BRIDGE: https://zoom.us/j/661303200?pwd=TFdRd0c2MTJUem8xa252UGJHTE1Mdz09

Passcode: 209247

We will start our meetings by mentioning the project's Antitrust Policy, which you can find linked from the LF and project websites. The policy is important where multiple companies, including potential industry competitors, are participating in meetings. Please review and if you have any questions, please contact your company legal counsel. Members of the LF may contact Andrew Updegrove at the firm Gesmer Updegrove LLP, which provides legal counsel to the LF.

Attendance is taken purely upon #info in Zoom Chat 

Zoom Chat Log

 06:59:37 From Ranny HAIBY (Samsung) to Everyone : #info Ranny Haiby, Samsung
07:00:30 From Timo Perala (Nokia) to Everyone : #info Timo Perala, Nokia
07:00:42 From Andreas GEISSLER (DT) to Everyone : #info Andreas Geissler, DT
07:00:43 From Dong Wang (China Telecom) to Everyone : #info Dong Wang, China Telecom
07:01:16 From Srini Addepalli (Intel) to Everyone : #info Srini Addepalli, Intel
07:01:32 From SaiSeshu MUDIGANTI (Huawei) to Everyone : #info Seshu, huawei
07:01:46 From Eric Debeau to Everyone : #info Eric Debeau, Orange
07:01:56 From Olivier Phénix (Bell Canada) to Everyone : #info Olivier Phénix, Bell Canada
07:02:49 From Jason Hunt to Everyone : #info Jason Hunt, IBM
07:03:16 From bin.yang@windriver.com to Everyone : #info Bin Yang, Wind River
07:07:21 From Ciaran Johnston (Ericsson) to Everyone : #info Ciaran Johnston, Ericsson
07:07:37 From Catherine Lefevre to Everyone : #info, cath. lefevre ATT
07:08:57 From Kenny PAUL (LFN) to Everyone (in Waiting Room) : @ciaran & caterine - thanks
07:19:39 From Kenny PAUL (LFN) to Everyone (in Waiting Room) : Link to OWL wiki: https://wiki.onap.org/x/5KvQAw
07:28:17 From Tracy van Brakle (AT&T) to Everyone : www.o-ran.org
07:38:15 From Catherine Lefevre to Everyone : @Kenny
07:38:18 From Catherine Lefevre to Everyone : we can not share
07:38:27 From Catherine Lefevre to Everyone : but can you share the agenda?
07:39:42 From Kenny PAUL (LFN) to Tracy van Brakle (AT&T)(Direct Message) : super exciting stuff! :-D
07:41:38 From Vijay Venkatesh Kumar (AT&T) to Everyone : +2. It will be devastating to lose this intel/windriver lab. It has been used heavily for both dev/test (atleast by DCAE project)
08:02:19 From Srini Addepalli (Intel) to Everyone : Sorry. Need to drop now.
08:05:10 From Alla Goldner to Everyone : #info Alla Goldner, Amdocs
08:05:40 From Alla Goldner to Everyone : really sorry for being late, had internal meeting with my management...
08:07:53 From Kenny PAUL (LFN) to Everyone : @alla thanks
08:21:25 From Jason Hunt to Everyone : my apologies, I have to drop a few mins early
08:26:00 From Ciaran Johnston (Ericsson) to Everyone : Me too, apologies
08:27:48 From Catherine Lefevre to Everyone : Dear TSC members (and proxies), we will perform an offline vote for the sign-off since we are losing TSC members
08:28:09 From Catherine Lefevre to Everyone : Thanks to the whole ONAP Community
08:36:30 From Kenny PAUL (LFN) to Everyone : we are still at quorum at the moment.
08:41:49 From Kenny PAUL (LFN) to Everyone : still at quorum.
08:43:43 From Kenny PAUL (LFN) to Everyone : #vote Does the TSC approve the signoff of the Honolulu release with the public announcement to be made on May 11? 
08:43:58 From Alla Goldner to Everyone : #vote +1
08:44:00 From Andreas GEISSLER (DT) to Everyone : #vote +1
08:44:01 From Olivier Phénix (Bell Canada) to Everyone : #vote +1
08:44:01 From Timo Perala (Nokia) to Everyone : #vote +1
08:44:05 From Ranny HAIBY (Samsung) to Everyone : #vote +1
08:44:07 From Eric Debeau to Everyone : #vote +1
08:44:08 From SaiSeshu MUDIGANTI (Huawei) to Everyone : #vote +1
08:44:11 From Catherine Lefevre to Everyone : #vote +1 with the condition that remaining release note finalized by May 5th (Thomas
08:44:44 From Catherine Lefevre to Everyone : #vote +1 with the condition that remaining release note finalized by May 5th eOD (Thomas - let me know how to help) and honolulu release maintenance
08:45:04 From Kenny PAUL (LFN) to Everyone : #endvote
08:45:12 From Catherine Lefevre to Everyone : Congratuations to ALL OF YOU !!!
08:45:20 From Olivier Phénix (Bell Canada) to Everyone : Congrats everyone!
08:45:35 From Kenny PAUL (LFN) to Everyone : #AGREED the TSC approves the signoff of the Honolulu release with the public announcement to be made on May 11
08:46:01 From SaiSeshu MUDIGANTI (Huawei) to Everyone : Congratulations!!!

Zoom auto-transcript service - These are often translated incorrectly and can be misleading. They are NOT Authoritative!   Information as to why .
They are included here as a time stamp cross-reference for the recording only!  The notes above this line and the actual recordings are authoritative. 

07:04:39 Okay transcript should be turned on.
07:04:48 At this point, and went through the normal housekeeping stuff. Once I moved a little transcript thing out of my way.
07:05:00 Meetings being double recorded.
07:05:03 Participants muted when you come in please stay muted unless you're speaking if you're on a phone line you can use star six to unmute yourself.
07:05:09 If you send a private chat messages, who the meeting host they will become part of the public record when it is cut and pasted into the meeting minutes.
07:05:22 And we will start our meeting by mentioning, our anti trust policy. You can find this link from the elephant project websites as policy is important where we have multiple companies, including potential industry competitors participating in these meetings.
07:05:35 Please review if you have any questions please contact your company's legal counsel, member of the lF may also contact Andrew up to grow up at the firm Gizmo appt Grove LLP which provides legal counsel to the LL.
07:05:53 So first up on the agenda today we'll be talking about Iran sc.
07:06:01 We'll do a quick thing for the normal infrastructure, then we'll focus on releases
07:06:13 and
07:06:17 get to a TLC to see if we can get that wrapped up.
07:06:23 So with that, Tracy, you should have presenter rights.
07:06:34 Let me unmute myself.
07:06:39 You should be seeing my two beautiful Siamese cat I hope my audio is okay, yes and yes. Okay.
07:06:50 Can you see what I'm sharing.
07:06:53 Slide.
07:06:54 Yes.
07:07:02 Okay.
07:07:07 Okay so okay to begin.
07:07:14 All right, a lot of you are aware that Orion. Orion software community own app, have been doing joint proofs of concepts and plug fests for the past. Well going on three years.
07:07:32 I'm here to present is the concept for the third global around POC pod fest.
07:07:40 And like it's two predecessors, it will include important elements and contributions from own app.
07:07:48 Open Daylight Of course. And our goal is to have this test bed and the concept defined established, such that we can support a series of interrelated but mostly independent demonstrations of multiple operator ran as well as 5g ai enabled telemedicine,
07:08:12 and oh that's a mouthful but it's a huge area of interest. Within around.
07:08:17 We're doing a catalyst project until a management forum.
07:08:21 As I already indicated, a lot of what we.
07:08:24 A lot of what we have to a lot of what we have today that realizes the concept of small service management orchestration layer, non real time Rick other Orion entities have been realized with components from own app open networking foundation is working
07:08:46 on near real time Rick in parallel with the near real time Rick project in around software community.
07:08:55 I pause the integrated public alert and warning system.
07:08:58 I'm sure there are similar initiatives in other regions across the around the globe.
07:09:07 And then last but not least, you know, how do we deliver, not this year not next year but probably five years, maybe 10 years, secure interested, 5g, and AI powered telemedicine.
07:09:21 So I'm going to go through this fairly quickly. I don't know if Ben Chung is going to be able to join today. But, Martin Skorupski will explain the particulars about, you know, how we expect to get this to work.
