15:33:08 <daviddavis> #startmeeting Pulp Triage 2019-11-08
15:33:08 <daviddavis> !start
15:33:08 <daviddavis> #info daviddavis has joined triage
15:33:08 <pulpbot> Meeting started Fri Nov  8 15:33:08 2019 UTC.  The chair is daviddavis. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:33:08 <pulpbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:33:08 <pulpbot> The meeting name has been set to 'pulp_triage_2019-11-08'
15:33:08 <pulpbot> daviddavis: daviddavis has joined triage
15:33:20 <ggainey> #info ggainey has joined triage
15:33:20 <ggainey> !here
15:33:20 <pulpbot> ggainey: ggainey has joined triage
15:33:41 <bmbouter> #info bmbouter has joined triage
15:33:41 <bmbouter> !here
15:33:41 <pulpbot> bmbouter: bmbouter has joined triage
15:33:43 <dkliban> #info dkliban has joined triage
15:33:43 <dkliban> !here
15:33:43 <pulpbot> dkliban: dkliban has joined triage
15:33:47 <ipanova> #info ipanova has joined triage
15:33:47 <ipanova> !here
15:33:47 <pulpbot> ipanova: ipanova has joined triage
15:33:52 <mikedep333> #info mikedep333 has joined triage
15:33:52 <mikedep333> !here
15:33:52 <pulpbot> mikedep333: mikedep333 has joined triage
15:34:02 <daviddavis> !next
15:34:03 <pulpbot> daviddavis: 4 issues left to triage: 5678, 5673, 5668, 5666
15:34:04 <daviddavis> #topic https://pulp.plan.io/issues/5678
15:34:04 <pulpbot> RM 5678 - jonasbartho - NEW - flatpak functionality update?
15:34:05 <pulpbot> https://pulp.plan.io/issues/5678
15:34:21 <ttereshc> #info ttereshc has joined triage
15:34:21 <ttereshc> !here
15:34:21 <pulpbot> ttereshc: ttereshc has joined triage
15:34:43 <mikedep333> 5678 would have been lovely at my last job; managing a fleet of Linux workstations :)
15:34:54 <daviddavis> there is no story tracking flatpak so I can convert this into a story?
15:34:57 <ggainey> change to story?
15:34:59 <ggainey> heh
15:35:00 <dkliban> yeah
15:35:02 <ggainey> gmta
15:35:02 <dkliban> +1
15:35:03 <ipanova> convert to story? 'as a user I can manage'..
15:35:03 <ttereshc> yeah +1 to story
15:35:09 <daviddavis> +1
15:35:17 <bmbouter> +1
15:35:20 <daviddavis> #idea Proposed for #5678: convert to a story
15:35:20 <daviddavis> !propose other convert to a story
15:35:20 <pulpbot> daviddavis: Proposed for #5678: convert to a story
15:35:31 <daviddavis> #agreed convert to a story
15:35:31 <daviddavis> !accept
15:35:31 <pulpbot> daviddavis: Current proposal accepted: convert to a story
15:35:32 <pulpbot> daviddavis: 3 issues left to triage: 5673, 5668, 5666
15:35:32 <daviddavis> #topic https://pulp.plan.io/issues/5673
15:35:33 <pulpbot> RM 5673 - osapryki - NEW - Resource reservations are not cleaned up if worker is killed
15:35:34 <pulpbot> https://pulp.plan.io/issues/5673
15:35:46 <mikedep333> I just want to make sure we don't do "cookie licking" (I think that's the term.) And make the user think we're working on it, if we will not work on it anytime soon.
15:35:56 <mikedep333> (for 5678)
15:35:58 <bmbouter> agreed
15:36:14 <daviddavis> mikedep333: I'll add a comment saying that there are no firm plans to work on it but we'll update the story if that changes
15:36:21 <mikedep333> daviddavis: Thanks
15:36:22 <dkliban> we should put a comment on there that we are looking for someone to do this work
15:36:27 <daviddavis> +1
15:36:28 <bmbouter> yes
15:36:33 <bmbouter> and also we should try to maybe seed it?
15:36:41 <bmbouter> this is the advice I've been getting from GDK
15:36:55 <daviddavis> I can maybe take a look after dev freeze
15:37:48 <ttereshc> #idea Proposed for #5673: Leave the issue as-is, accepting its current state.
