14:00:08 <[1]JonasB> #startmeeting FuelOPNFV No 8 14:00:08 Meeting started Wed Apr 22 14:00:08 2015 UTC. The chair is [1]JonasB. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:08 Useful Commands: #action #agreed #help #info #idea #link #topic. 14:00:08 The meeting name has been set to 'fuelopnfv_no_8' 14:00:09 #info Szilard 14:00:12 #info Fatih 14:00:17 <[1]JonasB> #info Jonas 14:00:35 <[1]JonasB> Dan go again, forgot to start the meeting 14:00:45 #info Daniel Smith 14:02:19 #info Szilard 14:02:23 :) 14:03:37 we waiting for others are we? 14:04:18 * stefan_berg here now 14:04:42 <[1]JonasB> Daniel: Waiting for Michal and Jose 14:05:49 <[1]JonasB> Jose, pound in your name 14:06:25 #info Stefan Berg 14:06:36 * stefan_berg knows he's not Jose 14:07:08 :)) 14:07:27 <[1]JonasB> SInceJose is in the wrong chat room, lets wait 14:07:32 Im here! 14:07:38 sorry 14:07:56 <[1]JonasB> #info jose_lausuch 14:08:04 <[1]JonasB> Great Jose - welcome! 14:08:12 Hi 14:08:27 <[1]JonasB> So should we start with testing 14:08:49 <[1]JonasB> #topic testing 14:09:01 sorry - is there an agenda that i should be following ? 14:09:03 #info Bryan Sullivan 14:09:24 <[1]JonasB> #info Jose will help out with getting tepest testing 14:09:36 I have one question for the agenda: where is the guide to what to do with the ISOs once we have them? We need next steps. 14:09:46 @ bryan 14:09:46 lmcdasm: Error: "bryan" is not a valid command. 14:09:52 on the wiki.. i will send the link.. one sec 14:09:57 yes, I'll be working on that now 100% of time 14:10:11 take a look at this bryan 14:10:11 https://wiki.opnfv.org/_media/get_started/opnfv_montreal_lab-revc.pdf 14:10:14 the idea is to get the tempest test suite 14:10:20 connected to jenkins 14:10:25 <[1]JonasB> Daniel is it ok that Jose continue to use BL9 14:10:34 ya 14:10:36 I have been looking into BL9 already 14:10:38 <[1]JonasB> Cool 14:10:49 afaik i havent changed lab assignments since you set them last friday 14:10:54 <[1]JonasB> Fdegir - can we connect Jenkins to BL9 14:10:58 so should all be the same (excet for Andrews is free). 14:11:17 [1]JonasB: I can ping aricg 14:11:21 and check with him 14:11:35 <[1]JonasB> lmcdasm: Didnt follow that 14:11:43 please, that would be nice 14:11:47 ya.. fatih we are talking BL9 in our lab 14:11:49 not in LF 14:11:56 yep 14:12:01 <[1]JonasB> Yea - that is right 14:12:02 ok.. so aric? 14:12:05 aricg needs to help us 14:12:11 what is the ip 14:12:14 hmm. can you explain? 14:12:15 https://wiki.opnfv.org/_media/get_started/opnfv_montreal_lab-revc.pdf 14:12:16 so I can login and do prep 14:12:31 <[1]JonasB> #action fdegir will connect BL9 to Jenkins 14:12:32 192.75.89.194 14:12:32 its in there - for BL9 all the fino (FUEL, range for env, etc) 14:12:35 no 14:12:48 thats not the IP Jose - that the IP of the GW - please refer to the Mtl Lab layout 14:12:50 can I get sudo on that? 14:12:55 sorry, yes 14:12:59 fati 14:13:01 FUEL IP : 10.118.34.211 14:13:04 I need to create a functional ID there 14:13:07 named jenkins 14:13:08 its int he document :) 14:13:23 all the acces info, range, password, etc are all there for each BLade/LAB 14:13:26 ok 14:13:35 but getting back to Jonas question 14:13:45 how does Aric fit here (just so i understand) 14:13:48 lmcdasm, I don't see any info in that PDF about what to do when Jenkins has handed you an ISO. But I can wait till this is the topic of the meeting. 14:14:15 bryan -ok ... how about you setup a quick 15 min with me andill gothrough your questions 14:14:22 <[1]JonasB> Aric needs to configure a jenkins slave into the master I guess 14:14:37 ok - that makes more sense.. i thought it was "our" jenkins 14:14:39 [1]JonasB: correct 14:14:46 thx for clarifying - 14:14:47 [1]JonasB: We already have a jenkins slave. It's building the ISO. 14:15:07 so you are going to want external inbound access from the internet for that blade then? 14:15:27 or are you going to forward it through the SSH GW (not ideal) 14:15:35 <[1]JonasB> davidmichaelkarr: Yes but this one will be focused on Jenkins 14:16:08 <[1]JonasB> lmcdasm: A good point, what do we want to do fdegir 14:16:15 lmcdasm, I guess that means it's not written down yet - so yes, a session in which I can write it down based upon what you can tell me would be helpful. 