Jenkins Google Summer of Code Office Hours, 10 Feb 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Jenkins in GSoC 2022, Feb 10, 2022 Office Hours

Description

Google Summer of Code 2022 office hours, including overviews of project ideas.

00:00 Introductions
04:30 Plugin health score project idea
20:00 Git cache management project idea

A

Hello uh welcome for this second office hour for google summer of code. um It is today the 10th of february uh 2 o'clock, utc and we'll try to. I will make the the most uh the most to finish on time. We failed to start on time because we had some link issues and needed to set up a few things, but we'll finish by respect for everybody on time in half an hour. This is the first meeting on this time zone and it will thus be a quite open discussion.

A

So we have uh around the table here um uh several uh mentors. uh I am the org admin for the google summer of code for jenkins, supervising the general process and making it all work together. The purpose of this meeting is to be able to meet the mentors ask whatever questions clarifications you have and bring whatever topic.

A

You want clarifications on so first a little presentation. We first have mark waite, who is joining in from the west of the united states, so very early bird.

A

We have and he's mentor for several proposed projects we have dhiraj, who is also mentor, I believe, and he's joining in from the indian time zone, and we have chris, who is also a mentor and part of the org admin team and he's located in a chinese time zone.

A

I am located in brussels, belgium, so the middle of europe between the united states and asia.

A

So we have proposed an agenda proposal here so before we start digging into the things I have a.

A

Very general question is: do you have general questions, doubts clarifications or comments to do. Don't hesitate to open your mic and be here between friends. So let's go.

A

I think it's quite impressive to talk like that, so maybe the questions will uh appear afterwards. uh General suggestion before we start is. I recommend that you listen to the recording of the pacific or asia.

A

Office hour that we hold in the middle of the night for europe uh they're also very interesting uh things that are uh shared there. So we we try to share the knowledge, a transfer on on both channels.

A

Don't hesitate to listen to that in the notes of the previous meeting. We added also here some useful links and uh information. Don't forget to look at them and.

A

Get ideas and probably also questions uh coming from the reading of these documents.

A

So do we have a question?

A

I see that the raj is.

A

Thanks, so can we also discuss here the project that haven't been selected yet but are under uh discussion for now? The short answer is yes, we can discuss that. The longer question is: will uh I I'm following this project quite closely and it's a project that is near to my heart on that, so I'm able to answer to that uh subject.

B

Raj well, and- and it may be worth clarifying that there may be a misperception there- that it's worth noting the a project has not been selected. What happens? Is candidates submit their proposals for the project, so there are.

B

All of these projects are possible ideas that you might choose as a potential contributor to submit as your proposal and what you do is you have to turn our idea into a real proposal and converting our idea into a real proposal means creating a plan for it and what the steps you think would be and how it will work. So there are many many steps to get from any one of these ideas to to a real proposal. That is your proposal.

B

Yeah john mark go ahead. Sorry it just for me, that's one of those. It's not a simple any contributor picks up a project id and it's ready to go. The idea with google summer of code is each person who is submitting their application, brings their proposal and outlines it. They they've thought about what steps should be taken, which changes should be made. They've got they've done some experiments themselves and then they bring a project proposal that is their own proposal.

A

Yeah, very, very true, and and very to the point: it's it's a two-way process.

A

It started by the mentors who started to describe a project idea and the participants or contributors can provide additional ideas, questions or or it's it's a work that we do together, uh especially once you have shown your interest or or get interested to one or the other topic.

A

uh Then you we can start the process of uh discussing with uh the mentors or in the in the project group where several communication channels are set up. uh Where you can, you can have a mailing list. A getter is generally the easiest way to interact with uh with the project, teams or participants being mentors and so on. So don't hesitate to propose ideas or or say well. This could be. This is how I understand the task. This is how I would I would propose that as an edition. So a very good point uh mark.

A

Any other general question about mythology or.

B

Oh yes, since since dhiraj asked for that topic, first I'd say: let's, let's spend some time on that one already john mark. I think you can you can give some overview. It's not been discussed before in any other session. So this would be a good thing to have recorded, as these are the concepts here, the ideas, etc.

A

Right, okay, so first thing I tend to be enthusiastic, so please this is not a sales pitch, so I'm not selling the project. This is this.

B

Is a sales pitch and we john mark is trying to sell you on why his choice is the best surprise. At least mine will be. When I talk about my ideas.

