16:01:33 <trevor_intel> #startmeeting OPNFV Pharos
16:01:33 <collabot> Meeting started Wed Dec  2 16:01:33 2015 UTC.  The chair is trevor_intel. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:33 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:01:33 <collabot> The meeting name has been set to 'opnfv_pharos'
16:01:52 <trevor_intel> bertrand_orange1: welcome
16:02:05 * fdegir_ is having trouble with irccloud
16:02:25 <fdegir> trevor_intel: yes
16:02:28 <fdegir> writing the mail
16:03:26 <trevor_intel> fdegir: ok I will talk to jack today, he may come on this meeting
16:03:27 <fdegir> #info Fatih Degirmenci
16:03:35 <narindergupta> #info Narinder Gupta
16:03:40 <trevor_intel> #info trevor_intel
16:03:50 <wshao> #info Weidong Shao
16:04:07 <bertrand_orange1> # Bertrand Le Lamer
16:04:32 <trevor_intel> Per yesterday release meeting we need to focus on support for CI/release
16:04:36 <bertrand_orange1> #info Bertrand Le Lamer
16:04:44 <trevor_intel> agree?
16:04:58 <trevor_intel> What other topics?
16:05:25 <fdegir> nothing that is more pressing than B-release resources
16:05:28 <fdegir> for me
16:05:45 <fdegir> if the time allows, we can chat about MAA Pilot
16:05:47 <fdegir> MAAS
16:05:50 <fdegir> and
16:05:51 <fdegir> LAAS
16:06:00 <trevor_intel> Ok lets start ...
16:06:07 <fdegir> and before that, perhaps POD config stuff
16:06:10 <fdegir> forgotten that
16:06:29 <trevor_intel> #topic Lab support for CI/releaseB
16:06:41 <trevor_intel> oops
16:07:07 <trevor_intel> fdegir: POD config stuff first?
16:07:17 <fdegir> yep
16:07:20 <trevor_intel> ok
16:07:38 <trevor_intel> #topic POD config opens
16:08:08 <trevor_intel> fdegir: Status of LF PODs?
16:08:56 <fdegir> trevor_intel: I meant yaml based POD config files
16:09:00 <fdegir> but anyway
16:09:07 <fdegir> here is LF POD status
16:09:30 <ashyoung> Sorry I'm late. I had to use the web IRC client
16:09:46 <trevor_intel> ashyoung: NP
16:09:47 <fdegir> #info LF POD2 is operational
16:10:05 <fdegir_> #info pbandzi is working on reconfiguring LF POD 1 back
16:10:54 <fdegir_> #info 4 new LF  blades will be taken into production
16:11:12 <fdegir_> I think that's all for LF HW
16:11:32 <fdegir_> ashyoung: I'm having trouble with irccloud as well
16:11:53 <trevor_intel> #info I am working on settign up training for support of UCS
16:12:01 <fdegir_> +1
16:12:27 <fdegir_> trevor_intel: have you talked to pbandzi?
16:12:43 <fdegir_> he has this simulator which might be useful for playing with it
16:13:21 <trevor_intel> fdegir: I haven't yet ... I will but I think we need something more formal too
16:13:43 <fdegir_> that's too
16:13:52 <trevor_intel> can we move to next topic?
16:14:12 <fdegir_> yes
16:14:24 <trevor_intel> config files?
16:14:32 <fdegir_> yes
16:15:06 <trevor_intel> Jonas made a proposal
16:15:53 <trevor_intel> bertrand_orange1: Do you have any update on config file templates or thoughts about this?
16:17:16 <fdegir_> #link https://gerrit.opnfv.org/gerrit/#/c/3583/
16:17:16 <trevor_intel> fdegir: This is also somethign we will disuss on Thursday as I undertand
16:17:46 <fdegir_> I think it would be good if pharos can comment on it since there are many things to discuss tomorrow
16:17:54 <fdegir_> not sure if we'll have enough time
16:18:48 <trevor_intel> fdegir: help us here ... ok what can we agree or decide on now?
