22:03:22 #startmeeting Reddwarf 22:03:23 Meeting started Tue Nov 20 22:03:22 2012 UTC. The chair is vipul. Information about MeetBot at http://wiki.debian.org/MeetBot. 22:03:24 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 22:03:25 The meeting name has been set to 'reddwarf' 22:03:46 Don't have an official agenda link as of yet 22:04:13 looks like there may be some stuff here 22:04:16 #link http://wiki.openstack.org/Meetings/RedDwarfMeeting 22:05:10 #topic Action Items 22:05:27 So going off the agenda, let's review the action items from previous week 22:05:32 vipul: i added that agenda 22:05:37 sounds good. 22:05:56 want to do action items by person or the list above? 22:06:04 hub-cap: suer 22:06:22 grapex: any update on the launchpad site? 22:06:39 ive done all the necessaries for the 3 launchpad sites 22:06:52 reddwarf, python-reddwarfclient and reddwarf-integration 22:07:07 Sounds good, hub-cap. 22:07:23 I've already started using the reddwarf-integration one. 22:07:30 hub-cap: any reason to use you specifically for driver instead of reddwarf-drivers? 22:07:31 sweet 22:07:38 link plz 22:07:54 https://launchpad.net/reddwarf 22:07:57 there is likely not agood reason 22:08:04 just prolly forgot to change 22:08:09 I'll do it now then 22:08:15 changed 22:08:24 sweet 22:08:25 sry didnt see u were changing :P 22:08:42 also, the futurestack designation for reddwarf, does that still make sense? if so, does reddwarfclient need to be added? 22:08:46 #info reddwarf-integration launchpad site: https://launchpad.net/reddwarf-integration 22:09:03 spiffxp: im not sure what the futurestack thing is... jaypipes added it back in the day 22:09:13 maybe he can chime in :D 22:09:36 spiffxp - i don't think it matters that much, supposedly as long as it's a subproject of openstack then everything is cool 22:09:43 hub-cap: looks like the pre-incubating project area 22:09:45 seems like ceilometer and heat aren't using it? 22:09:52 ya its likely just old 22:09:52 #info Reddwarf client binding launchpad site: https://launchpad.net/python-reddwarfclient 22:09:57 that page was added a LONG time ago 22:10:01 thx SlickNik for those infoz 22:10:18 hub-cap: long time ago... 22:10:20 the next item was for grapex to own CI - assume this means getting the integration tests in? 22:10:20 ok I'm going to remove that if no objections 22:10:23 if yall see any issues w/ them hit me up 22:10:27 it's dead, AFAIK 22:10:32 jaypipes: lol ya, do u know is tha tfuture....ok :) 22:10:35 heh 22:10:39 lets remove it then 22:11:23 or should i refer to him as GrapeX2 22:11:42 @hub-cap: np 22:11:53 grapex, GrapeX2, GrapeX3, GrapeX4 22:12:00 grapes 22:12:01 lol 22:12:01 depends on how many tiems hes logged in to irc 22:12:25 yeah hes away right now i think 22:12:33 vipul: lets chat about that in a bit 22:12:38 its one of the agenda items 22:12:52 its big enough to warrant further discussion than a action item 22:13:03 Yep, that's next, if we're done talking about the launchpad site 22:13:20 thats all the action items? 22:13:31 i'm going by person 22:13:34 AH 22:13:44 http://wiki.openstack.org/Meetings/RedDwarfMeeting 22:13:49 that has the next item as CI 22:14:01 lets skip it in the ActionItems portion and talk about it after all the action items have been done 22:14:16 sounds good 22:14:44 sounds good, let's move on to the next item. 22:15:14 yar. lets wait for GrapeXY 22:15:20 whre Y is a number > 0 22:15:38 k, so let's move on to the next action item 22:15:44 hubcap: devstack integration status 22:15:52 that got passed to jcooley 22:15:57 see item #2 22:16:02 :P 22:16:13 3 is done 22:16:14 haha 22:16:34 4 is done, and i havent been online to talk to lifeless or devananda about the image stuff, but its on my list 22:16:47 Greetings 22:16:51 #action hub_cap talk to lifeless or devananda about the image creation 22:16:51 hub-cap: o/ 22:16:53 jcooley was offline a good chunk of last week as well 22:17:00 lifeless: im on paternity leave still :) 22:17:05 ugh i should use the list on top 22:17:16 vipul: likely thats the best way 22:17:26 Yeah, I think 2 is still on jcooley's plate. 