14:31:59 #startmeeting weekly irc sync 14:31:59 Meeting started Wed Jun 3 14:31:59 2015 UTC. The chair is gzhao. Information about MeetBot at http://ci.openstack.org/meetbot.html. 14:31:59 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:31:59 The meeting name has been set to 'weekly_irc_sync' 14:32:19 Good morning gzhao 14:32:22 #info tbachman for GBP 14:32:29 #info regXboi for neutron northbound 14:32:31 #info anipbu for usc 14:32:34 phrobb: gzhao regXboi good morning :) 14:32:54 Good morning tbachman 14:32:56 * regXboi quips "heck it's almost afternoon" 14:32:56 #info LuisGomez1 integration 14:32:58 regXboi: I seem to have problem with internet speed. 14:33:08 #info lori for lispflowmapping 14:33:20 #info Hideyuki for VTN 14:33:34 are we expecting colindixon? 14:33:37 #info gzhao for release USC 14:33:48 * regXboi has a simple process question (smile) 14:33:53 #info ttkacik for yangtools and controller 14:33:59 #chair phrobb1 regXboi 14:33:59 Current chairs: gzhao phrobb1 regXboi 14:34:05 uh? 14:34:07 #info odlcasey for ODL 14:34:17 * regXboi wonders what he did to deserve that 14:34:33 gzhao: you might want to chair tbachman too… :-) 14:34:39 * tbachman hides 14:34:40 #chair tbachman 14:34:40 Current chairs: gzhao phrobb1 regXboi tbachman 14:34:41 lol 14:34:44 darn it! 14:34:45 lol 14:34:45 ;) 14:34:46 lol 14:34:48 * regXboi spreads the wealth/pain 14:34:55 #chair tbachman 14:34:55 Current chairs: gzhao phrobb1 regXboi tbachman 14:35:05 now I’m double-chaired 14:35:08 #info Thanh 14:35:15 * tbachman wonders what extra powers that gives him 14:35:30 * regXboi tells tbachman to calm down - chairing is idempotent 14:35:36 lol 14:35:39 PUT, not POST 14:35:44 :) 14:36:11 #topic blocking issues 14:36:23 btw — was RC0 ever cut? 14:36:36 * regXboi isn't sure ... 14:36:58 fyi colindixon sent a note saying his internet is flaky on him today, so he may not be able to join 14:37:05 #undo 14:37:05 Removing item from minutes: 14:37:20 #topic RC0 and branching, version bump 14:37:34 zxiro, can you give us an update on RC0? 14:37:58 Yes, I'm still trying to get a build going 14:38:15 latest attempt will be build 15 which i'm starting shortly. 14:38:34 I had to disable snmp4sdn this morning though as they have deleted all the code in their repo 14:38:34 #info zxiiro says he's still trying to get an RC0 build going… he's on attempt #15 14:39:02 snmp4sdn concerning patch https://git.opendaylight.org/gerrit/#/c/21747/ 14:39:38 I'll note that if we cut RC0 right now, we'll need an RC1 14:39:40 shall we go for RC1 directly, because RC1 is scheduled on 6/4 14:40:21 makes sense 14:40:29 I discovered yesterday that NN is missing a couple of pieces of it's northbound connections to OpenStack 14:40:40 * regXboi has to file a bug on that 14:41:04 * regXboi has to then get the bug closed 14:41:27 zxiiro: Anything else blocking RC0/1? Is there anything this group can do to help? 14:41:56 I think first we need to get RC build successful, then all projects need to complete branch cutting and version bump 14:41:57 #info regXboi says Neutron Northbound s missing a couple of pieces of it's northbound connections to OpenStack 14:42:27 #info regXboi wonders how to handle this situation because its not precisely covered in the API waiver process 14:42:34 #info phrobb1 asks if there’s anything else blocking RC0/1? Is there anything this group can do to help? 14:42:41 at the moment I don't see any issues. We'll know in ~6 hrs after build 15 runs if there's any new issues that pop up but so far I've been able to open bugs and submit patches to projects when issues appear 14:42:46 #info gzhao says we need to get RC build successful, then all projects need to complete branch cutting and version bump 14:43:25 zxiiro: and all projects are being responsive to your patches? 14:43:31 phrobb1: yes 14:43:35 cool! 14:43:44 #info there still 6-7 projects haven't completed branching and version bump 14:43:57 who do I talk to about getting versions added to a product/component in our bugzilla? 