14:32:13 #startmeeting Pulp Triage 2018-10-19 14:32:13 !start 14:32:13 #info dawalker has joined triage 14:32:13 Meeting started Fri Oct 19 14:32:13 2018 UTC and is due to finish in 60 minutes. The chair is dawalker. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:32:13 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:32:13 The meeting name has been set to 'pulp_triage_2018_10_19' 14:32:13 dawalker: dawalker has joined triage 14:32:17 #info daviddavis has joined triage 14:32:17 !here 14:32:17 daviddavis: daviddavis has joined triage 14:32:20 first! 14:32:20 #info asmacdo has joined triage 14:32:20 !here 14:32:20 asmacdo: asmacdo has joined triage 14:32:25 dang 14:32:51 daviddavis++ 14:32:53 asmacdo: daviddavis's karma is now 122 14:33:04 #info dalley has joined triage 14:33:04 !here 14:33:05 dalley: dalley has joined triage 14:33:57 if that's the way we're gonna be...*I'm* first :D 14:34:04 :( 14:34:21 lol, no you can keep the lead, daviddavis 14:34:25 :D 14:34:29 !kirby 14:34:29 (>'-')> <('-'<) ^('-')^ v('-')v(>'-')> (^-^) 14:34:43 Alright, let's do this 14:34:47 !next 14:34:48 #topic Licensing Options fo Plugin Writers - http://pulp.plan.io/issues/4080 14:34:49 dawalker: 7 issues left to triage: 4080, 4081, 4083, 4084, 4085, 4086, 4092 14:34:50 Issue #4080 [NEW] (unassigned) - Priority: Normal | Severity: Medium 14:34:51 Licensing Options fo Plugin Writers - http://pulp.plan.io/issues/4080 14:35:05 We skipped this last week waiting on feedback from legal 14:35:12 bmbouter, update? ^ 14:35:21 we're still waiting on feedback from legal. is this maybe a task though? 14:35:36 Yeah, I think so 14:35:50 +1 task 14:36:07 #idea Proposed for #4080: convert to task 14:36:07 !propose other convert to task 14:36:07 dawalker: Proposed for #4080: convert to task 14:36:11 Probably already overloaded, but also seems like a good pulpcon topic 14:36:47 Agreed, it's really interesting, bmbouter shared a little bit about it the other day. 14:37:04 our schedule is full but if we can drop something else then +1 from me. 14:37:17 #info milan has joined triage 14:37:17 !here 14:37:17 milan: milan has joined triage 14:37:35 probably won't need a full time slot, can probably fit it in with other discussions somewhere... 14:37:45 #agreed convert to task 14:37:45 !accept 14:37:45 dawalker: Current proposal accepted: convert to task 14:37:46 #topic Recommend usage of pip-compile for production deployments of Pulp - http://pulp.plan.io/issues/4081 14:37:47 dawalker: 6 issues left to triage: 4081, 4083, 4084, 4085, 4086, 4092 14:37:48 Issue #4081 [NEW] (unassigned) - Priority: Normal | Severity: Medium 14:37:49 Recommend usage of pip-compile for production deployments of Pulp - http://pulp.plan.io/issues/4081 14:38:13 I'll put licensing on the planning pad 14:38:15 at least 14:38:27 +1, thanks 14:38:53 !propose accept 14:38:53 #idea Proposed for #4081: Leave the issue as-is, accepting its current state. 14:38:54 asmacdo: Proposed for #4081: Leave the issue as-is, accepting its current state. 14:39:08 +1 14:39:13 +1 14:39:26 +1 14:39:29 oh also, story or task 14:39:44 story? 14:39:47 i think 14:39:50 sure. 14:40:21 #idea Proposed for #4081: accept and convert to story 14:40:21 !propose other accept and convert to story 14:40:21 dawalker: Proposed for #4081: accept and convert to story 14:40:29 FYI, I think we ought to consider pipenv instead. AFAICT it's the future of piptools 14:40:55 asmacdo: want to leave a comment to that effect? 14:40:58 ok. Do you want to help ensure this gets converted to story and that conversation is started in the comments? 