22:18:07 remind me of #2 and I'll give an update. 22:18:12 to discuss devstack integration 22:18:20 oh, hai 22:18:33 as in configuration of reddwarf via devstack jcooley 22:18:36 hai devananda 22:18:41 yes. discussed with mordred, need to discuss with devstack folks. still outstanding. 22:18:46 hey devananda. 22:19:02 #action jcooley to discuss reddwarf devstack integration w/ devstack folks 22:19:16 Sweet, thanks jcooley! 22:19:27 Ok, we're on 5 I believe.. Reddwarf-core completed 22:19:37 6 is related.. so done 22:19:43 np 22:20:03 we did a great job w #7 22:20:05 7 - Make sure reviews have blueprints or bugs 22:20:11 goodjob to everyone on that one 22:20:12 yep - we're getting better 22:20:17 yeah now that I've tripped over that twice I'm watching for it 22:20:21 Nice. Thx. 22:20:35 8 - grapex to own CI and make sure it gets accomplished 22:20:40 oh yeah i mentioned that earlier :-P 22:20:41 grapex is having connection issues 22:20:52 i didnt realize until today we had these meetings... 22:20:55 ok we'll wait for his response offline 22:20:57 so hes doing his best to connect up 22:21:12 Sorry guys 22:21:33 is that grapex? grapen? 22:21:37 Ah, welcome! Np. 22:21:38 lol 22:21:57 We're traveling back from San Antonio today. I've been using Nirmal's cell phone with a pretty decent connection, but we got weird issues trying to join freenode just now. 22:22:09 Oh ure on da bus 22:22:19 grape[n] 22:22:33 rnirmal: thx for allowing grapeKERJ#H$#KXNK#J!@@#$%%%1223456 to use your cell 22:22:36 Where "n" is the number of times I tried to connect. 22:22:43 lol 22:22:48 heh 22:22:49 so youre up, CI status 22:23:07 By the way, this is the real GrapeX and not a hacker. Now that I've assured you... 22:23:24 hub-cap: Well we've got a huge chunk of tests running in CI. 22:23:36 We still need to run tests in real mode somehow. 22:23:52 awesome. so they are running as part of the jenkins gate job now? 22:24:19 Yes, because its tox it naturally plugs into the Jenkins gate stuff. 22:24:38 As part of this, the client is now in PyPi. 22:24:38 hub-cap grapen - they are running currently as unit tests would 22:24:41 awesome. so some work accomplished, def some more to go 22:24:43 mordred and I co-own it. Apparently all the uploads to PyPi happen manually. 22:24:45 anyone have a link to the jenkins box handy? 22:24:50 Yeah 22:24:57 HP has done a lot of work getting the public RDLI environment to run. 22:25:02 #link jenkins.openstack.org 22:25:05 #link https://jenkins.openstack.org/view/Stackforge/job/gate-python-reddwarfclient-python26/ 22:25:06 esp1: its in all the reviews in gerrit as well 22:25:23 thx :) 22:25:36 ok we are for sure falling behind :) grapen anythign more to add? 22:25:49 hmm. pretty. 22:26:04 9 - hub_cap make sure the client launchpad page is up to date w/ the series like the reddwarf one 22:26:04 well except the thunder clouds 22:26:11 vipul: check 22:26:12 The next thing we've got to do is find some way to get CI for the RDLI tests running in real mode on a Jenkins box that can plug into Gerrit. 22:26:29 and we need to figure out the image creation for that too 22:26:30 Yeah, a few of us here at HP were looking at running the tests in real mode. Will take that offline with grapex… 22:26:40 hub-cap vipul fwiw I just went through and configured support and answers on that page to use launchpad too 22:26:47 grapen: I think that's what the whole devstack integration work is going to turn into 22:26:49 ok i guess that bleeds us in to the next section, and we can just summarize it then eh? 22:26:52 spiffxp: sweet thx 22:27:01 it will involve adding a redstack-gate job and run integration tests in real mode 22:27:38 k, done w/ action items? 