18:32:46 #startmeeting Networking FWaaS 18:32:47 Meeting started Wed Aug 6 18:32:46 2014 UTC and is due to finish in 60 minutes. The chair is SumitNaiksatam. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:32:48 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:32:50 The meeting name has been set to 'networking_fwaas' 18:32:58 #info agenda https://wiki.openstack.org/wiki/Meetings/FWaaS 18:33:08 SridarK: thanks for update the meeting wiki page 18:33:15 SumitNaiksatam: so try it - you can chair and unchair me :) 18:33:22 SumitNaiksatam: no worries 18:33:37 regXboi: in think unchair might work 18:33:45 regXboi: we probably need the chairs option 18:33:59 #chair SumitNaiksatam SridarK regXboi 18:34:00 Current chairs: SridarK SumitNaiksatam regXboi 18:34:04 and there you go 18:34:16 regXboi: awesome, we are good now! :-) 18:34:22 and now you can unchair me and I'll go away :) 18:34:26 or actually 18:34:29 #unchair regXboi 18:34:30 Current chairs: SridarK SumitNaiksatam 18:34:34 regXboi: you are welcome to hang in there 18:34:36 I'll unchair myself :) 18:34:44 no - have to go fight another battle :/ 18:34:46 later all 18:35:05 #topic Action item follow up 18:35:22 meeting with the DVR folks - do we need one? 18:35:39 SumitNaiksatam: i think it will be good to do that nevertheless 18:35:52 SridarK: sure, f2f or virtual is good? 18:36:01 SumitNaiksatam: hopefully by next week we will be in a decent shape 18:36:04 SridarK: perhaps we can start with webex to begin with 18:36:17 SumitNaiksatam: yes exactly and we can make the trip if needed 18:36:21 SridarK: okay sure, should we put something on the calendar for next tuesday? 18:36:28 SridarK: i mean WebEx 18:36:31 SumitNaiksatam: sounds good 18:36:38 i can set that up 18:36:39 i am able to configure the dvr as a centralized 18:36:52 on a single node 18:37:00 applied the firewall rules 18:37:11 as we expected they do not applpy on snat name space 18:37:18 badveli: hang on 18:37:28 badveli: lets do the update in the dvr topic 18:37:36 ok 18:38:15 #topic WebEx meeting with DVR team for Aug 12 18:38:19 #undo 18:38:20 Removing item from minutes: 18:38:30 #action WebEx meeting with DVR team for Aug 12 18:39:18 we had another action item “SridarK badveli to setup FWaaS DVR support wiki page, provide spec/impl/dependency links, send email to -dev with pointer to this wiki page” 18:39:33 SumitNaiksatam: still lagging on that 18:39:36 sumit as part of it i had got some info 18:39:47 SridarK badveli: np 18:39:53 will sync up with sridar and update 18:39:57 #action SridarK badveli to setup FWaaS DVR support wiki page, provide spec/impl/dependency links, send email to -dev with pointer to this wiki page 18:39:59 badveli: thanks 18:40:00 SumitNaiksatam: i think would like to get some basic setup going and then update with some meaningful info 18:40:13 yes 18:40:14 SridarK: even placeholder for now is good 18:40:30 i had action item to follow up with mlavalle 18:40:34 SumitNaiksatam: ok sounds good then we will do that 18:40:35 on tempest, which i did 18:40:49 so, i think we are good on that 18:41:07 we will discuss tempests test for service objects in that corresponding topic 18:41:18 yes 18:41:33 sumit:we will discuss there i have some update 18:41:55 #topic FWaaS support for DVR 18:42:08 badveli: please go ahead with what you were stating earlier 18:42:36 sumit: I was able to configure dvr in a single node 18:43:16 dvr_snat option 18:43:33 i saw that snat name space 18:43:39 badveli: thats great 18:43:53 does not get the firewall rules applied as we expected 18:44:15 sumit: we were planning to change that 18:44:42 the firewall gets applied on the internal router 18:45:43 SumitNaiksatam: badveli and i discussed and i am trying to set up on a server to eventually get to a multinode so we can also validate on the distributed router - running into some tangential issues 18:45:45 sumit: since the compute, network everything exists in the same node 18:46:09 SridarK badveli: ah ok 18:46:45 badveli: i believe you mentioned earlier that you were going to try a two node set up? 18:47:11 yes sumit, i have another server but it needs