Jenkins Google Summer of Code Office Hours, 9 May 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: GSoC 2020: External Fingerprint Storage project kick-off

Description

Project page: https://www.jenkins.io/projects/gsoc/2020/projects/external-fingerprint-storage/ . Meeting notes: https://docs.google.com/document/d/1_0lH_s5NpV860NjLmZT8cKd26Z4GrtXpgkBydDt103M/edit#heading=h.cmwmgfq0y4gk

A

I'll welcome to the external fingerprint storage of making. This is one of our google Summer of Code projects, and we have some meat on the call. It is student and we also have a number of mentors and potential stakeholders, and my suggestion would be to have short introductions before we place it with the project's agenda and after that we can just discuss what's next project.

A

So meet, would you like to start.

B

Hi everybody, my name, is Smith I am the district student for this project, external fingerprint storage and it's great to meet you guys and.

C

Thank you shows next I'll go next. Some like surreal way: I work at swabbies, my first time being involved with google Summer of Code as well as mentoring, but I'm looking forward to I. Think it's gonna be really good experience and I particularly heard this project seems pretty well thought through and thought out to me.

C

Thank you. Andre.

D

Andre Falco I have been a long time, Jenkins user. This is my second time mentoring for google Summer of Code for for Jenkins last year it was you know, Kafka teachers for for Jenkins that I was a mentor and, given my background in managing systems that carry state I felt that I could potentially help out with this one. This year,.

A

I'm trying to make notes in peril, but yeah it's everyone gets welcome to it. In the Google. Look good grip there introduce myself. My name is a liquid nitrogen.

A

History stories for Jenkins, we have delivered some of the stories, but not all of them, and external fingerprint storage was one of the items which is true in Jenkins and as a former mayor of docker, traceability, plug-in and other plugins. We shaking the work with finger things I'm, really interested in finger. Prints, API and I would be interested in external storage so that we could have Monali gigs.

A

We would have more effective implementations for fingerprints, so I am happy to work with this project and we thanks a lot scimitar of your proposal and for all the pages you have already submitted, though the junk is called for this area in q2.

B

For hiding it written that.

A

Rick, would you like to introduce yourself.

E

haha Rick contributor of Jenkins community- this is my second year to participate in this talk.

E

I have proposal this year. It's custom case distribution, service.

E

Chicken storage I accidentally, it's a very interesting topic.

E

In my opinion, in my experience, especially in my team, we want to story store.

E

Many things to my external storage, so yeah.

D

E

A excellent project, thank you.

A

John so yeah welcome everyone and hopefully we will have mastic holders and contributors. Do you think the community bonding, because I believe for this project is really important for junkies administrators, so hopefully communications.

A

Okay, so I guess it's time to share a Google document and go through probably a bit boring agenda because the first meeting, so we need to stop communication channels. We need to discuss how we would work together, but it's really important to have for this discussion so that we can this project trying to expose them in the future.

B

A

Or if you prefer, I can share it on my screen.

B

Well, I'm fine either way so.

A

Since you're on the meeting, it may be more convenient if you share this thing,.

B

So I think I have to just give some system references.

B

Could you do I think I have to restart zoom to enable this preference so I can like I guess? Could you do it this time so I, just okay, fine yeah share my screen.

A

Okay, yeah, okay, by the way, it would be great if you grant right access to mentors, but we can edit that this google walk to show you. We are not interested to vandalize it.

A

So yeah, meanwhile, we have a number of topics and the first topic we have for the project.

A

So historically we use charts like guitar, etc jae-seok project and my question to participants is whether you concede it is comfortable or whether you would like to consider other channels.

A

A

B

May be hiding gradable, we will at least from me plus one.

A

So right now what we have mint in the project description. We have a link to this cloud native special interest group. Richard I. Do not think that it's best thing for this project because, firstly, all native seek is Dortmund and I have an action item to restore that I already have a male in my drafts, etc. So I will try to just start it today or tomorrow, but assuming that the disiike who gets recovered most likely there will be high traffic yeah.

A

What is a lot of traffic and it related to your project plus I added another topic to agenda I. Do not think that alternative seek is actually am a good place for pluggable storage, so my suggestion would be to introduce a separate little channel because we don't have special channel for Jenks core development, so my preference would be to creating channel specifically for the project I.

C

Think I mix the motions as well.

A

Andre, what's your preference.

D

Yeah I think we did that that last year said that I think the getter channel work well, but yet having it a separate channel. So that's not the regular noise with another interleaved channel would be my boat yeah.

A

Okay, so you'll need a separate channel. I can just create something like external fingerprint storage.

