15:30:36 <fao89> #startmeeting Pulp Triage 2019-11-22
15:30:36 <fao89> #info fao89 has joined triage
15:30:36 <fao89> !start
15:30:36 <pulpbot> Meeting started Fri Nov 22 15:30:36 2019 UTC.  The chair is fao89. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:30:36 <pulpbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:30:36 <pulpbot> The meeting name has been set to 'pulp_triage_2019-11-22'
15:30:36 <pulpbot> fao89: fao89 has joined triage
15:31:28 <mikedep333> #info mikedep333 has joined triage
15:31:28 <mikedep333> !here
15:31:28 <pulpbot> mikedep333: mikedep333 has joined triage
15:31:43 <bmbouter> #info bmbouter has joined triage
15:31:43 <bmbouter> !here
15:31:43 <pulpbot> bmbouter: bmbouter has joined triage
15:31:47 <ipanova> jjeffers: i don't think this is considered nil value "proxy_password" : "",
15:31:59 <fao89> !next
15:32:00 <pulpbot> fao89: 4 issues left to triage: 5775, 5769, 5767, 5760
15:32:00 <fao89> #topic https://pulp.plan.io/issues/5775
15:32:01 <pulpbot> RM 5775 - mdepaulo@redhat.com - NEW - pulp-operator CI outputs the several-MB .iso rather than downloading it to disk
15:32:02 <pulpbot> https://pulp.plan.io/issues/5775
15:32:20 <fao89> #idea Proposed for #5775: Leave the issue as-is, accepting its current state.
15:32:20 <fao89> !propose accept
15:32:20 <pulpbot> fao89: Proposed for #5775: Leave the issue as-is, accepting its current state.
15:33:04 <jjeffers> ipanova: when I did pass a nil value to (runcible), the end result was a "" value in the db. It's possible that there's a better way to express "set this to no value"
15:33:43 <ipanova> jjeffers: i'm also not sure if we handle the situation of nullifying the password values, updating yes, but not setting to empty
15:34:08 <ipanova> fao89: +-1
15:34:10 <ipanova> +1
15:34:16 <ipanova> #info ipanova has joined triage
15:34:16 <ipanova> !here
15:34:16 <pulpbot> ipanova: ipanova has joined triage
15:34:33 <dkliban> !+1
15:34:33 <pulpbot> dkliban: Error: "+1" is not a valid command.
15:34:36 <dkliban> lol
15:34:40 <dkliban> #info dkliban has joined triage
15:34:40 <dkliban> !here
15:34:40 <pulpbot> dkliban: dkliban has joined triage
15:35:03 <fao89> #agreed Leave the issue as-is, accepting its current state.
15:35:03 <fao89> !accept
15:35:03 <pulpbot> fao89: Current proposal accepted: Leave the issue as-is, accepting its current state.
15:35:04 <fao89> #topic https://pulp.plan.io/issues/5769
15:35:04 <pulpbot> fao89: 3 issues left to triage: 5769, 5767, 5760
15:35:05 <pulpbot> RM 5769 - mdepaulo@redhat.com - NEW - ansible molecule CI seems slower than it should be
15:35:06 <pulpbot> https://pulp.plan.io/issues/5769
15:35:36 <mikedep333> FYI: These are part of an Epic of misc CI improvements (some badly needed). The idea was that fao89 would work on them.
15:35:55 <dkliban> cool
15:36:11 <bmbouter> +1
15:36:22 <ipanova> +1
15:36:27 <dkliban> let's accept
15:36:33 <fao89> I will work on them as soon as I finish to find the leak on rpm sync
15:36:38 <fao89> #idea Proposed for #5769: Leave the issue as-is, accepting its current state.
15:36:38 <fao89> !propose accept
15:36:38 <pulpbot> fao89: Proposed for #5769: Leave the issue as-is, accepting its current state.
15:36:47 <fao89> #agreed Leave the issue as-is, accepting its current state.
15:36:47 <fao89> !accept
15:36:47 <pulpbot> fao89: Current proposal accepted: Leave the issue as-is, accepting its current state.
15:36:48 <fao89> #topic https://pulp.plan.io/issues/5767
15:36:49 <pulpbot> fao89: 2 issues left to triage: 5767, 5760
15:36:50 <pulpbot> RM 5767 - mdepaulo@redhat.com - NEW - Ansible roles fail to evaluate pulp_settings_file when dynamically included
15:36:51 <pulpbot> https://pulp.plan.io/issues/5767
15:37:24 <mikedep333> fao89: We were going to hold a meeting. There we can plan them based on (soft) dependencies, your familiarity, how long they'll take, how badly needed they are, etc.
15:37:41 <fao89> nice!
15:37:45 <fao89> mikedep333++
15:37:45 <pulpbot> fao89: mikedep333's karma is now 41
15:37:56 <mikedep333> :)
15:38:03 <fao89> #idea Proposed for #5767: Leave the issue as-is, accepting its current state.
