17:00:10 #startmeeting 17:00:10 Meeting started Thu Jan 16 17:00:10 2014 UTC. The chair is colindixon. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:10 Useful Commands: #action #agreed #help #info #idea #link #topic. 17:00:20 #topic overview and roll call 17:00:36 #info Andrew Grimberg for infrastructure support 17:00:54 #link https://docs.google.com/spreadsheet/ccc?key=0AoSzir1BfjyWdDQyVElWNG9mcWxhblREckZjbjFxUVE#gid=1 is the document we're working from 17:00:55 #info Ed Warnicke for controller 17:01:03 #info michal_rehak for openflowplugin 17:01:04 #info Madhu for ovsdb 17:01:10 #info Abhijit Kumbhare openflowplugin 17:01:10 #info Prasanna for openflowplugin\ 17:01:12 #info Chris Wright 17:01:32 #info Anees / opendove 17:01:34 #info Dana for bgpcep 17:01:44 today, we should do some following up with individual projects, I'm open to suggestions for how to best do that quickly 17:01:53 #info oflibMichal for openflowjava 17:02:24 thoughts? 17:02:25 Quick walk through of their spreadsheets to see if they have folks assigned to tasks? 17:02:26 #info Robert for bgpcep 17:02:35 colindixon: for quickly...perhaps asking to fill in ETA for per-project tasks 17:02:55 #info Konstantin for defense4all 17:02:58 and reviewing status for ETAs that are near (like day before?) 17:03:00 that's not a bad idea 17:03:06 where are those ETAs listed? 17:03:12 it's DONE or ETA 17:03:14 or where should they be listed? 17:03:15 for status 17:03:20 for each tab? 17:03:22 ok 17:03:24 in per-project tab of workskeet 17:03:29 fair enough 17:03:29 sheet even 17:03:46 #topic how to collect status from projects quickly 17:04:09 #info cdub suggests that we just have it filled in on the project-specific sheet of the progress tracking in the status column 17:04:29 #info we also have ttkacik_ and janmedved in attendance 17:04:45 Hey all, drafts for the user guide and developer guide templates were sent to discuss about 8 hours ago 17:04:59 #topic documentation 17:05:02 this is important 17:05:11 do you have the link? are the on the wiki? 17:05:16 * edwarnicke cheers RobDolin and team :) 17:05:37 #info Phil Robb here as well 17:06:08 @ColinDixon - I plab to post on wiki after merging any feedback. 17:06:36 #link https://lists.opendaylight.org/pipermail/discuss/2014-January/001067.html is the e-mail RobDolin sent to the discuss list containing the first draft of the user manual and developer guide 17:06:55 do you have an ETA for that? 17:07:31 Ah yes that link. Thanks @ColinDixon 17:07:43 RobDolin: is there an ETA to post to the wiki? 17:08:07 or maybe we could just do that and allow others to put things in and/or you can do it later? 17:08:27 One other thing I noticed, we don't have the log levels linked in from the Hydrogen Release Work page... could we get that done (and pointers to other collateral)? 17:08:41 @ColinDixon - ETA to wiki about 8-16 hours depending on volume of feedback / conflicting feedback. 17:08:56 sounds good 17:09:17 #action RobDolin to post the user manual and developer guide templates to the wiki in the next 24 hours 17:09:30 edwarnicke: I can do that 17:09:59 #action colindixon will link to the log level stuff and any others from the Hydrogen Release Work page 17:10:09 cool, great work RobDolin 17:10:13 RobDolin: Could you link those in here when you do: https://wiki.opendaylight.org/view/CrossProject:Hydrogen_Release_Work#Recommendations_and_Templates 17:10:30 colindixon: I already did: https://wiki.opendaylight.org/view/CrossProject:Hydrogen_Release_Work#Recommendations_and_Templates 17:10:42 cool 17:10:42 Realized I had the link and it might be easier to do it than to complain ;) 17:10:43 thanks 17:10:43 @EdWarnicke - Sure 17:10:59 RobDolin: can you mark that progress on the spreadsheet? 17:11:08 RobDolin: Please feel free to split Recommendations and Templates into seperate sections if that makes sense to you 17:11:12 the fact that it should be posted by 1/17 17:11:24 #topic download page 17:11:29 Question: How do we want to handle our readout to the TSC? 17:11:34 I'm just going to paste what shague said earlier 17:11:50 shague: Hey guys, for the release status at 9:00aPST, I can't make it. 17:11:50 shague: I wanted to give status for item 30 in case it is covered today. 17:11:51 shague: General layout of download page is identified. 17:11:52 shague: Base edition artifacts have been identified and gathering information about the the other editions. 