20:03:00 #startmeeting log_wg 20:03:02 Meeting started Wed May 13 20:03:00 2015 UTC and is due to finish in 60 minutes. The chair is Rockyg. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:03:03 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:03:05 The meeting name has been set to 'log_wg' 20:03:48 I want to highlight meetings that involve log stuff at the summit. 20:05:13 I worked on the etherpad for oslo.log -- https://etherpad.openstack.org/p/YVR-oslo-log-plans 20:05:14 #link https://libertydesignsummit.sched.org/event/a03e8481405d57075c0a6d08c5601051#.VVOuYPCam1o 20:05:40 It's a Cisco session in Ops that would be good to garner info from 20:05:54 Thank you bknudson 20:06:18 mystic logs are bad 20:06:50 #link https://libertydesignsummit.sched.org/event/764d77baafe13caaad8ff1badabb9b9a#.VVOux_Cam1o 20:07:03 I think it's a sponsored session (the mystic logs...) 20:07:32 The second one is the state of log efforts. Ops. Strategy, status, etc 20:08:40 what's the goal of this session? 20:08:44 "How do we fix logging?" 20:09:02 I want to have this googledoc available for folks to flesh out: #link https://docs.google.com/spreadsheets/d/1XTncfK_droY8E-Uy2icVuU-z9ya38ZBK_ZIRvGfPOXc/edit#gid=0 20:09:41 bknudson: It's what we have done so far, whether we are doing the right stuff, and what do we do next, plus long term 20:10:01 what have we done so far? 20:11:36 the oslo log plans etherpad bknudson highlighted goes with #link https://libertydesignsummit.sched.org/event/fbfceb17bc4927136c0aa778d38586d1#.VVOv6fCam1o 20:12:17 Well, we've got a spec in review (that will have the next iteration out before the summit starts. Even if I do it on the plane) 20:12:58 API group has standardized around rfc6801 for time stamps (which is also syslog timestamp format) 20:13:19 log guidelines got published for dev 20:14:05 so, *we* haven't done much, but stuff got done. oslo.log graduated, got a global config in for syslog format, and got lots of shiny, new, comprehensible documentation in 20:14:24 not bad 20:15:24 Yeah. And the wg has done lots of learning and organizing so we can get more done next cycle. I got this group started much too late. My bad. but, it's going now 20:15:43 also, there is a spec in review for request IDs that the ops guys have been asking for 20:16:16 And thank you dhellmann for all the oslo work you've done. You're great. And the oslo team 20:16:22 Rockyg: I think you mean ISO 8601, not RFC 8601 (the latter is not related to date formats) 20:16:44 Rockyg: thanks! I hope we can keep the momentum going 20:17:04 dhellmann: Oops. dyslexed that one.... 20:17:17 well, the # was right, the standards body was wrong :-) 20:19:07 While at the summit, dhellmann and I (and maybe bknudson) are going to try to corral Morgan Fainberg to get the license changed on a chunk of github resident code for making python logging more syslog compliant. and maybe suck it into oslo.log 20:19:21 ISO8601 isn't really specific enough. 20:19:28 since it allows all sorts of formats 20:20:03 bknudson: i suspect they mean https://docs.python.org/2/library/datetime.html#datetime.datetime.isoformat 20:20:57 y, that's one format that ISO8601 specifies 20:21:12 bknudson: Yeah, but I think we can have oslo.log always generate a specific format if everyone follows ISO 8601. And, we can add specific ISO format to the guidelines 20:21:34 Plus, I think the API spec that merged specified the preferred format 20:23:13 So, for summit, there are also API working sessions. I need to talk to etoews about which one(s) might include a focus on logging. 20:24:17 bknudson, nkrinner, can you check the google spreadsheet I posted above and fill out any known logs and info you are aware of from your dev work? 20:24:55 I haven't gotten the keystone log files in yet. But I know they are (or mostly are) some apache format. 20:25:06 dhellmann: what am I forgetting? 20:25:30 what is this spreadsheet supposed to be showing? 20:25:32 the default configuration? 20:26:37 Rockyg: I added a couple of proposed changes to https://etherpad.openstack.org/p/YVR-oslo-log-plans , did you see those? 