21:00:34 #startmeeting quantum 21:00:35 Meeting started Mon Mar 4 21:00:34 2013 UTC. The chair is danwent. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:36 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:38 The meeting name has been set to 'quantum' 21:00:48 hi! 21:00:50 hi all! 21:00:52 #info agenda: https://wiki.openstack.org/wiki/Network/Meetings 21:01:00 #topic announcements 21:01:09 #info congrats to nati_ueno on becoming a member of quantum core! 21:01:17 Thanks! ALL! 21:01:33 #info we are now in full grizzly lock down… one week away from the G-RC1 drop 21:01:38 hi 21:01:51 hi 21:01:51 he 21:01:56 RC1 should contain all fixes we believe ware needed for the grizzly release, such that if no blocker bugs are found, it becomes the grizzly release. 21:01:58 I mean i 21:02:50 #info starting tomorrow, the bar for merging a bug becomes even higher: unless it is something that we would merge into a stable maintainence branch, don't merge it into RC1 without a discussion with the core team. 21:03:19 #info this week, our focus should be on knocking out the last of the bugs, testing more, and spinning up our documentation efforts. 21:03:27 any other announcements? 21:03:54 #topic Quantum RC1 feature freeze exceptions (danwent 21:03:55 ) 21:04:04 good news on this front. 21:04:21 #info both FFEs for the quantum team have merged. thanks for the hard work from yong + mark 21:04:39 thanks for the review team. 21:04:40 #info we could still use eyeballs on a devstack patch for lbaas: https://review.openstack.org/#/c/23382/ 21:04:59 #info for testing, i updated: https://wiki.openstack.org/wiki/Quantum/LBaaS/HowToRun 21:05:43 great. I just want to know something about Lbaas today. 21:05:44 gongysh: looks like https://review.openstack.org/#/c/23396/ merged, so no need to ask for more reviews. 21:06:05 yes 21:06:15 thanks for salvatore and amotoki. 21:06:17 gongysh: we'll talk about the sqlalchemy session exhaustion bug in the next section. 21:06:26 gongysh: did you have a question about lbaas? 21:06:41 no, not yet. I will have after testing. 21:06:46 gongysh: ah, ok 21:06:57 #topic Grizzly-RC1 Bugs (markmcclain ) 21:07:18 markmcclain: danwent: https://review.openstack.org/#/c/23406/ 21:07:46 garyk: yup, i think that's on the list, correct? 21:08:04 I think markmcclain is going to go through all open bugs targeted for RC1 quickly, and confirm we have someone working on it. 21:08:05 danwent: hope so :) 21:08:15 ok, sorry for bargin in 21:08:27 garyk and danwent it's on the list 21:08:37 markmcclain: tx 21:08:46 ok we've got 14 open bugs 21:09:09 hehe, actually, seems to have gone iup in the past few hours :) 21:09:09 we've got two that need to have work started: 21:09:26 https://bugs.launchpad.net/quantum/+bug/1084355 21:09:28 Launchpad bug 1084355 in quantum "Tap interface does not automatically get an IP address upon a reboot" [High,Confirmed] 21:09:40 https://bugs.launchpad.net/quantum/+bug/1136064 21:09:42 Launchpad bug 1136064 in quantum "quantum server consumes all of sqlalchemy sessions" [High,Confirmed] 21:09:43 i found at least one issue in nova releated to that a while back, still need to commit 21:09:53 (i'm talking about 1084355) 21:10:06 I will make sure that one gets taken care of. garyk is actve on it as well. 21:10:25 ok sounds like we two cores on it 21:10:31 1136064 21:10:50 the SQL exhaustion issue that gongysh raised earlier 21:11:00 gongysh: any update on https://bugs.launchpad.net/quantum/+bug/1136064 21:11:03 Launchpad bug 1136064 in quantum "quantum server consumes all of sqlalchemy sessions" [High,Confirmed] 21:11:11 seems to be our most critical bug out there, at a glance 21:11:45 agreed 21:12:22 ok, gongysh may be offline temporarily. let's circle back later to this bug. 21:12:31 ok 21:12:33 no, I am here. I am thinking. 21:12:36 markmcclain: i am using qpid so the behavious is a little different. i'll try and reproduce tomorrow 21:12:37 ah, ok :) 21:13:00 i definitely saw that issue testing with nvp plugin 21:13:21 I will write a program to reproduce it when quantum server is alive. 