17:12:07 Awesome :) 17:12:18 #info shague said the general layout of the page has been worked out, the base edition artifacts have been worked out, and they're working on the other editions 17:12:28 Do we have a link to that so folks can review it? 17:12:40 #topic log level recommendations 17:12:56 #link https://wiki.opendaylight.org/view/Draft_Syslog_Level_Settings 17:13:28 #info ryan moats (regexboi) posted the revised version with the logback levels 17:13:41 edwarnicke: No link yet. The mock up was done on white board. The LF web team is working on mock ups 17:14:00 Could we get a link to shague 's work on artifacts to be included? 17:14:12 I asked him for a link and he didn't respond 17:14:16 it would be nice if we could get one 17:14:34 #action phrobb or shague to provide a link for the progress on the download page and particular artifacts per edition 17:14:35 On the logging recommendations... could we change bundle to module? 17:14:50 I agree with that 17:14:53 any objections? 17:15:11 can we also add raw messages (for protocols) to trace? 17:15:17 (I can think of several examples of a bundle with multiple logical modules that 'start' and 'stop in it) 17:15:43 rovarga: You mean like messages received from the wire? 17:15:47 rovarga: can we just say "and others" or "etc."for trace? 17:15:54 edwarnicke: yes 17:16:17 +1 rovarga netty log handler facility is nice. 17:16:29 any objections? 17:16:31 +1 adding messages to the trace category 17:16:45 colindixon: a few examples from different problem domains are helpful 17:17:03 #agreed change "bundle" to "module" in the log level recommendations 17:17:16 Any objection to me updating the log recommendation page to indicate the bit about messages and to change bundles to modules? 17:17:42 #agreed expand the scope of trace to include items such as logging messages received and others 17:17:50 sorry, I may have jumped the gun on those two 17:17:53 but I don't think so 17:18:25 ok 17:18:27 regarding INFO 17:18:31 #link https://wiki.opendaylight.org/view/Draft_Syslog_Level_Settings page updated, anyone see issues? 17:18:37 bundle start and stop INFO is a stretch 17:18:39 it is useful 17:18:50 but under a DEBUG 17:18:58 Madhu: The recommendations are not 'you must log this' but rather 'if you are logging this, use this level' 17:19:00 INFO is something that user really likes to know 17:19:08 I understand. 17:19:19 but it can quickly become a lot of INFO :) 17:19:56 Madhu: I would think that a functional module has finished starting or stopped is important user information (not always, but if you are going to log it, it's probably pretty important to the user) 17:19:57 anyways. thats a minor comment. 17:20:08 +1 for DEBUG, INFO should not be cluttered during normal operation 17:20:27 ok, can we hash this out on the mailing list? 17:20:32 if people care 17:20:40 sure. 17:21:00 #info Madhu and rovarga say that they think module start/stop should be DEBUG and not INFO, but this will be taken to the mailing list 17:21:02 but in case there is lot of INFO from particular bundle, we can limit that with log settings on that class 17:21:02 thanks 17:21:23 #topic system.out.println() 17:21:33 can we mark "Test to check for System.out.println" as DONE? 17:21:51 also can I get a report from the projects represented here about progress 17:22:14 there is 1 open question 17:22:22 Well... I've filed bugs... 17:22:29 And indicates the grep line to check 17:22:32 Does that count? 17:22:36 is system.out.println applicable for UT / IT code ? 17:22:54 Also, Gal Mainer from lispflowmapping correctly pointed out we should also look for System.err and printStackTrace 17:22:55 i would say... just leave UT / IT out of this 17:22:55 Madhu: I think the answer is that it's a project by project choice 17:23:08 does anyone disagree? 17:23:17 i agree on System.err and e.printstacktrace 17:23:30 and project by project by choice to UT/IT is a +1 17:23:37 What about using System.out in the CLI part ? 17:23:55 Konstantin: good point. but it shud be ci.print 17:23:57 not system.out 17:23:59 :) 17:24:39 Madhu: I agree if you are talking the OSGI console, was that what you meant Konstantin ? 17:24:47 CommandInterpreter.println 17:24:51 #agreed individual projects can choose whether or not System.out.println is allowed in their Integration and unit tests, but in production running code it should not be used 17:25:12 No, i'm speaking about CLI running in the command propmt. 17:25:14 sysout is enemy to cbench testing 17:25:30 Openflowplugin - we have changed the System.out.println for the reason michal_rehak mentions 17:25:37 Konstantin: u mean the osgi> command prompt ? 