15:37:48 <ttereshc> !propose accept
15:37:48 <pulpbot> ttereshc: Proposed for #5673: Leave the issue as-is, accepting its current state.
15:37:53 <daviddavis> +1
15:37:55 <bmbouter> also
15:38:00 <ipanova> +1
15:38:03 <dkliban> +1
15:38:08 <bmbouter> we're planning to fix this post branch and before GA
15:38:17 <daviddavis> bmbouter: agreed
15:38:17 <bmbouter> so maybe add 3.0 blocker and label bugfix
15:38:29 <bmbouter> this was our agreement w/ AH since they implmemented a temporary workaround in their environment
15:38:55 <daviddavis> #idea Proposed for #5673: accept and add to 3.0 milestone
15:38:55 <daviddavis> !propose other accept and add to 3.0 milestone
15:38:55 <pulpbot> daviddavis: Proposed for #5673: accept and add to 3.0 milestone
15:38:59 <dkliban> cool
15:39:02 <ggainey> +1
15:39:07 <ttereshc> how do you label a bugfix? I thought that the fact that it's an "issue" signifies that it will be a bugfix
15:39:12 <daviddavis> me too
15:39:22 <bmbouter> that's what I mean
15:39:25 <bmbouter> I agree w/ you
15:39:28 <ttereshc> ok
15:39:34 <bmbouter> I hadn't looked at the type
15:39:36 <daviddavis> bmbouter just likes to confuse us
15:39:41 <bmbouter> !friday
15:39:41 <pulpbot> ♪ It's Friday, Friday, gotta get down on Friday ♪
15:39:44 <daviddavis> :D
15:39:47 <bmbouter> sing it with me!
15:39:52 <dkliban> lol
15:39:57 <daviddavis> I'm already singing it
15:40:05 <dawalker> !dance
15:40:05 <pulpbot> ♪┏(°.°)┛┗(°.°)┓┗(°.°)┛┏(°.°)┓ ♪
15:40:12 * bmbouter moonwalks
15:40:21 <daviddavis> #agreed accept and add to 3.0 milestone
15:40:21 <daviddavis> !accept
15:40:21 <pulpbot> daviddavis: Current proposal accepted: accept and add to 3.0 milestone
15:40:22 <daviddavis> #topic https://pulp.plan.io/issues/5668
15:40:22 <pulpbot> daviddavis: 2 issues left to triage: 5668, 5666
15:40:23 <pulpbot> RM 5668 - mdellweg - NEW - Using futures in sysncpipeline basically deactivates batching
15:40:23 <ttereshc> fun fun fun fun
15:40:24 <pulpbot> https://pulp.plan.io/issues/5668
15:41:08 <ggainey> interesting - I'm a big fan of code that helps avoid deadlocks, to be sure
15:41:23 <ggainey> and there's already a PR - accept, add to sprint?
15:41:24 <bmbouter> we should accept at least. x9c4 and I are going to meet to discuss this either todya or monday
15:41:28 <ggainey> kk
15:41:36 <bmbouter> also add ot sprint is good
15:41:37 <dkliban> +1 to accept
15:42:01 <ttereshc> +!
15:42:03 <ttereshc> +1
15:42:05 <daviddavis> #idea Proposed for #5668: Leave the issue as-is, accepting its current state.
15:42:05 <daviddavis> !propose accept
15:42:06 <pulpbot> daviddavis: Proposed for #5668: Leave the issue as-is, accepting its current state.
15:42:13 <bmbouter> +1
15:42:21 <ggainey> kk
15:42:23 <ggainey> +1
15:42:31 <dkliban> +1
15:42:31 <daviddavis> #agreed Leave the issue as-is, accepting its current state.
15:42:31 <daviddavis> !accept
15:42:31 <pulpbot> daviddavis: Current proposal accepted: Leave the issue as-is, accepting its current state.
15:42:32 <pulpbot> daviddavis: 1 issues left to triage: 5666
15:42:32 <daviddavis> #topic https://pulp.plan.io/issues/5666
15:42:33 <pulpbot> RM 5666 - ironfroggy - NEW - Failed imports leave orphaned artifacts blocking subsequent attempts
15:42:34 <pulpbot> https://pulp.plan.io/issues/5666
15:43:15 <bmbouter> this is pulp_ansible specific
15:43:28 <daviddavis> wouldn't this affect uploads as well?