a update 18:47:38 we were having an esxi older version i need to update 18:47:56 so i cannot use it 18:48:07 badveli: ok good, perhaps worth spending time doing that first then on the single node 18:48:27 that said, is DVR a misconfiguration on a single node? 18:48:36 fine sumit looks good 18:48:50 the dvr wiki says we can configure like that 18:48:56 SumitNaiksatam: the single node enables one to "kick the tires" on DVR 18:49:08 SumitNaiksatam: as stated on the DVR wiki 18:49:45 SumitNaiksatam: makes sense as a single node which could compute nodes 18:51:08 SridarK badveli: so FWaaS should still work on that right? 18:51:17 or are we not going to allow that? 18:51:27 SumitNaiksatam: we should allow 18:51:39 by putting it on the correct NS 18:52:12 yes 18:52:17 SumitNaiksatam: i think we should view this as a DVR with a single node 18:52:28 as opposed to the non DVR mode 18:52:37 SridarK: yes agree 18:52:44 where we will go to todays implemenation 18:52:51 yes we should put it in teh correct name space 18:53:03 SridarK: but i was asking in the context of badveli stating that it does not work 18:53:27 SridarK badveli: so our solution is going to be different for single node DVR versus multi-node DVR? 18:53:40 SumitNaiksatam: i think it will be same 18:53:56 SumitNaiksatam: with single node - we will be on the service node 18:54:04 i believe 18:54:14 sumit: on a single node since the compute,network are in the same 18:54:14 good point for clarification with the DVR team 18:54:39 node i was trying to think which name space we should apply the firewall 18:55:21 sridark: agreed we need more clarification 18:55:47 SumitNaiksatam: i think once we have the setup we will prototype on that quickly - badveli and i will close on this 18:56:15 SridarK: great 18:56:21 SumitNaiksatam: we would like to do this before we discuss with the DVR team 18:56:34 so we can have a meaningful conversation 18:56:51 i am just trying to make sure that we have not hit a technical roadblock in terms of getting this to work on a single node 18:56:57 SridarK: absolutely 18:57:21 SumitNaiksatam: agreed - we may be able to get this validated on a single node faster 18:57:48 on the multimode - i worry about the basic setup and some other tangential issues we could run into 18:57:58 *multinode 18:58:15 SridarK: completely agree 18:58:35 SridarK: however it might make sense to get to that sooner that later 18:58:55 SumitNaiksatam: yes agreed we will spin that in parallel 18:59:08 fine sridark and sumit 19:00:10 since the wiki is saying there is an option to configure on a single node, i started with that 19:00:22 since it is a service,network and compute node 19:00:29 badveli: makes sense to start simple 19:00:58 thanks sridar we can get a feel of it 19:01:03 badveli: i agree as well 19:01:09 badveli: SumitNaiksatam 's point is since the multinode is the final solution we should not lose sight of that 19:01:14 SridarK: do you have access to a couple of servers? 19:01:34 SumitNaiksatam: sigh - i have one - i am trying to get another one 19:02:24 SumitNaiksatam: will figure something out 19:02:53 SridarK: ah great 19:03:17 the reason why i asked was i wanted to check if we could start the multi-node setup in parallel at your end 19:03:26 SumitNaiksatam: agreed 19:04:04 Sridar and Sumit, I will also try that by bringing the second server 19:04:19 after updating it 19:04:45 badveli: great 19:05:03 badveli SridarK: so we know what the solution is for getting this to work on a single node? 19:05:23 SumitNaiksatam: yes we have a fair idea of what needs to be done 19:06:04 SumitNaiksatam: will feel more comfortable on trying some stuff out - we will target that by end of the week 19:06:34 Sridar looks like the same solution should be fine 19:06:50 for either the multi node or single node, right? 