14:41:09 I'll comment on it, just suggesting that it's not ready to be groomed 14:41:15 ah 14:41:18 cool 14:41:26 Ok, in that case, shall we skip for now? 14:41:42 no, we should convert it at least to a story or task 14:41:48 I think it's fine to accept as an ungroomed story 14:42:03 asmacdo: can you convert it when you go to comment? 14:42:06 ok 14:42:19 yup, I'll get this one 14:42:28 great 14:42:35 #agreed accept and convert to story 14:42:35 !accept 14:42:35 dawalker: Current proposal accepted: accept and convert to story 14:42:37 dawalker: 5 issues left to triage: 4083, 4084, 4085, 4086, 4092 14:42:37 #topic Non-sensitive information is not being displayed for last_override_config - http://pulp.plan.io/issues/4083 14:42:38 Issue #4083 [NEW] (unassigned) - Priority: Normal | Severity: Medium 14:42:39 Non-sensitive information is not being displayed for last_override_config - http://pulp.plan.io/issues/4083 14:42:51 just accept this 14:43:22 #idea Proposed for #4083: Leave the issue as-is, accepting its current state. 14:43:22 !propose accept 14:43:23 dawalker: Proposed for #4083: Leave the issue as-is, accepting its current state. 14:43:35 isn't "non-sensitive" info going to carry plugin to plugin? 14:44:05 not sure I understand? 14:44:08 whitelist seems right, but plugins may also need a way to extend the whitelist 14:44:18 yea 14:44:34 makes sense 14:44:41 cool +1 accept, I'll add a comment 14:44:58 thanks again 14:45:01 asmacdo++ 14:45:04 dawalker: asmacdo's karma is now 62 14:45:08 #agreed Leave the issue as-is, accepting its current state. 14:45:08 !accept 14:45:09 dawalker: Current proposal accepted: Leave the issue as-is, accepting its current state. 14:45:10 dawalker: 4 issues left to triage: 4084, 4085, 4086, 4092 14:45:10 #topic When Pulp checks for remote files, reword the log message to sound non-fatal - http://pulp.plan.io/issues/4084 14:45:11 Issue #4084 [NEW] (unassigned) - Priority: Normal | Severity: Medium 14:45:12 When Pulp checks for remote files, reword the log message to sound non-fatal - http://pulp.plan.io/issues/4084 14:45:59 I vote accept 14:46:20 +1 I'd prefer debug level 14:46:32 * in case you know, have to debug stuff ;) 14:46:38 hehe 14:46:41 +1 14:46:42 -0 I think I (as a user) would want to know when things didn't download 14:46:46 !propose accept 14:46:46 #idea Proposed for #4084: Leave the issue as-is, accepting its current state. 14:46:47 dawalker: Proposed for #4084: Leave the issue as-is, accepting its current state. 14:47:09 Info seems appropriate to me, rewording would be my preference 14:47:28 asmacdo, it would show in the debug level 14:47:48 or we could do both: debug level with codes and details 14:47:57 info level for user 14:48:04 overkill? 14:48:18 maybe. Either way we can accept. 14:48:23 well, the submitter didn't want "failed" and "404" in the output 14:48:37 +1 that's totally reasonable 14:48:44 Also, if you'd contribute these thoughts to the comments for it to track the discussion, that'd be great 14:48:45 yeah OT +1 accept discuss on issue 14:49:04 * asmacdo has a lot of homework 14:49:11 +1 14:49:17 And I think this sounds like once the discussion is done, maybe this would be a good candidate for vlada or ppicka ? 14:49:23 lol asmacdo 14:49:31 or both since it's in multiple plugins as noted in the comments 14:49:32 good call dawalker 14:49:45 #agreed Leave the issue as-is, accepting its current state. 14:49:45 !accept 14:49:45 dawalker: Current proposal accepted: Leave the issue as-is, accepting its current state. 