15:43:33 <dkliban> that's what i thought
15:43:38 <ttereshc> me too
15:43:41 * ttereshc listens
15:43:52 <daviddavis> suppose I upload an artifact to make a content unit and it creates the artifact but not the content unit
15:43:55 <daviddavis> and then I try again
15:44:15 <bmbouter> mmm that could be but then we have 2 issues
15:44:44 <daviddavis> I'm ok to convert this to pulp_ansible
15:44:52 <bmbouter> let me look for a sec
15:44:53 <daviddavis> and wait until someone reports a bug with the upload
15:44:55 <daviddavis> ok
15:45:21 <bmbouter> https://github.com/pulp/pulp_ansible/blob/master/pulp_ansible/app/galaxy/v3/views.py#L147-L150
15:45:39 <bmbouter> I agree we could have a similar issue in pulpcore
15:45:50 <bmbouter> daviddavis: that plan sounds ok to me
15:46:07 <daviddavis> I can look to see if there's a pulpcore issue and file one if there is
15:46:16 <ggainey> cool
15:46:16 <bmbouter> that would be great!
15:46:42 <daviddavis> #idea Proposed for #5666: daviddavis to move to pulp_ansible and investigate pulpcore
15:46:42 <daviddavis> !propose other daviddavis to move to pulp_ansible and investigate pulpcore
15:46:42 <pulpbot> daviddavis: Proposed for #5666: daviddavis to move to pulp_ansible and investigate pulpcore
15:46:47 <ttereshc> +1
15:47:00 <ggainey> +1
15:47:09 <bmbouter> +1
15:47:16 <daviddavis> #agreed daviddavis to move to pulp_ansible and investigate pulpcore
15:47:16 <daviddavis> !accept
15:47:16 <pulpbot> daviddavis: Current proposal accepted: daviddavis to move to pulp_ansible and investigate pulpcore
15:47:17 <bmbouter> +1 with thanks :)
15:47:18 <pulpbot> daviddavis: No issues to triage.
15:47:21 <daviddavis> :)
15:47:23 <daviddavis> OPEN FLOOR
15:47:25 <daviddavis> !friday
15:47:25 <pulpbot> ♪ It's Friday, Friday, gotta get down on Friday ♪
15:47:31 <bmbouter> !pulp
15:47:31 <pulpbot> 🍊 Yay, Pulp! 🍊 Go team go! 🍊
15:47:46 <daviddavis> I'll go first
15:48:08 <dalley> so, I've not worked with a plugin that uses publishers lately, but my gut feeling is that this looks wrong: https://github.com/pulp/plugin_template/blob/master/templates/docs/docs/workflows/publish.rst.j2#L13-L20
15:48:11 <daviddavis> I am merging the content origin changes after this meeting
15:48:40 <mikedep333> re: open floor: We need to start testing Pulp's installation on Fedora 31 and Ansible 2.9.
15:48:54 <bmbouter> dalley: our docs are much out of date that's going to be the big push here in the next 2-3 weeks
15:49:13 <bmbouter> dalley: want to file and issue, label it docs, and add 3.0 blocker label?
15:49:27 <daviddavis> mikedep333: when is F31 coming out?
15:49:27 <dalley> mikedep333, I have changes on my machine to add F31 to pulplift and ansible-pulp
15:49:39 <mikedep333> daviddavis: It came out a week ago.
15:49:41 <bmbouter> it's out is my understanding
15:49:50 <daviddavis> ah ok wow
15:50:00 <dalley> mikedep333, I also filed this https://github.com/geerlingguy/ansible-role-postgresql/pull/101
15:50:04 <dalley> which we're blocked on
15:50:10 <mikedep333> Ansible 2.9 isn't in Fedora yet, but it is released, and on PyPI.
15:50:51 <mikedep333> dalley: We can get in touch with geerlingguy easily. He has high code standards for his role, but I suspect he'll quickly merge that.
15:51:38 <dalley> bmbouter, should I just not bother with fixing the docs for typed repositories right now then?
15:51:42 <mikedep333> I say we nag him if he doesn't reply by the end of your workday.