19:07:12 badveli: slight changes for the multi node 19:07:18 depending on the node 19:07:26 badveli: for diff NS 19:07:34 but change will be minor 19:08:12 badveli: we can work thru this to have good confidence level once we try out some simple tests 19:09:16 yaeh i meant the same, we understand what we need to do, will try 19:09:51 SridarK: can we document this in the wiki? 19:09:59 SumitNaiksatam: ok will do 19:10:17 SridarK: i mean, something very short, at least as a starting point - in terms of the differences between the two approaches 19:10:45 SridarK: since you seem to have thought through at least a little bit of this 19:10:48 SumitNaiksatam: sounds good - we allude to this in the spec but will good to capture the salient points 19:11:17 SridarK: ah good - i meant in the context of the single node deployment versus multi-node 19:11:30 in a more crisp manner - yes will do 19:11:46 ok got it 19:11:48 SridarK badveli: so what is our target in the coming week? 19:11:52 SridarK: thanks 19:12:06 can we have a bare bones WIP patch posted? 19:12:17 SumitNaiksatam: we will get some basic prototyping done and try to have a 2 node setup up 19:12:39 Sridark: agreed, I think we need some prototype 19:12:46 SumitNaiksatam: we can shoot for that but i think we may get to next week on posting a WIP 19:13:22 SumitNaiksatam: but we will see if we can get some success with a proto we will try to get a WIP out earlier 19:13:45 Yes Sridark i think that makes more sense 19:14:35 SridarK: at this point i am thinking that we dont necessarily have to tested out the patch 19:15:01 SumitNaiksatam: ok 19:15:04 SridarK: as long as we can put an initial version which is along the lines of what is stated in the bp 19:15:11 this will of couse be WIP 19:15:19 SumitNaiksatam: ok we will look at that in those terms 19:15:19 but will open it up to whoever wants to review 19:17:08 Sumit,Sridar i am able to get some thing about the driver side and the agent side 19:17:16 badveli: sweet 19:17:40 badveli SridarK: hate to push this, but can we target those WIP patches for the coming monday? 19:17:51 that way we can provide an update in the neutron team meeting 19:17:59 SumitNaiksatam: no worries we will defn shoot for that 19:18:49 Sumitt as you had mentioned we can try to code as specified in the spec, also we have some info from the dvr team 19:20:19 but functionality we are not tested 19:20:49 sumit: functionality testing we are not doing is it fine? 19:21:17 badveli: we have to do that, but i am saying does not have to be finished by monday 19:21:28 badveli: that is an ongoing effort 19:22:17 sumit: so our goal should be code as per the spec and improve on it based on our testing 19:23:29 i am asking this because we are not yet finished having the set ups 19:23:31 badveli: yes 19:23:43 would like to make sure that people get a chance to review a lot ahead of time 19:24:04 sumit: fine with me 19:24:09 #action SridarK badveli to attempt WIP patches by Aug 11th 19:24:22 SridarK badveli thanks much! and sorry for pushing! 19:24:36 this will also help for our tuesday meeting i guess 19:24:39 SumitNaiksatam: no worries - we are fast nearing end of the release 19:24:53 ok anything more to discuss on this front? 19:25:03 SumitNaiksatam: nothing from me 19:25:17 badveli: ? 19:25:18 nothing from my side too 19:25:37 SridarK badveli: great, thanks! 19:25:48 SumitNaiksatam: no worries 19:26:00 #topic Service Objects 19:26:06 badveli: again :-) 19:26:17 badveli: how is it coming along? any progress with tempest? 19:26:27 Sumit: Yes added tempest test 19:26:35 but need one small help 19:26:43 badveli: thats cool, yes sure 19:26:55 is there any way we can do a break while running tempest 19:27:00 the tests are done 19:27:23 but needed to find some failure issues 19:27:49 as we can do a debugging with the usual unit test suites 19:28:00 can we do the same with the tempest? 