21:02:53 #startmeeting Networking 21:02:54 Meeting started Mon Mar 24 21:02:53 2014 UTC and is due to finish in 60 minutes. The chair is markmcclain. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:02:55 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:02:58 The meeting name has been set to 'networking' 21:03:15 #link https://wiki.openstack.org/wiki/Network/Meetings agenda 21:03:22 #topic Announcements 21:03:39 #link https://launchpad.net/neutron/+milestone/icehouse-rc1 21:04:10 We're are in the process of trying to branch RC1 21:04:38 if you have a new bug that is not critical please tag it icehouse-rc-potential 21:05:06 any bug that merges via the normal review process before RC1 is cut will be added to the RC1 milestone 21:05:44 if you discover a critical or high bug please tag the bug and contact me prior to targeting to RC1 21:06:47 At this point we should be less than 10 bugs (and we're 20+), so we'll need to make sure we can iterate on the reviews swiftly 21:07:21 Also reviewers should be on the lookout for string changes 21:07:45 as revising strings this late in the cycle add extra work for the translators 21:07:55 Any questions? 21:08:07 I dont see an owner for this one: https://bugs.launchpad.net/neutron/+bug/1244255 21:08:09 Launchpad bug 1244255 in neutron "binding_failed because of l2 agent assumed down" [High,In progress] 21:08:28 #topic Bugs 21:08:30 uvirtbot: thinking about the same 21:08:31 emagana: Error: "thinking" is not a valid command. 21:08:44 emagana: that was on my list to discuss 21:09:03 looks like uvirtbot is back after a few weeks away :) 21:09:19 I think the bot comes from a totalitarian country 21:09:43 * mestery welcomes uvirtbot back. 21:09:52 rkukura did't you request this bug to reopened last week 21:09:53 ? 21:09:53 mestery 21:10:15 I welcome the bot as well, even if he urges us not to think 21:10:21 or she for what I know ;( 21:10:24 markmcclain: I was just suggesting that the merge changing the heartbeat times may have helped it 21:10:33 salv-orlando: :P 21:10:43 oh so I can close this again? 21:11:03 Does it still show up markmcclain? 21:11:39 salv-orlando: haha 21:11:45 from what I recall it does not, but the issue was that the patch did not actually fix the underlying issue? 21:12:07 mestery: looking at the logstash query no 21:12:15 even if it tweaked things in a way that the problem does not occur anymore in gate/check tests 21:12:31 but I have blurred memories, made worse by beer 21:12:46 I'm closing and if it reappears we'll open a new bug with a better query 21:12:57 salv-orlando: you mean scotch 21:13:03 markmcclain: +1 21:13:05 +1 to that markmcclain 21:13:22 I second motion! 21:13:27 We have 2 critical bugs 21:14:00 jlibosva: how is the grenade work progressing? 21:14:17 mestery: any chance to look at https://bugs.launchpad.net/bugs/1291535 ? 21:14:21 Launchpad bug 1291535 in neutron "'Unable to retrieve OVS kernel module version' when _not_ using DKMS openvswitch module " [High,Triaged] 21:14:35 markmcclain: I plan to look at this afternoon after the meeting yet. I'll have an update in a few hours. 21:14:41 markmcclain: tempest tests fail after upgrade due to empty vif_details column 21:15:15 mestery: sounds good 21:15:26 jlibosva: ok so we have an incomplete migration? 21:15:31 I wrote an email on openstack-dev earlier this day how to deal with it 21:15:38 ok thanks! 21:15:40 jlibosva: ah I missed that 21:16:03 markmcclain: no, it's because vif_details column is new in Icehouse in ml2 port bindings table 21:16:13 scheme changed successfully 21:16:23 jlibosva: Is this after running the migration tool? 21:16:52 rkukura: no, but I guess it wouldn't help because havana was already using ml2 21:17:05 afaik the tool is lb/ovs -> ml2 21:17:16 jlibosva: Or just upgrading ml2 from havana->icehouse? 21:17:36 rkukura: yes, upgrading ml2 havana -> icehouse 21:17:40 jlibosva: the tools should be ml2 -> ml2 21:17:53 My fix for https://bugs.launchpad.net/bugs/1276391 actually should fix that 21:17:54 Launchpad bug 1276391 in neutron "ML2 MechanismDriver.bind_port() called inside transaction" [High,In progress] 21:17:58 great, I probably missed the latest patchset 21:18:16 the fix isn’t quite ready for review yet 21:18:32 rkukura: still targeting rc-1 for that? 21:18:58 it is, and is getting very close to ready for review 21:19:14 so we're not cutting until this is ready? 21:19:38 ideally we should have code posted for this 21:19:50 rkukura: when can we get this into review 21:20:20 I keep saying tomorrow 21:20:39 Once I classify this as a blocker the release management team the release management team will want a hard commitment 21:21:29 if you're short bandwidth to work on this we'll need to find someone else to pick up the task 21:22:01 also how what is the expected size of this fix? 21:22:08 I will get it into review tomorrow 21:22:30 hmmm :/ 21:22:41 no chance of getting something today? 