15:52:02 <bmbouter> dalley: yes I think the code is most important atm
15:52:07 <dalley> bmbouter, ack
15:52:08 <mikedep333> dalley Thank you for submitting it though!
15:52:11 <mikedep333> dalley++
15:52:11 <pulpbot> mikedep333: dalley's karma is now 182
15:52:24 <bmbouter> mikedep333: if no reply today I'll reach out to geerling
15:52:29 * bmbouter takes AI
15:54:11 <bmbouter> daviddavis: +1 to merging CONTENT_ORIGIN after this
15:54:24 <ggainey> +1
15:54:28 * ggainey cheers wildly
15:54:31 <daviddavis> bmbouter: ok, plz review the pulp_ansible changes
15:54:40 <dalley> +1 to content_origin
15:55:26 <daviddavis> bmbouter mikedep333 I'm not totally clear on https://github.com/pulp/pulplift/pull/53 but it sounds like we have to set it in the pulp settings for now
15:55:28 <mikedep333> +1 to merging CONTENT_ORIGIN
15:55:37 <daviddavis> otherwise we need to add logic to the installer to figure it out?
15:56:09 <mikedep333> daviddavis: The ansible-pulp logic for this is in the example vars file, rather than the role itself. Putting it in the role itself is more work.
15:56:12 <bmbouter> my concern is that if we need to add those vars to pulplift, then users only using the installer (not pulplift) will also set them which I don't think is ok
15:56:20 <mikedep333> I left the ansible-pulp PR in an RFC state for a few weeks because of that.
15:56:50 <bmbouter> mikedep333: does a user need to set CONTENT_ORIGIN when they use the installer atm?
15:56:53 <mikedep333> https://github.com/pulp/ansible-pulp/pull/185/files
15:57:13 <mikedep333> bmbouter: Yes, just as they need to set pulp_settings.secret_key
15:57:42 <bmbouter> mmmm true
15:57:57 <daviddavis> that's what I had expected
15:57:58 <bmbouter> but it uses "http://{{ ansible_fqdn }}" I thought that would handle it automatically?
15:58:13 <mikedep333> So that is set in the example vars files.
15:58:19 <mikedep333> pulplift does not use the example vars files.
15:58:27 <bmbouter> ic
15:58:55 <bmbouter> \quit
15:59:00 <bmbouter> lol brb
15:59:07 <daviddavis> I'll merge the pulplift change and we can revisit later
15:59:12 <daviddavis> anything else for open floor?
15:59:14 <mikedep333> Great
15:59:42 <daviddavis> bmbouter: we just agreed without you to merge the pulplift change and revisit later
15:59:44 <daviddavis> :)
15:59:55 <bmbouter> I'm +1 on that anyway
15:59:58 <mikedep333> I need someone with github admin perms to help me rename the repo: https://pulp.plan.io/issues/5619?issue_count=48&issue_position=4&next_issue_id=3809&prev_issue_id=5642
16:00:15 <ttereshc> I can do that
16:00:17 <ttereshc> mikedep333,
16:01:01 <mikedep333> ttereshc: Let me doe some last bits of research on how to make things simpler for users, and then I'll follow up with you in like an hour. Is that OK? Or too late?
16:01:18 <daviddavis> mikedep333: if ttereshc is not around, I will be
16:01:27 <ttereshc> mikedep333, it's ok, I'm working next 4 hours
16:01:36 <mikedep333> Thanks
16:01:43 <daviddavis> last call for open floor
16:01:50 <bmbouter> what do folks this about this question: https://www.redhat.com/archives/pulp-list/2019-November/msg00010.html
16:02:36 <ttereshc> is it to allow one publication per repo version?
16:02:46 <ttereshc> do I understand the question correclty
16:03:00 <daviddavis> I think it's to enforce one pub per repo version
16:03:09 <dkliban> the question is 'what is the use case for creating a bunch of publications from the same repo version?'
16:03:26 <daviddavis> I think there's no guarantee that two publications from the same repo version will be the same
16:03:35 <daviddavis> like suppose the metadata has a timestamp
16:03:46 <dkliban> yeah ... that's true
16:03:47 <bmbouter> that's true
16:03:54 <bmbouter> but is that useful?
16:04:02 <ttereshc> why would you want to have multiple publicaitons?
