tpetry

Transactions that haven‘t been written to the WAL yet are also lost when the server crashes or you run pgdump. Stuff not in WAL is not safe in any means, its still a transaction in progress.

They probably detect this. Thats why the problem is solved after one to ten minutes according to the article. There's probably nothing they can do which wouldn't stress the disks more.

Most probably Amazon wanted to push ebook prices further down - or a time-limited deal for Bloomsbury about less ebook fees expired.

The interesting part about the enterprise quality disk misinformation is how so wrong it is. The core idea of ZFS was to detect issues when those drives or their drivers are faulty. And this was more happening with cheap non-enterprise disks at that time.

Its really easy for them, and this also in line how they would operate:

Add a new Airtag v2 protocol to the next iPhone and sell new Airtags only using that protocol. Why should you buy them? They could have different improvements you would like.

Start deprecating Airtag v1 in 3-4 years - and only sell new ones. There are now 3-4 iPhone generations that can handle the new version.

The next iPhone in 6-7 years doesn't support Airtags v1 anymore as it is obsolete now for many years.

Voila, they killed Airtags v1 in less than 10 years without killing the entire product line by switching to a new version. Is that unrealistic? No, thats their normal way how they deprecate stuff. It still works but only with old hardware or by not getting new updates anymore (iOS, macOS).

Culture is not important for their revenue, most of the company is not. The ads business is their cash cow which finances the rest of the company. They can have horrible culture and mismanagement everywhere and revenue will still grow - if they don‘t fuck up their teams working on ads.

Because AMD and Nvidia don‘t have their on Fabs.

I invented a new way of visualisi g composite ones when I wrote my book about indexing. You can scroll down on the landing page to the fundamentals chapter to see it.

https://sqlfordevs.com/ebooks/indexing

The problem is typically day 1000 problems: The database broke, nobody really understands all the stuff and dependencies by the kubernetes helm chart and still you have to fix it.

Downtime is now calculated in days and not hours.

Letting a static analyzer do these checks is a sane approach! The runtime detection of code files is really weird and would fail with bundling etc.

The most important part is understanding that downtimes are not a problem!

Cloudflare is again down and your service is not available for an hour? You cant do anything. And nobody unsubscribes.

Most downtimes are because another service is down and you cant do anything.

Sometimes its really your issue. Your service is down for an entire day? Some customers may be pissed at the moment. But in the end they dont care if you are not a critical service - like this one.

Learning: Downtimes are not an issue. When such a message pops up just ignore it and look at it when you have time. This is strange first as we try to be perfect. But the world will continue to turn.

The problem is that these type of very much needed but boring new features don‘t grant you a promotion. But if you work on a GenAI product right now the odds are in your favour.

There are examples: Directly at the bottom of the black frame where you have to input the data.

Oh totally forgot to mention: There‘s also an API to skip the manual steps needed on the website. So I plan on integrating it into database tools like DataGrip etc.

LightVM is stating a VM creation of 2.3ms while Firecracker states 125ms of time from VM creation to a working user space. So this comparing apples and oranges.

But then also monitoring gets harder. With containers you could see (and monitor) all those processes running in container easy on the host.

With VMs? You now need a way to get data from within the VM which in most lightweight VM implementations just not possible.

You still have to wait for the end of the streamed response until you can continue with your task.

Most probably because ZFS still has the most extensive data correctness guarantess of all filesystems on Linux. Yes, bcachefs will have checksums too but it it is still in beta.

They‘ve stated clearly that you pay the same. Its just calculating stuff differently that you can now remove a dedicated machine mid-month and get the remaining time‘s money back. Which wasnt the case until now.

And when you have performance issues you still need a DBA. Because RDS only runs your database. It is up to you to make it fast.

Thats because EBS is a network block device and not a network filesystem - that would be EFS. And with network block devices you get the same perf and better compared to EBS.

Ordering that amount of amount of servers takes about one hour with hetzner. If you truly want a complete rack on your own maybe a few days as they have to do it manually.

Most companies don‘t need to scale up full racks in seconds. Heck, even weeks would be ok for most of them to get new hardware delivered. The cloud planted the lie into everyone‘s head that most companies dont have predictable and stable load.

Postgres Rust extensions is a completely different topic than rewriting PG in Rust. The problem is that a crash in any extension will take down the full PG worker. So you can havoc the robustness of your PG database with any extension. But with a memory-safe language like Rust which forces you to handle all error cases this can‘t happen anymore.

I hope for that too. But I believe it wont. The complete Fake News topic just shows that everyone is happy trusting information that matches their beliefs. There is never any critical thinking or verification if it is actually true or just made up.

And its enough when such a technology is 80% right. Nobody will verify every answer if it is correct most of the time.

Exactly. I‘ve seen it sold that anyone has less books to carry and they could do more digital stuff - but nobody had any real plans. In the end one of ten text books was a PDF and you didn‘t gain anything. Its worse, now your textbook can have no battery and there are no electricity plugs near you…

They are only doing unmetered DDoS for HTTP traffic - for websites. File storage etc. is not allowed. And git falls into that realm.

Any of the capabilities needed by the exploits is in linux only available to root users. And are now granted to a container.

So these ”escapes“ are basically giving a container full access to the host and using that. None of them are enabled by default.

1. I‘ve used a phpscript for the PDF version (themsaid/ibis on GitHub), pandoc for the ePub and the TailwindUI docs template for the online version

2. I‘ve saw him sharing the progress of redesigning the landing page for a framework I use on Twitter snd immediately reached out

3. Yes, everything is written and refined by myself in countless hours. No AI involved.

I've been helping devs for 2 years to work better with databases. Switched to full-time consulting this year.

The book took a very long time to create. Everything calculated together about 400 hours. So yes, I couldnt have done this in my free time. I just took the time this year as this year was planned as some kind of programming sabbatical - a do whatever I like thing. That it was such a great success and paid off all my time is just a bonus.

Author here. Didnt expect someone posting my article here so fast after I published it.

Ask me anything you want to know.