20:27:15 What log files are generated, their format, the default location they are written to, which project they belong to, what they are supposed to communicate, whether they use oslo.log or what other, etc. 20:27:53 If you think there is a characteristic missing, or one listed is useless, post a comment or add the column 20:27:59 I believe that all of the projects except swift are using some form of the oslo logging code, either the incubated module or the library (I haven't checked to see if all projects have adopted the lib yet) 20:28:17 dhellmann: Yes. I see 3 specs? 20:28:27 Rockyg: that's right 20:28:40 I edited from 2 different computers, so got different author colors 20:28:51 the green and yellow are both me 20:28:57 Ah 20:29:12 Rockyg: i can add some information tomorrow, will see what is still missing then 20:29:17 Do we need to reference the codes spec or the request id spec? 20:29:35 kewl 20:30:03 Rockyg: you can list those, but I don't think we're going to spend a lot of time in this session talking about them because they aren't finished. It's probably more appropriate to discuss those in an ops session. 20:30:42 Also, dhellmann, do we want to perhaps consider a new, 4th one if morgan fainberg agrees to the license modification? 20:31:11 Maybe put a placeholder in the etherpad, but not put it on the agenda list? 20:31:23 that was for a formatter, right? it's still not clear we actually need that code 20:32:47 Right. although the placeholders for missing fields was a big ops gripe. They lke the number of fields in the log messages to stay the same, even if they're not used 20:33:14 it would be good to have a bug or something to track that, but I think that's easy enough to fix 20:33:34 Cool. I'll file a bug for olso.log? 20:33:47 Then put its link in the etherpad. 20:33:55 Rockyg: one thing I'd like to learn more about is how big of a deal it is if we do end up changing the default format for logging -- it'll break whatever operators are using for parsing now, but if it's easier is that ok? 20:34:09 Rockyg: sounds good, thanks 20:34:10 In fact, I'll put a link in for all the bugs for oslo.log so folks can review. 20:35:20 Last summit, the ops present all were willing to take the pain for format improvements. Also, infra has a .erb file that does conversions of various formats that lots of folks grab. Infra would also like to see the file get smaller. 20:35:48 ok, that's good to know -- we worry a lot about backwards compatibility 20:35:57 I'll get the summary of Paris sessions done and posted somewhere (wiki page? Since it won't change) before the summit 20:36:08 Hopefully. 20:37:10 At least, I'm getting good info from both sides now about the stuff I need to highlight from those sessions :-) 20:37:22 Gives me a clearer direction 20:37:44 ++ 20:38:03 #action rockyg to add oslo.log bugs link to log-plans etherpad 20:39:37 dhellmann: would it make sense to use part of the working session or state-of session to file bugs against oslo.log for the highest priority ops gripes? We can move the bugs elsewhere if they don't fit in oslo.log, but it might be a great way to capture consensus 20:41:03 I'm not sure if that's the best use of f2f time. Will we have an etherpad with a list by then? We could discuss from that. 20:41:17 OK. 20:41:37 I see you are adding to the work to do on the etherpad, too. 20:42:14 What about swift and oslo.log? dhellmann: are you aware of what the swift logging issues are (I know infra has some) 20:42:32 Rockyg: you'll have to talk to the swift team about those, I don't know 20:43:16 OK. I plan on talking to infra, and ops about it. Once I know more, I'll be able to talk to swift. I know they don't use oslo.log, though. 20:44:21 OK. anything else we need to discuss before f2f? 20:45:11 * dhellmann can't think of anything 20:45:39 Ok. Since no one has any other issues, I say let's call it a meeting. Second? 20:47:00 Hearing no objections (and no second), this meeting is baked. see you at the summit! Those who can make it 20:47:06 #endmeeting