22:27:41 ok anything more to add? 22:28:01 not here 22:28:01 #topic CI / Image Updates 22:28:03 nay 22:28:27 I can speak a bit to the image portion of this 22:28:33 cool 22:28:52 hub-cap: Nope, sorry. Getting those tests to run in the main repo with Tox took forever. 22:28:52 We're currently working on building out a precise image with percona bits... 22:29:05 ps we also have a script we might be able to adapt that sets up a bunch of the image stuff. i can find out internally 22:29:16 hub-cap: I agree with this rnirmal character. 22:29:23 vipul: are u using the same process that devananda is? 22:29:26 :) 22:29:35 we're going to fix the image buildling code in redstack as the first step - and eventually make it so that you can build hp/rax images 22:29:49 ok great 22:29:51 hub-cap - not yet.. this is the interim step to getting things working on precise 22:29:57 that was a bit of a hack job :P 22:30:06 vipul: what are u using then? i suggest debootstrap! 22:30:18 we're also going to look at Devananda and tripleo's image builders and hook into those 22:30:29 ok cool 22:30:29 hub-cap we are in the process of creating something 22:30:50 vipul: Quick question, do we want to stick with Ubuntu 12.4 or go to 12.10? 22:30:56 juice__: cool. when i get back from paternity leave id like to help out 22:31:11 we reviewed the image creation currently in the redstack scripts 22:31:27 grapen vipul we shoudl stick to what nova uses dont u think? it seems funky to have diff version requirements for things that _could_ be deployed in the same env 22:31:27 grapen: We are targetting 12.04 for now, we might want to consider jumping ahead 22:31:46 12.04 please 22:31:49 waht does devstack target? 22:31:59 we'd like to stick with LTS unless there's a strong reason not to 22:32:00 12.04 hub-cap 22:32:00 I think you want a LTS which is 12.04 22:32:01 cuz if we start usign diff versions / etc tis gonna be a nightmare 22:32:06 12.04 it is 22:32:14 devstack targets 12.04, iirc since it is LTS 22:32:28 good, lets target that too, i think we are mostly in agreement eh? 22:32:32 yup 22:32:36 yep 22:32:40 #link https://github.com/tripleo/baremetal-initrd-builder 22:32:41 Sounds good. 22:32:47 grapen: did u have any reason for 12.10? 22:32:49 this is the tripleo image builder that we want to target eventually 22:33:37 hub-cap can I reach out to you with a couple questions on the current image building process? 22:33:38 also, i understand that 12.10 has the correct openstack clients 22:33:41 hub-cap: No, I thought devstack targeted the latest one. My mistake. 22:34:25 juice__: plz do, but grapen can help as well if im not very useful (ill be back officially monday) 22:34:28 but i respond to emails for sure 22:34:43 id say email us both 22:34:49 #info using Ubuntu 12.04 because it is LTS 22:34:56 thanks hub-cap 22:35:19 Ok moving on.. 22:35:21 #info need to work on image building / realmode tests 22:35:26 any CI updates? 22:35:29 i dont think we info'd much for that :P 22:35:43 oh i thought u meant the next topic totally 22:35:53 the second part of this topic 22:36:11 anyone want to own this for the next wk? i think we need to make progress w/ ci 22:36:16 grapen wink wink 22:36:48 One thing i'll point out here is we are trying to get the redstack-gate stuff sorted out - an additional gate on Jenkins that would do real-mode tests 22:36:51 hub-cap: I will try to help w/ devstack integration but I don't think I can own 22:36:58 I'm setting up a VM for the rdli stuff, but getting the fake mode stuff to work in tox took forever. 