A

Don't make sense. Oh now it renders correctly. I was creating another channel. Today this layout was a mess.

A

External fingerprint storage, I.

A

Will just plate this one.

A

Just join this channel reckon to not postpone these action items. Okay, so we have a channel and we can replace it to the next topic. So sweet wanted to discuss, meeting frequency and timing, so common recommendation and most of all, is to take at least one meeting per week.

A

Recommendation is there to have two meetings, but we did some experiments. Do you think, previous years? So, for example, one meeting could be ensured so that we meet in a video on the one Spivak and do the second under one. We need to do knowledge transfer set session. So we have a lot of flexibility there and I'm prepared to follow the decision by others share your preferences.

B

That sounds great and I think we can start with something and then maybe find what suits better.

B

Yeah, maybe Meleager that sounds good like one video and one chart, or maybe two videos is also good, I, guess I think your experience with past will be some indication to. Maybe that is better. So.

A

Also options for clicky fine, so what we really need to discuss is what would be the meeting times, because currently we have four different time zones, so the week is based in China and based in Europe, Michael and Andre are based in a United States right, yeah California for me also you're. Even on the west coast, we have finding a good meeting. Slaughter is going to be a challenge and my suggestion would be to start dude or whatever to find such slot.

B

Okay, so I think we can have two to last for the first meeting and the next well.

A

You can just create a one meeting, because you specify slots, you do multi selection and then, if you have two slots where you have a lot of responses. Okay, these are two meetings. Okay,.

B

A

Why it's better to start practicing because, as you might have noticed, in managing community, we had created a lot of doodles or similar voting's and other services.

A

Historically, it works pretty well.

A

Okay, so the next item is so: you want to request yeah.

B

Because I have no.

A

Well, so, assuming that everything is fine, when would you expect this exams in July right.

B

So basically, what is happened in India is that Commission for universities has said that we recommend July as a good option, but it's up to the universities to decide what to do so.

B

So so that's what so that, even if they happen between the coding period, like I, can have a good advantage of starting early in the community bonding. And then you know it's just better to be safe than you.

C

A

They find especially if you have something worth now. We could easily start earlier. Yes,.

B

The onion yeah.

A

We still need to discuss the design and what we would like to do. I already have a lot of pate analysis in your proposal, so I do not think it will take long to discuss a bit and maybe we start I need to start creating tickets etc yeah. Maybe we could all register to do something in one week or so.

B

So yeah, so what I thought was that so basically, I was looking back at the proposal and I saw that mostly for the phase one. The major chunk is I is the changes we have to. We have to add the fingerprint interface to the to tank in school, so I design, for that would probably be a big challenge.

B

B

So what I was thinking was that maybe next week we can take for some discussions around that and maybe the week after that we could give a short our you know, starting or if we are still behind on the designs. Then we can start the next week. Something like that. So.

A

What's your variability.

B

A

The next week.

C

I'm available pretty flexibly next week, the week after that will be out like Wednesday Thursday Friday. It's like the days. Those are.

B

C

The 20th, through the 22nd but otherwise available and I'm, used to working with people, one a bunch of different time zones, so sure we can find a slot that I can attend.

A

D

Actually I should be available I'm on call for like my job. So you know if, if I get paged or something or I have a thing to deal with, it might not be able to make it. But you know, I can just plant plant for the best and I should be able to make morning time slots on the other west coast. I.

A

Think what about you would you like to participate in these discussions and, if so, what would be your preference.

E

I'm not quite sure about where timeslot person current the time should be open. For me,.

E

A

Yes, oh my keys. My next week is quite parked because we have to meet ups. We are trying to create everything for some hackathon meet me, so most likely so I will be quite busy on Monday. Tuesday I will be still able to find some time to review reviews and provide feedback, but starting from Wednesday I will live to make sure that another able for the meeting the earrings are quite tough but I'm ready to move for the meeting city needed, and here one of availability topics for me. So yeah.

D

A

25Th to 29th again now there will be your ex hackathon all. We can actually good to hasten that, for example, for Sony. If you would like to have more onboarding with a fingerprint I could race, etcetera, definitely improving for UI UX for fingerprints could qualify, but in general they will be able to administrative tasks and other stuff for me to be handled this week,.

A

Yeah it's similar, but in next week, should be fine.

A

We should also have a number for tasks which could be created so, as you said so meter, the first objective is to create a kind of API and the change, but the cpi-u also implies having a reference implementation. We don't commonly marsha in new ApS unless that was a reference implementation for them.

