18:31:57 <SumitNaiksatam> #startmeeting Networking FWaaS 18:31:57 <openstack> Meeting started Wed Feb 25 18:31:57 2015 UTC and is due to finish in 60 minutes. The chair is SumitNaiksatam. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:31:58 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:32:01 <openstack> The meeting name has been set to 'networking_fwaas' 18:32:06 <SumitNaiksatam> #info metting agenda https://wiki.openstack.org/wiki/Meetings/FWaaS#Agenda_for_Next_Meeting 18:32:42 <SumitNaiksatam> #info kilo-3 feature proposal freeze is March 5th, and milestone date is March 19th 18:33:16 <SumitNaiksatam> any other info anyone want to share? 18:33:40 <SumitNaiksatam> #topic Bugs 18:33:51 <SumitNaiksatam> #link https://bugs.launchpad.net/neutron/+bug/1418196 18:33:52 <openstack> Launchpad bug 1418196 in neutron "fwaas: driver base class is stale" [Undecided,In progress] - Assigned to yalei wang (yalei-wang) 18:34:04 <SumitNaiksatam> is badveli here? 18:34:32 <SumitNaiksatam> since #link https://review.openstack.org/#/c/153930/ affects their patch 18:34:56 <SumitNaiksatam> sorry, not patch, affects the varmour driver, is what i meant 18:35:59 <SumitNaiksatam> anyone else got a chance to review the patch? 18:36:17 <SridarK> SumitNaiksatam: i did look at this - more on the method args 18:36:27 <SumitNaiksatam> SridarK: yeah 18:36:46 <vishwanathj> will try to look at it later today or early tomorrow 18:36:50 <SridarK> SumitNaiksatam: i was not so sure if this was mandatory 18:37:32 <SridarK> SumitNaiksatam: but as such the changes are fine - should not cause impact - not every one will have agent_mode as something significant 18:37:43 <SridarK> it is more to do with the ref implementations 18:38:17 <SridarK> logically i viewed this as the base class methods as stating some basic requirements 18:38:33 <SridarK> but this is more a python thing with this patch 18:39:00 <SumitNaiksatam> SridarK: okay 18:39:19 <SridarK> SumitNaiksatam: i think this is fine 18:39:24 <SumitNaiksatam> i am not able to find the commit id that this patch references 18:40:13 <SridarK> Also it did not get linked to the bug ? 18:41:00 <SumitNaiksatam> “The change set of f69ed3b3, Change-Id: Iba78e534ccf347ea6270aabc939a489dd40a7b9e” 18:41:05 <SumitNaiksatam> i could not find this 18:41:09 <vishwanathj> I see the bug id being referred in the commit message 18:41:25 <SridarK> SumitNaiksatam: yes but usually we also see a review url 18:42:32 <SumitNaiksatam> my point is that the bug is referring to a commit ID (and hash) that supposedly made a change to the fwaas driver base 18:42:41 <SumitNaiksatam> but i cant locate that bug id 18:42:45 <SumitNaiksatam> sorry commit id 18:43:04 <SumitNaiksatam> the linking from the review to the LP bug is a separate thing 18:43:12 <vishwanathj> ok 18:43:20 <SumitNaiksatam> sometimes it happens that way that the link does not get established 18:43:57 <badveli> Hello All, sorry a bit late 18:44:09 <SumitNaiksatam> i dont undertand what consistency is the author referring to 18:44:26 <SumitNaiksatam> i am guessing this is the root wrap changes 18:46:00 <SumitNaiksatam> okay anyway, i will try to track this offline 18:46:17 <SumitNaiksatam> badveli: can you comment on the review: #link https://review.openstack.org/#/c/153930/ 18:46:24 <SumitNaiksatam> as this affects your driver 18:47:13 <badveli> Yes sumit, will review it 18:47:16 <SumitNaiksatam> or this might be the DVR agent mode 18:47:35 <SumitNaiksatam> i am not sure why the driver base class has to be changed 18:47:39 <SumitNaiksatam> anyway 18:48:00 <SumitNaiksatam> vishwanathj: you are looking at #link https://bugs.launchpad.net/openstack-manuals/+bug/1419498 right? 