07:09:37 Everybody knows what pooling and slicing is what's new here is it's pulling multi vendor network resources across multi operator ran slicing their millions of definitions of slicing.
07:09:53 We're trying to clarify a few specific definitions of slicing start with the easiest to demonstrate definition and get more complicated more sophisticated, as we go along.
07:10:08 Now, the you know the issue here from an operator perspective, and I'm speaking from the operator perspective is so far there is no mechanism whereby operators forget about vendors can agree upon resource pooling.
07:10:25 But this is an evolving request from for various public safety agencies, there will be city scale disasters that will necessitate two operators who are competitors direct competitors to set aside their differences and pull network resources to get through
07:10:43 the disaster.
07:10:46 So, basically real high level.
07:10:49 The concept is, first of all, look at normal operation everything's fine handovers restricted to a single operator based on the UE subscription, and probably operator a, it could be a TNT and operator be could be Verizon or vice versa and North America
07:11:08 could be, you know, I'm just making this up, you know Docomo and or KDDI in Japan.
07:11:14 So, each operator is on data using AI to optimize to OVQOSC or traffic, you know, do all the good things that we do, but it's only on an individual operator basis.
07:11:27 Now after a city scale disaster.
07:11:30 And, you know, we've had a few. You know we in the United States recently there was a city scale disaster in Nashville, Tennessee.
07:11:38 So you have your towers functioning. You know, we want to be able to pull these are the network resources that survived the disaster across operator a an operator, be it could easily be, it could be more than two operators.
07:11:54 We want to hand over prioritize to the nearest available access point. We don't want operator, a customers.
07:12:04 You know, loss of communication because, you know, they happen to be close to an AP that from operator B's network that they could just as easily. You know, hand over to AI artificial intelligence will now assist decision makers and optimize across this
07:12:26 this pool to network and naturally slices, can be, let me just close this, but I close this
07:12:36 slices can be formed in advance. Let's say we have, you know, a forecast that there's going to be a, you know, a tsunami, God forbid, we could set up slices in advance so that we're ready.
07:12:50 In anticipation of a weather a catechism, a cataclysmic meteorological event. And there have been quite a few of those as well in multiple regions around the globe.
07:13:01 So I'm going to go through this real quick.
07:13:07 Pre disaster operator and everything's fine operator be everything's fine and traffic is spread across all available network nodes in the range of Beuys post disaster, you can see that this cataclysmic event destroys six of operator A's access points
07:13:25 three of these, but nine of the 11 remaining it APS are still alive and selected for pulling and slicing based on location.
07:13:35 And of course, you know, not all US will necessarily have access to the slice, or in some cases, there may be multiple slices. So you'll see this color coding is this yellow color coding means this is public safety slice formed from the nine selected
07:13:54 APS, and of course I'm not showing how this is orchestrated how this is done, but there are a couple slides that will touch upon that.
07:14:04 So the US and range, are no longer operator specific they have access to all APS contained in the slice.
07:14:12 Clearly this is a simplified view.
07:14:16 For our first demonstration, which will be last week of June.
07:14:21 This is the simplified slicing approach that will be taken for the next demonstration, which will be the third global Orion POC pod fest the last week of November, will do actual ran slicing and that will necessitate, maybe public safety having a service
07:14:39 management orchestration layer that sits on top of operator, a an operator be service management orchestration layers.
07:14:48 So after everything's said and done, it's time to normalize it's time to repair normalize the intelligence collected and analyzed by our very sophisticated artificial intelligence planner will be made available for operator a, and operator be to repair
07:15:07 damaged network resources or in some cases replace those that are utterly destroyed.
07:15:14 So forget this life is life. Well no, go back to the slide this is multi domain, this is, you know, obviously, in network, it's multiple operator ran its multiple network domains within each operator network.
07:15:32 And it's also, There's the human elements, there's the infrastructure.
07:15:39 There are the objects that are specific interest of first responders, all of those will be modeled in some way. And, again, there's at present, no clear way of how this will all work together.
07:15:53 So, this is going to be a series of experiments and demos, and hopefully bring together these different domains, such that the network can quickly adapt to the first responders requirements, as well as the human element.
07:16:11 When there's a city scale disaster there are casualties there are injuries, there are
07:16:21 people with head trauma, serious head trauma, who needs to get to the head trauma surgical center before people with cuts and bruises.
07:16:27 So, again, our first demonstration will be, you know, simulated conceptual, but as we get farther along will integrate more closely with evolving standards in public safety.
07:16:46 Of course we're in North America so we'll focus on our region, but other regions around the world have very similar interests and public safety initiatives.
07:16:57 So how this will work.
07:16:59 Martin I hope you were able to join this idea.
07:17:05 Okay, so keep it easy.
07:17:10 Yeah. Okay, so here we are going to instantiate the solution for those experiments in open violence lead. And that's an open lead from from owner.
07:17:33 The beauty on this lab is that we have licensed frequency for research and experiments. And so we really can do something outdoor with us experiments, and he is the basic layering so we have an intelligence layer on top we have the control layer for the
07:17:40 And here is the basic layering so we have an intelligence layer on top we have the control layer for the cloud contour for network control and for controlling the user equipment and IoT devices which are part of the experiments and that supports the first
07:17:52 responders are basically looking for right they really want to, to control their vehicles and drones and all that stuff.
07:18:01 This is course on an infrastructure with all the different network functions, 35 virtualized network functions physical network function.
07:18:10 And then we have the network itself, but transport run one network.
07:18:17 And those are also then going via the internet.
07:18:22 Somehow, so that we have here, for example the controller of an IoT device is then connected via internet to the solution.
07:18:32 And what we are doing here is, and then the basis of these user equipment down there.
07:18:44 So this is what we are going to set up, and if you go now to the next slide. Before we go to the next slide, Martin I'm sorry to interrupt but I just want to do.
07:18:53 The link to L is right here.
07:18:58 This is also, and here this is this is from an organic perspective, this is also an Orion testing and Integration Center.
07:19:07 It's also inter worked with vendor labs. This is inter worked with Nokia's 5g lab in Murray Hill.
07:19:17 That's one example. It's also inter worked with at&t Labs in Bedminster and Middletown, New Jersey. Last but not least, this is one of two recently designated FCC innovation zones, which means we have a perpetual experimental license from the FCC.
07:19:39 So now we need to translate this layer into some some architecture, maybe, and this is just the way on how to do it so you see these artifacts from the previous slide here.
07:19:55 Again, a little bit differently modified and the flow between you ease and the UE control over around network and the core network. And if you know, go to the next slide.
07:20:15 Today, you see the old on it, so it fits very well.
07:20:33 We have small, we use artifacts, and also, on se artifacts in artifacts in order to construct the service management at orchestration layer and below then we have all the different components from from vendors and but also emulators.
07:20:40 We will have some measurement capabilities in order to really measure the end to end flows over five g core network.
07:20:53 And, yeah, and in order to really validate that we have a complete run and all the experiments which the first responders requirement.
07:21:07 And the next slide is just, again, I mean, I always about that.
07:21:11 She's, you know, Bo one controller. This is where what aren calls it, it's based on SDN Cc Cc SDK.
07:21:22 Likewise, non real time Rick has also there's a lot of the work with non real time Rick and open source has taken place within all that, you know, and we'll be using several other own app components, as well as borrowing and importing from other open
07:21:44 source communities, you know as necessary.
07:21:48 If you look to the go back. Okay, sorry.
07:21:52 No problem.
07:21:53 So if you if you go back you see all the the owner terms right or as policy so
07:22:02 so anti message because of course the schema, and we integrate them currently between sem hour and the non new turmeric implementation by john Kynaston.
07:22:15 Okay.
07:22:16 And this is where it becomes real. You see, again, this this line between the US and, and to control and it goes away a small data center which we have established in this this lab to do the realization and and to do our experiment.
07:22:37 That's it.
07:22:39 Any questions, unless I just included some of the contributors who have
07:22:47 who have said Yes, go ahead. I'll put my name on the slide.
07:22:51 There are other contributors, whose names are not here, and that's just because they have to, you know, run it up.
07:23:09 these series of events.
07:23:11 So okay. That's it.
07:23:15 Comments, questions,
07:23:21 chocolate covered goodness.
07:23:24 And it's exciting stuff thanks for sharing.
