Ilker ManapandClaude Opus 5.5 d864438272 A model that failed to start is placed elsewhere next ...

The same inputs place a replica on the same node and card, so an on-demand
model that failed to load there was started there again after every
backoff until it was parked. The reconciler remembers where a model's
replicas failed and places the next away from them when anything else
fits (Scheduler.Away); with nowhere else, there again.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

2026-11-10 09:22:11 +01:00

..

cpu_test.go

Reserve CPU cores and require memory for CPU models

2026-09-30 17:43:13 +02:00

disk_test.go

Refuse a node with no room, and take back only what llmgw put there

2026-09-18 09:51:05 +02:00

facility_test.go

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

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

facility.go

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

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

mig_test.go

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

placement_test.go

A node given a role takes only what asks for it

2026-10-19 17:52:30 +02:00

placement.go

A model that failed to start is placed elsewhere next

2026-11-10 09:22:11 +01:00

topology.go

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

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