16:19:11 <fdegir_> did you have chance to look at it to the proposal?
16:19:53 <fdegir_> when I said we should comment, I meant comment on the patch directly
16:19:55 <trevor_intel> fdegir: yes but not sure next step
16:20:08 <fdegir_> if you're happy with it, you can +1 to the patch
16:20:38 <trevor_intel> but I think people who are workign hands on with the labs/installs need to comment
16:20:57 <fdegir_> I agree with that
16:21:07 <fdegir_> but at least we can say this is the way to go
16:21:14 <fdegir_> the details must be sorted out
16:21:24 <fdegir_> if that patch sits there, it will not move on
16:21:30 <trevor_intel> So who else on this call agrees?
16:21:39 <trevor_intel> It can't just be up to me
16:22:01 <fdegir_> you represent the majority
16:22:23 <fdegir_> no other pharos committer has been in this meeting if I'm not terribly mistaken
16:22:51 <fdegir_> since that file will end up in pharos repo
16:23:24 <fdegir_> perhaps we can pull morgan_orange to here
16:23:28 <fdegir_> morgan_orange: are you there
16:23:43 <morgan_orange> fdegir not far
16:23:43 <trevor_intel> So what I can do is prod the Pharos committers
16:23:59 <fdegir_> prod?
16:24:18 <trevor_intel> push
16:24:41 <bertrand_orange1> trevor_intel: sorry Trevor no update in lannion at the moment
16:24:43 <fdegir_> that would be good
16:24:59 <fdegir_> morgan_orange: I've seen you commented on config patch sent by Jonas
16:25:00 <trevor_intel> I am not comfortable being the only one to make important decisions
16:25:17 <fdegir_> are you ok with the pod config files which falls into pharos area
16:25:49 <fdegir_> morgan_orange: ^^
16:26:03 <morgan_orange> you mean pharos has to work on the pod_config.yaml?
16:26:17 <fdegir_> morgan_orange: pod config files will end up in pharos repo
16:26:26 <morgan_orange> yes
16:26:28 <morgan_orange> makes sense
16:26:31 <fdegir_> morgan_orange: based on example provided by Jonas
16:26:39 <fdegir_> so if we get an agreement regarding that
16:26:46 <morgan_orange> I think it was somehow what was imahined in the Jira asking for a json description...
16:26:51 <fdegir_> we can request the POD owners to create those files for their labs/pods
16:26:57 <morgan_orange> that is pending for a while
16:27:01 <fdegir_> jira says json/yaml
16:27:22 <morgan_orange> but yes, it shall be up to each lab owner to provide such file
16:27:27 <fdegir_> yes
16:27:30 <morgan_orange> the difficulty is to agree on what we put into
16:27:39 <morgan_orange> if we consider MAAS export there are lots of lines...
16:27:50 <fdegir_> the thing is, the jira stuff has been there for about 2 months
16:27:59 <fdegir_> and proposal from Jonas is the only thing we have
16:28:40 <fdegir_> https://jira.opnfv.org/browse/PHAROS-30
16:28:58 <trevor_intel> fdegir: perhaps we set a deadline for comments or alternative proposals?
16:29:11 <fdegir_> so Jonas kind of fixed that story partly
16:29:19 * DanSmithEricsson seems n. america IRC Is back online
16:29:28 * DanSmithEricsson wave
16:29:37 <fdegir_> and this one
16:29:38 <fdegir_> https://jira.opnfv.org/browse/PHAROS-34
16:31:14 <trevor_intel> Pharos-30 is POD config
16:31:24 <trevor_intel> Pharos-34 is deploy config
16:31:36 <fdegir_> yes
16:31:43 <fdegir_> that's what Jonas proposes
16:31:49 <fdegir_> templates
16:32:04 <trevor_intel> There will be buil config and test config too ... but I don't think those belong to Pharos?