07:23:32 The one thing that I would ask is, I think, Kathleen you and others are already working on this.
07:23:40 We want to make sure that this is a, a truly joint.
07:23:50 Goal Global Hawk plug for something I'm seeing from an organic perspective.
07:23:55 And when I say Park proof of concept I mean it's, there's a strong software focus here obviously. So we want to make sure that Linux Foundation and overran figure out how they want to word press releases blog postings, how we handle contributions from
07:24:19 own app member companies who may not be around member companies, and vice versa.
07:24:26 So all that stuff. And I think there's plenty of time to figure that out because we've got from now until the last week of November.
07:24:44 Yeah, hi, excuse me, yeah.
07:24:46 Excuse me, if this is kind of a naive question but just curious if you have a network configuration, that's only used in emergencies.
07:24:59 How, how do you verify that that configuration is, you know, operational on a continuous basis so when you, when you need to deploy it that it's actually functional.
07:25:16 In other words, if you go, you know, years without deploying it or months, or whatever.
07:25:23 How do you, how do you have confidence that it's going to actually work, when you need it.
07:25:29 That's That's an excellent question, and I would imagine that the public safety sector will want to do, non invasive non disruptive firebrands, which is actually what they do today with their iPods system.
07:25:46 And of course I pause is their integrated public alert and warning system.
07:25:51 Just to be sure absolutely sure that that will work when there's a disaster, they tested every six months. I'm sure they'll do something similar with the five g emergency network that has to be up and alive during a city scale disaster.
07:26:15 Okay, thank you
07:26:22 Kathleen where you're going to have something or no i. First of all, I just wanted to thank you Tracy and Mustaine to join our own app Jesse call.
07:26:33 Because I believe it's extremely important.
07:26:37 Collectively, to bring awareness about any button shared breakfast. In order to demo that both communities planning to do.
07:26:48 So, if there is any additional people requested to prepare this demo, let us know.
07:26:54 Because it's not only about making the marketing is when it's successful. It's also prepare correctly. These even in advance, and make sure that everything will be working fine, like any other club effects that you have been driving up to now.
07:27:17 On top of that, you're not going to have expressed, and I hope I'm speaking.
07:27:25 I'm speaking for the community here, based on the feedback that they've received for the unit going to express to establish kind of formal agreement between the community and the organ is see, because we see the value we see really the value of bringing
07:27:43 these two open source community together today you demonstrate.
07:27:49 Additional use games. While when we work together. So this discussion is still ongoing at the TASC.
07:27:58 But since you are part of the LNAC technical group. It's already extremely great to see that it's not only discussion in the arrows.
07:28:12 Business Strategy discussion. It's real. That's really the point I wanted to make it three these two open source community are working through that either.
07:28:21 We just need to formalize.
07:28:24 What we kick off a couple of weeks.
07:28:27 So, we skinny we try to to move forward in this direction, but definitely with all the positive feedback, I have received from PFC members, the door to a non Casey members about the importance of all the formal agreement.
07:28:45 I'm sure at one point it will happen.
07:28:48 And finally, as I said, we need to have Okay, I want to come back on your request to make the marketing to Can I give an action item to Kenny to sister back with Brandon, to see how we can make it happen.
07:29:06 So they are, they are pre, pre area of my discussion, one, we need to continue to establish this formal agreement between the two open source community for only, we need to understand the type of support.
07:29:22 The owner committee we need to provide although I understand a lot of companies are engaged to make it happen, and served, how we can use the marketing segue to demonstrate the great collaboration between these two open source going to move it makes sense.
07:29:40 Does it make sense. Casey and Muhsin.
07:29:43 Oh absolutely, I put in the chat window. The link to the around public facing website.
07:29:52 And if you look at news and blog postings. There's a drop down menu at the top, as well as the virtual exhibitions, you will see that there have been many examples of successful collaborations between own app members, using own app components, who also
07:30:18 to be around number of companies and showcasing Owens specifications.
07:30:21 With own app components.
07:30:24 So this has been going on on officially.
07:30:28 For the past two years.
07:30:31 Certainly 2019 and 2020.
07:30:44 You will see concrete examples. And yes, Catherine if it flies from the first two years. But it was all kind of unofficial and informal.
07:30:49 I think that's the first time something was discussed on this was maybe at the Beijing release
07:31:03 PDF and in Santa Clara may have been the first time that the not not this specific work that you shared today, but that the
07:31:16 that the the collaboration and the plug Fest, have been taking place, you're absolutely correct. And it was, I was there actually, and that's what we talked about ran information model, zero.
07:31:30 Now ran information model zero has now become a formal overran information model data model specification, quite a, quite a bit of that has recently been standardized by three GPPSA five.
07:31:49 So, there has been no collaboration.
07:31:55 For some time now.
07:31:57 Across the open source community, as well as over to the, to the standards development organization that Iran is very tightly aligned with 3g Pp 5g open industry standards and emerging of an industry standards.
07:32:20 So it's great, we've got a and Catherine I see you typing that out Brandon's on the call so just checking to verify that.
07:32:30 If Brandon heard all of this.
07:32:33 Yeah, I did. Thanks Kenny and Tracy Nice to meet you. Thank you for the great presentation.
07:32:39 I work to coordinate, oh nap, marketing, under the networking umbrella, and I'd be happy to work with you about doing some cross promotion of these words.
07:33:00 I'm here because the work that Catherine and own app is doing. I feel a little guilty because we've been borrowing it without proper attribution for too long. And I want to say thank you to Ben Chang who have been some was not here.
07:33:11 We happily borrowed his PNF plug and play.
07:33:16 Work products. I don't know if antimony is here we borrowed heavily from SDN CCC SDK. So, as I said I'm I feel a little guilty and I'm glad we're having this conversation.
07:33:34 I wouldn't. But, glad that we're, we're talking now sounds like this is a nice confluence of events and look forward to learning more and learn more about the demo, as well.
07:33:46 That's exciting.
07:33:48 Now this is this is super exciting stuff Tracy thank you and Martin Thank you so much.
07:33:55 We've got him in the minutes, I see a looks like a screenshot of the presentation.
07:34:05 If you could
07:34:09 upload a copy of the, of the presentation to the MIT, to the minutes crazy.
07:34:16 Yes. Would you.
07:34:17 Is it okay if you give me one more days so I can make sure I include as many contributors as possible. I know absolutely. So it'll be like before the end of this week.
07:34:27 Okay.
07:34:30 Thank you so much.
07:34:34 Thank you.
07:34:36 I'm sorry to repeat myself So two things, just wanted to seize the opportunity to highlight that Heidi JB have accepted to be your own up max representative.
07:34:48 So then definitely possible thank you so much to a need to step up on the whole it's not the only hold that honey is playing with this one is really important to continue to promote all the good work especially when we do these activities in collaboration
07:35:07 with all the open source and Tracy and Mustaine not totally sure to capture Carnegie Hall, we can help you to support you for the next block fest so feel free to let us know if additional support will be required.
07:35:26 I guess no new development is necessary at this stage.
07:35:34 So far so good you.
07:35:37 What do we plan on using the owner release, I stumble, but we'll do what we've been doing all along, which is, you know, bring our questions and and request to bench songs.
07:35:55 You know, we use case realization meeting I don't know why he calls it these days but will participate in that.
07:36:02 And if there's any other, you know, form of participation you need from us just let us know.
07:36:12 We, we will share the preview or we stumble, a roadmap, or I'm not totally sure I too much, your expectations too we probably need to review this item as well.
07:36:28 Okay.
07:36:30 So thank you again Tracy and Martin. Before we go to, to a review, because today the final at them for the another release and I'm already extremely excited about that because I can see the progress we have been made, but until that is your pool with
07:36:49 is done. I want to give the floor to the panel, which will be our next topic, and then we shift to the owners in the review. So, Ethan. Are you with us, because we wanted to ensure that the owner committee is aware that the, the luck will be shut down
07:37:10 by the end of this week, he can I give you the floor.
07:37:16 Yeah. Can you hear me.
07:37:19 Hello, Yes yes we can.
07:37:21 Okay, awesome. Yeah, thank you for adding this one to the agenda. So, so as I said in the email last Friday, because I'm not sure if everyone is the communication.
07:37:35 So I would like to come in here and to communicate. So, so we have been discussing with the community in the, in the past month or two months ago, about the lab shut down.