19:28:39 badveli: are u talking of something like a pdb.set_trace() 19:28:49 yeah i was just going to ask 19:28:52 yes sridark, i meant it 19:29:15 nothing mentioned in wiki 19:29:51 badveli: so which part of the code do you need to put the breakpoint in? 19:30:12 badveli: somehow my experience with this on tests is a bit mixed - i think i don't understand this as well as it is running in a venv and have not been very successful 19:30:20 it works great on the actual code 19:31:01 badveli: as a workaround - i have resorted to adding logs into files - not the best way 19:31:48 SumitNaiksatam: oops looks like we are at time 19:32:06 not sure if there is another mtg 19:32:09 SridarK: yeah 19:32:15 i dont think there is one 19:32:20 but we should wrap up 19:32:23 oh ok 19:32:25 Sumit: looks like this works in regular unit tests i wanted to break in the point where i am posting the command 19:32:40 badveli: on the tempest side on the neutron side? 19:32:45 tempest side 19:32:59 it does not work 19:33:00 badveli: if its on the tempest side, you can send a quick email to mlavalle and check it as well 19:33:10 fine sumit i will do 19:33:17 another issue is 19:33:42 i completed the driver side patch also 19:34:02 so i have almost all the patches 19:34:16 when running the db migration for the config,api 19:34:23 oslo.config.cfg.NoSuchOptError: no such option: mysql_engine 19:34:37 i enter into this problem 19:35:01 sumit.sirdar: do you have any idea 19:35:05 its holding me 19:35:23 it started recently 19:35:57 badveli: i think its an issue with the new oslo packages 19:35:59 badveli: there have been some changes in the db migration - perhaps u have something there 19:36:08 badveli: perhaps refreshign might help 19:36:21 badveli: also i think HenryG is a great person to check with on this 19:36:25 varmour@ubuntu-1-104:/opt/stack/neutron/neutron$ neutron-db-manage --config-file /etc/neutron/neutron.conf --config-file /etc/neutron/plugins/ml2/ml2_conf.ini revision -m "service_group_ext" --autogenerate 19:36:48 badveli: i can help u get in touch with HenryG - if u are not able to track him down 19:37:03 I am easy to track down :) 19:37:12 HenryG: :-) 19:37:53 badveli: there u have HenryG - u can send him an email too - i will send u his email off line if needed 19:38:04 Didn't I just +1 a review to fix that? 19:38:08 thanks HenryG,Sumit 19:38:40 HenryG: thanks, whats the link? 19:38:48 looking ... 19:39:37 https://review.openstack.org/112171 ? 19:40:55 thanks HenryG, I will change that file 19:41:17 HenryG: awesome thanks! 19:41:23 badveli: Let me know if that does not work for you 19:42:03 thanks HenryG, you saved my research. thanks for your help 19:42:40 HenryG: u have set the bar for response times now - wish all would even get to a fraction of this. :-) 19:43:19 SridarK: lol 19:43:23 +1 19:43:27 ok anything else on service objects? 19:43:35 since we are running short on time 19:43:39 Sumit> I will try to target to post all the patches 19:43:44 including the tempest 19:44:23 thanks for your helps 19:44:39 badveli: awesome effort in making progress 19:44:44 #topic open discussion 19:44:50 anything more today? 19:45:22 SumitNaiksatam: sent out bug scrub mail (i need to send one to myself too) - nothing much more to add 19:45:40 SridarK: that was great 19:45:48 SridarK: i did not see anything critical in taht 19:46:02 SumitNaiksatam: no worries - yes nothing critical 19:46:02 SridarK: i mean marked as critical 19:46:12 SumitNaiksatam: yes 19:46:20 SridarK: great thanks for going through the list! 19:46:32 SumitNaiksatam: no worries at all 19:46:34 SridarK: i think SridharG’s patch was merged? 19:46:41 i had +2’ed it earlier 19:46:43 SumitNaiksatam: yes that was great 19:46:54 SumitNaiksatam: yes 19:47:39 garyduan: anything you wanted to chime in on? 19:48:09 regarding flavor? 19:48:15 garyduan: yes 19:48:27 I talked to enikanorov 19:48:58 there is a neutronclient patch 19:49:02 beside neutron patch 19:49:22 oops looks like we lost SumitNaiksatam 19:49:22 but, first of all, we need get spec approved 19:49:36 :-) 19:50:05 i think my network connection if flaky again 19:50:05 Sumit is back. 19:50:23 ok 19:50:26 SridarK: please feel free to end the meeting if i am bounced out again 19:50:32 ok 19:50:37 will do 19:50:45 so, for flvar, waiting for spec to be approved 19:50:55 garyduan: yeah 19:51:15 garyduan: ok, you feel comfortable jumping on it for fwaas support, once its approved? 19:51:47 yes 19:52:00 garyduan: sweet 19:52:06 on that positive note 19:52:12 lets call it a wrap! 19:52:18 thanks all for joining 19:52:21 bye all 19:52:22 lots of work to do! 19:52:27 ok bye all 19:52:28 bye all 19:52:31 #endmeeting