Ilker ManapandClaude Opus 5.5 cf2edfd995 A model evicted for being idle is no longer wanted ...

flux evicted to make room for triposr was placed again two seconds later:
still on the wanted list, and first in the catalog, it took back the card
it had just given up, and the 3D job went on waiting. Its last replica
evicted, it leaves the list until it is asked for again.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

2026-10-10 21:11:40 +02:00

..

adminapi

The client API described in OpenAPI, and both APIs in Swagger UI

2026-10-10 20:22:48 +02:00

agentsvc

LoRA adapters trained on the nodes: llmgw ctl train

2026-10-08 20:25:57 +02:00

catalog

A custom node's own range decides the packages it names

2026-10-10 18:54:15 +02:00

fabrics

Watch each fabric's subnet managers, and keep a fabric one through a failover

2026-09-29 07:47:03 +02:00

fabrictest

Measure two nodes' fabric rail by rail: llmgw ctl fabric-test

2026-09-29 21:29:23 +02:00

facility

Facility-aware placement: racks, PDUs, holds and power caps

2026-10-05 01:17:49 +02:00

federation

Federation: another llmgw cluster as a provider, its models as PEER/MODEL

2026-10-05 01:21:30 +02:00

fleetmetrics

MIG: a card in MIG mode is placed as its instances, each a GPU of class <card>-mig-<profile>

2026-10-05 00:01:14 +02:00

inventory

Facility-aware placement: racks, PDUs, holds and power caps

2026-10-05 01:17:49 +02:00

modelserve

Qwen3Guard is the judge, and judges on a GPU and on CPUs are asked in turn

2026-10-08 10:22:47 +02:00

modelsvc

A profile may load weights of its own

2026-10-10 17:15:38 +02:00

nvlink

Find multi-node NVLink domains, place across them, and show them as racks

2026-09-29 09:07:59 +02:00

placement

Nodes show their engines' versions and GPUs' capability; a profile can ask for an engine version

2026-10-07 13:06:57 +02:00

quota

Guarantee and limit capacity per part of the tenancy tree

2026-09-30 18:15:38 +02:00

reconciler

A model evicted for being idle is no longer wanted

2026-10-10 21:11:40 +02:00

topology

Map every InfiniBand fabric, draw it, and place replicas by it

2026-09-29 08:46:23 +02:00

train

LoRA adapters trained on the nodes: llmgw ctl train

2026-10-08 20:25:57 +02:00

control_test.go

Give every engine a key only the router holds

2026-09-30 00:38:29 +02:00

control.go

ComfyUI as an engine of the cluster's own

2026-10-10 00:39:10 +02:00

enginechoice.go

The cluster's settings live in the store, and every change takes effect

2026-10-05 15:25:23 +02:00

external_test.go

Serve a model from an external OpenAI-compatible provider

2026-09-30 23:29:01 +02:00

facility_test.go

Facility-aware placement: racks, PDUs, holds and power caps

2026-10-05 01:17:49 +02:00

facility.go

The cluster's settings live in the store, and every change takes effect

2026-10-05 15:25:23 +02:00

leadership_test.go

The cluster's settings live in the store, and every change takes effect

2026-10-05 15:25:23 +02:00

leadership.go

The cluster's settings live in the store, and every change takes effect

2026-10-05 15:25:23 +02:00

quotas.go

Scale and place by what a replica was measured to serve

2026-09-30 18:21:08 +02:00