A

Okay, uh uh let's uh here, let's do it uh in an honest way, so the the health uh score idea is that I've I've been working with jenkins administrators and being have been in that duty for a while and when you want to install the plugin, you need to do that judgment. That's it.

A

What is the quality of the the plugin? What is the kind of risk that I take? Is it well-maintained plug-in? Is it trustworthy or will it tear down my my jenkins instance, because it has memory leak or whatever?

A

So there is all this reflection, what makes a good plug-in or not, and it's also a guideline for plug-in maintainer, uh meaning that well here here are the things to watch. This is a well-kept plug-in, and this is a plug-in like a garden that you didn't tend correctly and you have wild uh wild grasses all over the place and- and so it's also a hint for plug-in maintainers. What needs to be done?

A

There's several ideas and some work being done on trying to qualify what is a good plug-in and- and this is a very interesting uh uh subject to think uh about- and we're very interested to hear the opinions of participants to that. So there is this work.

A

Another very important work is: let's say that we decide that this is a measurable automatically, that you can automate criteria. That shows that the plug-in is uh a good quality and you can use it without too much worries.

A

So these criterias we want to try them out. So there is so so we want to try out some proof of concepts. How do you automate the validation?

A

uh How can you automate, uh for instance, are, are the the dependencies maintained and up to date, uh how many open issues are are opened and we want to run these probes on the 1800 plus plugins that we have so that we see what is the general distribution on the plugin heard that we have to see?

A

Is this criteria good and where are we going to put uh the the levels so summarize? First, what do we want to measure? Then? We need to do some uh pocs and further of implementing these probes and then the fir. The third step is. We need to think: how are we going to propose uh these results to the public?

A

uh How are we going to uh modify uh because the the we we tend to go in the direction of providing this information uh on the the plug-in update site uh and and what will be the most efficient way to present these information in a useful useful way? So this is for administrator, and then you have also how to show these results to maintainers so that they also see well. Okay, I should maybe do some house cleaning there and do some maintenance uh on that.

A

So to summarize three main axes and three main work topics where they're small items that we can work. So it's not a huge huge thing that will do it as a monolith.

A

The first one is work on probes to define the probes and invalidate and describe them, testing the probes and observing what is the statistical distribution of those probes, and so that we verify that the initial assumptions are correct. Third work domain is presenting the results of that, so it's a very complete project. There is some preparation work to be done. There are some quick short live work items that can grow into complete probes and then there is also the display.

A

So I believe it's a very rich project where there's lots to be done and we're not standing in front of a big mountain and say: where am I going to start with it so mark? Was that a good summary of it.

B

I think that was great, very good. Yes, now now I I think we should ask a question now to dhiraj dhiraj did that did john mark's answers which cause you to have any additional questions.

C

Yes, so thanks a lot for this really really great description, because you explained with the help of examples as well, so that puts things into picture and uh so the three steps you mentioned as documented by mark as well. uh I have a question about uh what are the attributes of a well maintained plug-in? That is the first one like which attributes do we want.

B

B

Keep going mirage well he's.

D

Yeah so dress has background noise. Okay go ahead. Yes,.

C

So the question is uh for the attributes. I was thinking that uh maybe like we can refer the contributing to the open source document by mark. So is this a good place to look for these attributes in order to determine the development.

B

Plugin, I think that's one one valid source: it's certainly not the only valid source, but it is. It is a valid source of of some data points. We might consider john mark your comments. Yeah.

A

uh Indeed, so there there are various data sources, there's also some work. uh A couple of people have already started.

A

I think it's important that we start consolidating that and and uh get the various sources uh uh together uh well and and because it's part of of the things I like, I would like to use that in order uh that we start explaining and writing uh some documents to help the people that it's not just saying well, this is not achieved, or this is under the expected level. But how do you get there? How? How can you uh do that? So there are multiple facets to it. Come back to your your comment.

A

uh The rush there are multiple sources contributing to open source is, is one uh it's the modernization uh workshop too there's a lot of very important insight in there good point.

B

To rash- and there may also be other hints in things like the jenkins health advisor from cloudbees- has some has some interesting pointers in it. um Conversations in the jenkins developer mailing list many other locations where we could find hints about what what could or could not be done.

B

A

A

Do you have other questions or is somebody else in the audience having uh questions or interest in this.

C

I do have some questions, but I don't think this would be the right place, because um my questions are more into the implementation part of it. So I think I'd be. It would be better to ask it on a channel.