16:32:12 <trevor_intel> build
16:32:14 <trevor_intel> agree?
16:32:18 <morgan_orange> just the pod description
16:32:32 <morgan_orange> if you look at Jonas patch https://gerrit.opnfv.org/gerrit/#/c/3583/
16:32:32 <fdegir_> PHAROS-31 and PHAROS-35 are for the real files based on templates provided by PHAROS-30 and PHAROS-34
16:32:44 <morgan_orange> only pod_config.yaml should be managed by lab owner
16:32:53 <fdegir_> morgan_orange: that's right
16:32:55 <morgan_orange> as Jonas cannot guess all the lab configs
16:33:04 <fdegir_> yep
16:33:14 <fdegir_> so it is important for installers and lab owners to look at the patch
16:33:41 <DanSmithEricsson> fatih - a question though
16:33:47 <fdegir_> DanSmithEricsson: yes
16:33:49 <morgan_orange> the other config files are installer dependent, first one is generic to all, 2 and 3 are installer specific (that is why it is in a folder)
16:33:50 <trevor_intel> ok so I can hasstl teh lab owners about this and give a deadline ... is that good enough?
16:34:02 <DanSmithEricsson> about this patch.. am i to understand that this config will still be held in a pharos or fuel repo somewhere?
16:34:14 <DanSmithEricsson> since i cant really put the IPMI/ILOM IP and MACs in a public place
16:34:56 <fdegir_> trevor_intel: good. what is the deadline you're thinking?
16:34:56 <fdegir_> DanSmithEricsson: pod config will be in pharos repo
16:34:56 <fdegir_> DanSmithEricsson: deployment config could end up in releng
16:34:56 <fdegir_> DanSmithEricsson: or genesis
16:35:03 <fdegir_> DanSmithEricsson: that is not decided yet but Jonas was mentioning releng since the CI is driven from releng repo
16:35:15 <fdegir_> DanSmithEricsson: so we can pass the deployment config to CI from releng easily
16:35:23 <DanSmithEricsson> fair enough - thanks.. i think and important consideration would be that from a standpoing of the lab
16:35:24 <fdegir_> DanSmithEricsson: but I don't really care tbh
16:35:28 <trevor_intel> fdegir: 1 week
16:35:33 <DanSmithEricsson> the configs can only go so far as the Jumphost in terms of public info.
16:35:53 <fdegir_> trevor_intel: ok, then we discuss during next pharos meeting
16:36:12 <DanSmithEricsson> the format and structure sure, but we cant have a central place where all the lab info is avaiable,, since if someone cloned the repo they only need one key and we are open - sorry - we got attacked yesterday so its prominent on my mind.
16:36:17 <trevor_intel> fdegir: ok
16:36:18 <DanSmithEricsson> great!
16:36:36 <fdegir_> DanSmithEricsson: would be good to see this comment on the patch as well :)
16:36:43 <DanSmithEricsson> will add it
16:36:44 <trevor_intel> DanSmithEricsson: hmmm, ok
16:37:17 <fdegir_> that kind of sensitive info should be excluded from config
16:37:21 <DanSmithEricsson> the challenge is that while the PODs are isolated off, theere is still an uplink to a switch thus surrounding infra is vulnerable to someone who gets in
16:37:23 <DanSmithEricsson> yup
16:37:35 <DanSmithEricsson> and now, since we were hit, i will have to enforce this
16:37:43 <fdegir_> DanSmithEricsson: I got you now
16:37:44 <DanSmithEricsson> and clear out our lab infra ip info off OPNFV now
16:37:46 <DanSmithEricsson> :)
16:37:48 <fdegir_> DanSmithEricsson: you're a lab owner!!!