18:48:01 <openstack> Launchpad bug 1419498 in openstack-manuals "Networking services in OpenStack Security Guide - FWaaS Section Updates" [Low,In progress] - Assigned to vishwanath jayaraman (vishwanathj) 18:48:26 <vishwanathj> yes, I uploaded a patch set and have got one +1 and one +2 18:49:03 <vishwanathj> the review is at link https://review.openstack.org/#/c/158943/ 18:49:31 <SumitNaiksatam> vishwanathj: oh wow, that was fast ;-) 18:49:56 <vishwanathj> SumitNaiksatam, thanks to your tips and guidance yesterday, that made it go faster 18:50:14 <SumitNaiksatam> vishwanathj: happy to help, and thanks for acting on this quickly 18:50:24 <vishwanathj> it was a low hanging bug with less changes 18:50:34 <SumitNaiksatam> vishwanathj: sure 18:51:29 <SumitNaiksatam> i dont see an update on #link https://review.openstack.org/#/c/147396/ 18:52:11 <yushiro> SumitNaiksatam, sorry. this patch needs following patch. 18:52:21 <SumitNaiksatam> yushiro: ah ok 18:52:30 <SumitNaiksatam> yushiro: thanks for stopping by in the meeting 18:52:45 <SumitNaiksatam> yushiro: one use case was that we actually wanted to share the firewall policies 18:52:59 <SumitNaiksatam> yushiro: across tenants 18:53:12 <SumitNaiksatam> yushiro: do you think this can be better handled by changes to policy.json? 18:53:36 <SumitNaiksatam> yushiro: by this i meant the patch you have posted 18:54:22 <SumitNaiksatam> yushiro: we have the shared attribute in the firewall policy 18:54:45 <yushiro> SumitNaiksatam, yes. firewall-policy has the attribute 'shared' 18:54:55 <SumitNaiksatam> which is set to True by default 18:55:12 <SumitNaiksatam> which means that the firewall_policy created will be shared by default across tenants 18:55:28 <SumitNaiksatam> hence you are seeing the behavior 18:56:35 <yushiro> SumitNaiksatam, sorry to late response. I'm not good at writing english 18:56:44 <SumitNaiksatam> yushiro: no worries at all 18:57:16 <SumitNaiksatam> yushiro: we can take it offline if you prefer that (over ML and #openstack-fwaas) 18:57:36 <yushiro> SumitNaiksatam, Yes, I'd like to. thank you. 18:57:42 <SumitNaiksatam> yushiro: sure 18:57:51 <SumitNaiksatam> yushiro: and thanks for your work 18:58:01 <vishwanathj> +1 18:58:04 <SumitNaiksatam> SridarK: badveli: anything else on the bugs to discuss today? 18:58:06 <yushiro> SumitNaiksatam, Thank you too :-) 18:58:15 <SridarK> SumitNaiksatam: nothing from me 18:58:19 <badveli> nothing much from my side 18:58:26 <SumitNaiksatam> ok moving on 18:58:32 <SumitNaiksatam> i know vishwanathj has to leave early 18:58:40 <SumitNaiksatam> lets try to get the firewall insertion in before that 18:58:52 <SumitNaiksatam> #topic Firewall Insertion 18:59:00 <SumitNaiksatam> #link https://review.openstack.org/152697 18:59:10 <SumitNaiksatam> SridarK: over to you 18:59:19 <SridarK> #link https://review.openstack.org/#/c/152697/ 18:59:35 <SridarK> next rev of patch posted on Sun 18:59:50 <SridarK> folks pls take a look (thanks pc_m ) 19:00:06 <SridarK> basic structure in place 19:00:26 <SridarK> working on validation code to test if routers specified can be associated with fw 19:00:45 <SridarK> ie, same tenant and router is not already allocated to another fw 19:00:53 <SridarK> testing that 19:01:00 <SridarK> other thing that remains is UT 19:01:23 <SumitNaiksatam> SridarK: okay great, thanks 19:01:26 <SridarK> i have fixed most of the existing UT, working thru agent 19:01:41 <SridarK> the other thing is to gauge vendor code impacts 19:01:47 <SridarK> freescale is fine 19:02:08 <SridarK> will need discussion with badveli & vishwanathj for their vendor code 19:02:19 <SumitNaiksatam> vishwanathj: what about your plugin/driver? 19:02:53 <vishwanathj> SumitNaiksatam, SridarK and I will have an offline conversation later today 19:03:06 <SumitNaiksatam> vishwanathj: okay sure 19:03:17 <SumitNaiksatam> badveli: you are fine with the changes so far? 19:03:48 <badveli> sridar and myself will need a talk 19:03:59 <badveli> we planned to discuss today 19:04:10 <SumitNaiksatam> badveli: have you reviewed the patch? 