22:37:10 our guys can work with grapen on this 22:37:11 Hopefully I'll have some progress made tomorrow. 22:37:13 #linkhttps://github.com/openstack-ci/devstack-gate/blob/master/README.md 22:37:16 err... 22:37:18 #link https://github.com/openstack-ci/devstack-gate/blob/master/README.md 22:37:24 I can work with grapen on this… 22:37:41 ooo I one other thing in CI actually 22:37:45 vipul: id like someone from each group to work together 22:37:54 so SlickNik or spiffxp? 22:38:19 I would say SlickNik, I can pester as appropriate 22:38:24 kk 22:38:25 #action SlickNik and grapen to work on redstack-gate 22:38:35 okay, sounds good spiffxp. 22:38:37 I knew enough to get reddwarf-integration into stackforge, but I don't think I have the bandwidth to own 22:38:46 :D cool 22:38:52 hub-cap: mind adding a "this is deprecated" commit to https://github.com/hub-cap/reddwarf_lite-integration? 22:39:10 w/ pointer to https://github.com/stackforge/reddwarf-integration 22:39:11 gonan do it 22:39:18 if u check out the meeting notes 22:39:18 Mod rax/reddwarf && hub-cap/reddwarf_lite to point to stackforge 22:39:21 is on it :) 22:39:26 ahh k 22:39:34 ok let's move on, got a lot to cover 22:39:40 #topic blueprinting 22:39:42 and very little time! 22:39:45 So when we say redstack-gate, I think the first job should be to just get any of us running the tests in a VM. After we do that, then lets look at incorporating it into Openstack CI. 22:40:02 lol we moved on grapen!!!!!! youre too late for input!! 22:40:03 grapen: yep, that's the idea, it would be part of the core CI process 22:40:20 hub-cap what did you want to discuss aobut blueprinting 22:40:34 2 thinkgs 22:40:39 grapen, let's discuss in #reddwarf after the meeting, when you have some time. 22:40:42 vipul: Agreed, just wanted to set everyone's expectations. 22:40:49 SlickNik: Ok 22:40:56 1) should we shoudl be using our own space or openstack's space? <-- gerrit reviews go to openstack, not reddwarf 22:41:12 if u look @ grapens tox test commit, it cant find the blueprint 22:41:27 yeah i noticed that for robert's commit earlier 22:41:44 gotta take dog out brb 22:41:53 i was wondering if there is some configuration we could do for gerrit that would make it point to reddwarf 22:41:57 i wonder if that's just an issue with gerrit config 22:41:58 I wonder if thats because its a StackForge process. 22:42:03 *project* 22:42:11 we all wondered the same thing 22:42:21 might be, I say we action item someone to followup w/ #openstack-infra 22:42:25 i'll look into seeing if thats possible 22:42:38 i'll compare with some of the other proejcts like glance or swift 22:42:50 #action cp16net to look into why blueprints in Reddwaf aren't closed with gerrit reviews 22:43:25 anything else to add to this? 22:43:52 #topic Stackforge 22:44:34 #action hub-cap to Mod rax/reddwarf && hub-cap/reddwarf_lite to point to stackforge 22:44:35 #action hub_cap to modify all rax resources to point to stackforge 22:44:38 LOL 22:44:47 oops, thought you were still away... 22:44:49 wb. :) 22:44:54 related to Stackforge move, there are a couple of reviews in the pipe that resolve some of the naming issues with RDL and D 22:44:57 I've got two outstanding reviews to s/reddwarf_lite/reddwarf as well 22:44:57 RD 22:45:13 ya lets remove the lite thing 22:45:17 all over 22:45:29 i like reddwarf and reddwarf-integration 22:45:31 k, just wanted to double-check that wouldn't mess anything up on your end 22:45:32 along these lines, can the hub-cap repos be modified to be read-only? 22:45:48 vipul: yup and as per the action above im going to point them to stackforge 22:45:59 #action hub-cap also make them readonly 22:46:28 Other resources may already point to stackforge 22:46:48 i tried to mod everything i saw, but there may be others in regard to launchpad 22:46:54 feel free to update anything