16:04:06 <bmbouter> as in is there a use case that would benefit that?
16:04:06 <ttereshc> exactly
16:04:13 <bmbouter> I can't think of one
16:04:19 <dkliban> same here
16:04:21 <daviddavis> maybe, suppose there's some options when doing a publish
16:04:25 <ttereshc> maybe if there are publishers
16:04:34 <ttereshc> and different configuraiton
16:04:36 <daviddavis> you want to publish the same repo version but two different sets of options
16:04:43 <dkliban> yeah
16:04:48 <bmbouter> mmmm that makes sense
16:04:54 <daviddavis> can plugins set a unique constraint today?
16:04:59 <daviddavis> if they choose
16:05:15 <bmbouter> let's look
16:05:15 <ttereshc> but that's in theory, do you have a specific use case? what kind of options?
16:05:27 <ttereshc> different checksum types for metadata?
16:05:35 <ttereshc> in rpm case
16:05:36 <bmbouter> yeah or different signing keys
16:05:39 <daviddavis> filtering out certain content types?
16:05:52 <ttereshc> yeah, good examples
16:06:37 <bmbouter> I can reply to Bin w/ this info if that' shelpful
16:06:43 <daviddavis> +1
16:06:48 * bmbouter takes AI
16:06:52 <daviddavis> bmbouter++
16:06:52 <pulpbot> daviddavis: bmbouter's karma is now 207
16:06:59 <daviddavis> ok, we're 6 min past 11am
16:07:04 <daviddavis> anything else for open floor?
16:07:10 <bmbouter> yes...
16:07:29 <bmbouter> just to advertise plugin writers look at this PR to see how they will be needing to adjust their code (asking for feedback on it)
16:07:42 <bmbouter> https://github.com/pulp/pulp_file/pull/307/files
16:07:54 <bmbouter> that illustrates the sync and upload work
16:08:13 <bmbouter> the "modify" endpoint is blocked until typed repos is resolved b/c it reworks that code significantly
16:08:23 <daviddavis> does this introduce breaking changes?
16:08:42 <bmbouter> it does in two ways
16:08:55 <bmbouter> 1) plugin writers need to always create the RepositoryVersion and finalize it
16:09:05 <bmbouter> you can see that in core here with s/repository/respotiory_version/ https://github.com/pulp/pulpcore/pull/369/files#diff-fe9949c7b0bac1298ede788d05d7dc15R18
16:09:12 <bmbouter> which pairs w/ that pulp_file PR
16:09:20 <daviddavis> ah I see
16:09:34 <bmbouter> that was (2) ^ (the param change)
16:10:10 <bmbouter> with the amount of change we need to get the limited set of plugins we are aiming to GA w/ I don't think we can wait until PRs are available for all plugins to merge
16:10:53 <bmbouter> as in plugin CI's will likely break and be broken for a few days
16:11:04 <bmbouter> and we should prioritize the plugins we are aiming for first
16:11:34 <ttereshc> bmbouter, I think it's fine, if we notify pulp-dev
16:11:42 <ttereshc> and katello
16:11:44 <bmbouter> pre-merge notification will go out
16:12:01 <bmbouter> and I'll follow up w/ katello also
16:12:19 <ttereshc> bmbouter,  do you plan to merge it today or on Monday?
16:12:50 <bmbouter> dalley: wdyt because I could merge today but then typed repos will need some conflict resolution
16:13:35 <bmbouter> ttereshc: short answer is I don't know yet
16:14:00 <bmbouter> my goal was to bring some awareness and ask for feedback from plugin writers, so lmk later if you've got suggestions
16:14:05 <bmbouter> this accomplishes my open floor needs
16:14:26 <dalley> bmbouter, I don't want it to interfere with gmbnomis
16:14:40 <bmbouter> I agree w/ that 100%
16:14:43 <ttereshc> bmbouter, ok, I'm asking mostly because it's easier to fix it in plugins when it's merged than to work on top of the chain of PRs but it's not critical
16:15:06 <bmbouter> dalley: let's plan to wait to merge until after typed repos on that bases
16:15:07 <bmbouter> basis
16:15:53 <daviddavis> last call for open floor
16:16:11 <daviddavis> #endmeeting
16:16:11 <daviddavis> !end