19:04:17 <SumitNaiksatam> so far 19:04:39 <badveli> just went through the code changes, 19:05:12 <SumitNaiksatam> badveli: okay, might be good to have a good understanding of the proposed changes, so that you can have a more meaningful interaction 19:05:27 <SumitNaiksatam> all please note that we are getting very close to cut-off on the features 19:05:40 <SumitNaiksatam> so we need to wrap up on the pending items at the earliest 19:05:40 <badveli> yes, Sumit, 19:06:18 <SumitNaiksatam> so i would very much request cooperation and deligent/speedy feedback from the entire team on this 19:06:33 <SridarK> Folks thanks in advance 19:06:41 <SumitNaiksatam> there was a suggestion that i made to SridarK in the morning 19:06:57 <vishwanathj> SumitNaiksatam, SridarK, sure, we will review and provide feedback and impact if any 19:07:10 <SumitNaiksatam> regarding the default behavior to keep it consistent with the prevailing semantics 19:07:32 <SumitNaiksatam> the suggestion was to check if no routers are specified during create 19:07:55 <SumitNaiksatam> and if they arent, then apply the firewall on all routers 19:08:03 <SumitNaiksatam> which is what is done today 19:08:38 <vishwanathj> ok 19:08:46 <SumitNaiksatam> if you need more granular application, you can provide the list of routers 19:09:10 <SumitNaiksatam> if you provide an empty list, then the firewall is not applied to any routers 19:10:17 <SridarK> SumitNaiksatam: that will the case when the FW will remain in PENDING_CREATE 19:10:36 <SumitNaiksatam> SridarK: correct 19:10:38 <SridarK> SumitNaiksatam: and in the case where the attribute is not specified and there are no routers in the tenant 19:10:51 <SridarK> SumitNaiksatam: yes agreed 19:11:23 <SumitNaiksatam> anyone else has thoughts on this? 19:11:25 <SridarK> SumitNaiksatam: also one thing that i also just realized is that the help string associated with the --router-ids can also convey this 19:11:34 <SumitNaiksatam> SridarK: perfect! 19:11:39 <SridarK> SumitNaiksatam: so users are more aware 19:11:59 <vishwanathj> seems reasonable to me 19:12:21 <SumitNaiksatam> my main motivation in suggesting this is that it makes this entire patch a non-disruptive change 19:12:23 <SridarK> So i will query if there is no attribute, i will query for routers on the tenant and apply that 19:12:43 <SumitNaiksatam> so that we dont have transient gate breaking issues 19:13:02 <SumitNaiksatam> also existing clients dont have to be changed immediately, and will continue to work 19:13:11 <SumitNaiksatam> SridarK: yeah 19:13:32 <SridarK> Sounds good - i will add that 19:13:53 <SumitNaiksatam> SridarK: AFAIK, that should work, but i dont know if things have changed since i last used the “ATTRIBUTE_NOT_SPECIFIED" 19:14:18 <SridarK> SumitNaiksatam: i will test this out first too 19:14:19 <SumitNaiksatam> vishwanathj: badveli pc_m: any thoughts on this default behavior semantics? 19:14:54 <vishwanathj> like I said earlier, the defaulth behavior seems reasonable to me 19:15:18 <vishwanathj> SumitNaiksatam, I got to leave, I will review the meeting minutes later and catch up, bye all 19:15:23 <pc_m> SumitNaiksatam: sorry, don't know enough about it... :( 19:15:27 <SridarK> vishwanathj: bye 19:16:06 <badveli> currently PENDING_CREATE 19:16:22 <badveli> state is changed when the firewall becomes active 19:17:22 <SumitNaiksatam> badveli: ACTIVE is also a state 19:17:24 <badveli> we are saying that if there is empty list the fw will remain in PENDING STATE 19:17:33 <badveli> and will never change, correct? 19:17:46 <SridarK> badveli: an UPDATE will change it 19:17:51 <SumitNaiksatam> badveli: empyt list implies that that fw should not be applied on any router 19:17:59 <SumitNaiksatam> badveli: yes 19:18:41 <SumitNaiksatam> badveli: however, if you dont specifiy the router_id list (which is different from specifying an empty list), it will be applied to all routers, as is the case today 19:19:52 <badveli> fine with me 19:19:58 <SumitNaiksatam> badveli: ok 19:20:03 <SumitNaiksatam> pc_m: no worries :-) 19:20:43 <SumitNaiksatam> btw, i posted the CLI patch for this insertion change: #link https://review.openstack.org/#/c/158118/ 19:20:50 <SumitNaiksatam> SridarK: thanks for the update 19:21:09 <SridarK> SumitNaiksatam: no worries, thanks for the input on this 19:21:12 <SumitNaiksatam> #topic FWaaS L3 agent refactoring/restructuring 19:21:21 <SumitNaiksatam> SridarK: thanks for the update on this earlier 19:21:37 <SumitNaiksatam> SridarK: pc_m: so i believe no immediate changes required at this point? 