16:37:56 <fdegir_> DanSmithEricsson: so you have to comment on it
16:37:59 <DanSmithEricsson> cause if someone hit our lab, then they can ge to the switch, they "could" get further
16:38:00 <DanSmithEricsson> yu
16:38:08 <DanSmithEricsson> yup.. will do.. no more comment needed today
16:38:10 <DanSmithEricsson> sorry ..
16:38:27 <fdegir_> DanSmithEricsson: sorry for what! happy to have you here
16:38:41 <DanSmithEricsson> overriding the flow of the meeting with side comment
16:38:43 <fdegir_> and your comments are always welcome
16:38:43 <DanSmithEricsson> :)
16:38:44 <DanSmithEricsson> thx
16:38:55 <trevor_intel> +1
16:39:02 <fdegir_> trevor_intel: I think we covered as much as we can for config stuff
16:39:10 <trevor_intel> Can we discuss release support now?
16:39:14 <fdegir_> +
16:39:19 <fdegir_> 1
16:39:20 <DanSmithEricsson> 1
16:39:54 <debrascott> +1
16:39:56 <trevor_intel> #topic CI/release lab support
16:40:21 <fdegir_> the idea with release support to scale our infra so we can have capability to test different combinations
16:40:34 <fdegir_> apart from supporting this
16:40:48 <fdegir_> ease the work done by test projects by protecting them from non-working deployments
16:41:07 <trevor_intel> This is current allocation of labs https://wiki.opnfv.org/pharos_rls_b_labs
16:41:12 <fdegir_> that results in separating installer CI stuff from platform/test/release CI
16:41:23 <fdegir_> and just this one doubles the needed hw
16:41:35 <fdegir_> if we forget about all the combinations
16:42:09 <fdegir_> we have 4 installers and we need 4x2 PODs; 1 for installer CI and 1 for platform CI, per installer
16:42:23 <fdegir_> and if you add combinations, then the number grows
16:42:36 <fdegir_> obviously we will try our best to utilize what we have as much as possible
16:42:51 <fdegir_> so that's why I put the number 8 out there
16:42:55 <fdegir_> 8 PODs
16:43:33 <fdegir_> and this is the proposal for Brahmaputra CI process: https://wiki.opnfv.org/octopus/brahmaputra_ci_process
16:43:39 <fdegir_> it doesn't go into details
16:44:02 <trevor_intel> fdegir: can yuo clarify again the meaning of installer CI and platform CI?
16:44:07 <fdegir_> ok
16:44:18 <fdegir_> currently we run stuff like this
16:44:31 <fdegir_> we first build installer using latest in installer git repo
16:44:41 <fdegir_> and then we try to deploy using the latest built ISO
16:44:52 <fdegir_> and if it fails, it stops test execution
16:45:10 <fdegir_> the problem here is that test projects depend on installers in this setup
16:45:23 <fdegir_> and since we always try to deploy using an unknown version of the installer
16:45:28 <fdegir_> we fail time to time
16:45:43 <fdegir_> the first picture on the wiki: https://wiki.opnfv.org/octopus/brahmaputra_ci_process
16:46:05 <fdegir_> what is proposed there is not to use the latest on the pods used for testing
16:46:13 <fdegir_> but use a known/working version of the installer
16:46:23 <fdegir_> so test projects can have results next day when they come into work
16:46:24 <trevor_intel> Platform CI is what test projects use for release testing?
16:46:35 <fdegir_> trevor_intel: yes
16:46:59 <fdegir_> the installer CI will give feedback to installer projects
16:47:09 <fdegir_> currently both are mixed
16:47:14 <trevor_intel> Installer CI is what is used to build and deploy the installers
16:47:14 <fdegir_> used by installer and test projects
16:47:21 <fdegir_> trevor_intel: that's right
16:47:33 <fdegir_> platform CI will start from deployment and do the rest
16:47:42 <fdegir_> if there is no recent version of the installer
16:48:03 <fdegir_> we can perhaps skip the deployment totally (this depends on test projects not messing up with the deployment)
16:48:17 <fdegir_> this shortens the time
16:48:27 <fdegir_> reduces the risk of having non-working deployment
16:48:36 <fdegir_> and opens up space for us to do feature testing perhaps
16:48:48 <fdegir_> if you think a failure in deployment
16:49:00 <fdegir_> that costs at least 8 hours
16:49:17 <fdegir_> 4 lost hours during the night and 4 lost hours for redploying the pod for testing
16:49:19 <trevor_intel> Question: all the PODs beingused by testing (yardstick/functest) for development ... why can they not be repurposed as production CI?