A

As I see there's a lot of interest, especially if you'd raj and we try to get a few other participants in there, we need to get to get the pot cooking and brewing and the ideas start to come and and will will work uh uh together on that one. But I see that at least you and probably other people start to see where we are they it's it's very interesting because you can.

A

Participants can have an impact on uh on on that and contribute even small pieces, but it's very interesting a project and that would really help the community to improve and better use in a better way jenkins.

A

So this said mark because we're going to run out of time.

A

So I this was what I feared that I would get a little bit too enthusiastic about that subject, but it was worthwhile giving the idea so there other subjects um where, uh during the last session, there were quite some answers given to the automatic space specification generator for the rest api.

A

This is one that I would like to move ahead, because I've been struggling with rest documentation, so I really would like. So I, if you're interested in that project, listen to the recording of last week.

A

uh uh Chris gave some good answers and explanations about the jenkinsfile runner so as chris is on the call. If somebody has questions on that subject,.

A

Please go ahead.

B

So I one of the things that we didn't get to in the last call was we didn't get to plug in health score you've reviewed that we also didn't get to get cash management, and I was going to shamelessly ask if you'd be willing to. Let me spend a few minutes john mark on the get cash management topic. Chris has very nicely put the hyperlink there and, if you'll, just click that hyperlink, if you're, if you're willing to let me, spend some time on that project idea, yep go ahead.

B

So if you could, if you could make this make the text bigger for us my eyes, don't read that well yeah, so the work with the I'm used to control plus being used the screen sharing it doesn't work. Oh okay, that's alright! It'll! Be fine as it's. No no worry.

A

I'll use perfect that oops, are you still there? We are yes.

B

Thanks, that's perfect! That's great! I.

A

Had a I had a power failure, I hear the alarms in the neighborhood, so I thought I'd listen.

B

Okay mark go ahead so so jenkins. One of its primary roles is to check out source code from repositories and get that source code then to be operated on by build processes by compilation by execution of tests.

B

One of the things that it needs to do frequently is to use a get a copy of that repository, and then it will do additional work on that repository. Other copies.

B

That means sometimes jenkins, has caches of git repositories and caches of git repositories are helpful because they reduce the amount of data transfer that has to come from the central git server to the jenkins controller.

B

That's a nice help, however.

B

The the implementers of git did exactly the right thing in that they optimized git for short-term operations and accepted that the user would periodically do some maintenance activities like garbage, collect the repository or repack it, or do other removed, stale branches, remove stale tags all operations that human beings, when they interact with their repository, can do pretty easily but jenkins as an automation.

B

Server does not have a facility right now that will automatically clean up caches, cached git repositories properly, so we've had cases where ci.jenkins.io our our primary jenkins server for the jenkins project has had slower than necessary repository caches because they hadn't been well maintained. It had been created a year ago and just continually updated and no one had ever run the maintenance operations.

B

This project idea suggests we need a facility in jenkins that will allow it to automatically and periodically perform cash maintenance. So that's the idea we're almost out of time, so I think I should stop there unless there are questions from people about the idea.

B

Great, so, okay, if you'd like to learn more about it, there's a command called, get maintenance, that's included in recent versions of git, and you can read the man page about git maintenance and you can see what it offers as techniques that are already built into command line. Get to do these kind of things. That's it for me. Thanks john mark.

A

Right, okay, mark sorry to have pushed you a little bit trying to finish on on time here. So a last advice or tip. I would give to everybody think on this write down some questions and reach out to us either through.

A

Community.Jenkins.Io or via the getter channels and we'll take, will we going to take particular to make a particular effort to answer these questions publicly.

A

So please avoid pinging us directly as as much as I would like to help everybody. This does not serve the community. Every question is worthwhile and is useful to everybody joining or contemplating to join the gsoc program. So don't hesitate to ask your questions publicly. There is no shame to ask a question.

A

Every question is, is good and will take the time, and especially I take a lot of it. May it's very important to me, and I want to answer that as good as I can publicly so that everybody can benefit from these answers, so we're now at half of the hour.

A

Thank you very much for participating uh to this call to listening in and to the questions that have been asked.

A

uh We're going to have the next meeting in this time slot that I think is good for uh india time zone uh we're going to have it in two weeks, we're alternating uh every week between what I call the pacific or asia office hour and the european middle east africa office hour.

A

So thank you very much for attending waiting for your questions either live or on gitter or on the discourse on community jenkins.

A