A

So it means that we can have two things in parallel and while Jenkins code itself is quite difficult to contribute to for your reference implementation and basically you can just start the local repository start, putting things there and we can work on getting him hosted. So creating a skeleton could be a good first task to table and you can basically work on that me easier because there is no commitment. There is no risk of breaking anything can be done. It's a depository.

B

B

If I, if I, am building a reference implementation, the changes I made in school, but so so would I be using a local. Basically, my changes in Jenkins code would be local or would I be doing that in the beta class. Well,.

A

You can, you would be still doing canonical base. You would create a pull request.

A

So yes, I think that we will like not much this pull request immediately, but we will look wait for some more deference implementation so that we can review EP is together right now we have a way to merge ETS early, because now we have better annotation for EPS way. We can say that the lis CPI is here, but we don't guarantee but value the compatibility for that.

A

So we can still move quickly, but so it's, unlike the Jenkins, caught right away, documentation.

D

We have to potentially break it up. Ass, like we would, it seems like well, will depend of like a reference implementation that and refine. It then probably have to break it up into multiple changes. Do you think, or one big request.

A

And so I, don't like a big pool request to be honest, so yeah there will be still like the initial big booty quest, which introduces but API and parallel. There would be reference implementation again. It might be a blob where we can still do some pull requests, etc. But once this initial implementation is merged, we can do it in action and device.

A

Does it make sense from your point of view, seen it yeah.

A

Yep, it's a plugin implementation.

B

So if they do start out with.

B

So if you do start out with the plug-in implementation, I think we will have to, because there were some points in the proposal that we had to discuss regarding the plugin implementation in community bonding, especially which, which platform of databases are we going to start it out on so.

A

Yeah so my suggestion, for that would be to first start with discussing with a synchro monster, so you have your proposal. What we could do. We could just copy this proposal to design dog and stop the reviewing page or just to keep working with the proposal, though my preference would be to create a new document and this document eventually it will become foundation for Jenkins enhancement proposal. You reference the law, but right now it can be just our boilerplate document. I wish we had review together, discuss on communication and yeah.

A

Let's try to are getting on major beats earlier.

A

So if you create this design dog, you can just send it for the views and maybe start reviewing the basically over the weekend next week, so that by the next meeting we have some up-to-the-minute results. Okay,.

A

So we are running out of time. The one question at all of you: do we want to place it for another five to ten minutes to discuss communication channels or do we want to take this meeting, tie sync mode.

C

A

C

Off for another meeting, but I can cut this I'll follow up and look at this at the meeting notes you see mm-hmm.

A

Okay, what's your weight, okay, yeah.

D

I could stick around for a few more days. Okay,.

A

Let's try to discuss these communication channels, it's good to meet everybody Thanks.

A

So the garden Zita. What was your questions made?.

B

So I mean I have to start with. So basically my question basically.

B

Do I need to start creating, maybe tickets or so what's the process of using you know. Basically till now I will be solid. Jira tickets, I've never created I, mean here.

A

So my recommendation how we usually approach that we create some ethics, for example, based on deliverables, so you have a sequence of deliverables in your project, let's say initially: API implementation, etc. You can use this scopes basic. Yes, you can create an epic of wish such a deliverable, and in this epoch you can create tickets.

A

So, in the case of this proposal we will have to use majira because for other projects many of them must wish to github wishes. But here we have gently score and Jenkins code has jiron well, we could potentially using github wishes, but it would be a bit complicated. So if everyone is finally that I would just suggest to start from JIRA.

A

What do you think.

D

Thumbs up for me.

A

So let's go get it then and yeah me too, you can already start creating tasks. Okay, even if we decide that we don't do something we can easily close to them, plus the most of tasks near proposal. We will have to implement them in some way.

D

A

So do we need to knowledge transfers for this project? So that's the question.

B

I, don't know I mean I would like your opinion on it. I contribute to the a little bit to the things. Only I'm, not completely boggled, by a fingerprint. The code base around fingerprints so like I, have some idea. I believe.

A

In a few weeks, you don't know if fingerprint it better than anyone, because for me, this knowledge is already a bit Ruston. We had another view during the application phase right.

B

Yeah? Yes, yes, yes,.

A

So we cannot have four molecules: the knowledge transfer.

B

So if you think that that is fine, maybe so I can like whenever I do get stuck, maybe we can iterate exact.

A

So if you don't need to schedule anything right now, let's proceed focus on design, and then we will see why we need to discuss your schedule inhalation answers. For example, it is highly likely that some of knowledge transfers will need to happen recently because he has more experience with databases for than me.