19:21:50 <pc_m> SumitNaiksatam: correct 19:22:10 <SumitNaiksatam> pc_m: okay, thanks for the discussion with SridarK on this 19:22:12 <pc_m> seemed like not many changes may be needed anyway, so can defer 19:23:14 <SumitNaiksatam> pc_m: okay great 19:23:22 <SumitNaiksatam> #topic Service Objects 19:23:26 <SumitNaiksatam> badveli: over to you 19:23:44 <SumitNaiksatam> any updates you want to provide? 19:23:47 <badveli> still working on the neutron patch 19:23:54 <SumitNaiksatam> badveli: okay 19:24:02 <badveli> able to retrieve my code changes 19:24:15 <badveli> some how my previous review is not there any more 19:25:34 <SumitNaiksatam> badveli: its probably abandoned 19:25:46 <badveli> yes might be 19:25:58 <badveli> need to get from some where locally 19:26:22 <badveli> which was very old one and need to modify 19:26:52 <SumitNaiksatam> badveli: okay 19:27:09 <SumitNaiksatam> thanks for the update 19:27:15 <SumitNaiksatam> #topic Open Discussion 19:27:22 <SumitNaiksatam> anything else we need to discuss? 19:27:27 <banix> have a stupid question that will probably waste couple of minutes of your time but if that’s ok, I will ask :) 19:28:04 <SumitNaiksatam> banix: we have exactly a couple of mins 19:28:08 <SumitNaiksatam> banix: so shoot :-) 19:28:10 <banix> any plans to include a group of ip addresses in service object or in firewall rules at some point? 19:28:33 <banix> i see references to grouping of addresses as a possible extension in the original design doc 19:28:34 <SumitNaiksatam> banix: we have address objects on the road map 19:28:40 <SumitNaiksatam> banix: yes 19:28:53 <SumitNaiksatam> banix: we prioritized service objects over that 19:29:00 <banix> SumitNaiksatam: i see. great. thanks for the info. 19:29:04 <pc_m> I mentioned to SridarK this morning, but I was looking at the API reference doc, and VPN doc is gone. Don't know when that happened. I'm not sure if FWaaS have API reference documentation too, but there is nothing there on the web page. http://docs.openstack.org/developer/devstack/ 19:29:16 <SumitNaiksatam> banix: but service objects has been in review for over a year now! 19:29:24 <pc_m> Just a heads up... 19:29:44 <SumitNaiksatam> pc_m: ah good to know 19:29:45 <pc_m> Wrong link... http://developer.openstack.org/api-ref-networking-v2.html 19:29:46 <banix> SumitNaiksatam yeah noticed that 19:30:08 <annegent_> pc_m: is there a WADL for VPNaaS in the openstack/api-site repo? 19:30:13 <pc_m> SumitNaiksatam: It has LBaaS 19:30:28 <SumitNaiksatam> annegent_: thanks for chiming in 19:30:32 <SumitNaiksatam> annegent_: we will check 19:30:37 <pc_m> annegent_: Not sure. What is a WADL? 19:30:52 <annegent_> The bug has been here since last fall: https://bugs.launchpad.net/openstack-api-site/+bug/1358179 19:30:53 <openstack> Launchpad bug 1358179 in openstack-api-site "Neutron Virtual Private Network as a Service (VPNaaS) API requires "v2.0" in URI" [Undecided,Triaged] 19:30:54 <SumitNaiksatam> annegent_: which exact location should we be looking at? 19:30:59 <pc_m> annegent_: In 2013 I added documentation as part of a commit. 19:31:01 <annegent_> Web Application Description Language 19:31:30 <annegent_> I'll triage that bug further 19:31:48 <annegent_> pc_m: it needs to go into https://github.com/openstack/api-site 19:31:50 <pc_m> annegent_: I committed, and it was on the web site, back in 2013 19:32:04 <annegent_> pc_m: the API reference? If so I can find it in history. 19:32:07 <SumitNaiksatam> to the best of my knowledge we had fwaas here as well, since i had added that patch (sometime in 2013 as well) 19:32:14 <pc_m> Used to be on the API reference pages. 19:32:16 <SumitNaiksatam> annegent_: okay thanks 19:32:25 <annegent_> SumitNaiksatam: will also look in history there 19:32:36 <SumitNaiksatam> annegent_: thanks, will try to follow up as well 19:32:39 <annegent_> thanks 19:32:56 <SumitNaiksatam> alrighty, thanks everyone, we are a couple of mins over time 19:33:03 <SumitNaiksatam> bye 19:33:08 <SridarK> bye all 19:33:12 <SumitNaiksatam> #endmeeting