21:22:57 maybe 21:22:58 hello all, sorry am late 21:23:25 Is the patch relatively small and contained… because otherwise we will need to take in account time for review, as well 21:23:28 rkukura: may it make sense to push draft patch so someone could pick it up? 21:23:47 or wip patch 21:24:29 I’ll do a WIP tonight if its not ready 21:24:34 rkukura: if you got time we'll chat offline after the meeting, so I've got the details for the release team 21:24:39 sounds good to me. 21:24:46 markmcclain: OK 21:24:48 Any bugs the team needs to discuss? 21:25:12 This one https://review.openstack.org/40296 ? 21:25:39 I think it may be too risky for icehouse. 21:25:42 tquite dangerous at this point, no? 21:25:57 I'm rebasing 21946, struggling with UT failure because this patch touches many plugins, but I think i can push new version in few hours 21:26:00 so please review it 21:26:08 after I push it 21:26:38 I think 40296 is one of those patches for which unit tests ensure correct functionality… but it's still rather large 21:26:58 also enikanorov_ you might want to add your views regarding the comments you had 6 months ago 21:27:26 not so blurred your memory is, eh? 21:27:28 :) 21:28:01 salv-orlando: as far as practice with drivers goes, I'm fine with this change 21:28:12 HenryG: let's keep working on this 21:28:20 create_all() causes lots of problems 21:28:32 * mestery has heard the horrors of create_all(). 21:29:03 Biggest problem wit 40296 is that quotas table migration is missing in most plugins. 21:29:08 nati_ueno: thanks for the heads up on 21946 21:29:09 also… technically the patch is not targeted to RC-1. So we can keep working on it, and defer the decision to when we'll see it as mature enough 21:29:10 the only thing I'm thinking about is revisiting those patches dedicated to fixing migration errors 21:29:15 and synching models with migrations 21:29:35 enikanorov_: I've seen those bugs in reviews 21:29:45 and we've been merging the fixes as their ready 21:30:09 ok 21:30:13 Any other critical/high bugs? 21:30:48 Any update on the VIF security fix? 21:31:36 rkukura: I'm rebasing, and facing many UT failures . IMO, I can push new patch in few hours 21:31:48 https://review.openstack.org/#/c/81124/3 is ready for review. 21:32:07 nati_ueno: can you remind me if this patch also has a nova counterpart? 21:32:10 nati_ueno: Great! I’ll review it ASAP 21:32:15 pcm_: I'll take a look after the meeting 21:32:22 markmcclain: thanks 21:32:29 salv-orlando: yes. This one has nova patch too 21:32:36 rkukura: Thanks 21:33:00 salv-orlando: https://review.openstack.org/#/c/44596/ 21:33:47 thanks nati-ueno 21:33:47 nati_ueno: I'm concerned this patch is abandoned 21:33:59 markmcclain: I'll restore just now 21:34:04 because that increases the chance it will fall off the Nova teams radar 21:34:20 markmcclain: you are right. sorry. I just restored 21:35:02 This patch will fail without neutron side fix, 21:35:06 RC1 are supposed to be cut _this week_ 21:35:08 so I focused on the neutron side 21:35:39 question on bugs/fixes: was there any bugs affecting floating IP connectivity that were found/introduced/fixed recently? 18-24 of march? 21:35:52 does anyone aware of a case? 21:35:58 enikanorov_: to my knowledge no 21:36:09 i see 21:36:20 not that my aware of 21:36:35 i'm asking because there was a spike in lbaas scenario failures 18-20 of march so we had to disable it in the gate. 21:36:53 but not changes were done to lbaas that could affect the test 21:37:34 enikanorov_: I thought there were changes to tempest that enabled a flaky test 21:38:27 ok I think we've covered critical bugs, so lets move onto docs 21:38:30 #topic Docs 21:38:32 emagana: hi 21:38:39 markmcclain: Hi There! 21:38:49 Good progress this week! I updated the wiki 21:39:08 the work around the re-structure of the Networking Installation Guide is going on 21:39:21 We need more eyes on: https://review.openstack.org/#/c/81983/ 21:39:21 ok 21:39:56 * Sam-I-Am wrote 81983 21:40:08 also working on the rest of the install guide 21:40:11 most of the bugs are WIP or will be covered by this BP: 21:40:14 ok so this is something that we can have multiple community members look at 21:40:22 https://blueprints.launchpad.net/openstack-manuals/+spec/networking-install-guide-improvements 21:40:34 and this one: https://blueprints.launchpad.net/openstack-manuals/+spec/improve-networking-cloud-admin-guide 21:40:35 * phil_h looking at it also 21:41:03 Sam-I-Am: thanks for working on the docs 21:41:09 phil_h: thanks for looking at it 21:41:15 marun is helping on the RedHat side! 21:41:28 markmcclain: no problem. i want the process to launch a VM using neutron to be as easy as possible 21:41:44 emagana: not by much :\ we're waiting on new RDO packages (in QA as of yesterday) 21:41:49 markmcclain: with an understanding of whats going on 21:42:33 emagana: looks like we're in good shape on the priority doc bugs 21:42:40 and the medium ones look to be good too 21:42:59 so far, so good! Thanks everybody! 