17:25:39 Konstantin: meaning some other CLI that is not OSGI that you are providing? 17:26:05 Yes, our cli speakong over REST with our server ))) 17:26:13 edwarnicke: yes, bgpcep also has two standalone jars which are normal CLI java apps 17:26:21 We ( defense4all ) are not bundle 17:26:41 Madhu, colindixon do you agree that a CLI that is *not* running within the controller process is something that should be at the discretion of the project on the System.out.println issue? 17:26:50 yeah I think I do 17:26:50 useful for manual validation, but not included in any of the distributions 17:26:51 yes. 17:26:56 So, here we are ok with System.out ? 17:27:17 Konstantin: I know you guys are not a bundle running in the controller process, which is why I raised the issue the way I did :) 17:27:20 avoid system.out.println for debug logging anywhere :) 17:27:39 code that does not get run by run.sh can use system.out.println at the project's discretion 17:27:54 does that sound like the right wording? 17:28:34 colindixon: That sounds either right or very nearly right to me, thoughts from other folks? 17:28:48 yes from me 17:28:51 run.sh might change ;) 17:28:57 so lets keep it abstract 17:28:58 works for us 17:29:03 and not specific to run.sh. 17:29:08 fair enough 17:29:27 code that does not get run by launching the controller in the normal fashion can use system.out.println at the project's discretion 17:29:34 Yes, our start-up script is not "run.sh" ) 17:29:58 ok . guess time to change the topic ? 17:30:34 I think first we need the #agreed 17:30:41 #agreed code that does not get run by launching the controller in the normal fashion can use system.out.println at the project's discretion, for example, to provide a CLI that runs outside the JVM 17:30:53 #topic per-project issues 17:31:06 I think I'm not going to spend another 30 minutes going through this here 17:31:21 instead I'm going to pound action everyone to fill out their spreadsheet 17:31:30 yeah for delegation 17:31:40 and we can do some prep to follow up with people that haven't filled it out in subsequent meetings 17:31:45 request for project owners to get this completed : https://docs.google.com/spreadsheet/ccc?key=0Anypq0z_rOtOdE1YOUFSejQ0TGk0TkFhbVdEZ0d0VHc#gid=0 17:31:46 Could we get a quick rollcall of spreadsheet filled out here though 17:31:47 and I'll send an e-mail to that effect 17:31:55 the version convergence is critical for stable testing 17:32:01 sure I can do that 17:32:08 A simple #info for whether all items in a given section that should have started have a person on them from each project? 17:32:42 #info in the goal of saving time during these meetings, the bulk of the communication for completion of per-project tasks will take place offline via the sheets for individual projects 17:32:57 #link https://docs.google.com/spreadsheet/ccc?key=0AoSzir1BfjyWdDQyVElWNG9mcWxhblREckZjbjFxUVE#gid=1 this spreadsheet has the sheets to fill in the info for 17:33:11 #info controller: All per project items in Global Code cleanup have a person signed up (thought its all edwarnicke right now, so the names may be updated) 17:33:19 edwarnicke: cleaned up the release recommendations page : https://wiki.opendaylight.org/view/Simultaneous_Release:Simultaneous_Release_Plan_2014:ReleaseRecommendations 17:33:28 please have a look / edit ? 17:34:00 #info Defense4all Jenkins job is done. No changes suggested. 2 WARN will be fixed till next submit. 17:34:04 #action colindixon to send an e-mail to discuss calling on people to update their sheets 17:34:19 Konstantin: can u pls change the status ? 17:34:35 Madhu: did you want to call out anything specific w.r.t. versions? 17:34:50 yes. it is important that we have it done ASAP 17:34:55 Madhu: Linked in here: https://wiki.opendaylight.org/view/Simultaneous_Release:Simultaneous_Release_Plan_2014:ReleaseRecommendations 17:35:20 a lot is dependent on consistent versioning 17:35:29 Madhu: Which you had already done, removed my dupe ;) 17:35:38 edwarnicke: yes :) 17:35:40 #action everyone please deal with their version synchronization as specified in the following link 17:35:56 #link https://wiki.opendaylight.org/view/Simultaneous_Release:Simultaneous_Release_Plan_2014:ReleaseRecommendations 17:36:09 #info versions status are tracked under : https://docs.google.com/spreadsheet/ccc?key=0Anypq0z_rOtOdE1YOUFSejQ0TGk0TkFhbVdEZ0d0VHc#gid=0 please edit it 17:36:17 thank you 17:36:56 folks please use #link instead of #info when adding a URL to the minutes 17:37:03 Madhu: It's marked DONE, is not it ? 