Quickstart

One-time setup

Create a DNS domain named after the project. Once per machine, and it needs sudo.

sudo container system dns create crawler-storage

That registers the store with the platform’s DNS as seaweedfs.crawler-storage, reachable from other containers and from the host under the same spelling.

Up and down

The same store comes up from inside a submodule (same project name, same volume), so there is no need to leave the consumer’s repository.

sh scripts/stack.sh up      # or  sh .upstream/seaweedfs/scripts/stack.sh up
sh scripts/stack.sh down

stack.sh looks at two things before it starts anything — whether container and container-compose are on PATH, and whether the DNS domain is registered. If either is missing it stops and names it.

Addresses

Port	From a container	From the host	Use
S3 API	http://seaweedfs.crawler-storage:8333	http://127.0.0.1:8333	the aws CLI. Normally this one
Filer	http://seaweedfs.crawler-storage:8888	http://127.0.0.1:8888	browsing the contents
Master	:9333	http://127.0.0.1:9333	/dir/status for its state

From the host, pick the 127.0.0.1 one. On macOS there are machines where a binary that is not Apple-signed (node, grpcurl, aws) cannot reach the containers’ subnet — there the DNS name does not get through.

To find artifacts in a browser: http://127.0.0.1:8888/ui/index.html — it lists them newest first, so the one you just captured is at the top. See Finding artifacts.

Buckets and keys

The buckets are listed in S3_BUCKETS in docker-compose.yml, and the entrypoint creates them at startup (with a bounded retry; there is no init container to wait on for ordering). The keys are spelled exactly like the bucket name, and identities are per bucket. Details in Configuration.

Read next: Operations (deleting, looking, resetting), Consumers (which repository uses what).