Garage - S3-compatible Object Storage alternative to Minio (garagehq.deuxfleurs.fr)
from kiol@lemmy.world to selfhosted@lemmy.world on 26 May 11:32
https://lemmy.world/post/30296165

Curious about thoughts on Garage as an alternative to Minio. It has been in development since 2020. Here is the project git. Documentation looks nice.

Curious what others think of it as a project that has been around for a few years and seems like a solid, open source contender now that Minio has removed most of their community edition functionality.

#selfhosted

threaded - newest

FlexibleToast@lemmy.world on 26 May 12:07 next collapse

I haven’t heard much good about Minio. I would also be curious about this project.

kiol@lemmy.world on 26 May 12:23 collapse

Minio has worked well for many years. Haven’t had any problems with it until now, but ready to migrate away.

FlexibleToast@lemmy.world on 26 May 14:44 collapse

Well, that’s good to hear. Maybe I only see people post when things aren’t working, because that’s when people are most likely to post. This project mentions geo replication, does Minio replicate as well?

rtxn@lemmy.world on 26 May 13:27 next collapse

Minio is about to get Redis’d.

danhab99@programming.dev on 26 May 13:52 next collapse

Love to see the flake.nix at the root of the project

johntash@eviltoast.org on 26 May 14:05 next collapse

I really like it. I don’t use it for much, but it’s super easy to have multiple servers in multiple locations and let it take care of replication.

It seemed like it was built more for the self hosting and homelab crowd and not enterprises.

rune@infosec.pub on 26 May 14:41 next collapse

Been using Garage for non critical stuff for a while, no problems.

suicidaleggroll@lemm.ee on 26 May 15:52 next collapse

I’ve always wondered - and figured here is a good a place to ask as anywhere else - what’s the advantage of object storage vs just keeping your data on a normal filesystem?

gottafixthat@mstdn.social on 26 May 16:03 next collapse

@suicidaleggroll @kiol Depends on the application, but portability and data safety are two a prime reasons. It's also much easier to scale with docker swarm or kubernetes for business applications, that way the data isn't tied to a particular host.

kiol@lemmy.world on 26 May 17:28 next collapse

One sentence answer: “Object storage manages data as discrete units called objects with unique identifiers and metadata, while file storage organizes data in a hierarchical structure of files and folders.”

fwiw, I see object storage used as a way to manage data regardless of the file system. It is designed to scale, as opposed to the file system, in large cloud environments.

Here is a recap from Google Cloud:

Object storage, also known as object-based storage, is a computer data storage architecture designed to handle large amounts of unstructured data. Unlike other architectures, it designates data as distinct units, bundled with metadata and a unique identifier that can be used to locate and access each data unit.

These units—or objects—can be stored on-premises, but are typically stored in the cloud, making them easily accessible from anywhere. Due to object storage’s scale-out capabilities, there are few limits to its scalability, and it’s less costly to store large data volumes than other options, such as block storage.

Much of today’s data is unstructured: email, media and audio files, web pages, sensor data, and other types of digital content that do not fit easily into traditional databases. As a result, finding efficient and affordable ways to store and manage it has become problematic. Increasingly, object storage has become the preferred method for storing static content, data arches, and backups.

dgdft@lemmy.world on 26 May 20:53 next collapse

Object storage is indeed a specialized filesystem in a trenchcoat.

Object storage is typically (but not always) associated with non-hierarchical key-value lookups, as opposed to the directory tree pattern most file systems use. Object storage systems are also typically (but not always) designed with sharding and distribution in mind.

Legume5534@lemm.ee on 27 May 23:53 collapse

I’m yet to find a project that actually uses it to good use where a regular filesystem wouldn’t be appropriate. The only situation it seems to be intended for is “big data” as in huge huge huge files that would otherwise trip up and choke standard filesystems.

For us regular folk it seems pointless. And I say that as someone who has used it for various things and just not found it worth it.

exu@feditown.com on 27 May 00:50 next collapse

I had the opportunity to check out a bunch of S3 servers for work. For a quick summary, Garage was much faster than Minio in my tests but lacks advanced S3 features like object locking, versions or retention. Be sure to check what you need before switching.

Also, it’s CLI only as far as I know, so the same as Minio will be.

borax7385@lemmy.world on 27 May 07:16 next collapse

For my simple use case (storing Velero backups), it works perfectly and with a resource footprint ridiculously low (~ 3 MiB memory when idle). In comparison MinIO used 100 times more memory.

nix98@lemmy.world on 30 May 20:19 collapse

How does this handle IAM policies? At work we have a blackened that relies on S3 and IAM policies. For local development we use minio as it seamlessly handles that. Does garage?

kiol@lemmy.world on 31 May 12:13 collapse

Sounds like a question to ask the project directly, via their contact on the site: garagehq@deuxfleurs.fr