14:05:08 <Jonas1> #startmeeting weekly fuel sync meeting 14:05:08 <collabot> Meeting started Thu Sep 24 14:05:08 2015 UTC. The chair is Jonas1. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:05:08 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic. 14:05:08 <collabot> The meeting name has been set to 'weekly_fuel_sync_meeting' 14:05:19 <Jonas1> #info Jonas Bjurel 14:05:35 <mskalski> #info Michal Skalski 14:05:45 <ruijing> #info Ruijing Guo 14:05:59 <s_berg> #info Stefan Berg 14:06:09 <Jonas1> Ok lets get roling 14:06:20 <Jonas1> #topic SR1 status 14:06:25 <lmcdasm> #info Daniel Smith 14:06:44 <Jonas1> So lets start with Build 14:07:05 <Jonas1> Everything I intended to do is in, s_berg? 14:07:19 <Jonas1> Hi Szilard 14:07:45 <s_berg> Jonas1: Depends if you had hidden intentions :) but yes, there's a repo cache in the iso now. 14:07:53 <jose_lausuch> #info Jose Lausuch 14:07:55 <szilard_> Hi Jonas 14:07:57 <szilard_> Hi Jose 14:08:02 <jose_lausuch> SR1 status is hot 14:08:03 <szilard_> Hi Stefan 14:08:04 <jose_lausuch> :) 14:08:14 <s_berg> So what's important now is to change the dea.yaml for those configs that we are using to actually use the local cache. 14:08:37 <s_berg> See https://gerrit.opnfv.org/gerrit/#/c/1951/ for details (in the commit message). 14:08:44 <s_berg> Hi Szilard! :) 14:09:02 * s_berg waves to all petrol heads. 14:09:13 <Jonas1> So Szilard needs to do a small tweak in DEA to start using the cache? 14:09:21 <s_berg> Correct! 14:09:22 <szilard_> ok 14:09:33 <Jonas1> Szilard: Did you catch that. 14:09:38 <s_berg> Super easy, just changing the repo URIs in the config part. 14:09:39 <Jonas1> Cool. 14:09:52 <szilard_> yepp I will see in the patch 14:09:54 <szilard_> how to do it 14:09:58 <szilard_> Stefan 14:10:08 <Jonas1> So the last patch (planed from stefan) regards virtual deployment, correct? 14:10:14 <szilard_> do you have a suggestion preference ? 14:10:24 <szilard_> how to name that caching parameter 14:10:25 <szilard_> ? 14:10:35 <s_berg> szilard_: It's already in the standard Fuel config. 14:10:55 <szilard_> ah ok 14:11:20 <szilard_> I'm also playing with MTU on interfaces 14:11:22 <s_berg> custom_repo_configuration 14:11:25 <szilard_> my colleague from SFC 14:11:31 <szilard_> Ferenc Cserepkei 14:11:39 <s_berg> Jonas1: Yes, testing that patch now, will push if it succeeds. 14:11:46 <szilard_> found out that you cannot SSH into the VMs 14:11:55 <szilard_> because of VXLAN overhead 14:11:58 <s_berg> Jonas1: Also putch a correcting patch on master, removing a few more not needed patches. 14:12:05 <s_berg> But Jenkins seems broken again. 14:12:06 <szilard_> have to increas the MTU 14:12:42 <s_berg> MTU for the base interfaces? 14:12:42 <Jonas1> Ahh, yes we should increase MTU, but it is not easy to do I believ. 14:12:52 <s_berg> That's in the standard Fuel config under networking nowadays. 14:12:54 <szilard_> I did some tweak 14:12:57 <szilard_> testing now 14:13:04 <szilard_> deploying in Montreal 14:13:08 <szilard_> to check if it works 14:13:13 <szilard_> Fuel allows to change the MTY 14:13:14 <szilard_> MTU 14:13:17 <szilard_> on interfaces 14:13:29 <s_berg> szilard_: Exactly! 14:13:46 <Jonas1> OK - great. 14:14:36 <Jonas1> #info We're good on build for the SR1 release, one more patch planned (For virtual deployment). 14:14:43 <Jonas1> Next deply. 14:14:48 <Jonas1> deploy. 14:14:59 <Jonas1> Szilard, what is the status? 14:15:06 <s_berg> (szilard_: Just FYI, look in DEA and search for "repo_setup", all the URIs are in there.) 14:15:08 <szilard_> deploy fixed now 14:15:14 <szilard_> Jose brought to my attention 14:15:20 <szilard_> that forwarding was disabled 14:15:25 <jose_lausuch> yes 14:15:36 <szilard_> thank you Jose, great job :) 14:15:39 <jose_lausuch> it was deactivated 14:15:41 <s_berg> In the jumphost? 