20:01:25 <jokke_> #startmeeting log_wg
20:01:25 <openstack> Meeting started Wed Jul 29 20:01:25 2015 UTC and is due to finish in 60 minutes.  The chair is jokke_. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:01:26 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:01:27 <jokke_> Agenda for Wednesday 22th of July (20:00 UTC):
20:01:28 <jokke_> Attendance & courtecy calls
20:01:28 <jokke_> Action items update
20:01:28 <jokke_> empty fields in logs
20:01:29 <openstack> The meeting name has been set to 'log_wg'
20:01:30 <jokke_> incubator using projects
20:01:33 <jokke_> proposed talks for summit
20:01:35 <jokke_> Ongoing specs
20:01:38 <jokke_> latest updates from req ids
20:01:40 <jokke_> Ops midcycle planning
20:01:43 <jokke_> proposals for log_wg meeting time
20:01:45 <jokke_> Open discussion
20:01:53 <jokke_> #topic attendance and courtecy calls
20:01:58 <jokke_> Courtesy meeting reminder: jokke_, rockyg, dhellman, johnthetubaguy, nkrinner, bknudson
20:02:06 <bknudson> hi
20:02:13 <dhellmann> o/
20:02:17 <jokke_> hi there
20:02:35 <jokke_> sorry for the flood at the start :(
20:02:53 <Nikolay_St> hi
20:02:56 <jokke_> #topic ongoing specs - req id's
20:03:05 * dhellmann is distracted with some release work
20:03:18 <jokke_> dhellmann: understood and appreciated
20:03:48 <jokke_> all, did you follow up yesterday's cross project meeting?
20:04:16 <bknudson> at the meeting they said they finished their analysis
20:04:21 <jokke_> we got quick update from tpatil ... seems that the spec is pretty much on track finally
20:04:29 <jokke_> and the update was really quick
20:04:30 <bknudson> and they thought there would be a problem with generators
20:04:53 <jokke_> I haven't seeb update on that spec yet, but hopefully it's something everyone can agee on finally
20:06:30 <bknudson> y, hopefully the spec will be updated with this new design
20:07:05 <jokke_> I'm not letting it land to glanceclient before it is :D
20:07:28 <bknudson> I don't know if they plan to do all the work themselves
20:07:38 <jokke_> this has been such a long fight to get it figured out that, not gonna give up now
20:07:43 <bknudson> maybe there will be something in oslo first.
20:08:24 <bknudson> there was an apiclient package for a long time but that petered out
20:08:27 <dhellmann> bknudson: I proposed a work-around for the generator issue on the mailing list
20:08:37 <jokke_> I haven't heard from Abhishek for a while so I'm not sure what's their current plan exactly
20:08:58 <bknudson> http://git.openstack.org/cgit/openstack/oslo-incubator/tree/openstack/common/apiclient -- it's still in oslo-incubator
20:09:46 <bknudson> it's mostly deprecated...
20:11:13 <bknudson> and it's still hanging out in keystoneclient -- http://git.openstack.org/cgit/openstack/python-keystoneclient/tree/keystoneclient/openstack/common/apiclient
20:11:20 <bknudson> I think we need a 2.0 release to actually delete it
20:11:45 <jokke_> :)
20:12:13 <jokke_> we're such hoarders ... can't throw anything away :)
20:13:56 <jokke_> rockyg: you found time after all? :)
20:14:23 <jokke_> ok moving on
20:14:30 <jokke_> #topic ops midcycle planning
20:15:17 <jokke_> rockyg: anything from your side, dhellmann had pretty much plan to let the ops lead
20:16:26 <rockyg> in defcore meeting, but yeah.  I'm kinda here.
20:17:28 <rockyg> What I proposed to tom for Ops midcycle is a session or two of bug triaging with a focus of getting top log bugs identified and top five bugs for each project logging or not
20:18:19 <rockyg> also possible to get a working group on fleshing out the sections of the error code spec that have comments on it
20:18:22 <jokke_> oh, sorry
20:18:46 <jokke_> cool, that sounds like good start
20:19:11 <rockyg> Nah. It's ok.  Just a bunch of folks debating what deserves getting a trademark
20:19:29 <bknudson> only the best loggers get a trademark
20:19:38 <rockyg> Yeah.  I think I talked about something else with him, but I forget.
20:19:43 <rockyg> bknudson: ++ lol
20:20:12 <jokke_> :)
20:20:32 <rockyg> might try a working session for other log wants, but we'll see.
20:20:33 <jokke_> I'd really like to see few of them joining us
20:21:17 <jokke_> those people seems to have lots of opinions and it would be great to get few of then joining us on these weekly talks
20:21:47 <rockyg> talked about a bugs triage weekly and sending email to ops mailing list with minutes/agenda/etc
20:23:06 <jokke_> sounds heavy
20:23:44 <rockyg> nah.  Mostly status on hot bugs and any new ones that should get priorities rased.
20:24:57 <rockyg> anybody else gonna be at the ops midcycle?
20:25:17 <jokke_> I can't
20:25:41 <jokke_> maybe on the next one
20:25:55 <rockyg> Yeah.  So I'll rep us and do more recruiting.
20:27:17 <jokke_> you need anything from us before that?
20:27:20 <rockyg> anything else on the ops meeting?
20:27:27 <jokke_> anything we could help with?