14:44:17 OK, gzhao shall we move to blocking bugs? 14:44:25 regXboi: helpdesk 14:44:28 #topic Blocking bugs 14:44:34 zxiiro: Is the Lithium Project Status spreadsheet accurate? 14:44:41 odlcasey: yes 14:44:47 regXboi: I expect that would be a request to helpdesk 14:44:50 odlcasey: i've been keeping it updated daily 14:44:57 Ok, thanks. :) 14:44:58 #link https://docs.google.com/spreadsheets/d/1KPpO9LH539Vlcoa4RvLa6PPCdLifi5JD-ihRhlybqeo/edit#gid=676729675 < -- current blocking bugs 14:45:13 * tbachman notes that he has to be careful when tabbing out odlcasey’s nic… odl_meetbot is just *too close* :0 14:46:03 regXboi: a helpdesk ticket. It's possible to grant some amount of power to committers of a given project to manage that themselves as well, I just need to know who wants it as I can't use the LDAP groups to grant the rights in bugzilla :( 14:46:42 tykeal: filed the ticket, but if you can grant me that power, I'll be less likely to bug folks (sorry for the pun) 14:46:44 tbachman: :) Maybe I should change it. 14:46:50 hehe 14:46:52 #info odlcasey asks if the Lithium Project Status spreadsheet is accurate 14:47:14 #info zxiiro says he’s been keeping it updated daily 14:47:33 so... I have a comment on the blocking bugs 14:47:35 odlcasey: :) 14:48:01 #info regXboi notes that the bugs being filed don't necessarily contain enough useful information to triage 14:48:06 I’ll try and be extra careful (tbachman already makes *enough* bugaroos in minutes, as it is) 14:49:54 regXboi: do you want to create a bug template? with reproduce steps etc. 14:50:14 gzhao: that's really up to each project 14:50:15 hideyuki: good news - found root cuase for that blocker - was able to have unit test reproduce it... 14:50:25 so I'm going to create a template for NN bugs 14:50:38 ttkacik1: Great news! 14:50:49 regXboi: thanks 14:50:50 ttkacik1: Thanks a lot! 14:51:19 hideyuki: could you info that in 14:51:23 gzhao: but that will apply *ONLY* to NN bugs 14:51:39 regXboi: we can use that as example at least 14:52:03 * tbachman reads scrollback 14:52:03 #action regXboi to create template for NN bugs in NN wiki 14:52:14 any projects have blocking bugs? 14:52:37 #link https://wiki.opendaylight.org/view/Weather#Blocking_Bugs 14:53:04 regXboi: Could you pick up blocking bug: https://bugs.opendaylight.org/show_bug.cgi?id=3368 14:53:37 edwarnicke: that's what led to my earlier comment about not having enough info and subsequent action item for a NN template 14:53:42 #info Tony found the root cause of the bug 3344, and it was able to have unit test reproduce it. https://bugs.opendaylight.org/show_bug.cgi?id=3344 14:53:56 regXboi: What info is missing? 14:53:58 #info abhijitkumbhare OpenFlow plugin 14:54:04 sorry got held up in traffic 14:54:16 #link https://bugs.opendaylight.org/show_bug.cgi?id=3344 blocker bug for VTN, which ttkacik recently found the source of and fixed 14:54:18 edwarnicke: in general I'd like the source JSON 14:54:23 call me picky 14:54:24 * tbachman looks for gerrit to point to 14:54:38 regXboi: Comment and ask for them 14:54:39 will be soon 14:54:40 regXboi: That said 14:54:45 need to write good commit message 14:54:57 edwarnicke: already asked for it on the neutron channel 14:55:00 so it's in the works 14:55:03 :) 14:55:39 tbachman: your info for bug 3344 is better :) 14:56:21 #topic RC1 14:56:27 hideyuki: just like to capture those with the #link option 14:56:32 was looking for the corresponding gerrit 14:56:38 ttkacik: do you have the gerrit for the fix? 14:56:41 (or is there one yet) 14:57:30 ah — sorry ttkacik… missed your message above 14:57:35 * tbachman was too eager :) 14:57:51 #info RC1 is scheduled 6/4 tomorrow 14:58:34 any issues people like to raise 14:58:48 I have a process question 14:59:01 I have a question... did we build an RC0? 14:59:08 edwarnicke: not yet 14:59:44 OK... so to my understanding the purpose of RC# is to have images to test with... so the question becomes... can we really have an RC1 if we never built RC0... 