Video

You look like a Europe east programmer that supports the library that runs behind scenes of everythink, very smart, I like the video and, the way you just get into the code!

Hey Good exercise but i have some questions about bentoml workers i implemented service with bentoml wit multi api-workers the model load once and this is magic and good benefit but don’t effect in performance when open multi worker in fastapi for example 2 workers the model load 2 times but the performance for handling requests was increasing but in bentoml the model load once because of sharing memory but I didn’t see that the performance was improving Did you know anything about that ?

Solid. By d way Holden it would be great if you can also add the PR urls in the youtube description so that it will be easy to navigate. I always search and get it though ;) But getting the PR urls below the video will be best. It is always great to see you code and explain it.

Yesterday, When searching some docs about Spark Debugging. I saw your presentation On Spark Summit, that was incredible! Now, Im reading your book about High Perf Spark - 15% on it! Very good book. Probably one of the best Spark Developer on the Earth! ❤ Keep it up!

very helpfull videos

interesting video

What keyboard do you use? Sounds really good. Apologies if you have already shared your desk setup before on this channel.

Thanks for this kind of workflow and deep dive content. I find it very useful!

this is awesome! Thanks for uploading this advanced content :)

Hello :) I want to apply to this project on my lab. Can you share with me yaml files? :)

Did you figure this out, Can you share github link for it?

seeing you debuging your code gives me motivation to keep debugging mine awesome :) keep posting

thx for sharing / uploading

Are you training a language model like chatgpt?

👌

Hi Holden, are you on some type of small treadmill?

thank you dude

Hey.

That's Awesome, you plan to do the book sessions on a schedule or random ?

NA trans are so disgustingly ugly, here in SA they do a little better

Hey Holden - Randomly looking for data quality libraries online for spark and landed on this video! I recall from the Spark summit this year that Nike is a Databricks customer. As an engineer using databricks in day job myself, one hunch I immediately have after looking at your PR is that to run spark/pyspark on databricks your in-house libraries cannot have direct dependency on spark/pyspark. It will conflict with the proprietary spark install that databricks provide. The other two missing libraries also happen to be available in databricks environment (and the delta lake lib also might have a proprietary version on that environment). That also kind of explains the delta lake dependency (and no easy flexibility for iceberg support etc. yet).

High Performance Spark 2e is comging up? 🎉🎉

Promo-SM 😒

Good Job Holden! keep trying!

Inspirational! Your dedication

Is this document available publicly somewhere? Didn't watch the video yet so apologies if you've mentioned it already.

I am using Minio & Rookceph & see the following error in Minio pod API: SYSTEM() Time: 12:26:16 UTC 07/10/2023 DeploymentID: e0986c7c-b40f-4f0f-ac6a-57034c20d078 Error: Storage resources are insufficient for the write operation .minio.sys/tmp/9286dad1-86b6-40aa-8f12-165c5d1371f2/25d31237-c15f-4830-ba5a-61a2da2e1320/part.1 (cmd.InsufficientWriteQuorum) 2: internal/logger/logger.go:258:logger.LogIf() 1: cmd/erasure.go:427:cmd.erasureObjects.nsScanner.func1() API: SYSTEM() Time: 12:26:16 UTC 07/10/2023 DeploymentID: e0986c7c-b40f-4f0f-ac6a-57034c20d078 Error: Write failed. Insufficient number of drives online (*errors.errorString) 10: internal/logger/logger.go:258:logger.LogIf() 9: cmd/erasure-encode.go:112:cmd.(*Erasure).Encode() 8: cmd/erasure-object.go:1190:cmd.erasureObjects.putObject() 7: cmd/erasure-object.go:954:cmd.erasureObjects.PutObject() 6: cmd/erasure-sets.go:767:cmd.(*erasureSets).PutObject() 5: cmd/erasure-server-pool.go:953:cmd.(*erasureServerPools).PutObject() 4: cmd/config-common.go:83:cmd.saveConfigWithOpts() 3: cmd/config-common.go:88:cmd.saveConfig() 2: cmd/data-scanner.go:227:cmd.runDataScanner() 1: cmd/data-scanner.go:80:cmd.initDataScanner.func1()

Promo-SM 💕

i keep wondering why is bash still used..maybe it's my fault that i fell short of learning it

Thanks holden. Have a good day

All the best on your book. 👍

looking forward to the second part

missing these sessions... just had a look at Distributed Computing 4 Kids. interesting stuff 👍

Big fan ❤of you Holden one day would love to meet u

Interesting stuff

😡 ᵖʳᵒᵐᵒˢᵐ

Can you please put good resolution in future. Its hard to see what you type. Thanks in advance.

Cool, can you make slides public if possible?

Love these end to end stuff. Please do more of it

I love these streams. Keep it up! 👍

It would be very useful if you could do a stream on ur vim/emacs tooling for ppl with less experience like me

AWESOME!!!

This is super cool 😎

After longtime. Hope your Ray book is done

I know its kinda fundamental but it would be very useful if you can do a live stream more focused on how you use emacs, packages, etc

Looking forward to this book

Hello, Why dont you use an IDE ?

awesome holden, will these documents be shared?

Wonderful... We have spark jobs running on dataproc .. is it really nice to run them on gke ?? Kindly suggest

Thanks. I've been spending the past 2 days trying to get PySpark to work on ARM Kubernetes. Gives me hope to see it actually working. I was building from source within Raspberry Pi to get the ARM image, but I'll try your QEMU way instead and build multi-arch image.