As an experiment I'm trying #immichapp as a photo manager. It doesn't support cloud storage as such, but I've mounted an S3-style bucket using #juicefs and it seems to work? Fun to see my raspberry pi have a petabyte available!
Ah, JuiceFS! Because what's better than turning your POSIX file system into a Rube Goldberg machine powered by the dynamic duo of #Redis and S3? 🤖🔧 Clearly, the only thing more convoluted would be creating a file system out of spaghetti and meatballs! 🍝🔍
https://github.com/juicedata/juicefs #JuiceFS #S3 #fileSystem #techHumor #RubeGoldberg #HackerNews #ngated
GitHub - juicedata/juicefs: JuiceFS is a distributed POSIX file system built on top of Redis and S3.

JuiceFS is a distributed POSIX file system built on top of Redis and S3. - juicedata/juicefs

GitHub

JuiceFS is a distributed POSIX file system built on top of Redis and S3

https://github.com/juicedata/juicefs

#HackerNews #JuiceFS #Redis #S3 #distributed #file #system #POSIX

GitHub - juicedata/juicefs: JuiceFS is a distributed POSIX file system built on top of Redis and S3.

JuiceFS is a distributed POSIX file system built on top of Redis and S3. - juicedata/juicefs

GitHub

🐘 Pigsty concept PGFS combines PostgreSQL + JuiceFS to store files in your database, allowing app consistent rollback of files to any point in time (ala database PITR). Loving all the validation the pigsty team has been doing!

https://pigsty.io/blog/pg/pgfs/

#postgresql #juicefs #pigsty #mysql #oracle #mongodb

PGFS: Using PostgreSQL as a File System

Using JuiceFS to treat your database as a file system, enabling file-level PITR and keeping files in sync with your database — CDP for the rest of us

The #JuiceFS seems to be a really nice distributed filesystem - but it has one problem - https://github.com/juicedata/juicefs/issues/4424 - it does not build/run on #FreeBSD ... yet :)
Does Not Build on FreeBSD · Issue #4424 · juicedata/juicefs

Hi. I really like this project but its not possible to built it under FreeBSD. Should be easy to fix as You already have Darwin supported. I used latest JuiceFS version from 'now' if that helps. //...

GitHub

@dmoonfire I'm also trying #SeaweedFS.

My main quirk so far it's lack of metadata preservation like file creation date.
It's not providing full #POSIX even when using it as #FUSE.

Maybe try mounting via #FUSE from #JuiceFS based on #SeaweedFS ?
I'll try it for #SQLite DB for #uptimekuma as a test.

Some options lacking in community edition of #JuiceFS is provided by #SeaweedFS.
So this pair seems like strong candidate for #HomeLab.

You can also try simpler #GarageFS:
https://garagehq.deuxfleurs.fr

The Garage team - An open-source distributed object storage service tailored for self-hosting

An open-source distributed object storage service tailored for self-hosting

Garage

Hey, is there anyone here with experience with juicefs? I am considering moving from s3ql (python-based) to juicefs (golang-based) and would like to hear some stories…

#juicefs

Introduction to JuiceFS | JuiceFS Document Center

JuiceFS is an open-source, high-performance distributed file system designed for the cloud, released under the Apache License 2.0. By providing full POSIX compatibility, it allows almost all kinds of object storage to be used as massive local disks and to be mounted and accessed on different hosts across platforms and regions.

Curious about #DistributedFileSystems?

#InfoQ brings you an insightful article by Changjian Gao, exploring the design principles, innovations & challenges behind 3 representative systems:
1️⃣ #GoogleFileSystem
2️⃣ #Tectonic
3️⃣ #JuiceFS

Gain valuable insights: https://bit.ly/3pvnNxt

#DistributedData #BigData

Comparative Analysis of Major Distributed File System Architectures: GFS vs. Tectonic vs. JuiceFS

Distributed file systems have emerged as dynamic and scalable storage solutions. We explore the architectures of three distributed file systems: Google File System (GFS), Tectonic, and JuiceFS.

InfoQ
Thinking further about the Nextcloud rollout:

Much of the stuff I was considering storing on
s3fs is already there. And if I switch to an alternate object-storage-based filesystem, such as JuiceFS, the migration process should be nothing more than a simple copy, followed by remounting the storage on the same mount-point. Nextcloud doesn't actually need to know how the backend works -- just that it looks like a filesystem.

So there's no risk to just sitting pat with what I have.

Also, even if I move it to another type of filesystem, I'll still have the filesystem level access to the data, so migrating to another system or integrating with other software should be simple.

In fact, migrating out of
Gitea should be a simple matter of doing a checkout onto the external storage. If the performance really does suck, I can then copy it to another external storage within Nextcloud (I think).

The rest of the material I want to store in Nextcloud is small enough that there is no real reason to put it into object storage. And there is one reason not to: surely it is more secure to keep private "office" documents on the same server (if I need to upload them at all, that is).

And finally, it appears that there isn't enough reason to keep using
ResourceSpace , since Nextcloud replicates the key features I would have been using it for. Instead, I should probably be looking at the available Nextcloud apps to see if I can fill in any gaps in functionality (mainly the ability to read and extract metadata and preview information from a range of file types).

So it looks like I may not even have to reinstall Nextcloud. Just invest some time in organizing and tagging.

#VirtualStudio #Infrastructure #Nextcloud #s3fs #JuiceFS #Gitea

Anybody played with #juicefs for persistent storage for misc container services? Curious what experience has been like for different use cases.

Curious what the most wreckless workload has been.