17:37:15 the two are virtually synonymous but #info doesn't hyperlink URLs 17:37:31 Konstantin: yes. 17:37:36 tykeal: ok 17:38:01 what was line 39 in the spreadsheet supposed to be 17:38:03 it's lost it's title 17:38:36 colindixon: "Write recommendation for cutting Release Branches and laying Release labels" 17:38:51 fixed 17:38:58 can we mark that as done? 17:39:02 I think we can, right? 17:39:18 if we can, I'm going to do a closing call for topics and then end the meeting 17:39:22 yes. 17:39:30 i will mark it done 17:39:36 thanks 17:39:40 editing this page gets challenging sometimes :) 17:39:43 its my FF maybe 17:39:49 #topic end of meeting 17:39:52 any final topics? 17:40:06 We need to decide what we are going to do to report out to the TSC 17:40:14 ah 17:40:20 Should be a short summary of state 17:40:42 #topic report to TSC 17:40:48 I think we can just keep it informal 17:40:55 I'd like to keep it high level 17:41:02 And that means probably a little thought 17:41:04 I think everything is actually on track except documentation 17:41:19 but I think that's caught up a lot in the last 24 hours 17:41:19 That's my sense as well 17:41:27 i would like to see an early code freeze and not on 01/27 17:41:28 Madhu: what's your thought and sense 17:41:37 colindixon: that's a good executive summary ;) 17:41:47 So: Global Code Cleanup is on track to plan. Documentation is about 3 days behind plan? 17:42:14 edwarnicke: cross project dependencies should be nailed ASAP 17:42:22 and shud not change often. 17:42:26 #info we'd like to keep the report high-level, most things seem on track, documentation seems to be a few days behind, but is showing progress now 17:42:37 to give some stable testing platform for atleast 2 weeks 17:42:40 Finalization Activities are still on track to plan 17:42:46 Release Reviews are still on track to plan 17:43:00 #info cross-project dependencies and versioning needs to be prioritized to be able to start testing ASAP 17:43:11 colindixon: thanks 17:43:22 #info Madhu would like to ask projects to try to get to code freeze before 1/27 if possible 17:43:29 I think that's a pretty good summary 17:43:50 I'm going to need to reach out to projects about release reviews and release notes 17:44:13 #action colindixon to follow up with projects about release reviews and release notes to make sure those things don't fall behind 17:44:14 Give me one second, putting up a tab on the spreadsheet for TSC report 17:44:26 edwarnicke: do we need that? 17:44:40 any other topics? people want to bring up while we're waiting? 17:45:12 edwarnicke: I was thinking we could just read out the notes for this last bullet 17:45:52 Are the branch recommendations done, they aren't marked on the Activites tab as such 17:46:03 they are done 17:46:13 Madhu was going to mark it as such 17:46:16 edwarnicke: am fixing it 17:46:52 Madhu, have you started looking at signing of artifacts (not beating you up about it, just lookign for status) 17:47:05 edwarnicke: no. starting today sir 17:47:30 ok 17:47:36 I'm going to call this meeting 17:47:38 going once 17:47:41 One moment 17:47:47 edwarnicke: ok 17:47:49 * cdub mutters madly at all the mailing lists 17:48:16 lol 17:48:20 cdub: we really need to figure out if there's a saner way to divvy up the mailing lists 17:48:27 colindixon: yes, yes there is 17:48:29 guys, how does this look for reporting to the tSC: https://docs.google.com/spreadsheet/ccc?key=0AoSzir1BfjyWdDQyVElWNG9mcWxhblREckZjbjFxUVE#gid=15 17:48:30 consolidate... 17:48:31 colindixon: it's call one list ;) 17:48:39 I'm all for consolidating them ;) 17:48:39 +1 17:48:45 cdub: one list might kill inboxes though 17:48:48 but I guess that's fine 17:48:49 colindixon: you can do topics if you don't know how to deal w/ a high (actaully low) volume list 17:49:07 cdub: Fast way to more traffic than anyone can follow... discuss is already there for common issues 17:49:15 it's pretty easy to ignore stuff you don't care about 17:49:20 But most folks have no interest in the inside baseball of projects they aren't following 17:49:21 and topics formalize that 17:49:39 edwarnicke: I might leave finalization activities as white 17:49:43 since it's really just getting up too 17:49:50 OK 17:50:05 #info tsc report is now a tab in the main sheet 17:50:09 #link https://docs.google.com/spreadsheet/ccc?key=0AoSzir1BfjyWdDQyVElWNG9mcWxhblREckZjbjFxUVE#gid=15 is showing the TSC summary 17:50:19 #topic mailing lists 17:50:34 #info some push to consolidate to fewer lists, but tabled to after hydrogen 17:50:37 #endmeeting