20:28:35 <rockyg> take a look at the etherpad for planning and comment, maybe?
20:28:50 <rockyg> Any reviews I should highlight there?
20:28:57 <jokke_> ok, mind to share the link to the notes?
20:29:11 <rockyg> gimme a sec....
20:29:43 <dhellmann> I think that's https://etherpad.openstack.org/p/PAO-ops-meetup
20:29:45 <jokke_> problem with etherpads, they are really difficult to find with google
20:30:00 <dhellmann> jokke_: yeah, that's on purpose to keep out spam bots
20:30:05 <rockyg> https://etherpad.openstack.org/p/PAO-ops-meetup  yup.  what Doug said.
20:30:30 <jokke_> #link https://etherpad.openstack.org/p/PAO-ops-meetup
20:30:51 <jokke_> #info please comment on the ops etherpad if anything specific in mnd
20:31:05 <jokke_> moving on
20:31:20 <jokke_> #topic proposals for log_wg meeting time
20:31:35 <rockyg> Tom has all the etherpad since the first midcycle on a wiki page:  https://wiki.openstack.org/wiki/Operations/Meetups
20:31:52 <jokke_> we had that discussion earlier on, not a single response on the mailing list, so I assume we stick with this time
20:32:13 <bknudson> this time works for me
20:32:35 <rockyg> I think we should make it earier, if possible.  But, I can poll the ops meeting.  and, we'll go an hour earlier when DST ends.
20:33:24 <jokke_> well give a poll, but as said no replies to the e-mail I sent to the dev list
20:34:43 <jokke_> #topic Open Discussion
20:34:59 <jokke_> Anything else on your hearts?
20:35:03 <Nikolay_St> well
20:35:39 <Nikolay_St> I'd like to introduce a logging spec adoption for murano: https://review.openstack.org/#/c/205200/
20:37:34 <jokke_> <3
20:38:43 <Nikolay_St> and that's all from my side, since we just migrate to oslo_log)
20:39:04 <dhellmann> thanks, Nikolay_St, I'll add that to my review queue
20:39:21 <rockyg> kewl.  I think I need more time to review, but this week?  Might want a reference to syslog RFC and Python log guidelines.
20:39:33 <dims> fyi - https://review.openstack.org/#/c/205779/ (Define TRACE logging level) in oslo.log
20:39:37 <rockyg> So, can we discuss again next week.
20:39:37 <Nikolay_St> dhellmann: thx)
20:39:48 <jokke_> dims: \o
20:40:29 <dhellmann> dims: ++
20:40:58 <rockyg> thanks, dims
20:42:05 <bknudson> dims: looks like it needs someone to wake up jenkins
20:42:12 <bknudson> maybe another +W
20:42:19 <dims> bknudson: y
20:42:26 <dhellmann> dims: we should add that to the guidelines in the cross-project specs repo
20:42:50 <dims> dhellmann: once we get a release out and g-r updated?
20:43:04 <dhellmann> dims: yep
20:43:06 <jokke_> ++
20:43:19 <dhellmann> bknudson: there's a bit of a backlog in the queue, but that job is being processed
20:43:20 <dims> will do dhellmann
20:43:59 <bknudson> what enables trace?
20:44:25 <bknudson> trace level... e.g., verbose=True enables debug
20:44:38 <bknudson> do I need logging.conf  level=5 ?
20:45:14 <bknudson> maybe change verbose= to allow verbose=trace ?
20:45:42 <dhellmann> there's actually a proposal to drop verbose entirely and make info the default, I think
20:45:52 <dhellmann> https://review.openstack.org/#/c/206437/
20:45:54 <bknudson> that would be fine
20:46:19 <bknudson> oh, verbose is INFO
20:46:22 <bknudson> it's debug=True
20:46:24 <dhellmann> yeah
20:46:29 <bknudson> so could be debug=trace
20:47:22 <dhellmann> I think the idea is you wouldn't want to trace *everything* so you would use --default-log-levels to control the modules where you do want tracing
20:47:35 <dhellmann> maybe just oslo.messaging, for example
20:48:01 <bknudson> that makes more sense
20:50:05 <jokke_> anything else?
20:50:07 <bknudson> what happens if I have both debug and verbose?
20:50:18 <jokke_> it's debugging
20:50:24 <dhellmann> bknudson: debug wins
20:50:32 <rockyg> as long as you put it globally.
20:50:54 <rockyg> I think you can still do project level configs, though.  On some projects, at least
20:51:01 <bknudson> keystone also uses debug=True to indicate to work in insecure manner.
20:51:06 <bknudson> should change that.
20:51:35 <rockyg> bknudson: uh, yeah.
20:51:52 <jokke_> :)
20:52:10 <bknudson> I can work on that. should be ez
20:52:41 <jokke_> and then as with any other OpenStack project, deployers runs it all the time on debug :D
20:53:59 <rockyg> so another argument for error codes +requestid is that with both of those, debug could most likely be turned off ;-)
20:55:47 <jokke_> most of the time yes
20:57:38 <jokke_> ok last minutes ... use them wise
20:58:04 <rockyg> Now I have a good justification for error codes!  Finally.  Something that is verbalizable.
20:58:40 <jokke_> ahaa-moment?
20:59:03 <rockyg> yup.
21:00:03 <jokke_> ok, thanks again everyone
21:00:08 <rockyg> Thanks!
21:00:12 <jokke_> #endmeeting