home.social

#glommio — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #glommio, aggregated by home.social.

  1. #aquatic HTTP and WS trackers now are buildable on modern systems by upgrading #glommio to v0.10 patch github.com/greatest-ape/aquati

    * aquatic is alternative to well known #OpenTracker #BitTorrent tracker, but written in #Rust; its UDP fork is also part of the #btracker ecosystem github.com/YGGverse/btracker

  2. #aquatic HTTP and WS trackers now are buildable on modern systems by upgrading #glommio to v0.10 patch github.com/greatest-ape/aquati

    * aquatic is alternative to well known #OpenTracker #BitTorrent tracker, but written in #Rust; its UDP fork is also part of the #btracker ecosystem github.com/YGGverse/btracker

  3. io_uring без розовых очков: 5 граблей, которые сожгли мне неделю, и где он реально быстрее epoll

    io_uring продавали как убийцу epoll. На деле на HTTP keep-alive разница 0-15%, иногда не в его пользу. Но на NVMe с queue depth 128 - в 3 раза быстрее. Честный разбор с бенчмарками, реальными граблями (SQPOLL, cancel race, partial recv) и почему Google отключил io_uring в ChromeOS.

    habr.com/ru/articles/1039820/

    #io_uring #epoll #asyncio #liburing #высокаянагрузка #ядроlinux #NVMe #tokiouring #glommio #syscall

  4. #glommio hard fork attempt, newsletter 1:

    - github.com/glommio/glommio/pul

    several small PRs were made (based on one big initial one) to clean up some lints and CI issues we inherited.

    Your reviews and ideas are more than welcome

  5. Hi Mastodon, we are trying to give glommio a second life: github.com/glommio/glommio

    it is not an official hard-fork, we only experiment with how it will go. If you used glommio before and will to give it a little bit of your time, ask me to add you to contributors!

    #rust #glommio #io_uring

  6. Each time I see posts about how Mastodon (the backend software) scale I start to architect a replacement in my head.

    Maybe a rewrite in rust using #Glommio and an event sourced core with storage in #PostgreSQL, projections in #ScyllaDB and #RedPanda for the queues.

    But that's very much not a weekend project so it'll never happen.