07:37:48 So one of the reason behind the lab shut down is, we need to upgrade the some of the security patches that have been put upon delays for quite some time.
07:37:59 And as we have been relying on the weavers won't go into security.
07:38:03 So we have to shut down lab to take our blood.
07:38:08 So the tentative kind of fun.
07:38:15 Kind of a shutdown date is this Friday, April 30, but I do understand that the release kind of lead is not only the state for for this release is, I think it's accurate today.
07:38:27 Is it, is it true.
07:38:33 So
07:38:34 that's independent of the car pool with or not, that the major had them to the Yes.
07:38:39 independent of the USC approve it or not, that the major had them to the Yes. Okay. Okay. So, um, so taking that into consideration if we have to make a few days kind of
07:38:49 after April 30, we definitely can can try to cater that just, just to make sure that everyone is begging that bullet data that will be needed for future project activities.
07:39:01 So I'd like to kind of hear the kind of feedback from the community.
07:39:05 So, at the same time, so we have some additional kinds of variations from, from our entertain on those servers, so actually the last service that we, we put it in for the on that open lab is pretty old is, we already know this some of the servers is having
07:39:23 issues. so we cannot.
07:39:28 We cannot make sure that they, they are still functioning well and not not breaking anything.
07:39:35 And the other consideration is, we do have some other high priority projects.
07:39:41 Coming, coming forward.
07:39:43 So, one of the decision we are trying to see is whether the community is still relying on the that resources, because we have an intention to kind of that.
07:39:56 So today we can end the life. This lab.
07:39:59 But before we make any kind of formal decisions we are to kind of get the feedback from the community. So, which means if we kind of shut down the lab is not going to come in back.
07:40:14 Now we, we, so I think I can answer on via audio not community developers has been coming done since the first CDs, and we definitely need to keep going.
07:40:30 Otherwise we will not be able to continue growing and developing odd. So, Yeah, don't don't the commission forever.
07:40:42 Because then we will have a massive issue.
07:40:46 Yeah, but the real issue is, so the service layer enough, and the phone into we don't have any budget right now to upgrade those servers.
07:40:56 So, so this is a kind of a big concern for us.
07:41:01 So, at this point, we cannot guarantee that is going to be continued serving the community, but we can always discuss it, there's some other options.
07:41:14 Yeah, on the other hands. So, I think I heard that the the the foundation already kind of setting up some AWS, and kind of other resources for for the, for the community.
07:41:27 Is it true or. This is something that that not being quite sure
07:41:37 what I really the anything that comes out of that, that would be set up on AWS is actually all comes out is, is paid out of the own app budget.
07:41:56 So that's difference here so so for clarity.
07:42:01 You do not anticipate the, the lab, being brought up again after it shuts down, tomorrow.
07:42:18 So is there any way that we can kind of graduate kind of shifting.
07:42:24 Out of the, the, on the open left to the arrest from the community perspective,
07:42:33 we would need to go to the LFN governing board and ask for additional budget.
07:42:40 Okay, I see a little bit of notice the fun. We saw that you were stopping the lab to build to rebuild it in alignment with what you have been discussing with Morgan so for us is a big new,
07:42:59 we didn't, we were not aware until today, even your email was not expressing this.
07:43:06 So we would like to request some time because otherwise we can send on the on a committee and went right.
07:43:18 Yes. Yeah.
07:43:18 Yeah.
07:43:18 Yeah. So, notify us now.
07:43:21 But we would appreciate you give us a couple of weeks, to see what the other option for the development team is thinking okay yeah so yeah because that also came to my surprise when my, my team kind of fun doing some additional kind of funny variations,
07:43:38 and then found some quite some kind of issues on the kind of those servers.
07:43:42 They already kind of some of the issues in the lab. So that is why I'm coming here to, to discuss with the community to see if that is available in advanced solution.
07:43:57 And because we don't want to kind of service in the lab, but at the same time cannot kind of funky are indeed the service level.
07:44:05 Yeah, so yeah so forth for Canada for the end of life and decisions, yet we definitely can't wait until community to, to discuss in to kind of figure out if there's some other options to to fill it up.
07:44:20 But for for the economy, fully shut down for security upgrades, I think, as we can we get it earlier. So it could take a couple of weeks.
07:44:29 So, so. So I think we, we probably should expect at least for the month of May.
07:44:38 That is not is not going to come in back.
07:44:41 While we are doing some, some activities there.
07:44:47 So it's really not.
07:44:48 Should we not delay, considering the information that we share with us to minimize the impact of the team can report on Shana lead delay this for another month for give us the opportunity to discuss with directors and board.
07:45:09 Above all, we this concern, and potentially build a lambda.
07:45:27 recovered.
07:45:28 Okay, so let me take it to, to entertain to see if they can still extend their Weaver, because their security we've been having house training for quite some time, because we knew we announced the release is coming up.
07:45:46 I will be inspired by this month. Yeah, so let me take it back to our, our leptin to see if the security Weaver can be extended. And I'll get back to you.
07:45:57 Okay.
07:45:59 Well it's it's not weird just thank you for that.
07:46:03 It's not clear
07:46:10 whether.
07:46:15 Does the community have a choice in this at all.
07:46:20 Or is already something that has been decided.
07:46:25 The lab is being shut down.
07:46:29 It sounds like it's it's the former.
07:46:33 So I want to make sure that we've got the expectation set correctly that if there's no chance of the lab coming back after it shuts down.
07:46:44 We need to be very clear that that's what's happening.
07:46:50 If it's a case of, well, maybe it can come back up maybe it can't, that's a different story but, but, Japan, if you know right now that the lab will not come back.
07:47:05 We need to know that.
07:47:08 Yeah, so from the, from the feedback from 1413 is so they're telling me that this lab is not coming back after kind of extensive evaluation.
07:47:20 So, I think that that is probably the kind of a kind of best answer that I have for now.
07:47:30 And I just want to make sure that you understand the situation here. We've been we've been told for, you know, about two months, roughly, that the, the lab was being, you know, modified, and that you are timing the shut down to avoid you know interfering
07:47:50 with own app. Release activity.
07:47:56 And now you come here and you say, Oh no, we're not modifying it. We're shutting it down and furthermore we're doing it tomorrow.
07:48:08 You know that that's that's a pretty difficult thing for you know this community to deal with it seems like Intel, could you know consider what you know the story that they've been telling you know nap community for some time.
07:48:30 And, you know, give us an opportunity to respond to, you know, this new set of circumstances, not instead of, you know, pulling the plug in 24 hours.
07:48:45 Yeah, as I said earlier, yet so we're not going to put the product in 24 hours.
07:48:51 Yeah, because we we we've learned that the release is today so I think this for for kind of for shut down for the because the first step that we're going to do is to do the kind of security patching that we intended to do.
07:49:07 Right. And that will kind of previously was was kind of scheduled for this Friday to shut down lab, and then do the kind of our security patching yeah and then some.
07:49:17 So, even for security patches. So, as I said earlier, we probably can kind of kind of fun extended for a couple of days. I can talk to the team to just make sure that all the data is affected up yeah and then some.
07:49:30 So, this is the kind of first, first part right in. In this is what I think Bob have been but Monkman have been kind of communicating to to you and community.
07:49:42 Right.
07:49:43 So the second thing that I learned, just a couple of days ago, when I kind of have kind of extensive discussion with our internal teams, after they do some additional kind of evaluation, that they found the that the servers in the lab is videos in sa.
07:50:01 Start with servicing some some issues.
07:50:03 So, so, and then I talked to our management and we we cannot find any budget to upgrade those servers.
07:50:12 So in that case, we which means we cannot guarantee the service level right, whenever it breaks down and then we also don't have the resources to quickly come fix it.
07:50:22 And this could eventually kind of impact the community.
07:50:27 So that is why I'm coming here trying to see if this is a valuable way right for into to kind of fall completely shut down the, the lab. After everything is is kind of transition smoothly.
07:50:42 So I think, I think a better plan would be what I would ask is for you to go back to your management and and explain to them that this is, you know, a you know a substantial impact on on own app, and to ask your management to propose the plan.
07:51:08 You know, for shut down that would, you know, give us three six months to prepare for that, you know, instead of saying, Hey, sorry guys we're going to shut this down in a couple of days.