14:49:47 dawalker: 3 issues left to triage: 4085, 4086, 4092 14:49:47 it's a hard change. I think nectar logs this 14:49:47 #topic ContentUnitSaver stage is vulnerable to race conditions. - http://pulp.plan.io/issues/4085 14:49:48 Issue #4085 [NEW] (unassigned) - Priority: Normal | Severity: Medium 14:49:49 ContentUnitSaver stage is vulnerable to race conditions. - http://pulp.plan.io/issues/4085 14:50:14 oh, ok, daviddavis I defer to you. I am unfamiliar with nectar 14:50:22 Oh that's a good point daviddavis. I've been spoiled by pulp3. 14:50:26 hehe 14:50:52 I vote accept on this 14:51:01 +100 14:51:10 maybe add to the Sprint too? 14:51:23 #idea Proposed for #4085: accept and add to sprint 14:51:23 !propose other accept and add to sprint 14:51:24 dawalker: Proposed for #4085: accept and add to sprint 14:51:35 +1 14:51:49 #agreed accept and add to sprint 14:51:49 !accept 14:51:49 dawalker: Current proposal accepted: accept and add to sprint 14:51:51 #topic ArtifactSaver stage is vulnerable to race conditions. - http://pulp.plan.io/issues/4086 14:51:51 dawalker: 2 issues left to triage: 4086, 4092 14:51:53 Issue #4086 [NEW] (unassigned) - Priority: Normal | Severity: Medium 14:51:54 ArtifactSaver stage is vulnerable to race conditions. - http://pulp.plan.io/issues/4086 14:51:55 +1 tho sounds like a feature ;) 14:52:43 Same here? 14:52:57 +1 14:53:05 +1 I guess 14:53:21 #idea Proposed for #4086: accept and add to sprint 14:53:21 !propose other accept and add to sprint 14:53:21 dawalker: Proposed for #4086: accept and add to sprint 14:53:24 they might be manifestations of the same issue tho 14:53:25 no, it's a bug in the existing stages imo 14:53:35 milan, a feature to have race conditions or a feature to save us from them/ 14:53:36 ? 14:53:57 a feature as one instance might have passed ;) 14:54:19 I'm fine with a story 14:54:48 jortel, is it the same issue or but in the existing stages? http://pulp.plan.io/issues/4086 vs http://pulp.plan.io/issues/4085 14:54:58 *bug 14:55:00 !here 14:55:00 #info jortel has joined triage 14:55:01 jortel: jortel has joined triage 14:55:17 There's 2 existing stages, both have the same problem 14:55:33 corect 14:55:48 cool, then probably it's going to be one patch fixing, no? 14:55:53 This is a known problem in the 2 core stages. I mainly wanted to document it. 14:55:53 Well 4 stages if you consider query existing and save to work together 14:56:12 hmm 14:56:37 They should be related, but I think separate is correct 14:56:37 anything doing query-save logic is vulnerable to the same race condition. 14:57:29 so...accept and add to sprint? 14:57:36 if any other stages than the 2 I identified are doing this we need bugs filed on them too. 14:58:16 jortel, the race condition is between the query stage and the save stage, which are separate 14:58:52 asmacdo: right, the issue is with the save stages 14:59:13 how about we just convert to an ungroomed story 14:59:18 mainly that save stages cannot rely on the query stage 14:59:23 We can add to the Sprint after we groom 14:59:32 +1 jortel 14:59:42 I think it's an issue, not a story. 14:59:47 sounds more and more feature to me, how that would be fixed? 15:00:29 jortel I agree it's an issue, but the story pipeline might be useful so we can groom it before someone starts on it 15:00:41 milan: the typical pattern for concurrent systems is to try: insert; catch the dup key and query to get the existing one. 15:01:10 asmacdo: agreed. the fix wrt bulk_create() will require a bit of thought. 15:01:25 solving for bulk_create() may be really hard. 15:02:04 but the main problem is that users will see syncs fail with random integrety errors which will be concerning to them. 15:02:07 jortel, right, the second query is susceptible to the same race, no (removed after added)? 