9:30 PDT on this channel. 14:16:19 [1]JonasB: "focused on jenkins"? What does that mean? 14:16:43 here is what I understood 14:16:52 <[1]JonasB> davidmichaelkarr: We will run Jenkins tests there 14:17:00 we need to connect BL9 to OPNFV Jenkins so Jose can do tempest work via OPNFV Jenkins 14:17:04 = bryan - we have lots of stuff - so think its a questino of what you need - if i understand you are looking for the FUEL installation guide on how to install Fuel node - yes - lets do this after.. 14:17:35 <[1]JonasB> jose: Do we have a list of Jenkins testcases to start from? 14:17:35 fatih - is the setup a PuLL (From the slave to master) or a push (from master to slave)? 14:17:43 is this correct [1]JonasB, lmcdasm, jose_lausuch 14:17:57 for now, no, we just run all, but that I Can check later 14:18:05 lmcdasm: we need to reach the internet from BL9 14:18:08 lmcdasm, no I mean when Jenkins has build the ISO of the OPNFV distribution, what's the next step to get it installed. But yes later is fine. 14:18:28 <[1]JonasB> fdegir: is what correct 14:18:29 ok.. bryan - yes.. yolu need an FUEL install GUIde- to take from ISO and setup FUEL 14:18:34 yes, let me doublechek it later 14:18:36 we will outline later 14:18:44 fatih the blaes have internet access (you can go out) 14:18:54 however the firewall on that lab doesnt not allow inbound push 14:19:00 What is "later"? I have to run an hour meeting in >1.5 hours. 14:19:05 so my question.. the setup to talk to OPNFV Jenkins.. is a pull or a push? 14:19:27 lmcdasm, we will pull 14:19:36 hmm.. too many threads at once.. my later was to bryan to talk about installs - but we are talking about testing status now in the meeting 14:19:41 <[1]JonasB> llmcdasm: We should be fine with outbound access! 14:19:47 good :) 14:20:32 <[1]JonasB> Jose, do you need help from any one? 14:20:48 I'll contact Morgan 14:20:52 he has been doing some work about that 14:22:10 <[1]JonasB> Ok - lets move on to the problem Daniel is experiencing 14:22:34 <[1]JonasB> #topic deployment issue 14:22:48 <[1]JonasB> Daniel can you summarize with info 14:24:48 #info retrying deploy on Baremetal - some info received from Michal today. I have resent the environment info showing successful verified networking and the deployment is running now. The patch that Michal outlined should be put into the release since it doesnt make sense that a deployment should fail - VIP doesnt answer - from the controller that is actually hosting it - the publicURLs 14:24:48 are populated with the VIP. 14:25:27 <[1]JonasB> Michal, are you here? 14:25:28 #info - once the deploy is done, if there are no issues, then we move foward, if it fails, then im hoping someone will point me in the right direction from questions this morning 14:25:33 #info - thats it or me 14:25:44 [1]JonasB: hi Jona, I'm 14:26:11 <[1]JonasB> mskalski: Have you looked at the logs Daniel sent you 14:27:13 [1]JonasB: I'm in the middle of doing that, but I have hope that after proper configuration of public subnet deployment will work 14:27:42 <[1]JonasB> Ok lets croos our fingers! 14:27:53 <[1]JonasB> Should we go to autodeploy 14:28:08 <[1]JonasB> #topic autodeploy 14:28:32 hmm. 14:28:33 <[1]JonasB> #info Our focus on bare metal will be on Eri Lab for now 14:28:34 ok.. 14:29:03 <[1]JonasB> Szilard, where are you in terms of autodeploy? 14:29:06 questino - doest that include a baremetal FUEL then or we still use FUEL VM? 14:29:14 Hi 14:29:25 <[1]JonasB> No I think we still go Fuel VM, dont you 14:29:40 either or - but lets give Szi time 14:29:41 I'm running and debugging my deployment script on Blade 13 14:29:53 #info I'm running and debugging my deployment script on Blade 13 14:30:27 #info I have adjusted the Python Autodeployment according to Stefan's script 14:30:46 <[1]JonasB> Ok - how far do you think from something working? 