21:43:03 marun: once phil_h and i have a working rhel/rdo environment, we will add ML2 specific items to the install guide 21:43:10 so far only ubuntu works 21:43:41 Sam-I-Am: understood 21:43:49 Sam-I-Am: good to know 21:43:58 emagana: any other doc items? 21:44:03 no more! 21:44:38 emagana: thanks for updating 21:45:23 markmcclain: n.p.! 21:45:32 Looks like Nova Parity has been updated and is mostly same as last week for time let's skip to Tempest 21:45:35 #topic Tempest 21:46:02 salv-orlando, rossella_: Have been studying the Full tempest job 21:46:22 mostly rossella_s 21:46:23 #link http://lists.openstack.org/pipermail/openstack-dev/2014-March/030907.html 21:46:45 rossella_ ? 21:46:49 yes! 21:46:59 I wrote everything in the mail that markmcclain linked 21:47:04 rossella_: thank you for sending out the latest update 21:47:14 markmcclain: np! 21:47:34 bottom line is that check failure rate is 25% but at least 45% of those failures are not because of failures in neutron 21:47:48 that puts us in a worst case scenario of 13% failure rate 21:48:07 next steps is to look at the bugs rossella reported to see which ones need to be addressed 21:48:28 ok 21:48:30 they are all known bugs 21:48:33 and push some sort of dummy patch to have an estimate which excludes errors in patches under review 21:49:38 oh so push more patches like translations 21:49:46 or something that does not change logic? 21:49:54 the first you wrote 21:49:58 ok 21:50:20 markmcclain: what would be an acceptable failure rate? 21:50:24 0%? 21:50:47 yes anything greater and we run a risk of causing problems in the community 21:50:56 anyway top 2 bugs in my list are tempest/infra bugs 21:51:02 so, sorry for bringing it up again, we have tempest scenario test for lbaas skipped due to some strange error. right now it's a slight suspicion that https://bugs.launchpad.net/neutron/+bug/1293818 may be responsible, which is fixed by https://review.openstack.org/#/c/81537/ 21:51:03 Launchpad bug 1293818 in neutron "Agents don't need root to list namespaces" [Medium,In progress] 21:51:08 we might need to compare to the baseline for other projects 21:51:41 enikanorov_: how would that have anything to do with it? 21:52:01 enikanorov_: not using root would simply reduce the use of rootwrap and its performance hit 21:52:06 marun: i can explain after the meeting 21:52:06 enikanorov_: we can discuss in open discussion or after meeting 21:52:10 enikanorov_: ok 21:52:19 want to focus on testing with remaining time 21:52:31 btw, we have a functional job that allows sudo invocation... 21:52:47 it's success rate is 100% since coming online this week. i'd like to move it to voting asap 21:52:54 marunb: Cool! That's awesome! 21:53:08 marun: awesome 21:53:10 (basically because it only runs 4 tests) 21:53:20 marun: :P 21:53:26 but it opens the door to all kind of investigative testing that tempest isn't good for 21:53:29 marun: have to start somewhere 21:53:34 so back to the full job 21:53:55 looks like rossella_ and salv-orlando have a good handle on it for now 21:54:18 when there are actionable items for the rest of the team feel free to ask in IRC or on the mailing list 21:54:54 markmcclain:thanks...I will keep monitoring the full job 21:54:57 rossella_ and salv-orlando thanks again for working on this as it is a big item for parity with the rest of the community 21:55:19 :) 21:55:21 mlavalle: I know we're running short of time anything else to highlight for tempest? 21:55:31 thanks rossella_ and salv-orlando! 21:55:46 the agenda is updated with the latest status, so we can move on 21:56:01 good progress last week 21:56:02 mlavalle: thank you for updating 21:56:12 good see the progress the team has made 21:56:48 #topic Open Discussion 21:57:05 Looks like the rest of the agenda has been updated with sub team reports 21:57:21 Any items that need to raised with the whole team in the time remaining? 21:58:20 might this be the first time we finish a meeting with two minutes to spare? 21:58:49 salv-orlando: I'm going to make you wait 2 minutes :) 21:58:53 two minutes are too short to start meaningful discussion 21:58:59 unless it's about scotch or beer 21:59:01 must be missing something... 21:59:04 Posted a couple of blueprints for review. 21:59:10 #link https://blueprints.launchpad.net/neutron/+spec/neutron-auditlogging #link https://blueprints.launchpad.net/neutron/+spec/neutron-port-template-framework #link https://blueprints.launchpad.net/neutron/+spec/neutron-port-template-policy 21:59:33 the ml2 migration script still needs reviews: https://review.openstack.org/#/c/76533/ 21:59:45 marun: thanks for raising that review 22:00:35 marun: I'll follow up with you after the meeting since we're out of time 22:00:42 markmcclain: ok 22:00:43 Bye 22:00:48 ciao 22:01:07 I know we discussed many bugs this week, but closing them is crucial to a successful to the RC 22:01:26 Thanks to everyone for their hard work on polishing this release 22:01:38 Have a good week and talk to everyone on the ML and IRC 22:01:41 #endmeeting