text stringlengths 0 1.49k |
|---|
• Importance of addressing sexual harassment in the industry, with a potential future episode on the topic |
• GopherCon conference and its Code of Conduct (COC) |
• Incident at GopherCon where someone violated the COC |
• Promotion of GopherCon's Call for Papers (CFP) and importance of submitting proposals early |
• Discussion of the gops project by Google, a tool to list and diagnose Go processes |
• Reminder that the CFP for GopherCon ends January 31st |
• Submission guidelines for talks at GopherCon |
• Qualifications for speakers and how to demonstrate expertise without revealing identity |
• Benefits of submitting a talk, including mentorship and travel compensation |
• Variety of presentation formats available, including tutorials and workshops |
• #FreeSoftwareFriday segment featuring projects or software recommendations |
• Discussion of specific open-source projects, such as oklog and OpenOCD |
• Redis 4.0 release candidate is out |
• Travis Jeffery discusses his experience with Redis and its uses |
• Brian Ketelsen shares his use of goa/gorma for designing APIs first |
• Discussion about using Go frameworks for web development vs standard library |
• Buffalo framework is mentioned as a popular choice for Go web development |
• Eric St. Martin mentions writing web applications without using a framework |
• Discussion of a Docker-backed web terminal and issues with WebSocket upgrades |
• Tradition at GopherCon of taking the first ticket purchaser out to dinner |
• Request to discuss the standard logging interface in Go |
• Proposal for a single interface for logging in Go, inspired by Java's Log4j |
• Concerns about current logging packages in Go, including lack of log levels and structured logging |
• Discussion of potential compiler changes to improve logging performance |
• Show wrap-up and thank-yous to listeners |
• Shoutouts to sponsors: StackImpact and Backtrace |
• Recap of show participation: Travis, Carlisia, and Erik St. Martin |
• Promotion of GoTime.fm resources: Twitter handle, email newsletter, and GitHub ping for guest suggestions |
**Erik St. Martin:** Hello everybody and welcome back for another episode of GoTime. Today's episode is number 31. Our sponsors for today are StackImpact and Backtrace. |
Today on the show we have myself, Erik St. Martin, we also have Carlisia Pinto... |
**Carlisia Thompson:** Glad to be here. |
**Erik St. Martin:** And Brian Ketelsen. |
**Brian Ketelsen:** Hello! |
**Erik St. Martin:** Brian and I have talked about Kafka and our love for it a number of times through a couple of episodes, and I think we've even mentioned the project, so today's special guest is Travis Jeffery, here to talk to us about Jocko, which is a Go implementation of Kafka... Mixing two worlds we love! |
**Brian Ketelsen:** I know, it's two great tastes to taste together... When you get your peanut butter in my Kafka, I couldn't be happier. |
**Erik St. Martin:** So you wanna talk to us a little bit about the project and your motivation behind it? |
**Travis Jeffery:** Yeah, sure. |
**Erik St. Martin:** Well, first tell everybody a little bit about yourself, that might be helpful, too. |
**Brian Ketelsen:** Yeah, introductions are always good. |
**Travis Jeffery:** Alright... This question's always a little funny. I grew up on a farm about two hours North of Toronto, Ontario, Canada, middle of nowhere. My parents were both entrepreneurs; that left me with a lot of time to watch movies, play video games, read books. They put me in a daycare called Teddy Bear da... |
When I was around 12 I picked up a book - I think it was The Pro Book, and the second book was a C primer book, and then I got "Hacking: The Art of Exploitation", because I thought I was gonna be a hacker. |
**Erik St. Martin:** I think I have the original version of that book, too. |
**Brian Ketelsen:** Signed... |
**Travis Jeffery:** Yeah, and a little bit after that DHH put up the How To Build A Blog In 10 Or 15 Minutes, and I was like "Holy crap, that's crazy", and that's how I got started making web software. |
Just before university, I started contributing to open source stuff; I contributed to Emacs and Vim and Django early on, and then Rails. Then between the first and second year of university, I started getting recruited by the big tech companies like Google. Around this time, when I first went to university I thought I ... |
Once I started getting those recruiting mails and I asked one of them, "If I don't finish my degree, can you get me into the U.S.?" and they were like, "Yes", and then I was like, "Okay, well I'm done." So I dropped out and started a startup with some friends and we ended up selling out to Shopify. |
After that I went and worked at Basecamp (37Signals) and that was pretty cool, to end up doing that. Then I wanted to do another startup again, so I talked to one of my friends, TJ Holowaychuk and I asked him where he was working, and he was telling me about this company Segment IO, which was like an analytics data sta... |