14:59:49 edwarnicke: still in progress, but once it is done, it will be RC1. current zxiiro said it took about 6 hours 15:00:07 gzhao: 6 hours sounds actually faster than I expected ;) 15:00:08 edwarnicke: RC0 info is in the meeting log… we're on attempt 15 and will know if it works in ~6 hours 15:00:41 edwarnicke: so we should still call it RC0 15:00:53 #info edwarnicke asks if the purpose of RC# is to have images to test with... so the question becomes... can we really have an RC1 if we never built RC0... 15:01:07 I guess my point is, we should discuss how many RCs we need to go through for the release... if we are fine skipping RC0 that's fine, but it materially means that we've cut off a third of our RC testing 15:01:10 I kicked off build 15 as RC0 15:01:23 but if that fails i can call it RC1 if we want or i can keep trying to build RC0 15:01:43 #info edwarnicke notes that skipping RC0 materially means that we've cut off a third of our RC testing 15:01:59 zxiiro: My personal gut tendency would be to call it RC0... but that's not well thought out on my part, and is just *my* gut 15:02:12 #link https://jenkins.opendaylight.org/releng/view/autorelease/job/autorelease-release-lithium/15/ <--current RC0 build 15:02:58 #info zxiiro says there’s an RC0 build underway, and should be available in ~6 hours 15:03:09 do we have any other blocking issues? 15:03:31 3 15:03:37 So I have an odd one 15:03:37 2 15:03:45 1 15:03:47 Because its not *quite* in odl repos 15:04:00 But we do have a blocking issue around the odl-networking not passing SG and SGR objects to ODL 15:04:05 Tony: https://git.opendaylight.org/gerrit/21791 15:04:15 Do folks have suggestions as to how to register that as a blocking bug? 15:04:22 Given that the code lives in stackforge? 15:04:33 edwarnicke: you mean openstack networking_odl? 15:04:40 tbachman: Yes 15:04:42 ah 15:05:05 tbachman: Without it, nobody in ODL can do SecGroups correctly 15:05:14 (you can kind of fake them a bit, but not do them correctly) 15:05:28 #info edwarnicke asks about a blocking issue related to security groups in the stackforge networking_odl project — and how we can track dependencies against it 15:05:31 regXboi: Thoughts? 15:05:35 edwarnicke: We don't get updates AFAIK thats the real issue 15:05:42 edwarnicke: wasn't this going to be done last week ? 15:05:46 alagalah_: That matches my understanding 15:05:58 edwarnicke: that's outside our scope 15:05:58 #info alagalah_ says the issue is that security group updates aren’t being sent 15:06:03 alagalah_: flaviof has been working it somewhat, among other things. I'm starting to look at it 15:06:08 https://review.openstack.org/#/q/status:open+project:stackforge/networking-odl,n,z 15:06:09 #info regXboi says that’s outside of scope 15:06:20 * edwarnicke does not recognize scope boundaries well ;) 15:06:29 regXboi: Who's scope is it in then? 15:06:47 regXboi: Not throwing stones at you mate, just want to understand 15:06:49 as soon as you say stackforge, I say openstack 15:07:07 zxiiro: I have a question if not all projects are on stable/lithium, will it cause tagging issues after RC build? 15:07:15 regXboi: Yes... but we are the ones who care 15:07:15 regXboi: Yes, but then is it the impl of Neutron NB responsibility? That means potentially 3-4 folks pinging them. 15:07:26 regXboi: I'm not trying to push this on neutron northbound 15:07:35 edwarnicke: it sure sounds like it 15:07:36 #info alagalah_ asks regXboi what’s in scope then? 15:07:37 regXboi: I'm just trying to figure out how we should record the issue in our blocking bugs 15:07:45 #info regXboi says stackforge is openstack scope 15:07:53 gzhao: tagging no. but we should have stable/lithium in order to tag otherwise they won't have a lithium stream once tagging completes 15:08:04 regXboi: Apologies if it sounded like I was trying to push this on neutron northbound 15:08:23 regXboi: My interest starts with 'lets record this is a blocking bug so we can track it' and ends with 'this is fixed on our OS driver' 15:08:25 gzhao: so process wise i say we shouldn't tag unless all projects have stable/lithium 15:08:30 regXboi: Neither was I.. .just, where do we go ? 