14:15:41 <Jonas1> Forwarding - where? 14:15:45 <jose_lausuch> and now we have again a network issue 14:15:51 <jose_lausuch> on the jumphost 14:15:56 <jose_lausuch> ip forwarding was disabled 14:15:57 <szilard_> again ? 14:16:02 <szilard_> I made a patch 14:16:04 <jose_lausuch> so fuel couldnt connect to internet 14:16:06 <szilard_> for setup.sh 14:16:16 <jose_lausuch> on the jumphost, enp7 is setup instead of enp7.300 14:16:19 <szilard_> to temporarily check and enable forwarding 14:16:22 <jose_lausuch> and for fuel we need it on enp7.300 14:16:47 <szilard_> enp7.300 14:16:48 <szilard_> ? 14:16:53 <szilard_> for private 14:16:57 <jose_lausuch> for management 14:17:00 <szilard_> ok 14:17:09 <jose_lausuch> pbandzi: can we look into this now? 14:17:17 <szilard_> fuelweb_admin ? 14:17:21 <szilard_> is not enough 14:17:25 <szilard_> PXE interface ? 14:17:31 <jose_lausuch> I can set it up myself manually, but I dont know what will happen next lab setup script 14:17:33 <Jonas1> jose_lausush: Is this because of the change trozet did yesturday - or is it something new? 14:17:39 <pbandzi> hi szilard, jose_lausuch 14:17:43 <jose_lausuch> Jonas1: not sure 14:17:45 <szilard_> hi Peter 14:17:59 <jose_lausuch> its not something new for sure 14:18:00 <szilard_> why do you need enp7.300 ? 14:18:21 <jose_lausuch> when you source the public OS_AUTH_URL 14:18:28 <szilard_> you don't need to connect Fuel VM to that 14:18:29 <jose_lausuch> if you run for example glance image-list 14:18:32 <jose_lausuch> it says: 14:18:44 <szilard_> ok so you want to access it from Jumphost 14:18:45 <szilard_> I got it 14:18:49 <jose_lausuch> Error finding address for http://192.168.0.7:9292/v1/images/detail?sort_key=name&sort_dir=asc&limit=20: ('Connection aborted.', error(113, 'No route to host')) 14:18:52 <pbandzi> jose_lausuch, szilard : what we are solvin now? in shor 14:18:54 <jose_lausuch> which is the management ip 14:18:57 <szilard_> aha 14:19:07 <szilard_> that is when you run rally on jumphost 14:19:16 <szilard_> you need to have access to management 14:19:17 <szilard_> ok 14:19:19 <jose_lausuch> due to a this bug I reported https://ask.openstack.org/en/question/68144/keystone-unable-to-use-the-public-endpoint/ 14:19:23 <jose_lausuch> yes 14:19:27 <s_berg> jose_lausuch: I don't really get the use case there. 14:19:40 <szilard_> ok 14:19:41 <s_berg> jose_lausuch: So you're logging into one of the nodes from the jumphost? 14:19:45 <trozet> szilard_: so that enp7s0.300 is not needed? 14:19:53 <jose_lausuch> in short: we need access to management network from jumphost to be able to run the tests 14:19:54 <szilard_> not for Fuel 14:20:04 <jose_lausuch> enp7.300 is needed in Fuel 14:20:05 <szilard_> for Jose's Functest 14:20:07 <jose_lausuch> but not in foreman 14:20:12 <jose_lausuch> well 14:20:14 <jose_lausuch> not in Fuel 14:20:16 <jose_lausuch> in the jumphost 14:20:17 <szilard_> aha 14:20:20 <szilard_> that 14:20:23 <jose_lausuch> to access the Openstack management network 14:20:24 <szilard_> is correct 14:20:26 <szilard_> yes 14:20:27 <szilard_> yes 14:20:33 <s_berg> jose_lausuch: Ah, ok, got it. 14:20:35 <szilard_> I know :) 14:20:37 <szilard_> let's see 14:20:44 <jose_lausuch> thats a limitation we couldnt solve yet 14:20:46 <szilard_> why we cannot access it 14:21:02 <jose_lausuch> the reason we cannot access now is because enp7 is setup instead of enp7.300 14:21:11 <jose_lausuch> and Openstack management on fuel is on VLAN300 14:21:15 <jose_lausuch> on foreman isnt on any vlan 14:21:21 <szilard_> I add it to my patch 14:21:27 <szilard_> I will create that VLAN interface 14:21:35 <szilard_> is that a deal 14:21:40 <szilard_> and then you will have 14:21:45 <szilard_> enp7.300 14:21:47 <jose_lausuch> I will modifiy it manually now 14:21:50 <szilard_> ok 14:21:57 <szilard_> what IP should I give to it 14:21:57 <jose_lausuch> well, the network-script is there 14:21:58 <szilard_> ? 