Call for glommio maintaners

[humb1t]

Initially I posted my thoughts as a comment to PR: #702 (comment)

TL;DR - let's figure out how to keep glommio alive.

[humb1t]

Those who support the idea, put 👍🏻 on this issue's description. Those who are willing to participate in maintenance - put ❤️

[snichme]

This is kind of funny, I was thinking of the status of glommio just earlier today. I've been following/looking at glommio for some time and now seeing other similar runtimes popping up (monoio and compio) and have done some research to try to compare them and most of the time I'm falling back to glommio. I think it's an interesting approach all these three are taking even though I've never use any in the real world.

Anyway, I would like to put a ❤️ on the description but have limited experience in rust so not sure how much I can help. At least put a 👍 on it

[nacardin]

There's a real need for a Linux io_uring focused/optimized async executor that makes no compromises. There seems to be a recent explosion of Linux desktop crates from the Cosmic DE and other desktop utils like launchers. It disappointing to see all of them tightly coupled to tokio. I think if glommio is maintained and active, we can build up a crate ecosystem around it, and make it a viable option for Linux oriented work.

[glommer]

Hi folks - first I want to deeply apologize for not being able to do anything here, but I have been spending 150% of my time on Turso, and would spend even more if it wasn't for my family. I barely get notifications here (or anywhere else).

I think because the project org is controlled by DataDog, it will be hard to get an official maintainer. I am no longer with Datadog for many years, but I know a lot of other people who were involved with Glommio inside the walls are either no longer there or moved to different projects.

My suggestion is to pursue a hard fork. If there is anything I can do to help, including writing a message on X or whatever putting my weight behind the fork, I will (so long as it is clear that there is a clear leader and that the project is going somewhere)

Glommio was very important to me and played a large role in my life, and it is great to see there's still huge interest

[humb1t]

@glommer thanks for answering and huge thank to everyone who put his/her/they time on glommio! Also exited to see you are working on really nice and interesting project.
I am not sure that I may be the best person to take the responsibility of leading the "new" glommio. What I propose is that I will make a fork and update the project to match and work on my machine (I use Arch btw) with latest io-uring lib version 2.14 at the moment of writing and will try to keep it always up to date (hopefully just by merging PRs of others contributors).
If there will be some people who will want to make more deep contributions, we will be able to form the maintainers core and, in that case, I think it will be time to make an announcement that new project is ready to take an "estafet" from this repository.
I will put links to new org and repo in this thread later today.

[humb1t]

@dahankzter I see your fork is well ahead of glommio. I suppose it is more experiments, but maybe some part of it can also be merged to fork already?

[humb1t]

I made an org and repo based on PR #702 https://github.com/glommio/glommio - as it seems to be working on new linux kernel and contain minimal changes at the same time.
We will see if it will fly or not in the nearest couple of months. I will make some clean up and settings on organizational side, but as not an expert in glommio under the hood, and I think we all definitely will appreciate those who are willing to move the project forward from the technical perspective. So please, ask me to add you to the team of the new organization if you are interested to give it a try, or just share your vision about how things can be done better.

[nacardin]

@humb1t Thank you! I think one of the first steps is to review existing PRs on this repo and apply them to the new one. Starting with the simple ones like updating deps, I can help.

[humb1t]

Great idea @nacardin - I sent an invite to add you to new org team. If you can start moving such PRs in small batches there, will be great!
Please notify me if you will need additional access rights in the new org.

[humb1t]

Hellou everyone, small update.
I took the accountability for rushing through all outstanding PRs that were ready to be merged and obvious about the ways to do that and by this, I am closing glommio/glommio#6 with some comments (about old drafts) and one issue (about musl support) waiting for your opinion.

The next 2 things I will work on today are:

1. Use static and "AI" analysis to review the codebase and make obvious fixes;
2. Create a simple statically generated website for glommio via zola, as I think it deserves more than a link to a blog post.

About contributions, I think it is worth:

• look through these PRs that I merged without reviews and maybe spot some issues with them
• start contributing new features
• make tests suitable to be run for musl target on CI

[humb1t]

The site is here: https://glommio.github.io/
The feedback and contributions are more than welcome, while I think in new era it will not be used as much as it would be before.

Also we need a logo for glommio and if someone wants to try draw it - please do.

That said, @glommer I would like to ask your opinion.
If you think that we can start "hard-forking" the project, we will need to migrate crates.io and docs.rs things for glommio to fork. If you want to wait and do it later, when the project will become more mature, or if you think it is better to publish it separately under a new name, please also let us know.

P.S. I put links to zulip chat, so will be nice if admins of it will share their thoughts on that matter.

[glommer]

hey sorry - it is best if you contact me at glauber at turso.tech.
I am happy to move crates to the fork as long as there is a clear leader. A lot of people recognize the name, so it is good to keep it.

[utilitydelta]

big ups to @glommer and the rest of the devs who got glommio to where it is; made a real impact moving use of io_uring and thread-per-core in rust more mainstream! I'm up to helping maintain the project and open to the hard fork idea. I use glommio in an OSS database I'm building right now (we don't use rust at my real job, sad)