13:03:04 <mvollmer> #startmeeting meeting
13:03:04 <zodbot> Meeting started Mon Sep 12 13:03:04 2016 UTC.  The chair is mvollmer. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:03:04 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
13:03:04 <zodbot> The meeting name has been set to 'meeting'
13:03:09 <mvollmer> .hello mvo
13:03:10 <zodbot> mvollmer: mvo 'Marius Vollmer' <marius.vollmer@gmail.com>
13:03:13 <andreasn1> .hello andreasn
13:03:13 <zodbot> andreasn1: andreasn 'Andreas Nilsson' <anilsson@redhat.com>
13:03:17 <dperpeet> .hello dperpeet
13:03:19 <zodbot> dperpeet: dperpeet 'None' <dperpeet@redhat.com>
13:03:58 <stefw> .hello stefw
13:03:59 <zodbot> stefw: stefw 'Stef Walter' <stefw@redhat.com>
13:04:13 <mvollmer> #topic Agenda
13:05:15 <larsu> .hello larsu
13:05:16 <zodbot> larsu: larsu 'Lars Karlitski' <lars@karlitski.net>
13:07:02 <dperpeet> * cockpit version number
13:07:15 <dperpeet> * log pruning
13:07:41 <dperpeet> * known issues
13:08:11 <mvollmer> * pcp (if time)
13:08:26 <mvollmer> alright
13:08:34 <mvollmer> #topic cockpit version number
13:09:03 <dperpeet> cockpit has dropped the 0.x
13:09:10 <andreasn1> #info https://lists.fedorahosted.org/archives/list/cockpit-devel@lists.fedorahosted.org/thread/UTSBRX6XQ2GJY4EHNY7DFOVE5OBITICN/
13:09:28 <dperpeet> nothing ground to a halt as a result so far
13:09:31 <dperpeet> from what I can tell
13:09:40 <andreasn1> so it went from 0.117 to 118
13:09:45 <mvollmer> people are missing the ".0" moment
13:09:48 <dperpeet> https://github.com/cockpit-project/cockpit/releases/tag/118 looks nicer than 0.x
13:09:53 <mvollmer> sorry, the "1.0" moment
13:10:16 <mvollmer> can we do something instead?
13:10:21 <mvollmer> code names once in a while?
13:10:32 <andreasn1> yeah, Josh brought that up in the thread, but I think that's a thing we can control to some extent as well
13:10:38 <dperpeet> cockpit 119 'funny monkey'?
13:10:48 <andreasn1> there can be a big bang if we make it a big bang :)
13:11:17 <mvollmer> or milestones like cockpit is default in fedora server
13:11:25 <mvollmer> cockpit has reachd debian unstable
13:11:29 <mvollmer> things like that
13:11:31 <dperpeet> are there any blogs/posts that need to be written?
13:11:36 <stefw> dperpeet, yup
13:11:38 <stefw> two things
13:11:43 <stefw> 1. what we define as stable
13:11:50 <stefw> 2. and a general post saying "yay"
13:12:16 <dperpeet> are there places where we will collaborate on any of these, or are you on it stefw ?
13:12:49 <mvollmer> i think josh was treating cockpit as a kinf of add-on application
13:12:57 <stefw> i don't have time to do it this week or next
13:13:02 <mvollmer> that people go and download and try out
13:13:55 <aweiteka> maybe this is part of the current discussion… anyone know when release 118 will be pushed to docker hub? the build failed: https://hub.docker.com/r/cockpit/kubernetes/builds/bj6njmsjrqsuxpwkx39vghg/
13:14:00 <dperpeet> the way I see it, we should go for the content of the "what we define as stable" post first
13:14:31 * aweiteka didn’t realize this was an official call. sorry
13:15:16 <dperpeet> aweiteka, I think 118 was only built for fc25
13:15:21 <dperpeet> we can tackle that after the meeting
13:15:57 <dperpeet> let's see if we can get the posts stefw mentioned started in bullet points this week
13:16:13 <dperpeet> I'm happy to flesh those out once we agree on the content
13:16:49 <dperpeet> and I'm glad Cockpit dropped the 0.x!
13:17:45 <mvollmer> yes, me too
13:18:14 <mvollmer> okay, next?
13:18:29 <mvollmer> #topic log pruning
13:18:54 <dperpeet> mvollmer reviewed and merged the changes to our log sink https://github.com/cockpit-project/cockpituous/pull/32
13:18:57 <dperpeet> thanks!
13:19:18 <dperpeet> now our logs on fedorapeople should be deleted after 30 days
13:19:23 <dperpeet> how often does the pruning run now?
13:19:31 <mvollmer> twice a day
13:19:36 <dperpeet> our logs folder went from 70G to 6.9G
13:19:42 <dperpeet> a lot more manageable
13:20:05 <dperpeet> I also saved the raw log files (along with index.html) to my local machines in case we need some statistics
13:20:10 <mvollmer> as an interesting side story, I found a change in the production code that wasn't in any commit...
13:20:11 <dperpeet> not sure yet if we'll actually use that
13:20:39 <petervo> mvollmer, cockpit or the sink?