16:50:17 <DanSmithEricsson> well
16:50:22 <DanSmithEricsson> thats an interesting question
16:50:25 <DanSmithEricsson> since for Yardstick for example
16:50:31 <DanSmithEricsson> i think (have to chekc with Ana for sure)
16:50:44 <trevor_intel> Yarstick is using at least 3 PODs today
16:50:45 <DanSmithEricsson> but they want to have Yardstick measuring everything else as it comes up and goes down
16:50:46 <DanSmithEricsson> ya
16:51:10 <morgan_orange> Functest uses only LF POD2 so far...
16:51:34 <fdegir_> apart from using them for real development work
16:51:35 <morgan_orange> we just started working to use Orange POD 2 (beta joid)
16:51:42 <fdegir_> some PODs might not conform Pharos specs
16:51:53 <fdegir_> some PODs might be incapable of fetching artifacts
16:52:07 <fdegir_> due to low bandwidth, access problems, etc
16:53:21 <trevor_intel> fdegir: makes sense ... but isn't it up to Yardstick to say they don't meet the requirements?
16:53:26 <fdegir_> btw, I'm not against this idea
16:53:44 <fdegir_> partly yardstick
16:53:50 <fdegir_> but if they can't access google
16:53:55 <fdegir_> there is nothing yardstick can do
16:54:06 <fdegir_> that's why they deployed arno sr1 in some labs in China
16:54:15 <fdegir_> and they don't redeploy
16:54:20 <trevor_intel> I get the impression our current resources are not being used as effectively as possible ... just adding more may not solve the problems?
16:54:38 <fdegir_> actually we are not asking for adding more
16:54:43 <fdegir_> we're trying to match the numbers
16:54:52 <fdegir_> morgan_orange already have the table
16:55:06 <fdegir_> and we need to fill that table using already available pods
16:55:25 <fdegir_> https://wiki.opnfv.org/brahmaputra_testing_page
16:55:40 <trevor_intel> fdegir: means the test coverage/dependency matrix?
16:55:48 <fdegir_> the first table on the page
16:55:52 <fdegir_> listing pods
16:56:13 <DanSmithEricsson> sorry
16:56:26 <DanSmithEricsson> but there is a requirement to all installers that they support internetless installations
16:56:38 <DanSmithEricsson> so the access to google, shouldnt come into play
16:56:43 <fdegir_> DanSmithEricsson: I think it changed
16:56:47 <DanSmithEricsson> lab should be able to implement either option
16:56:52 <fdegir_> DanSmithEricsson: scripted installations are allowed
16:57:02 <fdegir_> DanSmithEricsson: apart from that, we store installer ISOs on google
16:57:09 <fdegir_> DanSmithEricsson: so even if all the installers have ISO
16:57:17 <DanSmithEricsson> are allowed - but the requirement from SR1 is that a lab must support a private installation
16:57:22 <fdegir_> DanSmithEricsson: we need an artifact repo for China
16:57:31 <DanSmithEricsson> they threw this away am i to understand?
16:57:32 <fdegir_> DanSmithEricsson: or mirror in China
16:57:52 <fdegir_> DanSmithEricsson: there was a TSC decision to allow scripted installs if I'm not mistaken
16:58:10 <fdegir_> DanSmithEricsson: or at least a discussion with no major objection
17:00:53 <trevor_intel> fdegir: back to mrgans table ... so there are x4 PODs for each stable version of deployers ... is this what you would consider as platform CI?
