heed VS kube

At Meilisearch we are currently trying to add a better error handling in heed v0.20, our LMDB key-value store wrapper. Unfortunately, when there are a lot of generics it can become harder to play with…

There are issues and pull requests but I advise you to look at the milli folder in the Meilisearch repository, it’s where all the logic is done. We extensively use RoaringBitmaps, heed the LMDB wrapper and grenad when indexing.

I’ll continue to introduce new features and new safety guards until v0.20.0. Can you tell me more about your project? Or is it private?

LMDB support multiple readers and one writer at the same time. It is ensured by the library. Note that LMDB is a key-value store. You can use the heed library which is the most maintained Rust wrapper.

Note that heed is ensuring that you are not trying to use transactions, databases and environments in the right way. I have added much more work in that regard in the important update that I am working on too!

I am currently working on exposing the new LMDB encryption feature from heed the safe LMDB wrapper with the help of the Cryptography community.

Have you considered heed or even sanakirja?

rkv hasn't been updated for a while. I recommend using heed - https://docs.rs/heed

For wrappers around LMBD, I'd recommend RKV or Heed https://github.com/mozilla/rkv https://github.com/Kerollmops/heed

https://github.com/kube-rs/kube/blob/main/kube-client/src/api/portforward.rs ill create an issue and start working on implementing it. Thanks for the feedback

Interesting, I haven't tested that one yet. We use kube-rs (https://github.com/kube-rs/kube) under the hood, which does automatically token refresh when needed. But there might be some edge cases not considered.

So I used kube-rs and came up with this: https://github.com/mach-kernel/databricks-kube-operator

Stackable is building k8s operators in Rust for bringing up interconnectable, data crunching applications (being able to bring up a complete data platform on any k8s cluster is the goal). kube-rs is used under the hood and code is open source.

yes, we had a very small API to create who created role bindings for a CI. I chose Rust because it was easy and seems easily maintainable with the kube-rs https://github.com/kube-rs/kube-rs crate

I have experiencie on this. I'm a DevOps engineer that started before DevOps was a thing, and I want to write rust for a living. What I'm doing is writting rust for tooling and kubernetes controllers using https://github.com/kube-rs/kube-rs. Rust and DevOps are close specially if you need to write your own tools. My recommendation is if you have some space to solve problems at your job with rust go and do it, if not try some opensource projects that interest you to get more experience.