\[03:59\] I ended up joining there, and that's how I was introduced to Go, because originally Segment IO was built on NodeJS, and we started to scale up and it got to the point where Node's event loop would be blocked all the time processing JSON, and so that's how I started to introduce Go, and ultimately we ended up ... |
Today I am head of architecture at another analytics company called Taplytics. So that's what I'm doing now. In the future I would like to bootstrap my own company, maybe write some scripts... I wanna make movies one day, I think that would be cool, and do lots of writing. So that's what I'm about. |
**Erik St. Martin:** Nice. So what was the primary motivation for doing Kafka in Go? |
**Travis Jeffery:** So I've been using Kafka for a couple years now, and it's pretty awesome, I love Kafka a lot. There are some annoying things... For one thing, it comes with some baggage, like the JVM and Zookeeper. You know, they're not the nicest things, as you have to maintain Zookeeper and all that stuff... And ... |
For instance, you can configure a topic to have a certain amount of data on a broker, how much data it will retain, and let's say you then add another topic, and that topic gets assigned to that broker. It can then go over the amount of data that you wanted to save, because... Basically, what you wanna do is set a per... |
But basically, the whole idea behind Jocko was writing a Kafka that would be really easy to set up... So I would distribute a single binary, I wouldn't depend on Zookeeper, I would maintain protocol compatibility so that people that use Kafka now, they can just drop Jocko in and it would work the same, and it would be ... |
**Erik St. Martin:** I wonder whether we should roll back a little bit too and talk a bit about what Kafka is and what it's useful for, too. |
**Travis Jeffery:** Yeah, so Kafka is a replicated, distributed commit log service. Basically it keeps a sequence of things and then those things can be consumed by workers. You can use it as a message queue or any type of thing that you would want to stream, that's basically what it's used for. For instance, at Taplyt... |
**Brian Ketelsen:** The thing that I think is the most magical about Kafka is that you can use it as your system of record, and I think that's the power of Kafka for me - the idea that this crazy thing that looks and acts a little bit like Git, and the queue and the database, all at the same time can be your system of ... |
**Erik St. Martin:** Yeah, I mean it's really great for stuff like that... Brian and I worked together on some projects where... Depending on your use case, sometimes you want the same data in multiple data stores, so Kafka is kind of like a really great way of being a system of record and then having all those data st... |
**Travis Jeffery:** Yeah, you can think of it as like a data hub for all your data, basically. At Segment, a lot of times what we would do is that basically something would go into Kafka, a worker would pull that off, do some processing on it, and then put it back into another Kafka topic, and then so on and so forth, ... |
**Brian Ketelsen:** \[08:04\] Yep, very, very common pattern. |
**Erik St. Martin:** The nice thing for that streaming data workflow too is that it's great for services to come online and offline if a service crashes or you take it down for upgrades and stuff like that... The data is still being pumped into Kafka, the queue just backs up a little bit until the consumer comes back o... |
**Travis Jeffery:** Yes, exactly. For scaling up it's awesome, because basically things would just get queued up, and then you can just add more workers to pull them off quicker. |
Another awesome thing is basically managing your dependency graph, so rather than having services that communicate to each other directly, you can just have your workers, which don't know their relationship to each other - they just read from Kafka and then put something else back on to Kafka, so they have no idea abou... |
**Erik St. Martin:** Yeah, it's basically kind of like using goroutines and channels. For instance, you don't really care about the things that are consuming stuff from your channel, just that you're either pushing data to a channel that you're responsible for, producing data on, or you're receiving it and doing some s... |
**Travis Jeffery:** Yeah, and the same topic can also be consumed by various consumer groups. Let's say one group reads from it and they make that data immediately available. Maybe it will put it into a cache or something, and then another consumer can take that data and it will do something to it to make it available ... |
**Erik St. Martin:** So where are you at in the development of it? How is the performing comparison? Is it feature-complete? Is it just you working on this? |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.