i missed 22:47:12 #action Look over all references (including launchpad) to reddwarf and ensure they point to stackforge 22:47:31 who owns that, u? 22:47:46 #action vipul owns updating references 22:48:08 k.. next topic 22:48:12 wait 22:48:16 k 22:48:24 #action hub_cap to talk to mordred about why gerrit wont search reddwarf blueprint space 22:48:26 ok we can move on 22:48:27 hub-cap: i made comments to make the reddwarf-intergration PRs to gerrit instead 22:48:29 wanted to get that in 22:48:39 cp16net had a similar action 22:48:50 cp16net: cool from the public repo? <3 22:48:51 hub-cap: yeah i have something simiar 22:48:51 "_ 22:48:57 yeah 22:49:05 #topic os_admin vs root 22:49:22 vs is such a harsh word ;) 22:49:28 so i had a question for yall 22:49:39 #info background: os_admin user does not exist in public version of RD instances 22:49:43 does your agent need to access the database? 22:49:48 #link https://bugs.launchpad.net/reddwarf/+bug/1078981 22:49:59 Launchpad bug 1078981 in reddwarf/grizzly "The mysql user that the guest agent uses needs to be different for HP and RAX" [Undecided,New] 22:50:09 nice :) i like that 22:50:20 hub-cap: our implementation will not require login, n o 22:50:45 ok so maybe we should widen the discussion to _do we need a default login user_ ? 22:50:56 but im not sure thats good for the meeting 22:51:00 maybe a ML thing 22:51:19 yes - sounds like a larger discussion is needed 22:51:19 ML? 22:51:22 rax wants one for the users/dbs/root portions but hp doesnt need. i think we need to figure it out 22:51:24 mailing list cp16net 22:51:27 oh yeah 22:51:55 #action hub-cap start discussion about the need for the extra actions that require login in reddwarf 22:52:02 #info rax wants one for the users/dbs/root portions but hp doesnt need 22:52:03 i miiiight not get on that for another wk 22:52:14 but id like to work it out 22:52:28 id like a nicer way to turn on/off those things that we dont necessarily both need 22:52:46 yeah i agree it could be done multiple ways 22:52:50 hub-cap, yep, agreed, although for the public version, we may need to support both ways 22:52:58 ya so that brings us to 22:52:58 and when we deploy, we'd toggle a flag 22:53:07 shoudl we default the public version to root? 22:53:09 yeah its an option 22:53:13 and let our stuff flag os_admin 22:53:26 One idea that was floating around was sone sort of conf toggle. 22:53:44 grapen, rnirmal cp16net is there anything that would blow up if we default the user to root, and flag it to os_admin? 22:54:07 i tihnk it might require a change in sneaky pete but i dont think thats a deal breaker or anything 22:54:16 not that i can think of off hand right now. 22:54:19 specially since yall dont use it :D 22:54:36 yeah i am sure the same change will need to happen in sneaky 22:54:40 hub-cap: The only thing is when we reset root we need to make sure the agent can use the updated password, if its also logging in as root. 22:54:51 I think that's it, but I feel like I'm forgetting some important edge case. 22:55:09 grapen, we will still use os_admin, it woudl be a flag that we set in the guest conf 22:55:20 grapen: yeah good thing for tests :) 22:55:27 so sneaky woudl always still use os_admin for our needs 22:55:31 or whatever we want to use 22:55:34 #info os_admin can be a config so it works for rax/sneaky pete 22:55:47 it seems that we shooudl do this anyway, cuz a 3rd company might not want to use either of those 22:56:38 #action make user configurable in guest agent 22:56:38 does anyone in particular want to own this? 22:56:59 I can drive it on our end... 22:57:13 #action grapen work w/ vipul to update sneaky once its been pushed 22:57:20 kk, moving on then? 22:57:24 I can look at this as part of building the image yes? 22:57:34 must type faster :) 22:57:54 