21:13:31 ok, as long as we know it has gongysh's full attention, i think we can move on 21:13:43 gongysh: please let me know whenyou have it. 21:13:46 main goal is just to make sure that important bugs get the attention they need this week, and others are bumped out. 21:13:57 yeah.. we also have 4 other cores looking at it too 21:14:02 garyk: I always welcome your help. 21:14:11 :) 21:14:19 https://bugs.launchpad.net/bugs/1100282 21:14:21 Launchpad bug 1100282 in quantum "DoS through XML entity expansion (CVE-2013-1664)" [High,In progress] 21:14:35 The code has been submitted for this change, but needs a small change 21:15:16 https://review.openstack.org/#/c/23024/ 21:15:30 ok, looks like markmcclain and garyk are actively reviewing, we're good here 21:16:21 https://bugs.launchpad.net/bugs/1103840 21:16:22 Launchpad bug 1103840 in quantum "Deletion of DHCP should be notified to the security group rpc" [High,In progress] 21:17:03 nati_ueno: is this needed to fix broken functionality? can't tell from bug title 21:17:17 did not look really high to me as well 21:17:32 but it seem the patch is quite mature and self contained 21:17:41 danwent: We should delete iptables rule entries for DHCP, when the DHCP is deleted. 21:17:49 nati_ueno: ah, ok. 21:17:55 yeah, patch seems small enough anyway 21:18:05 danwent: so impact is about garbage collection 21:18:07 garyk + amotoki are reviewing 21:18:12 so I think we're ok 21:18:17 may be not high but it good to have it 21:18:20 i feel it's not high since it is not an usual usecase changing dhcp ip or toggling enable_dhcp. 21:18:32 yeah.. we're in good shape with most bugs having 2+ cores 21:18:41 the one I wanted to discuss was this: 21:18:47 yes. so I'll trying to get merged it in today otherwise wait stable 21:18:53 https://bugs.launchpad.net/quantum/+bug/1144426 21:18:54 Launchpad bug 1144426 in quantum "The name of the security group rule field "source_ip_prefix" is misleading" [High,In progress] 21:18:57 amotoki: thinking about agent schedule, the dhcp IP is changing often. 21:19:47 markmcclain: i'm in favor of the change 21:20:02 ok.. the review is here: https://review.openstack.org/#/c/23451/ 21:20:08 with egress rules, using "source" is confusing. In fact, I thought we had cleared this up with a discussion a while back, but the change never got in 21:20:09 1144426 impacts the API, so we should have it in G. 21:20:09 gongysh: thanks. I don't think dhcp ip *often* changes.... *sometimes* changes. 21:20:15 its important to fix this before the API is released. 21:20:30 danwent: +1 21:20:34 arosen, you had a chance to python-quantumclient. does this also require a change to nova proxy? 21:20:44 or maybe not, as nova doesn't support egress? 21:21:10 No change in the client for nova 21:21:31 k 21:21:31 https://bugs.launchpad.net/bugs/1104337 21:21:33 Launchpad bug 1104337 in quantum "Deleting a subnet that is added to a router leaves behind a port that cannot be deleted" [High,In progress] 21:21:34 the changes in order to drop source with remote for nova are here: https://review.openstack.org/#/c/23459/ 21:22:22 As we did not heard from the original dev for 1104337 I pushed a new patch for it 21:22:23 salv-orlando: looks like you're rescuing this bug. I think this deserves to go in. 21:22:26 yup 21:22:35 https://bugs.launchpad.net/bugs/1115999 21:22:36 Launchpad bug 1115999 in quantum "quantum-netns-cleanup does not stop metadata proxies" [High,In progress] 21:23:06 this one needs to be fixed and we have 2 cores on 21:23:07 https://bugs.launchpad.net/bugs/1119409 21:23:09 Launchpad bug 1119409 in quantum "Delete is possible of an in-use port" [High,In progress] 21:23:29 i personally think this doesn't need a fix in quantum 21:23:42 danwent: i agree 21:23:47 agreed.. the review has been abandoned 21:23:50 I'll move it out 21:23:56 i think we could add a fix in pytyon-quantumclient + horizon to require a --force or confirmation