Local wizard served by worker-agent setup · 127.0.0.1
The coordinator hands out work and collects results. Ask your cluster admin for its address.
For a served instance this is the address printed by coordinator serve.
Serve instances: one token for every worker.
Shared clusters: a key tied to your account.
The admin can copy it from coordinator token on the server.
Where tasks run and how the machine appears in the cluster.
Shown in the coordinator’s worker list.
Datasets and shard results live here. ~1 GB free space recommended.
Detect the machine’s CPU count.
Limit what this machine advertises.
Process this many shards in parallel on this machine. Each loop runs its own task runner subprocess.
Making sure this machine can reach the coordinator and run SciMesh workloads.
Configuration will be saved and the worker started as a background process.