15:08:40 #info gzhao asks if not all projects are on stable/lithium, will it cause tagging issues after RC build? 15:08:42 regXboi: TSC ? 15:08:51 zxiiro: ok, thanks 15:08:56 #info zxiiro says tagging no. but we should have stable/lithium in order to tag otherwise they won't have a lithium stream once tagging completes 15:09:10 zxiiro: gzhao What is the link for the branch cutting spreadsheet (I know you've got it somewhere up log... but could I get it here)? 15:09:16 alagalah_: regXboi edwarnicke is the stackforge bug truly a blocker?… ie we wouldn't ship Li without it being fixed? 15:09:24 edwarnicke, regXboi: going to TSC won't help 15:09:37 these are bugs that need to be filed on openstack and the ODL ML2 driver 15:09:37 phrobb1: Well, its really a question of whether we are supporting basic neutron L2 or not 15:09:43 yes we should file them 15:09:51 phrobb1: If we are comfortable not actually supporting basic neutron L2, its not a blocker 15:09:58 and heck, volunteers can be good citizens and fix them 15:10:01 phrobb1: But if you *actually* want to support neutron L2, you need it 15:10:09 but they are bugs *ON* the openstack code base, not the ODL code base 15:10:17 and that's my point 15:10:44 edwarnicke: and that's on the shipping release of Kilo for OpenStack? 15:10:59 … ie we would need to wait for an update to Kilo? 15:11:19 phrobb1: that's likely for individual folks to decide :( 15:11:28 phrobb1: My understanding is that the drivers evolve independently of the broader neutron 15:11:53 regXboi: ... and if we weren't two weeks away from our own release, I'd definitely pony up and give it a try 15:11:55 phrobb1: But the net net is we can't even *know* if we can support neutron L2 all the way without those objects coming down to us 15:12:18 #info regXboi says that these are bugs that need to be filed on openstack and the ODL ML2 driver 15:12:34 phrobb1: There were bugs fixed in ML2 in Kilo to *enable* us to do this in our driver 15:12:50 #info phrobb1 asks if the stackforge bug truly a blocker?… ie we wouldn't ship Li without it being fixed? 15:13:01 #info - I know this is a bit of a regression, but I'm not convinced that 3368 is a blocking bug 15:13:08 #Info edwarnicke says its really a question of whether we are supporting basic neutron L2 or not 15:13:14 For me, it doesn't sound like a blocker but instead a severe bug that we document, then update with a version identifier on Kilo/driver/whatever if/when the issue is fixed in OpenStack 15:13:41 #info phrobb1 says that to him, it doesn't sound like a blocker but instead a severe bug that we document, then update with a version identifier on Kilo/driver/whatever if/when the issue is fixed in OpenStack 15:14:21 phrobb1: Well... do we just say that we don't support L2 and SG then in our release notes? 15:14:27 phrobb1: Because this is *that* fundamental 15:14:46 phrobb1: I had thought that 'neutron support was the most important thing' and it doesn't get more fundamental to providing that support than this 15:14:50 edwarnicke: we say we support SG as much as openstack does 15:14:58 regXboi: But we don't 15:15:08 regXboi: We support them way less than OS does 15:15:23 edwarnicke: if the drivers to us don't send information then we are done 15:15:26 file the bug there 15:15:30 #info edwarnicke asks if we just say that we don't support L2 and SG then in our release notes? 