter has joined triage 14:34:29 !next 14:34:30 fao89: 3 issues left to triage: 5642, 5636, 5626 14:34:31 #topic https://pulp.plan.io/issues/5642 14:34:31 RM 5642 - mdellweg - NEW - Pulp client api is broken 14:34:32 https://pulp.plan.io/issues/5642 14:34:39 daviddavis, sounds like you don't 14:34:40 #info dawalker has joined triage 14:34:40 !here 14:34:40 dawalker: dawalker has joined triage 14:35:02 #info ppicka has joined triage 14:35:02 !here 14:35:02 ppicka: ppicka has joined triage 14:35:18 #idea Proposed for #5642: Leave the issue as-is, accepting its current state. 14:35:18 !propose accept 14:35:18 fao89: Proposed for #5642: Leave the issue as-is, accepting its current state. 14:35:18 #info ttereshc has joined triage 14:35:18 !here 14:35:19 ttereshc: ttereshc has joined triage 14:35:23 #info ipanova has joined triage 14:35:23 !here 14:35:24 ipanova: ipanova has joined triage 14:35:36 #info daviddavis has joined triage 14:35:36 !here 14:35:36 daviddavis: daviddavis has joined triage 14:36:09 #info dkliban has joined triage 14:36:09 !here 14:36:09 dkliban: dkliban has joined triage 14:36:38 accept and add to sprint 14:36:40 +1 14:36:52 #idea Proposed for #5642: accept and add to sprint 14:36:52 !propose other accept and add to sprint 14:36:52 fao89: Proposed for #5642: accept and add to sprint 14:36:56 #agreed accept and add to sprint 14:36:56 !accept 14:36:56 fao89: Current proposal accepted: accept and add to sprint 14:36:57 #topic https://pulp.plan.io/issues/5655 14:36:57 fao89: 3 issues left to triage: 5655, 5636, 5626 14:36:58 should it have a more specific title? 14:36:59 RM 5655 - iballou - NEW - RPM repo remote URL without trailing slash causes 404 14:37:00 https://pulp.plan.io/issues/5655 14:37:08 yeah 14:37:28 I think this one is not rpm specific, so I moved to to pulp 14:37:58 can we retitle it a bit also? 14:38:01 not every plugin expects the slash i think 14:38:24 shouldn't that be consistent? 14:38:26 I think ansible requires there to not be one 14:38:32 pulp_ansible that is 14:39:13 depends on the client dawalker 14:39:20 I think the problem is in removing everything after last slash 14:39:23 yeah agreed pulp_ansible needs all sorts of data there 14:39:29 think we should accept and move to rpm 14:39:56 I don't think rpm plugin removes anything from the url 14:39:59 #idea Proposed for #5655: accept and move to rpm 14:39:59 !propose other accept and move to rpm 14:39:59 fao89: Proposed for #5655: accept and move to rpm 14:40:04 that's why I moved it here 14:40:13 ok, I can move it back, fair enough 14:40:37 ttereshc: if there is something core can do we want to, but I'm not sure what atm 14:40:46 since it's interpreted in FirstStage 14:41:07 +1 14:41:17 bmbouter, and url is not modified/validated in a ny way before that? 14:41:45 not that I'm aware of 14:41:59 this: "Pulp chopped the 'x86_64' off" made me worriede 14:42:00 ok 14:42:02 ttereshc: just a simple validation of it not being None when a remote is created 14:42:07 we can move on 14:42:31 #agreed accept and move to rpm 14:42:31 !accept 14:42:31 fao89: Current proposal accepted: accept and move to rpm 14:42:32 #topic https://pulp.plan.io/issues/5636 14:42:32 fao89: 2 issues left to triage: 5636, 5626 14:42:33 RM 5636 - Timoses - NEW - Dependency version conflicts between pulp plugins and pulp installation with ansible-pulp 14:42:34 https://pulp.plan.io/issues/5636 14:43:06 we have not released docker compatible with rc7 14:44:06 I feel like this is unsolvable? 14:44:22 plugins could depend on different versions of core 14:44:34 i agree 14:44:46 GA should help mitigate this though 14:45:03 until the plugin api stabalizes it's unsafe for a plugin to allow users to receive the 3.y+1 14:45:09 bmbouter, ttereshc, daviddavis one thing that occurs to me is that "name" on repository, remote, etc. is universally unique, meaning you can't have a file repo named "foo" if you also have an RPM repo named "foo" 14:45:31 dalley: oof, let's revisit at open floor 14:45:36 +1 14:45:38 GA should help some though I agree 14:45:42 sorry 14:45:46 dalley: np at all 14:46:20 for this bug I think the resolution is for docker to release? 14:46:24 +1 14:46:43 yes, i will reply that a new release of docker is needed 14:46:48 +1 14:47:49 #idea Proposed for #5636: ipanova will reply that a new release of docker is needed 14:47:49 !propose other ipanova will reply that a new release of docker is needed 14:47:49 fao89: Proposed for #5636: ipanova will reply that a new release of docker is needed 14:47:51 #agreed ipanova will reply that a new release of docker is needed 14:47:51 !accept 14:47:51 fao89: Current proposal accepted: ipanova will reply that a new release of docker is needed 14:47:53 #topic https://pulp.plan.io/issues/5626 14:47:53 fao89: 1 issues left to triage: 5626 14:47:54 RM 5626 - Timoses - NEW - Sync Task fails with "Cannot resolve keyword '_created' into field" 14:47:55 https://pulp.plan.io/issues/5626 14:48:14 closed worksforme 14:48:19 +1 14:48:24 +1 14:48:39 #idea Proposed for #5626: closed worksforme 14:48:39 !propose other closed worksforme 14:48:39 fao89: Proposed for #5626: closed worksforme 14:48:43 #agreed closed worksforme 14:48:43 !accept 14:48:43 fao89: Current proposal accepted: closed worksforme 14:48:44 fao89: No issues to triage. 