We made our filesystem 47× faster by deleting it

(microsandbox.dev)

40 points | by appcypher 4 days ago

9 comments

  • gnabgib 57 minutes ago
    Title: How we made our OCI filesystem 47× faster

    Your coworker/other account, messed it up last time you submitted it too: We made our sandbox filesystem 47× faster by deleting it https://news.ycombinator.com/item?id=48195883

    > Otherwise please use the original title, unless it is misleading or linkbait; don't editorialize.

    You're linkbaiting.. the opposite of the guideline

    https://news.ycombinator.com/newsguidelines.html

    • makeboss 3 minutes ago
      fair point. that's the title we used on our X article and i copy-pasted. updated the article's actual title.
  • a_t48 51 minutes ago
    I've found that tar processing tends to dominate the time used to do anything with standard OCI layers. I have a more efficient format (that splits apart the layer into metadata+chunks) that I'm open sourcing soon if y'all are interested in using it.
  • andix 1 hour ago
    Isn't it really obvious that a user space fs will always be slow, and especially slow with small files?

    I don't know the purpose of microsandbox, but such an article doesn't give me great confidence in exploring it further.

    • toast0 25 minutes ago
      > Isn't it really obvious that a user space fs will always be slow, and especially slow with small files?

      Small files seem like the perfect case for a user space fs... depending on what you mean by user space fs. If you mean interfacing with a FUSE (or similar) filesystem by using syscalls in your program to context switch into the kernel, then context switch to the userspace FUSE layer, then send it back to the kernel and then back to your program ... that will be especially slow with small files where date bytes per context switch is small. OTOH, if you mean a user space fs where your program has a built in filesystem it can access without context switching, then that will be of benefit ... especially if the files are small enough that you can pack multiple files into a single page.

  • dspig 1 hour ago
    That's such a HN title!
    • nine_k 1 hour ago
      And the content. A number of smart design decisions, and analysis of what was wrong with the previous versions.

      Also, it's a great illustration of the benefits of layered, modular design that Linux sports: it allows to mix and match parts to build what you need.

    • menno-sh 1 hour ago
      Sounds like a title you’d read on the satirical version of HN in a TV show like Silicon Valley
    • nathanmills 1 hour ago
      [flagged]
  • nasretdinov 1 hour ago
    If your problem could be solved without FUSE, it probably should!
  • jmclnx 25 minutes ago
    The site is impossible for me to read due to the colors. I went to lynx and i looks like it is about a file system in a VM.
  • cassianoleal 1 hour ago
    Is anyone here using this software? How do you integrate it with your agent workflow? Do you run agents in editor (Zed, VS Code, Cursor, whatever)?

    Have you tried the sync feature?

    Edit: FML why is this being downvoted? At least have the decency of explaining, I'm happy to adjust my conduct but I can't do so if I don't know what I did wrong.

    • Enginerrrd 1 hour ago
      I can only guess at the actions of others, but I would guess it’s because your comment is a tangent and at best only vaguely related to the featured article?

      The article is really about solving a particular problem with the backend of their infrastructure. Discussion about VMs, Linux kernel syscalls, file systems (virtual, FUSE, etc) would all be relevant.

      Your comment is a question about whether and how people use the software itself, which is pretty unrelated to the article.

      It’s a bit like an article about Porsche identifying a particular engineering nuance in their fuel injectors, and how things didn’t work the way they thought at a low level, and how they solved it once they realized it. And then you come in with a comment about what people like to do with their Porsches. Like, sure, it involves the same company but what would that have to do with the underlying article on automotive engineering?

      Combine that with a growing disdain for the insistence of certain segments of the tech scene to make everything about agentic workflows, (an echo to the constant evangelism of cryptocurrencies or blockchain in the recent past) and you have a recipe for downvotes.

      • cassianoleal 1 hour ago
        This is pretty common on this forum though. Many times the comments section becomes mostly about things that are not necessarily directly related to the article but remain related to the bigger thing the article is about.

        Oh well. :) Thanks for your insight anyway.

        • Enginerrrd 1 hour ago
          You’re not wrong, and this is speculation, but I suspect you’re just missing the subtext I added in my edit: that some people are burnt out on the evangelism of agentic workflows in the same way they were about blockchain or whatnot.
    • wizzwizz4 1 hour ago
      People are generally sick of AI, and of people who bring AI up in every single comment thread. The downvoters may not realise that TFA is by an AI company, about an AI product, making your comment tangentially relevant, and not (strictly-speaking) an example of the behaviour they're fed up with.
      • cassianoleal 1 hour ago
        Heh yeah it very well could be. I am also fed up with AI everywhere but I don't go out downvoting everyone and everything who mentions it - and definitely not where the whole context is about it.
  • moralestapia 1 hour ago
    >Every file operation inside the VM had to bounce out to the host through FUSE

    Lol, yeah that was your mistake. FUSE is a phenomenal idea but anyone who has used it knows how slow it can be.

    • fh973 44 minutes ago
      FUSE is not slow. Our distributed file system pipes over 70 GB/a through a single mount point.
    • himata4113 1 hour ago
      I have learned first hand with my agentic workflow as it took 1 hour to compile rust instead of 6 seconds.
    • goneri 1 hour ago
      Especially compared to direct virtio access to physical volume.
  • PunchyHamster 45 minutes ago
    You made your filesystem 47x slower by NIHing it