15:15:33 regXboi: Because if I change a SG, I have to teardown and restand up *all* the VMs that impacts in order to reflect that change 15:15:41 #info regXboi says we say we support SG as much as openstack does 15:15:47 #undo 15:15:47 Removing item from minutes: 15:15:53 :) 15:16:02 regXboi: Or simply have the user silently not get the SG they thought they were getting for existing ports 15:16:03 #info regXboi says that all we can do is say we support SG as much as openstack does 15:16:08 regXboi: thx! 15:16:11 (which is a major major security hole in my mind) 15:16:34 #info edwarnicke notes that if I change a SG, I have to teardown and restand up *all* the VMs that impacts in order to reflect that 15:16:35 #info we don't support SG as much as OS does, but instead substantially less... entirely because of this bug in our OS driver 15:16:38 edwarnicke: I continue to fail to see how this is the right forum for this 15:16:45 edwarnicke: thx! 15:16:49 regXboi: Because we need to get it fixed before the release 15:16:49 edwarnicke: That makes sense to me. Regardless of the codebase where the bug resides, do we know who the logical people are to fix it?… ie is it someone from the ODL community who has been working in OpenStack?… or is there someone in OpenStack we need to work with to get it fixed? 15:17:03 phrobb1: flaviof has been poking at it 15:17:11 #info phrobb1 asks if regardless of the codebase where the bug resides, do we know who the logical people are to fix it?… ie is it someone from the ODL community who has been working in OpenStack?… or is there someone in OpenStack we need to work with to get it fixed? 15:17:24 phrobb1: mestery has been giving it best effort (which is kind of him, but he's very busy) 15:17:30 edwarnicke: so let me boil this down to the bottom line 15:17:32 phrobb1: I've started looking at it as well 15:17:50 edwarnicke: you are saying that without this patch to OS, *we don't have a Lithium release, period* 15:18:02 #info edwarnicke notes that mestery has been giving it as much effort as he can, but he’s very busy; flaviof has also been looking at it 15:18:14 mestery: We were just singing songs of you valiant attempt to help us get our OS ML2 driver to send SG and SGR objects... and lamenting that you are so busy ;) 15:18:26 #info regXboi askes if edwarnicke is saying that without the SG patch to OS, then *we don't have a Lithium release, period* 15:18:47 regXboi: Well, the call in my mind is really between these two choices: 15:19:01 1) Get this fixed, and make sure projects handle the fix correctly and claim OS support 15:19:08 #info regXboi notes that this feels like this is setitng both a bad precedent and is engaging in serious scope creep 15:19:10 2) Don't, and don't really have neutron support in Li 15:19:13 Pick one 15:19:27 regXboi: and edwarnicke, Ultimately, the TSC decides when the release is ready to ship. edwarnicke has a really good argument that the OpenStack bug is a blocker for ODL, particularly if we have the resources/ability to fix it regardless of code base. 15:19:36 #info edwarnicke says to him there are two choices: 1) Get this fixed, and make sure projects handle the fix correctly and claim OS support; : 2) Don't, and don't really have neutron support in Li 15:19:37 phrobb1: No disagreement there at all 15:19:44 phrobb1: Its entirely the TSC's call 15:19:57 #info phrobb1 says ultimately, the TSC decides when the release is ready to ship. edwarnicke has a really good argument that the OpenStack bug is a blocker for ODL, particularly if we have the resources/ability to fix it regardless of code base. 15:19:58 phrobb1: I am just trying to get this listed on the blocking bugs for tracking 15:20:05 edwarnicke and phrobb1: I will remind both of you that the TSC does not allocate resources 15:20:22 phrobb1: And understand very well that when the rubber hits the road, the TSC decides whether or not to ship with a particular blocking bug outstanding 15:20:28 regXboi: I know that better than most :) 15:21:01 #action project: CCFT, D4A, DIDM, Doc, NIC, Persistence, Reservation, SNBI, SNMP4SDN need to complete stable/lithium branching and version bump ASAP 15:21:09 just want to put that in the meeting minutes 15:21:15 phrobb1: I will note though, that generally while the TSC does decide which blocking bugs to ship with or not, historically its left it to folks to decide what to *list* as blocking bugs (which has worked really really well so far ;) ) 15:21:17 #info regXboi points out that the TSC doesn’t allocate resources 15:21:28 gzhao: Thank you :) 15:21:39 edwarnicke: https://docs.google.com/spreadsheets/d/1KPpO9LH539Vlcoa4RvLa6PPCdLifi5JD-ihRhlybqeo/edit#gid=1196332566 zxiiro added two columns, branching and version bump 15:21:45 regXboi: but they do determine minimum functionality and robustness for release. 