15:02:23 Yeah. We could just leave these as is, and add SerialSave stages along side, and leave it to the plugin writer to know what they need to do 15:03:09 how do folks feel about prio/sev? 15:03:19 let's start there for triage 15:04:16 given that users will see random IntegretyError .. 15:04:43 log.error -> log.debugg and call it a day ;) 15:04:57 I think this should be high priority. It also makes plugin development a pain 15:05:07 lol milan 15:05:17 +1 15:05:20 shall we call for discussion on issue? 15:05:21 milan: :) 15:05:55 dawalker, I think the discussion will emerge on its own ;) 15:06:02 So, accept and add to sprint? Do we want to change this one and the previous one both to high priority? 15:06:09 milan, true 15:06:11 suggest we set to high/med and triage. we can design on the issue before putting on a sprint. this should be marked as GA blocker. 15:06:25 +1 jortel 15:06:29 +1 15:06:47 +1 , provided it can be treated as documentation issue any time 15:06:49 ;) 15:07:08 !propose other high/med, accept, mark as GA blocker 15:07:08 #idea Proposed for #4086: high/med, accept, mark as GA blocker 15:07:09 dawalker: Proposed for #4086: high/med, accept, mark as GA blocker 15:07:55 Also +tag "painintheass" 15:08:07 lol asmacdo +1 :D 15:08:10 lol 15:08:20 fyi, this was noted on the stages pr with agreement to merge anyway. these issues are a follow up. 15:08:32 to be sure it 15:08:37 it's corrected before ga 15:08:54 aaahhh, K makes sense now jortel 15:08:58 #agreed high/med, accept, mark as GA blocker 15:08:58 !accept 15:08:59 dawalker: Current proposal accepted: high/med, accept, mark as GA blocker 15:09:00 dawalker: 2 issues left to triage: 4077, 4092 15:09:00 #topic pulpcore can't be installed in a Django site that also ships a a custom User object - http://pulp.plan.io/issues/4077 15:09:01 Issue #4077 [POST] (unassigned) - Priority: Normal | Severity: Medium 15:09:02 pulpcore can't be installed in a Django site that also ships a a custom User object - http://pulp.plan.io/issues/4077 15:09:05 where do I mark as ga blocker? 15:09:19 milan: yeah, that's what I meant by "known". 15:09:31 POST accept 15:09:51 #idea Proposed for #4077: Leave the issue as-is, accepting its current state. 15:09:51 !propose accept 15:09:51 dawalker: Proposed for #4077: Leave the issue as-is, accepting its current state. 15:10:11 +1 15:10:28 +1 15:10:35 #agreed Leave the issue as-is, accepting its current state. 15:10:35 !accept 15:10:35 dawalker: Current proposal accepted: Leave the issue as-is, accepting its current state. 15:10:37 #topic redirect_host defaults to socket.getfqdn() instead of None - http://pulp.plan.io/issues/4092 15:10:37 dawalker: 1 issues left to triage: 4092 15:10:39 Issue #4092 [POST] (evgeni) - Priority: Normal | Severity: Medium 15:10:40 redirect_host defaults to socket.getfqdn() instead of None - http://pulp.plan.io/issues/4092 15:10:49 that's mine <3 15:11:25 thanks Zhenech 15:12:09 nice bug report Zhenech 15:12:34 #idea Proposed for #4092: Leave the issue as-is, accepting its current state. 15:12:34 !propose accept 15:12:35 dawalker: Proposed for #4092: Leave the issue as-is, accepting its current state. 15:12:35 Zhenech ++ I think it's legit 15:12:45 +1 15:13:42 Yeah I think accept, and if necessary further discussion can be had on the PR. 15:14:04 #agreed Leave the issue as-is, accepting its current state. 15:14:04 !accept 15:14:04 dawalker: Current proposal accepted: Leave the issue as-is, accepting its current state. 15:14:05 dawalker: No issues to triage. 15:14:11 #endmeeting 15:14:11 !end