14:48:48 Open floor! 14:50:05 ok so detail model unique names 14:50:09 discuss! 14:50:33 there should be no problem with it at db level, so I'm curious where it is 14:50:38 how do we handle this with remotes? 14:50:50 we don't I don't think 14:51:24 I believe all detail names are stored in a table with a unique constraint so I think the db level is where the problem lies? 14:51:30 then I think we can ignore at least through the wed merge and handle separately 14:51:40 is it a problem though? 14:51:42 that's more of a tactical recommendation than a technical one 14:51:44 ttereshc, it's because each master/detail model is split across 2 tables, the "Repository/rRmote" table and the "FileRepository/FileRemote" table, and the "name" and other shared attributes are in the former table 14:51:53 it is a problem I think 14:52:16 what if we just make it not unique? 14:52:40 you could then have two rpm repos with the same name for instance. not sure if that's a problem I guess. 14:52:49 bmbouter: why do you see it as a problem? 14:52:58 yeah, I see it now, I thought it's the other way around, that name is stored n the detail table 14:53:02 we use HREFs basically everywhere which is unique. it wouldn't be a "problem" but it would be confusing to the user, which might be a problem 14:53:18 dalley: at least it won't be new to them 14:53:31 i think pulp2 works same way :) 14:53:39 yea 14:53:43 multiple (object) with the same name? 14:53:44 ipanova: it's not a big problem, just a usability problem 14:54:00 dalley: no the reverse, universally unique names across types 14:54:05 like two users using 1 pulp system, now at least 1 user looses to the other 14:54:12 ah 14:54:19 dalley: ou cannot have file repo foo and rpm repo foo 14:54:43 bmbouter: I could also see the opposite: two users each with the same rpm repo with the same name 14:55:04 s/with the same/with a/ 14:55:34 ipanova, I think now with dalley's changes you will be able to 14:55:53 I don't think so 14:55:56 hit the same problem 14:56:01 oh 14:56:15 ttereshc, no, my PR doesn't really change anything 14:56:20 in that respect 14:57:02 I mean that repos are now master-detail so they have the same issue 14:57:13 can we do unique_together for type and name? 14:57:23 never mind, I'll explain myself later t ipanova 14:57:26 we can 14:57:45 daviddavis, +1 14:58:00 i think i am inclined having 1 repo name across the types 14:58:40 yeah I think this is debatable about what is right 14:58:45 i like david's idea ... you can't list all repositories in one place now 14:59:16 so it would make sense to only have unique names per type 14:59:40 now = with typed repositories 14:59:49 I can file an issue for this and we can discuss further 14:59:55 we discussed last week that we might add such view to show all repos 14:59:58 yeah I like this idea ... in 3.1 15:00:11 dkliban: we might still add the ability to list all repos..later 15:00:11 in any case, +1 to unique_together 15:00:25 ttereshc: is faster :-P 15:00:28 +1 to file and we can chat it up 15:00:42 great 15:00:46 +1 15:01:52 +1 15:02:08 I've got a two topics 15:02:09 +1 15:02:11 both shorties 15:02:18 the same is true of remotes, same solution there? 15:02:26 yup 15:02:36 all master/detail ones? 15:04:06 distributions as well, can't think of any others atm 15:04:21 content guards? 15:04:50 yup 15:05:24 yup 15:05:41 publisher? exporter? :) 15:05:45 ok, I guess all 15:05:48 one of the thing I want to mention is I think conference submission deadlines are today is that right? 15:06:54 I applied for devconf cz booth 15:07:08 CfgMgmtCamp is till Sunday 3 November 23:59:59 CET 15:07:15 excellent! 15:07:29 I'm organizing my talk to there and will submit 15:07:39 fosdem is end of november 15:07:55 there are actully 2 rooms at FOSDEM that would work for Pulp 15:08:05 great 15:08:11 Dependency Management room 15:08:13 so for sunday's deadline who else is submitting again? 15:08:25 I am submitting for Sunday's deadline 15:08:35 I believ x9c4 is also and I'll email him 15:08:39 cool 15:09:55 do we need 4? or 3 is enough? 15:10:03 I was hoping for 4 15:10:44 same here 15:11:04 also we need to put in for a plugin writer workshop, was that the 4th? 15:11:47 I think it's separate 15:12:09 for the 3rd day there are 2 options - fringe and workshop 15:12:27 yeah .. we want to do a workshop on that da 15:12:34 we better go with 1 or 2 workshops 15:12:36 ok I can submit that 15:12:40 yes 15:12:43 yeah, it's 3 hours 15:12:47 while fringe is 8 15:12:53 yeah not fringe again 15:13:52 ok I'll submit one plugin writer workshop 15:14:12 +1 for the workshop and not fringe 15:14:16 bmbouter++ 15:14:16 ipanova: bmbouter's karma is now 206 15:14:36 anyone can join and collab, I just want to make sure we have a submission in 15:15:23 thank you 15:15:35 yeah, thank you! 15:16:04 #endmeeting 15:16:04 !end