14:22:09 <jose_lausuch> but just the ip, switch it to enp7.300 instead of enp7 14:22:13 <jose_lausuch> and enable enp7.300 14:22:16 <jose_lausuch> the same 14:22:21 <jose_lausuch> enp7 shoudlnt have an ip 14:22:24 <szilard_> aaah 14:22:25 <szilard_> ok 14:22:34 <szilard_> so use that IP 14:22:38 <szilard_> ok Jose 14:22:41 <s_berg> But guys, does it sit an IP on enp7.300? 14:22:46 <jose_lausuch> 192.168.0.66 14:23:04 <s_berg> So what would forwarding solve, that's in the same broadcast domain? 14:23:04 <Jonas1> So do we have a deal - Jose and Szilard fixes this outside this meeting? 14:23:14 <szilard_> ok 14:23:15 <jose_lausuch> yes 14:23:23 <jose_lausuch> forwarding is another different point 14:23:25 <szilard_> deal 14:23:26 <jose_lausuch> not related to this 300 14:23:39 <s_berg> OK. You seem to be on top on this. :) 14:23:47 <Jonas1> Otherwise deploy works OK? 14:23:50 <jose_lausuch> we are mixing things maybe, my fault :p 14:23:55 <szilard_> forwarding is now solved in the patch 14:24:03 <jose_lausuch> last deploy was ok 14:24:28 <jose_lausuch> but Im also expecting for the next switch to foreman and viceversa 14:24:32 <jose_lausuch> that lab reconfig script 14:24:33 <jose_lausuch> lets see 14:24:43 <jose_lausuch> foreman now will happen at 19:00 CEST 14:24:52 <Jonas1> #info deploy is working with one caveat, enp7.300 must be set-up as well as forwarding. Szilard and Jose are on top of this. 14:24:57 <jose_lausuch> so that gives us some time to troubleshoot fucntest problems 14:25:17 <jose_lausuch> sorry for not "infoing" :) 14:25:22 <Jonas1> So lets move to functest - where do we stand Jose? 14:25:52 <jose_lausuch> Morgan and I sent an email around with some concerns 14:26:21 <jose_lausuch> #info functest a bit blocked due to failed deployments, short time between deployments 14:26:31 <jose_lausuch> #info troubleshooting and fixing started already though 14:27:11 <jose_lausuch> #info functest would like to have a couple of days with Functest job succesfull 14:27:23 <jose_lausuch> and that leaves us a bit tight in time 14:27:32 <Jonas1> Ok, anything we can do to help-out apart from providing patches needed? 14:27:48 <jose_lausuch> you are already helping :) 14:28:03 <Jonas1> Ok, thx. 14:28:15 <Jonas1> Son next: Documentation 14:28:54 <Jonas1> I have uploaded patch sets for readme's, build instructions, release notes, etc. 14:29:42 <Jonas1> Remaining is: Deploy README, Szilard is there a need to update that? 14:30:14 <szilard_> no, I have updated 14:30:16 <szilard_> README 14:30:17 <lmcdasm> <sent you a question about tab naming in your last commit- are you sure its not .1.1 isteadn of arno.2.0?) 14:30:22 <szilard_> in deploy/ and ci/ 14:31:18 <Jonas1> lmcdasm: Just copied from trozet. We will see I need to update it with the test results when they are available anyway. 14:32:32 <Jonas1> #info All README's, the build instructions release notes have been updated. 14:32:56 <lmcdasm> cool thx Jonas1 14:33:20 <Jonas1> So remaining is to finish-up the install instruction that lmcdasm has already started. 14:33:47 <Jonas1> lmcdasm: I think we have an .iso to work from. 14:34:37 <Jonas1> lmcdasm: Can you make a manual install and update the instructions today/tomorrow? 14:34:53 <lmcdasm> sure.. point me to the iso you want used and ill do it! 14:35:10 <lmcdasm> or i can do a build today as well 14:35:22 <Jonas1> Can you build the iso from master? 