14:31:47 #info by tomorrow should be ready 14:31:58 <[1]JonasB> Great 14:32:00 #info tomorrow evening 14:32:21 #info it will work on Libvirt nested environment 14:32:34 #info because that's the one I have been testing 14:32:49 #info no ESXi, no HP baremetal yet 14:32:50 can i have second to clarify something? 14:32:54 yes 14:32:55 ask a question on that 14:33:02 sure 14:33:06 so that we understand order of operations 14:33:29 ok 14:33:37 when you say "libvirtd" for autodeployemt, you are talking about the whole thing (create a FUEL VM, boot from a Automatic deployable FUELISO0 14:33:44 create VMNs, boot them, etc. correct? 14:34:05 yes, basically I took Stefan's stuff, I mixed it with mine, and that's the result at this moment 14:34:07 as oppossed to just creating a AutoDeployable ISO (meaning no menus, config is passed, etc). 14:34:33 let me clarify more 14:34:42 ok - thats great.. i just want to caution about merging the two idea too far.. we want the Automatic ISO installation to be as deoup[led for the HW adapter as possible 14:34:47 this is how it runs at this moment 14:34:58 you run deploy.sh specify the location of the Fuel ISO 14:35:08 the deploy takes care about everything 14:35:15 deploy the Fuel VM 14:35:27 creates Blade VMs, virtual networks 14:35:32 and then start the Python Deployment 14:35:44 and wait until is finished 14:35:52 that's all 14:35:54 great 14:36:06 It's basically Stefan's idea, pythonified a little bit 14:36:10 and deploy.sh takes the DEA to know what type to do (libvirt, esxi, etc) 14:36:12 ? 14:36:21 yes 14:36:23 <[1]JonasB> great 14:36:23 but ... 14:36:46 if you want to run it on ESXi or HP 14:36:49 that's another story 14:36:51 <[1]JonasB> Daniel - you dont have something similar to a HP DHA adapter lying around :-) 14:36:59 for esxi? or 14:37:13 <[1]JonasB> No HP bare metal 14:37:17 well.. i mean, ya 14:37:30 the whole autodeployment will work for Libvirt because 14:37:35 but the HP DHA does that stuff (has the IPMI /ILO stuff to attach ISO,s etc) 14:37:49 Yes I have HP adapters, those are working 14:38:03 <[1]JonasB> Ok yes forgot 14:38:05 1 clarification 14:38:13 what i would suggest is the following: 14:38:24 don't have HP Adapter written for ISO attach 14:38:35 - szi finish on his work and delivery "whatever" is working to me 14:38:47 just only for stop, start blade, and set boot order 14:39:08 - i will create a second "smaller" baremetal setup in the lab and then I can start tinkering with the ISO attach .. its not as easy as in iDRAC 14:39:19 (Szi - thats great!). 14:39:35 That's what I need, baremetal, otherwise I'm not able to test it 14:39:41 from there, we can have the big system for system test and pooking, around and start doing interations of others 14:39:51 baremetal and netsted.. 14:40:10 and then when we are all happy wihth the big BM setup for release, we can turn the autodeployer on the big one 14:40:10 ? 14:40:39 <[1]JonasB> lmcdasm: agree 14:41:27 can I have at least 3 blades for baremetal testing ? 14:41:34 you can have two 14:41:38 1 control and 1 compute 14:41:42 and a FUEL VM that is wired to it 14:41:49 i just dont have any more blades 14:41:49 aha, okay 14:41:54 sure, good enough 14:41:56 once the big one is fine we can shrink it a tbit 14:42:14 <[1]JonasB> But first after we have performed successful manual bare metal deploy 14:42:15 and ya - it wont be long before we try it against the full blown one (end of the week). 14:42:23 yes JOnas - exact 14:43:22 <[1]JonasB> Anything more before we end? 14:43:35 that's all 14:43:44 time for AOB? 14:43:54 <[1]JonasB> #topic AoB 14:43:57 <[1]JonasB> Shoot 14:44:03 We can address the questions I had at the start. 