15:38:03 <fao89> !propose accept
15:38:04 <pulpbot> fao89: Proposed for #5767: Leave the issue as-is, accepting its current state.
15:39:05 <dkliban> +1
15:39:07 <fao89> #agreed Leave the issue as-is, accepting its current state.
15:39:07 <fao89> !accept
15:39:07 <pulpbot> fao89: Current proposal accepted: Leave the issue as-is, accepting its current state.
15:39:07 <ipanova> +1
15:39:08 <fao89> #topic https://pulp.plan.io/issues/5760
15:39:08 <pulpbot> fao89: 1 issues left to triage: 5760
15:39:09 <pulpbot> RM 5760 - iballou - POST - Cannot list repository versions using Ruby bindings
15:39:10 <pulpbot> https://pulp.plan.io/issues/5760
15:39:21 <dkliban> this should be in modified
15:39:25 <dkliban> i'll update it
15:39:41 <fao89> #idea Proposed for #5760: Leave the issue as-is, accepting its current state.
15:39:41 <fao89> !propose accept
15:39:41 <pulpbot> fao89: Proposed for #5760: Leave the issue as-is, accepting its current state.
15:39:46 <fao89> #agreed Leave the issue as-is, accepting its current state.
15:39:46 <fao89> !accept
15:39:46 <pulpbot> fao89: Current proposal accepted: Leave the issue as-is, accepting its current state.
15:39:47 <pulpbot> fao89: No issues to triage.
15:40:08 <fao89> Open floor!
15:40:58 <mikedep333> ipanova raised the issue today that CI needs to use the correct branch for pulpcore: 3.0 or master, depending on the plugin's needs.
15:41:52 <dkliban> yeah ... can we file an issue for that?
15:42:04 <dkliban> that's badly needed now
15:42:05 <mikedep333> I'm sure I can update plugin-template by the end of the day with this, but I want to figure out (with dkliban's advice?) whether we should try to automatically determine which branch to use (by reading setup.py?) or by having a travis template variable. And then each branch for the plugin would set that variable differently.
15:42:09 <mikedep333> Yeah, 1 sec
15:42:33 <dkliban> yeah ... we should just have it set in the the template_config.yml
15:43:07 <ipanova> mikedep333: thank you
15:44:27 <dkliban> mikedep333: i am about to file an issue to make change how pulp-operator generates teh settings file for pulp. it currently tries to create a Python file called settings.py. However, our template for that is better fitted to generate JSON.
15:44:48 <dkliban> since dynaconf can understand JSON, we should instead be generating a JSON file called settings.json
15:45:45 <mikedep333> https://pulp.plan.io/issues/5782
15:46:08 <dkliban> thank you!
15:46:11 <mikedep333> dkliban: sounds good. I prefer JSON anyway for easier yaml conversion.
15:46:53 <bmbouter> +1 to all this
15:47:14 <dkliban> bmbouter: what do you have for open floor?
15:47:57 <mikedep333> This does lead to a related point on my mind: In the near future, pulp-ansible and pulpcore.git/containers (to be moved to pulp-operator.git/something) will have different codepaths for 3.0 vs master. However, I will have to detect the version in 1 of 3 different ways:
15:48:16 <mikedep333> 1. Have the user specify it ahead of time as a variable.
15:48:42 <bmbouter> I don't have open floor topics atm
15:49:14 <mikedep333> 2. Detect it early on, before pulpcore gets installed, such as by reading from PyPI or a git repo, not just a folder on disk. This is needed (for example) for handling dependencies early on in the install; before pip env gets created.
15:49:39 <mikedep333> 3. Just detect it after pulpcore gets installed (not sufficient due to #2's description.)
15:49:50 <mikedep333> If anyone has any advice, let me know.
15:50:14 <dkliban> mikedep333: some mix of 1 and 2 i think
15:50:28 <dkliban> i just dont' want the user to have to specify too much
15:50:52 <mikedep333> dkliban: Yeah. I want an easy user experience, but don't want to kill myself. I'll probably have the variable set to "auto", but only auto-detect it when it is feasible in the short-term. In the long-term, I'll auto-detect it under more circumstances.
15:52:15 <mikedep333> I'm also concerned about plugins battling over the version of pulpcore. In case users accidentally specify a 3.0 plugin and a master branch plugin. We may want to lock/pin it (to use .deb terminology) or something.
15:53:12 <mikedep333> s/kill myself/overwork myself writing tons of pre-emptive logic./
15:53:17 <dkliban> yeah
15:54:41 <dkliban> it's kind of tough ... i think it's up to the user who is building the container to select the right versions of plugins
15:55:38 <iballou> hey, is there no modify for ansible repos? https://pulp-ansible.readthedocs.io/en/latest/restapi.html#tag/repositories:-ansible
15:56:27 <bmbouter> we need to add it
15:56:34 <bmbouter> iballou: can you file an issue so we can?
15:56:38 <iballou> sure thing bmbouter
15:58:01 <fao89> #endmeeting