14:35:23 <lmcdasm> and then try that if you guys think the .iso from build will work 14:35:26 <lmcdasm> i can use 14:35:27 <lmcdasm> yes* 14:35:37 <Jonas1> Cool. 14:35:38 <lmcdasm> will fire up a clean build VM 14:35:44 <lmcdasm> and do a new end to end. 14:36:32 <Jonas1> lmcdasm: Be aware of that there is a race issue that sometimes shows up when deploying nested (Stefan is working on a patch) 14:36:55 <Jonas1> So sometimes the deploy times out when building the ubuntu image 14:37:40 <Jonas1> When that happens, leave it for 30 minutes, go into fuel GUI and just pres deploy - and it will come up. 14:38:08 <lmcdasm> ok 14:38:11 <lmcdasm> im deploying baremetal 14:38:42 <Jonas1> #info Daniel will do an end2end install/deploy and update the installation instructions accordingly 14:39:16 <Jonas1> lmcdasm: Then I dont think you will have any issue (Lucky you) :-) 14:39:51 <lmcdasm> ok.. so anything special to note 14:39:58 <lmcdasm> for a BM buid - or just make all? 14:40:08 <lmcdasm> (im cheating and not reading the updated build II :P) 14:40:11 <lmcdasm> about to kick it off now 14:40:20 <Jonas1> Not what Im aware of, any one else - something Danie should be aware of? 14:40:45 <lmcdasm> well.. lets see! :) 14:40:51 <Jonas1> lmcdasm: One thing, you need all three ntp serves provisioned or Fuel gets sick. 14:40:57 <lmcdasm> ok 14:41:13 <lmcdasm> ill find that stuff as i go through 14:41:30 <lmcdasm> the real trick will be the steps for using the mirror on the FUEL serger 14:41:31 <lmcdasm> server* 14:41:43 <lmcdasm> as i understand it those steps are the same configurations as in the dea/dha? 14:41:58 <s_berg> (Also to note that underscore is not valid for hostname (both according to an RFC and fuelmenu), use hyphen instead.) 14:42:01 <lmcdasm> right - or are we just going to have them static in the II and give a screen shot of the manual config in the GUI? 14:42:01 <Jonas1> Yes, should be. 14:42:10 <lmcdasm> 10.20.0.2.. etc.etc. 14:42:33 <s_berg> lmcdasm: Yes. If you deploy manually, just a matter of changing URIs for the repos in the config tab. 14:42:34 <Jonas1> lmcdasm: Yes 14:42:56 <lmcdasm> great.. 14:43:17 <lmcdasm> so ill just confirm them when i set them .. .and in the II we will say " point yoru config to xyz" view screen shot below as reference 14:43:36 <Jonas1> lmcdasm: Exactly 14:43:37 <lmcdasm> and those values are in the README of your patchset that we just looked at right s_berg? 14:43:54 <lmcdasm> the repos will be up and available once Fuel is deployed 14:43:56 <s_berg> lmcdasm: They're in the commit message yes. 14:44:10 <lmcdasm> or is there anything to do to activate the container now (i dont think so, but wanna be sure) 14:44:24 <s_berg> lmcdasm: No black magic, they'll be in there. 14:44:31 <lmcdasm> ok santana.. 14:44:33 <lmcdasm> we will see 14:44:48 <Jonas1> So seems we have a plan! 14:45:00 <lmcdasm> build is ongoing.. so ill keep you up to date of the install on BM throughout the day and into tmrw 14:45:03 <lmcdasm> with II updates 14:45:14 <Jonas1> Anything else for SR1 right now? 14:45:19 <lmcdasm> and when its all done ill push the II as a patch review 14:45:38 <Jonas1> lmcdasm: Great 14:46:26 <Jonas1> #info Next thursday we will start discussing/Working on Fuel@OPNFV B-release 14:46:47 <Jonas1> Would it be worth while to have a goto-meeting then? 14:47:03 <s_berg> Good idea. 14:47:26 <s_berg> Also an Etherpad I think. 14:47:41 <Jonas1> Lets go for that, I'll send an updated call with goto meeting details + etherpad. 14:47:46 <Jonas1> 3 14:48:08 <Jonas1> 2 14:48:25 <Jonas1> 1 14:48:32 <Jonas1> #endmeeting