D

Was that initial knowledge, transfer and fingerprinting is that recorded by any chance? It.

A

Should be recorded to find recording, basically, it was a good day.

A

Okay, so I will find this knowledge, transfer and milk. It yeah shouldn't take long, but I will try to do it tonight or on the weekend.

A

B

How to break down the first task in we are discussion, so basically I, think design. Document creation is one thing that you're suggested and creating the epic is the other yeah.

A

Right, so let's try to do that. Well, if you see some obvious tasks right now, it's perfectly fine. So, for example, we need to create a drug TPA. We need to convert the existing system logic to this API extension.

A

We also need to check whether we have enough test coverage for fingerprints, el gusano, De Anza know so such tasks could be separated right now and yeah. Follow that one a reference implementation may be, it will require some additional discussions, but for playing skeleton. Definitely you can play the task.

A

A

So yeah I had a question about special interest group. So what I already communicated I will start a thread about recovering cognitive, sick.

A

About cloud native see what my vision for cognitive seek is a bit different, because historically it was really focused just a second, so it was focused on changing junk circuit action, pluggable storage, cognitive Jenkins, which basically boiled down to jinx Padron asked by bilities another things.

A

Configuration was quoted, praetor and the many of these changes were very low level and we have had difficulties with building a community in this thick mighty structures. We should is seek more to user specific use cases, for example, integration with was from C&C fa cai' system, especially monitors with monitoring tools like parameters, also to still keep working on jinkies, pronounce the patrollers to keep working on another areas, but with more focus on and users and for such changes.

A

Yes, we still need a cloud native storage, so this project, if it's this sequel, but yesterday, for example, we had a discussion- will be smart quite about creating music for core features and core functionality, because in the print storage a you don't have to be cognitive or for what it works. You don't have to be in the cloud to get benefit from this project.

A

So we were thinking about moving it to a separate sheet moving some projects from cloud, negative 6 and projects from platform seek and just creating a new C sub project for this core architecture, changes and core features. So it's yet to be discussed it's yet to be submitted, but there may be. We will try to create music around it and it will be great to know what is your opinion and what your opinions and rayon drink about whether it makes sense to proceed and such direction and, if so, how. We would like to much.

A

This project show special interest groups.

A

Yeah, it is like a lot of bureaucracy here, but sober otras is something I can tackle. What would be your vision and principle? Where would this project fit better? Oh.

A

We could we don't have to answer this question today, so we can really discuss it during community bondic but yeah. What would be my preference so that by the end of the community bonding, we have one presentation and whatever seek so that we start facilitating and feedback from wider community and just from the project team.

A

You can adjust the public meeting. There is no sig by this time, but let's say.

B

A

So yeah it gets a bit sit from me yep. If you have any other topics we you should discuss today what's discussed time.

B

So the public meeting during community bonding you mean the design documents that you will be creating get an opinion on. Yes,.

A

Oh just present the vision for the project.

A

Design doc, we can shade earlier so once you have an initial version, we can shade with the developer million tees so, but we don't need to wait for the meeting.

B

A

Basically, this meeting could be a forum for view the discussion, so we first a share, maybe accommodate some feedback. Invite those who are interested to join this project meeting- and maybe somebody will stay as a stakeholder in the project or just provide more feedback, show some weaknesses in the design which is really important.

A

Then we can proceed with that.

B

Project, if you do have a good design document in mind, that.

B

A

My recommendation would be to start from something like how jobs are written and I created. The number four chips before, to be honest, main keys, say: you've been creating jobs on Google Doc and then converting them to asciidoc, because Google Doc is much more convenient for discussions and initial feedback. So here, for example, you can find some dogs related to pluggable storage, for example, though this is sort of take open chips which actually active ones so, for example, external artefacts storage.

A

So this is the document which describes how it would look like what would be the API is you don't have spent so much time, my documenting that, because you can have a pull request and you can just reference this pull request until the installation, but some common sections like motivation like vision for your the implementation, it would be really important and then you can go to whatever low level in details, you're interested in so, for example, this job is quite top-level. The Objectivist actually suggest and justify the architecture decisions you are making.

A

So, for example, one important decision for you is a choice of the target database as a reference implementation.

A

Another one is just how to would integrate. Existing is called code base, so basically to these items would already create a good job. So here is another example which turn on block storage for pipeline, which is also quite easily turd in which you can rebuild. It includes some user cards etc, which one job two one zero, so yeah, it's external locks, storage.