13:20:43 <dperpeet> sink
13:21:02 <mvollmer> it was probably a change that I made when testing a new feature in production
13:21:08 <mvollmer> and then got interrupted etc
13:21:20 <mvollmer> so, example of worst practices
13:21:24 <mvollmer> hrr
13:21:39 <dperpeet> thanks for deploying the new sink, mvollmer :)
13:21:51 <mvollmer> i imported that change to the repo now
13:22:17 <mvollmer> https://github.com/cockpit-project/cockpituous/commit/4c23d67c619001c605423c4eb55bc756506ae067
13:23:06 <mvollmer> also, there are a couple of files in the logs dir that are now owned by "cockpit"
13:23:10 <mvollmer> those will never be pruned
13:23:25 <mvollmer> very few, so we can just ignore that, I guess
13:23:32 <dperpeet> we can take care of that manually
13:24:00 <dperpeet> mvollmer, you mean ones that *aren't* owned by cockpit
13:24:08 <mvollmer> yes, sorry
13:24:12 <mvollmer> "not"
13:24:16 <dperpeet> those are stef's
13:24:18 <dperpeet> I htink
13:24:19 <dperpeet> *think
13:25:06 <dperpeet> end of topic?
13:26:55 <mvollmer> #topic known issues
13:27:11 <dperpeet> just a quick thanks to stef for removing a bunch of old known issues today
13:27:30 <dperpeet> I removed a few recently and I'm happy to see that list actually get shorter for once
13:27:36 <dperpeet> yay!
13:27:47 <dperpeet> end of topic :)
13:28:21 <andreasn1> yay
13:28:23 <mvollmer> cool
13:28:36 <mvollmer> #topic pcp
13:28:45 <mvollmer> there was some renewed interest in pcp
13:29:12 <andreasn1> where?
13:29:16 <mvollmer> https://github.com/cockpit-project/cockpit/issues/4941
13:29:21 <mvollmer> sorry, had to dig out the link
13:29:37 <andreasn1> #info https://github.com/cockpit-project/cockpit/issues/4941
13:29:49 <mvollmer> some people are thinking of adding more pcp specifics to Cockpit, such as changing the default sampling rate
13:30:02 <stefw> Lets start from the basic point that:
13:30:19 <stefw> Our integration with PCP is currently implemented in such a way that goes counter to the way the rest of Cockpit is built.
13:30:33 <stefw> We normally only consider APIs as those that are remotable
13:30:42 <stefw> the PCP API is not remotable ... in our books it's not an API
13:30:48 <mvollmer> hmm.
13:30:59 <stefw> it's our big exception ... and we need to consider a plan for reconciling that
13:31:10 <stefw> before we dig ourselves deeper into that hole
13:31:28 <stefw> not to say such work is a blocker for fixing bugs
13:31:33 <stefw> but at least there should be a plan
13:31:37 <mvollmer> can you elaborate?  what is not remoteable?
13:32:00 <stefw> it is the only system monitoring/configuration/management subsystem by which we push bits onto the C stack
13:32:05 <stefw> and make actual C function calls
13:32:22 <mvollmer> ahh, I see
13:32:28 <stefw> As a general architecture, Cockpit interacts with the system from javascript
13:32:38 <stefw> and the binary parts are only there to facilitate that
13:32:55 <stefw> our binary parts typically have no knowledge of how to configure a system, or monitor a system
13:33:02 <stefw> the PCP, and to a lesser extent metrics stuff
13:33:19 <stefw> are the areas where we compromise on that
13:34:07 <mvollmer> yeah, makes sense
13:36:11 <andreasn1> what is the exact UI they are asking for? I didn't read through the entire thread
13:36:25 <mvollmer> nothing concrete
13:36:39 <mvollmer> https://github.com/cockpit-project/cockpit/issues/4941#issuecomment-246273069
13:37:07 <andreasn1> "a rich UI for optional pmmgr use someday"
13:37:23 <andreasn1> sounds like it's not critical right now at least
13:37:33 <mvollmer> i agree
13:38:38 <andreasn1> what was the connection with the original issue?
13:38:44 <andreasn1> why didn't it work properly?
13:39:06 <petervo> pcp crashing because of a missing / malformed file
13:39:26 <mvollmer> that's what I found, not sure if that was the original issue
13:39:34 <mvollmer> quite likely, though
13:39:56 <mvollmer> and malformed in this case meant: zero bytes
13:40:04 <mvollmer> err, I mean size 0.
13:40:28 <mvollmer> those files appear all the time after a unclean shutdown
13:40:42 <stefw> andreasn1, as far as "not-critical" ... our response should be
13:40:45 <stefw> make a remotable API
13:40:55 <stefw> and it'll be easy to build a rich UI for pmmgr use
13:41:09 <stefw> whether a DBus, REST, cmdline+JSON ... anything
13:41:15 <stefw> in 2016 C APIs aren't
13:41:35 <andreasn1> yes, and I think for the use case " the Red Hat customer support folk would like an easy way to query and set the default pmlogger recording interval for a given site"
13:41:59 <andreasn1> that sounds a bit outside the core cockpit use case, but could still be useful to make into a plugin that someone else maintains
13:45:00 <mvollmer> done?
13:45:20 <andreasn1> I think so
13:45:23 <mvollmer> #topic any other biz
13:47:34 <mvollmer> okay, that's it?
13:47:40 <andreasn1> I think so
13:47:43 <mvollmer> #endmeeting