14:44:08 yes 14:44:22 so.. to answer you better, can you tell me the following: 14:44:31 - where / what hardware you are trying to install this on 14:44:32 Where on the wiki (or wherever) is a guide to what we do once Jenkins has handed us an ISO? 14:44:53 <[1]JonasB> Hangon 14:44:55 there is a FUEL installation GUIDE that was published at the star of the project (im trying to find it - since none of our wiki pages are indexed) - but i know its there 14:45:04 We have a multi-blade environment with VSphere managing VMs that we will use to install. 14:45:19 ok.. so you are going to do a Nested enviorment or you are going to use vCenter in FUEL? 14:45:20 We have Jenkins VM slave building the ISO. 14:45:26 <[1]JonasB> #link http://artifacts.opnfv.org/ 14:46:22 [1]JonasB: So the "installation-instructions"? 14:46:23 more spefiically for bryan 14:46:23 http://artifacts.opnfv.org/genesis/fuel/docs/installation-instructions.html 14:46:25 I'm not sure what you mean by "use vCenter in FUEL" but we have what I said. A jenkins slave VM from which we will kickoff the install 14:46:33 and david.. 14:46:39 <[1]JonasB> davidmichaelkarr: correct 14:46:46 however gents - for Jonas and I - its important to understand a couple of things in terms of expectations 14:47:35 That link (instructions) looks like a manual install process. Is that correct? 14:47:48 as outlined in the meeting we are working on autodeployment nbow 14:47:50 so yes.. its manual 14:48:08 but i mean manual in that you click through some screen once the ISO si built 14:48:13 OK, that;s good enough 14:48:22 Starting from the ISO 14:48:34 can you and David outline your expectations from us in terms of support 14:48:39 We will let you know how goes. How to provide feedback and issues? 14:48:46 well. thats just to my point 14:48:48 Answer questions only. 14:48:55 <[1]JonasB> Important - as we're still in development - this document has not yet been reviewed! 14:49:21 Understood. If we run into an issue its an issue to be documented and fixes somehow 14:49:32 Will that be helpful to you? 14:49:42 [1]JonasB: And as you wrote it, what's your "confidence level" of accuracy then? 14:49:47 <[1]JonasB> Yep, why dont you check into this meeting 14:49:51 the II works fine 14:50:01 i wrote and tested it and its what was done in the video 14:50:07 however, it assumes a level of competency 14:50:19 <[1]JonasB> Pretty close, but it still misses so,me ODL stuff 14:50:28 Is the video linked in the doc, or otherwise on the wiki? 14:50:32 meaning that you know how Fuel works (have read the architecute) you have deep knowledge of openstack and pXE 14:50:57 So Daniel wrote it, but the doc says that Jonas wrote it? Are we talking about the same doc? 14:51:12 We both wrote it David 14:51:15 <[1]JonasB> Daniels video is here: 14:51:28 Ok. Doc just refers to Jonas. First issue then. :) 14:51:41 however, if you are doing a Nest ESXi environment, then i dont talk about what you need to do for vhv enable , or how to set your switches 14:51:48 or promiscius mode ont he vSwitch or anything liek that 14:51:58 <[1]JonasB> #link http://www.dasmlab.org/opnfv.html 14:52:00 OK, we will get as far a we can and when we hit a roadblock we will ask what to do next. I would assume the goal is that installers would not need to be deep experts in PXE etc - right? When do we expect to get there? 14:52:08 cause that is something that we (in OPNFV) i dnt think wanna comment on and its not really scoped in scuppotr 14:52:26 i disgaree bryan 14:52:39 we are not installing windows here, we are installing a Virtualization Infrastructe 14:52:51 so knowledge of switching concepts, and opensystems stuff is required. 14:53:10 we cannot take everything down toa single click - cause there are too mnay variations to deal with 14:53:23 Understood - but to get take-up of the OPNFV platform it can't be like pulling teeth to get it installed in every environment 14:53:29 the expectations in my docuemnts is that you are familar with Opensystems and have IP background 14:53:40 and its not 14:53:48 we have prvided screen shots and pictures adn videos 14:53:56 OK, how deep the background is needs to be somewhat level set in the guide 14:54:01 but if a person doesnt know how to install Ubuntu in a VM and get outo to the net with resolv.conf 14:54:09 then how are they going to install a whole cloud infrastrucure? 