I will work with vipul on it 22:57:55 juice__ the config would likely be derived at runtime 22:58:01 #topic bugchat 22:58:12 ok so an orthogonal issue 22:58:16 done moving on 22:58:23 #info ive added milestones and series to all the bugs 22:58:38 if i missed anything plz notify me or update it :) 22:58:52 who owns attaching milestones? reddwarf-core? 22:58:55 ive tried to put things that are smaller in g2 and that seem bigger in g3 22:59:12 vipul: not sure i mgith have set myself as the triage, since nova has a single person doing as well 22:59:25 i _thnk_ ttx only does that, but posssibly a whole team does? 22:59:39 do we need to change the way we've been doiing bugs so far? 23:00:02 hub-cap: looks like nova bug team is the manager 23:00:08 or supervisor of the bugs 23:00:15 hmmm we need to have a bug team then it seems 23:00:17 I say we have reddwarf-core triage informally 23:00:22 yeah maybe so 23:00:22 kk 23:00:24 makes sense 23:00:33 and take a formal policy up offline or next meeting 23:00:35 we can mod to do the bug team in the future 23:00:37 #info reddwarf-core to triage bugs 23:00:58 i dont mind triaging as well, so if u dont get around to it, ill try to spend a hr or 2 each wk doing 23:01:01 right now it would be ok to just have the reddwarf-core do it 23:01:20 #topic open-discussion 23:01:32 im chiangin them to reddwarf-core now 23:01:33 running out of time... any remaining out of band items? 23:01:35 is there a meeting after us? 23:01:42 not sure.. 23:01:45 * cp16net shrugs 23:01:51 i dont think there is a meeting after 23:01:58 but i dont really have any open items to add.... 23:02:07 hub-cap, grapen: waht's the best way to ping openstack-core when we want a RAX person to look at a review? 23:02:10 i just add that _in case_ we mised something 23:02:29 openstack-core? id say ping us in #reddwarf 23:02:30 I've started add redstack-core to open reviews, but I'm not sure if that's too noisy 23:02:41 spiffxp: im fine w/ that 23:02:50 spiffxp: I'm not worried about the noise. 23:02:54 if there is no meeting.. i'll give a quick status update on some things that were not part of action items 23:02:54 k 23:02:57 i did notice a review went in today? w/ vipul spiffxp review 23:03:19 Work being done to get redstack on precise 23:03:21 #link https://etherpad.openstack.org/reddstack-precise-12.04 23:03:32 ok cool vipul 23:03:43 spiffxp: sounds good 23:03:44 hub-cap: yah we missed that one I think, I think consensus in #openstack-infra is it's tough to customize that policy too much 23:03:45 hp cloud precise images are a bit different from what rax may have -- outline some of teh tweaks we've had to make 23:04:00 but lets still get some rax eyes on those since they affect our stuff too 23:04:21 #link https://review.openstack.org/#/c/16552/ 23:04:27 mabye i was wrong tho 23:04:56 We're also looking at unit testing holistically 23:05:21 #info HP going to be adding tests for python-reddwarfclient and guestagent 23:05:33 good we abandoned unit testing very quickly but id love to have them back in 23:05:59 python-reddwarfclient doesn't seem to have any, so we're working on that 23:06:10 and the guestagent probably is not well tested either 23:06:13 hub-cap: here's the review that made it in w/o RAX eyes fwiw, if you want someone to look at it https://review.openstack.org/#/c/16551/ 23:06:15 vipul: I'm always for more tests, but we may want to check the code coverage for python-reddwarfclient from the current tox tests in reddwarf 23:06:30 https://review.openstack.org/#/c/16573/ this one was reviewed by one of our guys but not one of our core 23:06:31 vipul: I omitted that directory in the tox.ini file, but you can take that omission out 23:06:33 id prefer them to be core 23:06:48 ya spiffxp thats the one i saw today via email 23:06:53