17:01:06 <fdegir_> trevor_intel: yes
17:01:08 <debrascott> Install methods available are part of CI configuration right? Is there a full matrix yet of pod configurations needed?
17:01:12 <trevor_intel> means today there are x4 platform CI PODs
17:01:27 <morgan_orange> planned...
17:01:39 <fdegir_> yep, that table shows the needs
17:01:40 <morgan_orange> only LF POD2 is performing automatic testing
17:01:46 <trevor_intel> So if you number of x8 is true ... we only need to allocate anohter x4 PODs to Platform CI?
17:02:06 <fdegir_> that's right if morgan_orange doesn't object
17:02:18 <fdegir_> and that's to start with
17:02:36 <fdegir_> if we see they're not utilized, we can surrender the ones we don't use
17:02:48 <morgan_orange> the difficult is "only need to allocate"
17:03:02 <debrascott> Ash offered up 5 pods yesterday too. Is anyone working yet to see if these meet config needs?
17:03:18 <ashyoung> Well, my definition of a POD was off :)
17:03:30 <ashyoung> It's one POD
17:03:38 <fdegir_> :)
17:03:48 <ashyoung> I put in a request to supply more
17:04:02 <fdegir_> if we speculate now
17:04:22 <fdegir_> apex uses intel pod2 and we're getting lf pod1 rebuilt
17:04:27 <trevor_intel> IF we can get 1 POD from Dell, Huawei (SC), Intel .... that is 3
17:04:38 <fdegir_> fuel uses lf pod2 and DanSmithEricsson will hopefully give another POD for fuel
17:04:50 <fdegir_> morgan_orange: you said joid will end up on orange pod
17:05:01 <fdegir_> and joid is on intel pod5 as well
17:05:17 <trevor_intel> orange pos is not bare-metal?
17:05:19 <fdegir_> compass is on Huawei SC
17:05:21 <morgan_orange> target is Intel POD 5
17:05:36 <morgan_orange> for CI
17:05:41 <fdegir_> morgan_orange: then we lack 2 PODs
17:05:49 <fdegir_> 1 for compass latest and 1 for joid latest
17:06:06 <fdegir_> apex and fuel have the pods for latest and stable based on my speculation
17:06:51 <ashyoung> BTW, the PODs I'm supplying would be via Supermicro, not Huawei
17:08:10 <fdegir_> don't know what supermicro is - as long as they conform pharos spec and we can deploy and test opnfv on it, that's fine
17:08:12 <trevor_intel> JOID has Intel PODs 5 (production) and POD 6 (dev/test)
17:08:40 <ashyoung> understood
17:08:47 <fdegir_> trevor_intel: adding this to table
17:08:54 <fdegir_> and the apex and fuel ones as well
17:09:55 <fdegir_> added
17:10:55 <fdegir_> the row latest on the table means that
17:11:12 <fdegir_> they will be connected to jenkins and produce stable ISOs every night
17:11:22 <fdegir_> and during daytime, they can be used for development
17:11:29 <fdegir_> anyone disagrees with this?
17:11:50 <morgan_orange> when is night and daytime?
17:11:51 <trevor_intel> I think we should explain that on teh page next to the table
17:12:00 <trevor_intel> What about compass?
17:12:08 <fdegir_> morgan_orange: that's a good question
17:12:15 <fdegir_> currently it is based on UTC
17:12:31 <fdegir_> trevor_intel: need to fix a POD for compass
17:12:39 <DanSmithEricsson> i am just waiting on aIP assignment and yes you have your POD
17:12:51 <trevor_intel> fdegir: Intel POD 8
17:13:14 <fdegir_> updating the table with this
17:13:33 <trevor_intel> We were doing to allocate that to STORPERF but maybe can make another plan
17:13:53 <fdegir_> left question marks on all of them
17:14:02 <fdegir_> need to make installer teams know about this
17:14:05 <trevor_intel> FYI POD 3 will be shared by VSPERF, OVSNFV, kvm4nfv
17:14:07 <fdegir_> and what this means for them
17:14:30 <fdegir_> trevor_intel: I might have 2 of your servers given to another project - but this is another subject
17:14:42 <morgan_orange> and a big, question once we will have put the stickers on the 8 PODs, who will install them...jenkins jobs are not ready and will require some tuning
17:14:57 <trevor_intel> In tomorrows meeting we shoudl review all of this and explain ... yes?