07:51:23 It's kind of outrageous, frankly, yeah, right. Yeah, so, yeah, yeah, I can further understand that. Yeah, it does make sense yeah so I can take this feedback to management to see if we can, we can work through that.
07:51:39 But in the meantime, I think, for long runs.
07:51:43 I think definitely, we should consider shifting away from the interest on that left.
07:51:50 Because, as I said earlier, so we don't have any budget right now right to upgrade the server to make sure it, it is, it's going to kind of sets us free service in the community going forward.
07:52:01 Yeah.
07:52:04 So I, I heard Bob's name come up through.
07:52:09 And you indicated that he's been in communication with somebody at the LS about this.
07:52:20 So, Bachman commands, his, his phone.
07:52:33 Okay guys, I think we can follow up with
07:52:48 I'll follow up with Brandon and I can probably take that as a action point. Yep. Yep.
07:52:57 That from my perspective what I'm understanding is that for security patching temporary shutdown is okay. At this time, but what we are, what the elephant that as a community we are requesting that Intel is that after temporary temporary security fixes
07:53:27 Yes, Kenny and I was already writing back until I have a chance to talk to you since you are first of all, not to see and you can probably bring additional input to your management because you know how intensively this lab is you and I have spoken about
07:53:45 the concern. So, if there is two ways. If the budget, the security patching takes one month. We have also, we are so in trouble right because I believe we might have a kind of honorary maintenance, which will happen in the coming weeks so we need to have
07:54:04 a better understanding about how much time it would take to do the security patching because of her for me bank that might take hold, hold me Mom, which probably also a major impact for us.
07:54:22 Now I mean I need to.
07:54:26 And I do not have a plan is taking care of this from our side.
07:54:44 Not all, but what I, what I'm understanding is I think I have a feeling that Intel will be very concerned about exposing infrastructure with some security vulnerabilities so I don't know how much choice we have, as I mentioned, maybe we can, we can have
07:54:48 a few days here and there to get the waiver, that's my, my experience in general not specific to overlap.
07:55:05 So I think that is fair I think what what the request is that, that we can keep it for a while, until some alternate elements can be made. And that could be anywhere between three to six months.
07:55:18 So, the ask is to basically, maybe a band myself can go to management and see that that we can be taught upgrading it to new servers, whether we can keep the same servers for a while.
07:55:32 And, and if so, what are the challenges. And so, so maybe, maybe you can let that maybe want to ask one thing which we can have internal discussion with the team on consultants to keep it for some, some time.
07:55:47 Yeah, sure.
07:55:49 Yeah.
07:55:55 So, So, in the meanwhile, so come into the western you had mentioned earlier. So, is the community okay for us to kind of to address the security patching, and then I'll try to confirm with a commitment with our left and right, to see if actually kind
07:56:14 of fun.
07:56:16 A couple of weeks is needed, or it can be kind of, you can be.
07:56:20 You can be sure that right.
07:56:22 Yeah, I mean before doing anything let us know because if you take six week.
07:56:29 We need to discuss with you for another plan. Right, so I would say come back to us with the option because we cannot put on all the labs for six weeks.
07:56:41 I'm not talking about six weeks, I think I'm kind of maximum like a four weeks.
07:56:46 But if everything is successful, actually, you will just take a couple of days or even just a week about because we are not sure what will break when we're doing the security badge and so that is reasons we are trying to kind of, to, to get some the way
07:57:05 just to make sure that we can kind of commit talk to our promise.
07:57:12 Okay, so there's there's.
07:57:15 That would be great. Thank you.
07:57:17 Okay.
07:57:33 Okay.
07:57:38 Yeah.
07:57:38 All right.
07:57:40 Nevertheless, I want to keep the, thank you everyone for joining us and to inform us about what we have in front of us and Cindy Thanks in advance for all the support that you can bring here, because we need it.
07:57:57 So I'm still quite excited to understand if we will sign off for today.
07:58:04 I will give the flow in a couple of minutes to David and also the integration team.
07:58:11 Om and documentation teams who are the key stakeholders from a candidate perspective.
07:58:19 But first of all I want to thank the community, the PTA and the subcommittee, the committee or the contributors, everybody the TSP the release manager, the elect, everybody.
07:58:32 Because again, we have been trying a new way of working together, we have, we are still learning and we still want to improve offense, but we we have accomplished significant new requirement 16 requirement in total, excluding the best practice and the
07:58:56 verbal requirement is the contain of Honolulu, looking at all the task, which was related to se do for the project, and also the requirement owner, everything has been completed.
07:59:11 For the testing of the requirements. The requirement owner, give us their go.
07:59:17 Nevertheless, we need to look now about the feedback from the integration team.
07:59:24 The Wizard from last night was extremely well to the a little bit less, but let's look at the detail with the with and Morgan now, but the kudos to all of you.
07:59:37 Please cascade this kudos to your company, your team members, because again it's not a great accomplishment.
07:59:48 and we have a lot of great items to share, and to provide gate via our marketing channels. Thank you.
08:00:02 Right. Thanks Catherine.
08:00:09 We don't see anything okay.
08:00:14 Maybe he lost your page, because I think this is not the latest version, I believe.
08:00:30 Okay.
08:00:34 So, as, as Catherine mentioned we made a lot of progress.
08:00:41 The.
08:00:44 We completed 15 requirements for the release. And
08:00:52 let's see, this is not the page I wanted.
08:00:58 Sorry.
08:01:17 You can stop it.
08:01:24 Yeah, seem to have that problem that links today.
08:01:38 All right, I'm just sorry there that.
08:01:43 Yes.
08:01:43 did you want to say something Catherine.
08:01:46 No one wants to go back to the city faster but you have it, that's fine.
08:01:50 Okay.
08:01:52 All right. So looking at the requirements.
08:02:01 We completed 15 requirements, not including the global requirements and best practices.
08:02:15 The.
08:02:18 See, milestone status.
08:02:43 Things seem to be loading slowly today.
08:02:51 Have it over here.
08:03:01 Wow just done like you know. Yeah, when, when things go wrong things go wrong, I guess.
08:03:12 So,
08:03:12 yeah, we can see if, see if you have a better connection me.
08:03:21 Okay.
08:03:21 So we were going to milestone status. Yeah,
08:03:37 Let's be an issue. Spinning, spinning very good. There we go. Alright.
08:03:40 So, you can see that we've scroll down a little bit. We've completed
08:03:48 all of the
08:03:52 requirements for the release requirements for the projects we've completed all of the requirements for the.
08:04:01 The requirement owners.
08:04:05 Anyway, if you want to flip back to the TASC page, Kenny
08:04:20 has been.
08:04:14 Sure.
08:04:21 Yeah, this is actually the page I was looking for earlier. And so you can see that we've completed testing for this set of requirements.
08:04:37 Back to the TASC page.
08:04:44 Yeah, go ahead.
08:04:47 And the daily.
08:04:57 Hello live
08:05:05 on the bottom.
08:05:09 Okay, so as as Catherine mentioned, this is declined a little bit from yesterday.
08:05:18 Can you do you want to pull up the previous.
08:05:28 So, At this point I'd like to invite more again, just to give us his take on x testing status.
08:05:43 This Morgan available.
08:05:48 If I prepare a slide actors to give a status on the rational, of those tuition.
08:06:09 Okay, sure. Let me give you a mission.
08:06:00 There you go.
08:06:05 Yeah.
08:06:05 Okay, so I just didn't. It's also it's not for the marketing you can see that they use a very nice style so just to give to the TC summer information on the last results on the different activities, and the integration responsibility.
08:06:24 So, first of all, I just added look at all the local versions, so that's something I'm used to track on the DVD of the day, you can see what has changed what has not changed in Honolulu so most of the components out there was a new component with CPS.
08:06:43 We have some components that even some are and maintain but some are not met mention as and maintain but there's not lots of activities, but I think we are we are good with the RC too so I finalize all the jar with readiness of some of the Docker.
08:07:00 And here we are for grad so it's a timeline view of the different results. So you have cat frog grass corresponding to the four categories so infrastructure categories.
08:07:17 So check, an app and check end to end test and security. I put a line not to to visualize what was the target at the beginning when we started.
08:07:24 So for the infra. It's about 55%, because there's one test that is expected to fail to check on the end to end is 100%, and for security just to be 70, it's not 60 because we remove.
