Ilker ManapandClaude Opus 5.5 5db920bf2c Slurm: a job that asks for no memory gets a share, not the whole node ...

Tried on the PoC: five one-second jobs on an 8-CPU node started 10 s
apart. Each asked for no memory, so each was given the node's all, ran
alone, and the next waited for backfill's pass. Not accounting, as
thought. The partition's DefMemPerCPU is now the leanest node's memory
per CPU.

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

2026-11-11 23:11:24 +01:00

..

account

People sign in with their Active Directory account

2026-10-18 10:59:50 +02:00

accounting

Rate limits counted across the control planes with router.active_active

2026-11-01 17:30:08 +01:00

agent

Slurm's ports for the cluster only, and srun in a range of its own

2026-11-11 22:10:37 +01:00

audit

Standbys that answer the client API themselves, behind router.active_active

2026-11-01 13:47:47 +01:00

auth

Requests missing the cache for the same key at once share one read and one allowance

2026-10-31 09:30:18 +01:00

bench

ctl bench reads a guarded engine's metrics with the engine's key, or says why it cannot

2026-10-05 00:20:25 +02:00

bundle

Offline sites and Kubernetes: llmgw bundle create, import and trust, offline.mirror, server init --bundle; a Helm chart for the control plane

2026-10-05 00:13:39 +02:00

canary

Canary routing: a model's canary takes its share of requests, a session stays on one side, x-llmgw-variant pins one, and the two are compared

2026-10-04 19:56:16 +02:00

cli

Slurm in the console: each node, and lend or take back from its row

2026-11-11 22:53:26 +01:00

clientapi

Pictures, 3D models, a vision model and workflows as tools at /mcp

2026-10-14 13:11:09 +02:00

clientip

Believe X-Forwarded-* only from router.trusted_proxies

2026-10-04 10:29:02 +02:00

comfy

ComfyUI models of one environment share one replica

2026-10-14 13:55:34 +02:00

config

Slurm in the first install: server init --slurm, and slurm.all_nodes

2026-11-11 14:50:35 +01:00

control

Slurm: a new leader with nothing kept waits for the nodes a while

2026-11-11 22:37:52 +01:00

directory

An organization's admin, and forms only for those who may send them

2026-11-08 11:38:06 +01:00

docgen

Slurm in the first install: server init --slurm, and slurm.all_nodes

2026-11-11 14:50:35 +01:00

doctor

A card in the wrong PCIe slot is said so

2026-10-29 08:49:16 +01:00

firewall

The firewall setting reaches the mirror machine and the store's members

2026-10-09 08:53:17 +02:00

gguf

How fast a model answers on the CPU, before it is served

2026-11-03 08:52:15 +01:00

guard

Outside services read answers redacted and whole; a card's own lanes

2026-10-29 13:22:39 +01:00

hf

A repository's files read from every directory, not only its top

2026-10-30 12:16:11 +01:00

join

Bound idle time, header size and request reading on every listener, and connections on the client port

2026-10-04 10:21:15 +02:00

kb

The gateway embeds the documents llmgw-kb sends as text

2026-10-29 21:56:05 +01:00

kvevents

Precise KV-index routing: vLLM KV cache events through the agent, routed by leading blocks held

2026-10-05 01:21:11 +02:00

mail

A forgotten password is set through a one-time link

2026-10-07 13:20:11 +02:00

mcpclient

An MCP client for the company's tool servers

2026-10-20 10:43:35 +02:00

mirror

A mirror download cut off carries on from where it got to, and is tried again

2026-10-30 17:47:32 +01:00

modelplan

The serve forms start at the longest context that fits

2026-11-08 14:38:41 +01:00

notify

Several keys for one outside provider

2026-10-17 08:56:30 +02:00

obs

Tracing beyond the router: reconcile passes, placement and a replica's start traced down to the node

2026-10-05 00:03:17 +02:00

onboard

A node's own readiness and its multi-node readiness are kept apart

2026-10-07 13:11:16 +02:00

pgha

A lost store machine is given up on in seconds, and a new leader waits for its judge

2026-10-15 20:33:13 +02:00

preflight

A machine too small for the models it would serve is said so first

2026-10-25 16:19:00 +01:00

promq

Show usage over time, per organization, team, project and person

2026-09-28 08:11:31 +02:00

release

Offline sites and Kubernetes: llmgw bundle create, import and trust, offline.mirror, server init --bundle; a Helm chart for the control plane

2026-10-05 00:13:39 +02:00

router

Rate limits counted across the control planes with router.active_active

2026-11-01 17:30:08 +01:00

safedial

gofmt what the security work left unformatted

2026-10-04 12:22:51 +02:00

scim

An organization's admin, and forms only for those who may send them

2026-11-08 11:38:06 +01:00

secrets

Sign-in is set up from the console, and a directory tried before it is kept

2026-10-19 12:55:27 +02:00

settings

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

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

sim

Onboard a node with a script, and find two examples that never worked

2026-09-16 12:28:36 +02:00

slurm

Slurm: a job that asks for no memory gets a share, not the whole node

2026-11-11 23:11:24 +01:00

sourcecheck

Keep the control plane's fetches off link-local addresses, and see through NAT64, 6to4 and proxies

2026-10-04 10:48:01 +02:00

store

An organization's admin, and forms only for those who may send them

2026-11-08 11:38:06 +01:00

strata

Strata's DeepSeek not offered on cards its engine was not built for, and refused there by the agent

2026-10-30 13:00:21 +01:00

tlsconf

Put etcd and Patroni's API on TLS, and move a running cluster there in maintenance mode

2026-10-04 12:07:17 +02:00

version

Switch the project language to English

2026-09-16 01:15:21 +02:00

vscodeext

VS Code's sign-in keeps its place; recordings checked for secrets

2026-10-29 11:40:19 +01:00

web

The serve forms start at the longest context that fits

2026-11-08 14:38:41 +01:00