before deleting a port with a device_owner or device_id set 21:24:12 fixing this would require us to change the nova model for using an admin context. So it's too much imo 21:24:30 salv-orlando: yeah, i actually don't think the api should ever change 21:24:47 https://bugs.launchpad.net/bugs/1136264 21:24:49 Launchpad bug 1136264 in quantum "Linuxbrige agent downs with tap device deletion timing issue" [High,In progress] 21:25:05 fix is here: https://review.openstack.org/#/c/23259/ 21:25:28 markmcclain: amotoki had some very good insight here. hopefully it will be addressed following his comments 21:25:48 ok.. looks like you both have this handled 21:26:02 ok, looks like amotoki and garyk have this one handled 21:26:13 https://bugs.launchpad.net/bugs/1143283 21:26:14 Launchpad bug 1143283 in quantum "Default behavior of Security groups egress handling has different semantics than ingress" [High,In progress] 21:26:44 make this change in grizzly makes sense I think, as it would be hard to change afterward if the api is already out 21:26:52 we had an discussion on the ML about this. 21:27:00 if anyone disagrees, lets discuss on the review 21:27:13 tomoe_: thanks for pointing this out 21:27:16 https://bugs.launchpad.net/bugs/1143569 21:27:18 Launchpad bug 1143569 in quantum "make quantum pep8 version + exceptions match other openstack projects" [High,In progress] 21:27:27 i think this is merging already 21:27:30 we need throughout tests across various plugins. 21:27:42 danwent: NP. i'm working on fixing tests now 21:27:52 please open your eyes on the patch if you are plugin maintainers. 21:27:55 hopefuly i can submit another patch soon 21:27:56 tomoe_: there's a review in progress already, right? 21:28:19 danwent: yes. but that one is not quite ready 21:28:23 tomoe_: i don't see it associated with laounchpad bug 21:28:29 review is in progress. we are wating for test codes. 21:28:43 https://bugs.launchpad.net/quantum/+bug/1116137 21:28:44 Launchpad bug 1116137 in quantum "Missing error handle for quota" [Medium,In progress] 21:28:51 review here: https://review.openstack.org/#/c/21197/ 21:29:23 cleans up api error reponses looks like danwent and salv-orlando are reviewing this one 21:29:24 tomoe_: i added it 21:29:41 https://bugs.launchpad.net/bugs/1126661 21:29:42 Launchpad bug 1126661 in quantum "move cisco qos + credential extension to plugin dir" [Medium,In progress] 21:29:52 danwent: thanks. 21:29:59 markmcclain: yes, we'll take care of it 21:30:13 danwent: was it really as easy as moving the files and fixing those few imports? 21:30:44 salv-orlando: i can't test their plugin, so arvind is test it. 21:30:54 k, thanks 21:31:06 https://bugs.launchpad.net/bugs/1136027 21:31:07 Launchpad bug 1136027 in quantum "improve unit tests for wsgi module" [Medium,In progress] 21:31:11 they will need to update their plugin path to include plugins/cisco/extensions 21:31:33 danwent: seems reasonable for vendor extensions 21:31:41 https://bugs.launchpad.net/bugs/1136027 21:31:55 markmcclain: yeah, it was supposed to always be that way, but we caught a few exceptions. 21:32:12 this is 100% unit tests and really just needs the commit message updated 21:32:15 wsgi tests patch is big, but all unit tests code 21:32:17 yeah 21:32:20 SumitNaiksatam has been leading this one 21:32:41 as long as it merges by tomorrow i'm fine 21:32:47 https://bugs.launchpad.net/bugs/1138757 21:32:48 Launchpad bug 1138757 in quantum "l3-agent is not working with Metadriver" [Medium,In progress] 21:33:13 just moves some configurations opts around 21:33:17 needs two cores 21:33:35 will review in background :) 21:33:36 I can be one 21:33:36 review is here: https://review.openstack.org/#/c/23462/ 21:33:44 I can be one too 21:33:48 I am reviewing it. 21:33:55 nati_ueno: you can get your first +2 :) 21:34:06 hi sorry to interject.. anyone seeing pep8 errors on Master when doing a ./run_tests.sh? 21:34:08 haha, ok amotoki + nati_ueno you take it 21:34:08 markmcclain: wow scary.. 