08:07:40 Recently, one of the tests that was not relevant anymore so that twice as we have one test less of a critique decreases.
08:07:48 So, category by category in fries really, As a result from the installation. So we can see that yesterday and the day before everything was pretty clear that we are the rich are prettier.
08:08:04 Today we have a slow decrease Oh, to test darling kids was it Why is that why go to 25 but that's because of the to test Allen kid, and if you look at the Oh, It's a state reset that is not fully completed so it's not a big arrow so I think we need to
08:08:18 quit it okay but yeah I'll check. Sam we were rich several times to Pretoria.
08:08:25 Today, we didn't reach it because we got all that is a known as that happens time that is referencing a JIRA. It's used to race condition on the SF DC at startup.
08:08:38 So, it happens at sometimes we go too fast for these DC and it did not finish an operation before starting a new one, then we got an error so that's what happened this morning.
08:08:48 For the basic onboarding but this test is replayed in the smoke and it was okay so so it's not problem I think we are all killed so from the perspective, from the end to end test, we got very good results yesterday and the day before as well.
08:09:03 Today same some rest condition because probably the installation was not completed as fast as a baby for.
08:09:10 And we have one arrow which remain which we, which is a basic their macro so it's a simple test using the macro the payment to to start.
08:09:20 This one is working finding good in, it's not working in only your master but done mention that there is a patch to fix that and there is a workaround so it's not blocking it's possible to to to to to, if we documented It's okay.
08:09:35 Until security. With the introduction of the waiver as a second waiver, we can see that we are even overalls as a crew to expect this morning, which also consequences of the fact that sometimes didn't work as it didn't work does not trigger some tests
08:09:51 that did not trigger the creation of birds that would have been in a limited listener. So that's why it's over, but I think we also met a quitter.
08:10:01 So from a daily perspective, I think we met the criteria. We got very good days of course I would like to have one seven days to have good Barry to to to to say that it's really stabilized but it looks relatively well civilized.
08:10:20 Father, smoke test.
08:10:23 As I indicated, we have one problem with one use case, but it's understood on the root cause is found there is a batch and gate. There is a workaround possible so we need to document it.
08:10:38 And we have also additional budget is suggested is based on the time of extension which is not very good because when you start to just extend timer. When you have lots of time as well you know, we need to find a better solution and I think when teams
08:10:55 suggest something more robust than just extending the timer, but it's fine, I would.
08:11:00 So from an integration perspective, I think it's green because it's either a meeting the criteria, or the issue unknown and can be documented. So we have one issue that can will be will be fixed very soon was a workaround.
08:11:16 And once it would be much but it's not blocking.
08:11:21 And one, two issues, also known issues but there have been announced by as one sec, that they will not fix them for for new rules so they've been moved to stumble.
08:11:31 That's a race condition in sec, and timeout.
08:11:36 Unexpected timeout occurring in so when we are especially when we're running tests with BPM and sometimes there's so request, keep on responding in progress I'm never finished.
08:11:48 So that's not an issue that already existed in Gideon, but we added more frequently in order, so there is some life regression in this area, but globally from a daily perspective, I think we okay on screen night for that.
08:12:05 Now we speak about the residency testing so we did some tests so in England, we didn't make any tests on in this area in the El Alto info, as well, in Atlanta, we used to have some light test which consists in restarting just a border.
08:12:23 And it was a doing a test restarted but checks that it's still working after the battery staff.
08:12:29 JD here I was on the, on the journey for the backup and restore west to to do something different. So to to crush Kubernetes, compute, or control nada.
08:12:41 So as you know, an app is deployed in a set of communities compute nodes. So, our 160 birds are distributed on different community sports that can be bare metal or innovation and communities control is also split into three nodes for ha mode.
08:13:01 So the idea was what happens if we stopped.
08:13:05 If we crash if we shut down, one of the control node, one of the computer. So that was our goal for us as part of integration work.
08:13:14 So we, we did that so far.
08:13:17 First of all, for the.
08:13:20 Yeah, well I put just the link if you want to have more details, etc. You will see several jobs associated.
08:13:26 What we can see that for the workers death as a system does not automatically survival so if you crush, one of the worker nada.
08:13:37 The some components will not restart on BET, there is a way to make it work.
08:13:56 The fact that there's a component, do not restart is truly because they are blocked by some in it stuff that it itself blocked because some of the components that are in the worked out that we destroy intimidating mode on, do not restart automatic.
08:14:04 As he spots as a state full corresponding to the state for setup.
08:14:08 So there is a way to force the deletion. And when you force these deletion manually, then it will restart this path. And usually, most of the components we stopped properly.
08:14:19 So we did several tests, it's relatively. It's not easy to reproduce because here we are speaking of test. When we for example we shut down a computer.
08:14:30 We should done about the five to 15 birds on simultaneously.
08:14:36 Okay, and the buds are never distribute the same way. So, from one test to another. You can be sure that you will affect exactly the same way so you cannot be fully deterministic in the, in the walker you destroy you may have some things that are more
08:14:53 straightforward set. So the fact the state full path of absolute was a database so it can have more impact. So it's very hard to to reproduce. But anyway, we have, we have done it several times in some some labs on the neurons labs, we confirm that with
08:15:11 We confirm that with this procedure of forcing the state for said deletion. Then, most of the time.
08:15:27 All the put out back on the test, are back again. So I would say that the premises, is, is almost done. As issue is mark on upside, because we are not fully, we have state will set that do not restart.
08:15:35 shoe is mine on upside, because we are not fully. We are state will set that do not restart. So if I stumble we need probably to address that and it's important to continue on this journey. But it means concretely, that today we have at least a procedure to explain that we can
08:15:45 explain that we can survive a crash of a node on even crash of two nodes depending on the number of worker you other, and for the controller path. I made it so test on that, and then it was transparent the HMO was was okay.
08:15:59 So it was, it was, I was able to keep testing, without was only two nodes and restarted with up with a problem.
08:16:11 But for that distance to documentation to be done and to be completed so there is a three page. I need to finalize it I was very busy with the test itself but it's my to do list to complete the page.
08:16:22 In order to indicate this positive.
08:16:26 So it means that our residency tests from integration perspective, I think we are good on.
08:16:32 It does not mean that it's perfect. It means that on up to date on it was a case before it never worked to be an ester. So when you have one of your workers that was crushed.
08:16:48 For all the way or on your resume if you have VMware West corrected or whatever, then wizard, this procedure to force the deployment of risk out of the stage was set it was not working on, even if you do that, it method that what we experience in the
08:17:02 The last test, I do it twice in a row so I destroy a walker. Then I saw that everything was evicted properly.
08:17:15 summer internship at some components may have some issue to to restart, but most of them as I say, are walking back.
08:17:15 For example just to illustrate that.
08:17:21 And I weighed that everything was up again. Then I destroyed another Walker, and it was bad luck but residency was on the first one on one, I'm the first one it restarted properly without an issue and the second one, then I got an invitation.
08:17:37 Was that an issue and the second one, then I got an invitation. But so that will create gr and see if we can improve. We have also some exception in particular case that we need to address.
08:17:51 But I would say that it's almost under control at this can be documented it's not perfect, it's not fully automated as well.
08:17:56 We could imagine to have a real framework, doing that, we don't have today but I think we are at least better than we used to have on most of the components that are provided by web so all the databases that are working well for for for all these operations.
08:18:12 Okay, so that's for the resiliency testing. So, when the test when we destroy things. Now let's move to the stability test down. So for the stability test we perform three tests now.
08:18:25 It was done on the weekend.
08:18:45 From the 19th of April, so it was not the last weekend. As you know, several touches have been metadata we can bet. We need to have a 72 hour or so.
08:18:44 We wanted to have reserved. We had to start a little bit earlier, we made three tests, one to validate stability of modal mobilization parcel densities a stimulating HTC mainly on to test regarding instrumentation, one very simple when we run continuously
08:18:59 one test on one with pilot in some tissue.
08:19:02 So we did the pilot instant session before the simple one of what's maybe a mistake but I was very confident at this stage.
08:19:13 So for this as the sea stability test. The test consistent five file and boring during 72 hours. So it means that we have five okay running the onboarding test, and they're simultaneously calling zS DC.
08:19:30 When all the five are completed, we restart a bunch of fiber, so we have series of five tests over 72 hours.