14:54:11 see my point? 14:54:12 :) 14:54:18 We can do that. It;s way above PXE 14:54:29 :) 14:54:33 <[1]JonasB> Lets see how it goes 14:54:48 by PXE i mean understanding how computer to computer hands off installation (jumpstarting) works 14:54:56 And we know openstack. ODL gets hairy but we expect that the installer will build the vswitches etc for us 14:54:59 not.. going down into the PXXE stack itself now 14:55:01 <[1]JonasB> Hint: dont forget /opt/opnfv/pre-deploy.sh 14:55:10 well.. ODL doesnt use switches at alkl 14:55:27 in fact the first page of the ODL guide in the installation shows to wipe out the existing OVS brige 14:55:33 cause then you setup vxlan tunneling.. 14:55:43 so.. again. i think that we really need to have a round table on expectations 14:55:46 from the community and users 14:56:00 cause there is a lot of flack about docuements and non-claritya nd ithink it comes form expectations 14:56:21 <[1]JonasB> So anything more to say? 14:56:27 cause when i hadn someone a lab for exampke, i expect they know nova,neutron, (ostack) have read the docuemtnation on ODL, and understand how OVS will interwork 14:56:27 I mean all the virtual networking stuff that it does do... that takes some real expertise. The installer should arrange all that as we are really cloud focused not SDN focused at this time. 14:56:33 well.. 14:56:46 bryan for what model. for what deployment for what VNF? 14:56:58 OK we will see how far we can get. Where to reach out for help? 14:57:08 you can always mail me 14:57:18 @stefan_berg can we talk in private ? 14:57:18 SzilardCserey_ER: Error: "stefan_berg" is not a valid command. 14:57:19 and the team (the is a Q&A on the get_started mpage as well i think)> 14:57:19 Any IRC channel? 14:57:30 well.. that just it 14:57:31 <[1]JonasB> bryan_att: Or check into this meeting 14:57:33 what are you expecting 14:57:47 like a support channel or just adhoc? 14:57:49 just some expert to help us over hurdles 14:57:52 adhoc 14:57:55 well. when you hi tht hurdle 14:58:06 mail us and we will arrange as needed 14:58:14 ok good enough 14:58:14 if you can - so we have some background 14:58:24 send us a diagram of what you are trying to build (high level) 14:58:38 just the OPNFV default end state platform. 14:58:47 as well - you mentioned ESXI - so i assumed you are doing nested setup? 14:58:48 We will deploy our VNFs on that 14:59:06 Yes, in a vSphere managed VM environment 14:59:10 and you know how to do nested switchin and hypervisor in ESXI (or you use vcenter) 14:59:19 We use vCenter 14:59:33 that was not my quetion .. vSphere is suimply the overlay to esxi (esxi is the hyperisor - vCEntere is like Horizon) 14:59:39 so.. on your blades 14:59:52 what is your intention? you will setup a FUEL VM and then baremetal the other (with esxi) 15:00:10 or you will create Virtuyal blades and do a whole nested environment? 15:00:28 or you will configure fuel to use vCEnter API (not in scope of what we are doing = but available at the openstack layer -) 15:00:29 ? 15:00:44 Yes, we will emulate bare metal using vCenter, creating VMs on blades that will be our compute/control etc nodes 15:00:51 ok. 15:01:06 we will create a diagram though to help 15:01:10 no sweat.. 15:01:29 thanks for your help. got to go to my copper meeting 15:01:30 just be awware that nest in ESXI has some caveats to your networking setup - but get started and i will help where i can 15:01:32 chhers 15:01:34 <[1]JonasB> So I have to end the meeting now 15:01:35 cheers* 15:01:45 <[1]JonasB> Thanks every one 15:01:49 i have to go eat anyay - 15:01:54 FYI - the deploy dailed on Baremetral 15:02:00 <[1]JonasB> #endmeeting