15:21:46 gzhao: :) 15:22:00 #info phrobb1 points out that they do determine minimum functionality and robustness for release. 15:22:09 #undo 15:22:09 Removing item from minutes: 15:22:27 #info phrobb1 points out that the TSC does determine minimum functionality and robustness for release 15:22:41 tbachman: edwarnicke: phrobb1: sorry for the delay. My focus has been on the summit and L3 in ovsdb netvirt. I tried to help mestery_ a bit on the SG stuff, but that has not been a priority for me. 15:22:43 phrobb1: Only to the extent that it can decide to not ship. The TSC controls no resources 15:22:58 flaviof: ACK :) 15:23:17 (and welcome flaviof ;) ) 15:23:19 phrobb1: The TSC has no capacity to say 'This will be the minimal functionality in the release' only to say 'We will not release a release without this functionality' 15:23:31 flaviof: Welcome :) 15:23:42 edwarnicke: regXboi for the record, I agree and support the statement that the TSC has no control over resources :-) 15:24:17 flaviof: Can you tell us what you know of the SG and SGR stuff? Is there a bug filed on the OS side, and if not, where should it be filed over there? 15:24:39 edwarnicke: ty. ;) the sad state of our openstack ci makes devel on the SG harder, in my humble and honest opinion. 15:24:55 #info edwarnicke asks flaviof if he knows whether there’s a bug filed on the OS side, and if not, where should it be filed over there? 15:25:01 edwarnicke: I think so. /me looking 15:25:25 honestly, openstack ci is somewhat orthogonal to this conversation 15:25:41 I concur witj regXboi 15:25:43 #link https://bugs.launchpad.net/networking-odl odl ml2 bugs 15:26:12 https://bugs.launchpad.net/networking-odl/+bug/1444112 15:26:26 * gzhao wonders should he end this meeting 15:26:53 #link https://bugs.launchpad.net/networking-odl/+bug/1450603 critical odl bug: fix ci 15:26:57 #info regXboi want to make the point that this is a change in scope 15:27:26 #info regXboi points out that we are now saying we are blocking on bugs that *ARE OUTSIDE ODL'S SCOPE* 15:27:28 #link : https://bugs.launchpad.net/networking-odl/+bug/1444112 bug against networking_odl security groups (in openstack) 15:27:50 #info regXboi says "and that's a *major* change" 15:28:00 gzhao, maybe ask the group if there are other blocking bugs that should be discussed here? 15:28:14 colindixon: are you reading my last infos? (as TSC chair) 15:29:49 phrobb1: I think this SG issue should be discussed in TSC meeting, it probably cannot reach agreement here, but it is a good discussion 15:30:11 all: I have a problem with that - 15:30:18 I'm not around on thursday to input 15:30:18 gzhao: agree 15:30:24 we doubled the meeting time already, so I am going to end this meeting, don't mean to end the discussion 15:30:26 phrobb1: agree 15:30:59 or shall I just add a topic Openstack SG 15:32:30 gzhao: Not sure what you are asking… Maybe it would be best to send a note to the TSC list right after this meeting to kick off the further discussion with the TSC on the ML. That way regXboi can contribute to the discussion 15:32:51 phrobb1: gzhao I've added the bug to the blocking bugs list 15:33:00 phrobb1: ok 15:33:11 #link https://wiki.opendaylight.org/view/Weather#Blocking_Bugs 15:33:25 * tbachman has to go 15:33:26 edwarnicke: thanks, let's put that but in info, then I will end the meeting 15:33:46 #info edwarnicke has added the SG and SGR issue to blocking bugs list 15:34:34 #action gzhao sends email to TSC for further SG/SGR discussions 15:34:37 #topic cookies 15:34:37 phrobb1: that would be a good idea 15:34:43 * regXboi giggles 15:34:49 #endmeeting