21:34:14 https://review.openstack.org/#/c/23462/ 21:34:23 absubram: let's talk about it in open discussion at end of meeting 21:34:37 oops.. wrong linkk 21:34:38 https://bugs.launchpad.net/bugs/1124117 21:34:40 Launchpad bug 1124117 in quantum "security group extension should be disabled when Noop Firewall driver is used" [Low,In progress] 21:35:06 thanks danwent 21:35:44 markmcclain: as i indicated in the review, i'm ok with that change, even though its on the kind of big side, as the changes are pretty trivial and repeitive, 21:35:57 and it reduces the risk of someone doing something insecure 21:36:29 last one... 21:36:29 https://bugs.launchpad.net/quantum/+bug/1144370 21:36:31 Launchpad bug 1144370 in quantum "log the configuration options for metadata agent and proxy" [Medium,In progress] 21:36:32 sorry for becoming the test code bigger... 21:37:07 i ran into a similar issue where i wanted to see logs for the metadata proxy, so this seems valuable to me 21:37:09 It will help me to answer metadata related questions in the furture. 21:37:20 looks like we need one more core dev on that 21:37:40 yeah… I've looked at this briefly, and will take a look 21:37:58 we also have https://bugs.launchpad.net/bugs/1144076 21:37:59 Launchpad bug 1144076 in quantum "Quantum CLI does not show the default Quotas" [Medium,In progress] 21:38:02 or did we talk about that already? 21:38:23 we did a little.. garyk raised it and forgot to come back to it 21:38:55 i think this one is valuable, assuming the patch is reasonably safe 21:39:02 I think it is good to see default quotas when tenant-per is not available. 21:39:40 yup 21:39:41 It is all about viewing. 21:39:46 danwent: ageed 21:39:50 ok, any other RC1 bugs we did not discuss? 21:40:07 that should be all of them 21:40:10 when are we planning the rc1 release? 21:40:12 markmcclain: thx 21:40:29 #info target for grizzly RC1 release is March 12th 21:40:36 thanks 21:40:43 Is that possible to fix this one too https://bugs.launchpad.net/quantum/+bug/1085276 ? 21:40:44 Launchpad bug 1085276 in quantum "The router interface cannot be deleted by using Quantum Metaplugin" [Undecided,In progress] 21:41:02 review is here https://review.openstack.org/#/c/17571/ 21:41:37 patch looks a bit larger than I would have expected, and changes from core code 21:41:48 Ok so let's wait stable 21:41:51 In the recent patchset , new bug fixes are added. I recommended it should be split. 21:42:09 nati_ueno: i'm not saying one way or another. we'll have to look at the review in more detail 21:42:25 I agree with amotoki that it should be split per bug 21:42:32 amotoki: ok, splitting this into more targeted fixes would be a good idea 21:42:41 also looks like there is code from other proposals mixed in 21:42:42 danwent: OK, so Akihiro, and Summit is reviewing this bug. so If this is not for RC1 please -1 21:42:50 we can target the resulting items for RC1 21:43:09 if any of the core devs reviewing feel like it is too risky for RC1, let's raise it for the team, otherwise, proceed. 21:43:18 ok, next topic 21:43:30 ony 15 minutes left, so need to hurry a bit 21:43:35 #topic python-quantum client status 21:43:40 gongysh & markmcclain 21:43:45 hi, 21:44:06 that is good for client side, we have a fix for source_group_ip. 21:44:18 others are unit test patches. 21:44:41 we can push 2.2.0 when server's RC1 is out. 21:44:51 that is it. 21:44:54 gongysh: we actually need to push before 21:44:59 gongysh: i thought we were planning on pushing before 21:45:00 yes 21:45:01 before? 21:45:12 b/c we need other projects to update their dependencies 21:45:15 gongysh: 2.2.0 is required before nova and horizon rc1 are shipped. 21:45:34 markmcclain: we were planning on march 6th? 21:45:42 i forget exact date 21:45:42 Ok, just wait source_ip problem. 21:45:50 nova requires source_group fix and horizon requires lbaas api update. 