08:19:38 So what we can see here in the. So I give lots of details in the GR Suchet was a test. But it's interesting for us is to notice that the duration of the onboarding is increasing, almost linearly with the time so tomorrow test, the market models we have,
08:19:57 you know, not enough time it will take to to execute the test.
08:20:01 And we can see that the success rate is decreasing progressively, so it's okay for for why then it's decreasing was a big step. At about 600 models.
08:20:15 And then we have when we have a 20 minutes 20% and only one test was okay, among the five done in parallel.
08:20:25 And the graph with a blue and the red dot show read that show that it was a failure. blue it's the best so we can see that we have two types of era we have some time out and we have some 500 corresponding Tuesday are at the bottom of the graph.
08:20:41 So we have always some era.
08:20:44 We don't have.
08:20:46 When I was 12 when everything is perfect, but we can see that the success rate is declining relatively slowly.
08:20:54 And to be honest 500 models in sec is already a big things I don't think we have 500 models of CNN cnn today twin brother.
08:21:03 So I think it's more dimension in tests and stability test and we can see that the beginning it's answering correctly. And at the end of the test case stuff processes or yellow face DC
08:21:16 connected to the platform and was able to detect some exceptions that were, he was not able to have with functional tests so he came back to the tweet stream and they will address that.
08:21:28 It's linked to access to the database as they can, and data management layer so that the middleware between sec and the database set was answering 500 after why.
08:21:43 So the conclusion for that is that we can send a documentation that you can reasonably onboard up to 500 models.
08:21:53 Success sites will be above 80%.
08:21:56 If you have lots of lots of models, then we have constraint that will try to improve that for next week's.
08:22:06 So that's what I did. Okay, so I think it's okay from from my perspective.
08:22:12 So I have also also graph so we can see that we did not detect during this test bigger issues right in memory. In Gideon were able to to visualize, so it was an instant session, a Memory of Light memory leak at CMC liver which was due to the Miss configuration.
08:22:30 But he can summarize in pulling was a Java so it takes everything we give to it. And we can see that
08:22:40 there's a limit is very isolated, we don't, we didn't see evidence of big memory of the other 712.
08:22:49 Now the simple stability and sensation. So, here we run a basic man test continuously so we just wouldn't one test, we wait for the test to be finished and then restart it.
08:23:00 And we replace this during 72 hours.
08:23:05 So as I said this test has been done after the test of testing that I will detail later. So it can be also exploration so what we, what we have.
08:23:16 We have a good ratio that the beginning.
08:23:33 Then we have the crash, and due to some Maya DBL but on the one of the note is in a recovery model, then it's rock, again, we some Ella anywhere but it seems to work again.
08:23:35 And then it crashed.
08:23:37 And it has crashed for very long time, and Maya WKR never recovered as you can see a put into stages of a Mayo, getting our four days a good thousands of free trial on the restart but it never comes to contacts.
08:23:55 So, there were also more than we used to have, in a in good enough, especially more timeouts, I think, and we can see that they were timeout was unexpected timeout on this was detected before we perform his Tesla, we saw it in a daily on gates on one
08:24:14 report it to a source of difficulty is that so is our country door.
08:24:19 And behind there are lots of possibility explain that.
08:24:22 So it's not necessary so it can be lots of things, but we have more unexpected timeout, the meanings that is so keeps on answering that the work is in progress, but it never finished.
08:24:35 So on our side we put a timeout of 30 minutes for the test and when we say that we don't have any answer. After 30 minutes. We stopped the test and we tried new one for another things that we already knew or so from good because it was already a JIRA
08:24:50 that were reported in Julian, is that there is
08:24:55 lots of difference between the same test for the duration. So just as a test on the same exactly the same, same code is executed.
08:25:03 And you can see that some text, takes about 500 seconds. when the same test will take more than 2006.
08:25:10 So it really depends on the way.
08:25:21 It was answering we are using basic VM which is using Alex FBPMM, so we are doing lots of requests while doing a request to create a se to create a VNF to create the visas module to delay the module to delay the VNFD Let's service.
08:25:29 So that's why we have also this kind of flights corresponding to, to the Delta time that is taken at every step, and we can see that if we have the raw scenario in each step, then we can have this very long duration for the sentence.
08:25:43 So, that's for this test, so it was a bit frustrating because after the crash of the Maya we didn't get lots of details here. And it was recovering the first time but it never.
08:26:00 We can see that from this.
08:26:03 minutes of from the figures of as a test cluster i is fine. I mean, we are requesting so much memory, and so much CPU, that you can see that it's CPUs and most ridiculous and memories on the lesson, 20%, or I say that didn't detect evidence of memory
08:26:27 leaks or for 72 hours, and 472 hours.
08:26:33 Eight hours. That's a graph from the full weekend. We can see that there was not something increasing alert.
08:26:41 We have some changes as he has density Well, when a restart but I see it's okay is yes so I think once Maya was out, we didn't have any traffic anymore because it was failing before castle is very high, but look through reasonably stable.
08:26:59 The only thing we can see so Maya the big area. So, the mechanics to clean up the memory is working we can see, but we can also see that we are exceeding naturally meet better request that we set, and when we reset then we start getting some some tuna.
08:27:18 But something was wrong with a mayor w getting audio during the test time, and was not able to recover. So I just made a table to compare on Hulu and get him, because it was exactly the same test we did.
08:27:31 So, including, we didn't have this crash of the GB. So we used to have about 76% of our success. And if we remove some some time window when we have failure but we don't face a failure in this case it was the as the emcee, the memory issue we need to
08:27:49 restart so sometimes it's happened during the night.
08:27:52 sometimes it's happened during the night. So, we removed and and we were reaching 86% we can see that today and it's much less than that.
08:27:59 But maybe bad luck with an idea who they are because father. Pfeiffer, I didn't put the other five in parallel, good.
08:28:17 We, but we didn't get any Maria DB Kelly is so maybe it was back like is a lot less expensive. So for the competition for that, it's it's a bit annoying meant to be able to have at least the same level of 10 billion.
08:28:26 There is currently discussion of a discussion during the PDL meeting about moving back to one not to apply walk around that will help, but then we look, we left the replication.
08:28:37 So it's.
08:28:40 There is no easy answer it okay so so so the balancing session is here so we can see that we were able to run
08:28:49 without Jeremiah the baggage issue. The first issue was as DNC that work at a restart. The second issue was an inpatient psych nursing to deal with an app.
08:28:58 So we have success rate of 65% here with five parallel testing. So it makes me think that if we will start with one one test I'm not five in parallel, probably could have reached the figure from good in modules in fact that we have more unexpected time.
08:29:20 time. Okay, so yeah that's to compare, we were able to do a $10 instrumentation going on, and the results were not so bad. So I think, compared to good in, we test was done was modeled on we have not a big crash, of course, and we have a bounty less.
08:29:40 The test was done was modeled on we have not a big crash, of course, and we have a bounty less. So timeout, then we experimented in Canada.
08:29:43 So the conclusion here is that.
08:29:47 Yeah, I would have liked to have something better for the stability testing, but it may be due to a new and I had a big area under training is not easy.
08:30:02 So the functionality is okay the residential care stability is not really okay but as a problem seems to be a DB component problem. It may be very long and complex to troubleshoot and you need to have this key people to do that typically is a question
08:30:17 Should we move. Remove the replication to add something more consistent, but then we will lose reputation. So I cannot open question.
08:30:29 I don't have the answer. And I'm pretty sure that it could take quite a long time to discuss.
08:30:34 And I'm pretty sure that it could take quite a long time to discuss. So I think it's not fully blocking up fully.
08:30:45 I will reinstall no no Lulu weekly tonight, and they will relaunch stability test.
08:30:59 think today, it's really blocking. And it's up to the TC to decide what they want to do with it.
08:31:03 And that's it.
08:31:07 So we're gonna actually just try one.
08:31:09 So you can just run this command, you have one more a bit is, we have had a issue with Maria DB Calera which is what might be causing the time more so.
08:31:22 So we are investigating further and single instance of morality might solve the problem of parameters.
08:31:26 Because as we have observed in Honolulu time frame we have increased internationally so with the TV.
08:31:33 And that is what we found could be the difference but IMO, which is not found before.
08:31:39 So, the latest patch which is what is already there in om might be a solution but we are still investigating further. So that's, I hope you're also concluding that when you take this center set right, including the Maria DB changes.