A

Also, we have Egypt for configuration, storage, API. We have a for external Gilligan. Can the Jenkins core? So this one is in the pipeline this one in the core, so you can start from them yeah you don't have to create long ships. So, for example, this one is saturated me wrong because it suggests a core API change with a lot of potential consequences, but yep you can place it as you want, and you don't have to create all subjects like here. Instead of that, you can probably create some schemas or whatever.

A

So, whatever works for you, it's just an example. It's not a guideline. How you shouldn't commended.

A

So the main advantage of following jet jab format that we can be easily converted to a job at some point.

A

A

So Andre, which topics would you like to see addressed.

D

Yeah, the only thing that that's on my mind, I think that'll be tricky. Is you know what kind of ORM layer we're gonna choose if we decide to like use a relational database or if we're thinking about a different type of storage,.

D

But yeah that that's something to think about, especially when, as we write the Jeff, what I would encourage is starting out with just an in-memory kind of database if you're prototyping, because then we can rework that and plug it into the external storage post. Prototype.

A

Actually, in-memory database is a good point, so what we have now, so we have a system database to be refactored.

A

In-Memory database.

A

We will definitely need at least one physical database and they in your proposal so submit it was post, go straight.

B

A

Bill databases.

A

D

Yeah I just expect like um in the like, when you share the design back and get feedback, you're gonna probably get asked like okay. Why not Postgres.

A

Yep, so how we handle it in jerk and jab, we have freezing ink so, for example, yeah they're all with some concerns, because if it's in the architecture change, so people have different opinions here, for example, there are some concerns about how to log in how to mirror works, etc, and we try to address it in the reasoning search. So, for example, if you just have opinion what did the BC you you would like to use, you can put it like that same way, even take a look at concerns.

A

What would be the backward compatibility input? Hopefully we won't have backward compatibility issues in this project, but I would see I'm totally willing to make changes have to so yeah. You can follow this format, but the andre is right. There will be a question about traditional database, so if you the relational database, somebody will be wondering why it's not elasticsearch. If you do elasticsearch, somebody will go wondering why it's not MongoDB or Postgres, so we'll be able NT to have these questions then.

D

Stakeholders are yeah, I mean there could be some organizations that don't don't run. My oh there's a be some organizations that don't run a certain types, so they're gonna, be if you feedback like. Oh no I want you don't do this, but XYZ yeah.

A

But that's why we designed the project, because one of the objectives is to create API. We do not have to create all the implementations, but providing good API is something we need to do so, even if we don't do all these reference implementations keeping kind of mind, this use cases is good. For example, there are some key, for example, if you use elastic soldier, then you have eventual consistency by design and eventual consistency for fingerprints for logging.

A

It might be a significant concern if you use post base, you don't have eventual consistency, but then you have to figure out how to create data from that, because if you just put XML or JSON, it might be quite slow or if you use educational data base, then again you have to think how you put data there, because Postgres is good, but Patrice offers much more features for structured data than my skill. Let's see oh MongoDB Yates, just key value storage and you have to figure out everything on your own.

A

It's totally possible, but as long as we provide API which potentially unlock these cases, we are fine.

A

Yeah but EPA can be developed incrementally. So that's why, when we start working on pluggable storage, so just to share some history. Well, last iteration of talk about storage when we were cantabile, basically with Carlos Sanchez on creating car support managers and understories. Just think we need is to create beta annotations and support, because we realized that, even with our expertise, we cannot create good EPS on the first iteration.

A

So instead of creating API spending too much time, we introduced an engine which supports which allows are doing good things implemented and now with heavily useless engine, and we show some it. We don't expect you to create a good epa from the first attempt, so we will get there, hopefully during the project or maybe even after but yep, it's perfectly fine to approach iterative.

A

A

Okay, so anything else for today.

A

Great and if anybody starts rambling about APA but okay,.

A

A

Do you have enough informations made for the next steps.

B

Mm-Hmm yeah I think I can kiss I think I have a few action items. I'll start with a bear boon design document which will most likely be a strong reference from my proposal.

B

Let's see how I can try to ID reasoning for anything I do so using these points and then look at epics, maybe- and you know, try to break down actually. First.

D

Don't hesitate post any like questions at all on get our channel we'll start watching it and try my best to answer.

A

A

Just to the chart.

A

Okay, so yeah. Let's say you saw this room and I right now with one beer detective, but yeah.

A

So yeah, let's keep working on this project and again welcome Samhita and yeah thanks a lot all the oil filter on this project, because it's really important and it's great to have any imaging this community. Thank.

B

You mentoring me so having this great first meeting really excited.

A

Thank you too. Okay, I will start with a stop the recording in and see you again next week and let's vote in doodle to see what exactly will be the time.

A