21:45:53 danwent: yes 21:46:31 #info python-quantumclient 2.2 release planned for march 6th, to be used by nova / horizon for grizzly 21:46:38 ok, anything else on client/cli? 21:46:53 not from me 21:46:58 amotoki: we'll go a bit out of order and do horizon + quantum next 21:47:03 https://blueprints.launchpad.net/horizon/+spec/quantum-lbaas 21:47:10 #topic horizon + quantum (amotoki) 21:47:11 Need feedback from LBaaS team before merging the patch and need testing with HAproxy. 21:47:31 amotoki: https://wiki.openstack.org/wiki/Quantum/LBaaS/HowToRun 21:47:45 maybe link horizon lbaas branch here and point people to test it? 21:48:05 that would be helpful for testing 21:48:09 I will add the link to the above page. 21:48:41 ok, anyting else for horizon + quantum? 21:48:49 no from me. 21:48:55 i will let kc know about the wiki page 21:49:08 thx 21:49:11 i will send an horizon status update later today. 21:49:17 #topic grizzly documentation (danwent) 21:49:53 I have updated the security group docs here https://review.openstack.org/#/c/20636/ 21:49:58 Ok, its the time in the release cycle when we need to start putting most of our energy toward documenting things, and then testing that the documetnatioin works. 21:50:06 tomorrow i'll update the vif's 21:50:18 arosen: thanks. 21:50:29 who can tell me how to test the isolated network metadata feature? 21:50:30 I've created a grizzly docimpact page for quantum: https://wiki.openstack.org/wiki/Network/GrizzlyDocImpact 21:50:54 please link doc bugs there, so we can see what we already have a bug for, and where we have gaps 21:51:09 i'll be coordinating grizzly doc efforts 21:51:25 why not use openstack-manuals directly? 21:51:42 if you file a bug against openstack-manual, and it's something that impacts the API too, please do not forget about openstack-api-site 21:51:51 gongysh: we will track items in openstack-manuals for admin guide items 21:52:11 gongysh: goal with the wiki page was just to collect items from each of the blueprints merged into quantum in grizzly 21:52:23 (i.e., to track everything with a doc impact) 21:52:38 once we know we have everything in openstack manuals, we'll work just off of the links. 21:52:47 #info grizzly doc impact wiki page: https://wiki.openstack.org/wiki/Network/GrizzlyDocImpact 21:52:51 it is ok, just don't want to duplicate efforts. 21:53:02 #info: openstack manuals bug page: https://bugs.launchpad.net/openstack-manuals/+bugs?field.tag=quantum 21:53:23 #info: openstack quantum api bug page: https://bugs.launchpad.net/openstack-api-site/+bugs?field.tag=netconn-api 21:53:56 gongysh: are there any changes to the CLI doc for grizzly? I'm not sure where those would be tracked. 21:54:19 I will find the location. 21:54:27 and update the wiki page. 21:54:30 expect to see emails from me as we identify doc bugs related to code you wrote :) 21:54:34 gongysh: thanks 21:54:44 :) 21:54:44 #topic open discussion 21:55:03 ok, i think we had one question during the meeting that was bumped to open discussion 21:55:27 i need to run right at the top of the hour though… have to get to another meeting 21:55:36 ok…. going once, going twice? 21:55:36 How about we have etherpad pages for Hanova summit presentation? 21:55:55 nati_ueno: we will, but I don't want people too focused on havana summit until we've shipped RC1 21:56:04 danwent: I agree. 21:56:04 summit is still more than a month away 21:56:21 so fear not, you will not be cut out of any planning if you have not started thinking about the summit yet 21:56:39 we'll start discussing topics, making etherpads, etc. once RC1 is out. 21:56:59 danwent: I got it. I just worring about the voting for the session. 21:57:06 ok, have a good week folks. great work getting grizzly wrapped up from a code perspective. 21:57:09 voting? 21:57:14 nati_ueno: technical sessions have no voting 21:57:21 those are only for the conference sessions 21:57:23 not the summit 21:57:25 aga,, what's a misunderstandings... 21:57:29 Thanks 21:57:35 we'll discuss as a team what our summit session priorities are 21:57:42 then allocate sessions. 21:57:49 ok, talk to you later folks! 21:57:53 #endmeeting