08:31:56 I'm not an expert so I understand your point is that's what we're discussing PDL.
08:32:01 That's true that there are lots of work on the tuning of the DB. I didn't mention here but if you look at the big consumption of memory. Today, the main consumption is done by Dr.
08:32:14 Cassandra dB. So as she means that, and maintain most depreciated also get some of us are using gigabyte of memory and Maria dB of course it has more interaction with so ICV also is using a Maya DB so let's see things to to to to improve in this area,
08:32:35 but it's clearly not a simple subject, there's not one switch to fix everything moving to one single murder is a key, but then you lose a replication.
08:32:51 So, as I said I'm not an expert of now if I'm not able to to to say that there is a good answer to that.
08:32:53 But probably, it could light the process and could give better as if that's such a possibility.
08:33:09 Sure I'm in my only question was during shooting back to when you're talking about testing today right i mean when you take the today's latest patch will include everything, so that we can have it tough for the test.
08:33:15 Right.
08:33:15 Yeah, I think it's not yet but if we have it in master and if it smelters that's something that I could test.
08:33:22 I planned on the future so nothing, no no no, because we are speaking of the sign up, but this test and so I plan to integrate them in the wiki chain.
08:33:32 So they will be the replay every weekend. In order to give the figure, not on these are five, the last days before the release, but all over the development lifecycle.
08:33:45 So we should be able to detect if we have a drop in performance.
08:33:48 So I plan to do that with a sweet test so I need to to to tune the figures because, as I say, I was very optimistic I plan to do 10 in parallel, for instance session but I saw that it was too much and we have lots of timeout.
08:34:02 So I decided to move back to fiber.
08:34:18 This test will be integrated into chain, hopefully, in order to give it back to to own to have a better control of the stability and not on the simpler integration campaign. Two days before the trip.
08:34:21 Okay guys, those guys will work together and work on this. Let's hope we can get some solution.
08:34:27 Thank you.
08:34:28 So thank you so much for the detailed information about the testing.
08:34:37 I think you, you make it clear, some homework then that we will probably.
08:34:45 We will continue to investigate deep dive any concern that has been in non defy, and I believe the fight that we are already considering to reduce men and please make this initially sent.
08:35:02 It might be as you said just a tuning exist side so while we might have some session with additional expert from the phone company to support to. We just need to become organized and tactical to monitor on this table for concern.
08:35:22 Understand, we have lost in a couple of CSE members I'm not sure we have fill the form.
08:35:29 But before we probably need to move the sign of vote offline.
08:35:38 I really want to gather the information from the TSP, but we the PST we're still on this bridge, is there any concern question that they would like to request.
08:35:51 Before we before
08:35:54 we do still have quorum.
08:35:57 Okay.
08:36:00 So, in the test you have any question concern for voting and we will continue to collect the vote of the people who have been not able to stay.
08:36:14 Definitely Katrina, maybe you can explain. These are last situation but the pending batches at us to.
08:36:27 Sorry.
08:36:28 No, no.
08:36:30 So just to say that, yes, we had,
08:36:34 we, we have funds to actually notice it.
08:36:38 Sebastian from Bell Canada has found that we are, we wanted following exactly the recommendation from can wound up for my edgy began era configuration so we have pushed that into Master, and we met, maybe a little bit too fast because actually, he made
08:36:57 CDs, not walking when. So the recommendation of my configuration for kaminer is networking where we see yes. So now we have a badge that has been merged that as CBS is using a separate database, or accessible database actually owns a buddy where it is
08:37:18 using its own database we have change the configuration.
08:37:23 Just to say that we saw that means that we have, we have.
08:37:28 So that means that we have, we have. We have some scenes that teaser could actually be staggered, unfortunately, as we had an issue with Garrett, and next year, with all the Nokia.
08:37:41 That means that we are not able to perform tested today and yesterday so we are not able to measure everything into another room.
08:37:49 So that means that we are in the middle of merging everything into an odd. It can be thought of maintenance carries, or can be part of your fisheries, a tree depends on duties.
08:38:10 Thank you for the additional feedback.
08:38:33 So is, is your recommendation for the release to be ago
08:38:25 so on.
08:38:33 Oh my recommendation, sir.
08:38:31 So I, I have no recommendation, again, what they're trying to do an on premise to whoever wins that master version of Word that is walking always winner, which is not the case today but was it games when you branch to a newbie.
08:38:44 And that's I'm Senior, we have to be a beta. And so that means that at any Dima, we should be able to do to decide that this is a vision for x y z or he needs it.
08:38:57 So, so my opinion is that we could have center for two weeks ago when months ago, yesterday or tomorrow, may go to the Emmys said, it's really up to the TC to decide if the questions that we are be stupid enough feature have enough and so he
08:39:20 he stood up, he's not perfect, far from that, but maybe we can make it perfect next week but maybe next week we will be. We can also move in a week more into an inter inter at the point of time you have to decide when you have to do this enough and I
08:39:34 this is reserved yes yeah but it's not, it's not that bad. It's not that good. It's, it's okay. I would say to to to to to sign up for today or issue is a GC since we want to do or some senior a lot more stable and maybe we can return is all week but
08:39:52 maybe it will not be sufficient and, you know, I think that that the point of time we need to decide that it's sufficient enough to to move it.
08:40:04 And we let us let me know so that we should have a maintenance ladies are in less than a month, I believe, or in a moment, but this light what I would say to resolve the issue that we have seen today.
08:40:18 And what I'm seeing and talking, especially in double process, you know, very simple one, when we see that established enough.
08:40:27 We present and say these are just crazy little nowhere documents and data.
08:40:40 So my opinion is let's move on
08:40:38 here but I would follow your recommendations, which is to move up, because the indicator he sees by the project team the requirement theme or the feedback you provided, and also the feedback provided by Morgan.
08:40:51 If we consolidate everything together.
08:40:55 It's not as bad as it looks like it's the opposite we constantly improve our firm the stability we try to push to the next limit.
08:41:03 And that's where we need to investigate Furthermore, originally we were doing 72 stability test sequential mode. So I would suggest based on the existing users to pull back the free, the release, as it is that will be my recommendation to the TASC it's
08:41:21 up to them to see us or not. But then if we definitely follow your guidelines advice to give an analogy, men Anthony, where we we continue to bring expertise.
08:41:35 Like, Maria dB, or anything else to continue to improve yourself as well. So that would be my recommendation to go with the condition that we have in order, you need maintenance within the next month, we will address.
08:41:51 What we know already, and the concern when we do this thing.
08:41:59 And we still look on King.
08:42:01 Yes.
08:42:05 Good to see members of you any quick question or concern, based on the recommendation that I've tried to formulate based on the indicator provided by the different key stakeholders,
08:42:22 making the exception in the feedback that we didn't check on the documentation team would I make the assumption that former release notes perspective.
08:42:33 We are also good to go.
08:42:35 Yes, I think we have just some few ones which have not provided a release notes. Unfortunately the tickets, which was raised with already closed because they have not been done.
08:42:50 And I think AI has an $8 the map has still not cherry picked their release notes to the Honolulu branch. But this is also some something which has to be done.
08:43:03 Accordingly, but Thomas I think we'll have a look at.
08:43:06 Yeah, so it's a small gun and sort of answered we are okay but not perfect. Yeah, but that's fine.
08:43:18 Okay, so we've got on okay recommendation from integration we've got an OK recommendation from Om. We've got an okay recommendation from the doc team.
08:43:32 And I think Gorman owner, and project team as well, important to mention that.
08:43:40 Okay, what do we do, I'm dropping, dropping a vote in
08:43:49 the TSA approved sign off of Honolulu release with the public announcement to be made on May 11.
08:43:59 Oh, does it in chat.
08:44:40 Gonna give it 10 more seconds.
08:44:56 I'm going to go ahead and end the vote, we are still at quorum,
08:45:07 and it carries.
08:45:11 Congratulations.
08:45:16 Congratulations, team. And don't forget the know how many items that we were highlighting and tomorrow, please let us know.
08:45:25 We need to finalize the release notes before the marketing lunch. So I said the deep on the fifth of May.
08:45:34 And please difficult to work with the ppl to make it happen.
08:45:41 Thank you.
08:45:45 And kudos to everybody who gets