17:15:02 <DanSmithEricsson> well i can help with installs if you need a  "grunt"
17:15:04 <fdegir_> trevor_intel: yes
17:15:12 <fdegir_> morgan_orange: I think for latest case, it is easy
17:15:15 <DanSmithEricsson> if there is access and you have a plan of what is to be installed where i can help with that
17:15:27 <fdegir_> morgan_orange: we already have apex, compass, and fuel running on those pods
17:15:37 <fdegir_> morgan_orange: in installer CI mode
17:15:45 <fdegir_> morgan_orange: I'll create new set of jobs
17:15:51 <fdegir_> morgan_orange: and send for review
17:16:02 <morgan_orange> ok and David is working with cnaonical boyes on joid
17:16:16 <fdegir_> the new jobs will be named opnfv-fuel-daily, opnfv-compass-daily
17:16:18 <fdegir_> etc
17:16:24 <morgan_orange> so goal to have 8 PODs up&running connected to CI before xmas break...sounds ...reasonable...
17:16:33 <fdegir_> and your test jobs will be hooked under these new opnfv-* jobs
17:17:09 <fdegir_> morgan_orange: jsut to clarify since DanSmithEricsson asks same thing
17:17:19 <fdegir_> morgan_orange: what you mean with install?
17:17:25 <fdegir_> morgan_orange: they will be installed by jenkins jobs
17:17:44 <fdegir_> morgan_orange: whenever we have new stable version of the installer
17:17:53 <morgan_orange> ok
17:17:56 <fdegir_> morgan_orange: if we want to deploy a different config, then this should be done using jobs
17:18:05 <fdegir_> we need to create the jobs based on what we know
17:18:16 <morgan_orange> 100% OK
17:18:36 <ashyoung> What's the gap on PODs? It sounds like you have it covered.
17:18:37 <fdegir_> to me the first task is cloning existing jobs under the names opnfv-*  so things are clearly separated
17:18:41 <fdegir_> and then work from there
17:19:03 <fdegir_> ashyoung: looks like it
17:19:23 <fdegir_> but we need plan B as well :)
17:20:42 <ashyoung> Just let me know. I will have 1 by end of the day. And I can probably have a couple more early next week. This is all latest HW.
17:21:02 <ashyoung> All using latest Intel SSDs too.
17:21:07 <fdegir_> ashyoung: will do, thanks
17:21:13 <ashyoung> Sure
17:21:18 * fdegir_ has some experiences with latest hw :(
17:21:31 <fdegir_> lf pods were kind of tricky...
17:21:41 <ashyoung> Also, 10Gb to each VM, if that makes any difference.
17:21:49 <ashyoung> Oh
17:22:15 <fdegir_> but anyway, we'll see what happens
17:22:16 <ashyoung> I'm working with Rebar folks today to put all of the bare metal provisioning in place
17:22:20 <ashyoung> Sounds good
17:22:30 <fdegir_> trevor_intel: ashyoung: morgan_orange: DanSmithEricsson: debrascott: are we happy then?
17:22:55 <ashyoung> I'm happy :)
17:22:57 <morgan_orange> yes
17:23:00 <fdegir_> me too :)
17:23:49 <trevor_intel> me too :)
17:24:08 <trevor_intel> Ok to end meeting?
17:24:15 <trevor_intel> Then pick up tomorrow
17:24:37 <trevor_intel> #endmeeting