mrsaynothing.devsystemd timer persistent — the missed 3am run
A systemd timer that misses its window while the box is off never fires, and nothing tells you. Measured: Persistent=true caught up at boot, 107 s late.
a stop job is running — the 90-second fuse nobody lit
Not slowness — a fuse. Measured on systemd 262: 90.20s stubborn, 10.10s chosen fuse, 0.03s graceful. Read TimeoutStopSec with show, set it with edit.
ollama context window — the silent 4096 wall
Default context window: 4096 tokens. Paste longer and ollama trims your input, journals a WARN, and answers anyway. The doors out and the RAM bill, measured.
why is google not indexing my site — the 199-page walk-back
Google quietly stopped indexing this site — dead slugs behind 301s. The audit, the 199-row alias map, the 410 graveyard, and the numbers ten days on.
systemd unit is masked — the /dev/null execution
systemctl says your unit is masked and won't start. The mask is one symlink to /dev/null — three commands cure it, and systemd-delta lists every mask.
ollama stop model — the five-minute squat
The chat ends and ollama still holds the model — 2.97 GiB for five measured minutes. Stop doors: ollama stop, keep_alive 0, --keepalive, OLLAMA_KEEP_ALIVE.
how to stop ai crawlers — the census of who obeys
Robots.txt is a sign, not a lock — our 28-day census caught GPTBot on 155 pages without reading it once. The five doors that actually stop AI crawlers, ranked.
KV cache quantization — the bill your context hides
q8_0 keeps the KV cache at 17/32 of f16 and q4_0 at 9/32 — at 32k context the bill is 4608 MiB against a 2382 MiB model. Receipts, crossover, speed tax.
systemd service restart on failure — the hidden breaker
systemd service restart on failure has a fuse: five crashes in ten seconds trip the start limit and park the unit. Bench: the trip, the dead slot, the cures.
what is llms.txt — the answer engines never came
Eight weeks of bot logs: robots.txt drew 343 fetches, our API 111, llms.txt 17. What is llms.txt for? The answer engines never came — receipts inside.
systemd restart service daily — stop cron-scripting it
Restart a systemd service daily with a timer, not a cron line: OnCalendar=daily, RandomizedDelaySec, Persistent. Bench-fired on systemd 262 — receipts inside.
systemd show reason for service restart
Show the reason a systemd service restarted: status shrugs, the journal confesses. Five death codes from a live bench — and the one death it calls success.
ollama speech to text — the ear is the picture door
Ollama ships an audio model that hears, but the API has no working ear. We fed real speech through the image door — receipts, garbles, full token bill.
ollama web search api — the search isn't local
Ollama's web search API shipped September 2025. We knocked keyless tonight: 401 in 0.13 and 0.15 seconds. The key, the limits, and the part that isn't local.
AI crawlers vs search crawlers: the 7-night census
Seven nights, 8,654 bot knocks, 247 human visits. AI crawlers vs search crawlers, measured: Perplexity beat OpenAI 9-to-1 and 65% knocked unnamed.
The Thinking Bill
A thinking model spent 866 tokens and 53 seconds to say “Bug”. The three switches that turn thinking off in Ollama — and the measured bill that says when to flip them.
The 7% Ghost
Best local LLM for coding on 8GB VRAM? Check the fit table first — an 8GB card holds 8.59 decimal GB but only ~7.3 usable GiB. Sizes measured today, KV math shown, one honest fit table — the 9B the tables love fails at 8k.
GGUF VRAM calculator: the three numbers that decide
Weights + KV cache + overhead: the three numbers that decide whether a GGUF quant fits — printed as formulas, checked against four public GGUF files today.
Run LFM3 on Linux with Ollama (voice included)
ollama v0.40.13 brings Liquid AI's LFM3 — text, audio, omni — to Linux and Windows. The 1.7B rung benched on a CPU VM: 17 tok/s, 1.2 GB, one honest catch.
llama.cpp v0.8.0: the same LFM3 week, one layer down
Ollama wrapped LFM3 for pull-and-run the same day; llama.cpp v0.8.0 is where that week actually landed — realtime audio endpoints, six new architectures, sparse attention on a fused mask, MoE experts cached on host RAM.
ollama d1 vs d3 — which decision model do you need?
Two decision models shipped into ollama five days apart: d3 (2b-27b, runs on Linux CPU today) and d1 (2b, Apple-only this week). Which one to call.
Does vLLM support GGUF? The GPU-border rule
Does vLLM support GGUF? Yes — on GPUs only. The honest vLLM–GGUF matrix: six green hardware rows, every CPU row red, and the border rule the guides skip.
Why is my Ollama so slow? Five usual suspects, ranked
Nine times in ten it's CPU offload you didn't notice — ollama ps shows the split in one line. Then: cold reloads after the five-minute keep-alive, an oversized context, a heavier quant than you need, and swap.
What happens when Ollama runs out of VRAM? The offload cliff
Ollama never refuses a model that misses VRAM — it splits it. Fitted layers stay on the GPU, the rest fall back to CPU, and tok/s falls with them. How to see the split in one line, and the three honest fixes, ranked.
ollama d3 refuses to chat: the CPU bench ledger
Ollama's d3 scores tickets instead of talking. The 2b and 4b rungs on an 8-core CPU box: yes 98% in 0.8 s warm, a 64 KiB door fuse, and a verdict that wobbles.
