Skip to content

RESUME_PROMPT.md

Latest commit

 

History

History
98 lines (79 loc) · 3.61 KB

RESUME_PROMPT.md

File metadata and controls

98 lines (79 loc) · 3.61 KB
Raw

You are continuing a Minecraft world generation task. Here is your context:

What's Happening

A full 1:50 scale Minecraft Earth map is being generated on machine alice using the world-generator pipeline. The pipeline has completed OSM preprocessing and QGIS image exports (64,800 tiles total). It is currently running the WorldPainter stage in legacy mode (8 parallel JVM workers), which generates .mca Minecraft region files per tile.

Current progress: ~9,049 / 64,800 tiles done (~14%).

Quick Status Check

Run this first:

ssh alice "echo 'Tiles done:'; ls /mnt/DataDisk2T/world-generator-50/earth_1_50/region/.tile_*.done 2>/dev/null | wc -l; echo 'Disk:'; df -h /mnt/DataDisk2T | tail -1; echo 'WP processes:'; ps aux | grep -c 'wpscript.js'; echo 'Latest log:'; tail -3 /mnt/DataDisk2T/world-generator-50/wp_resume4.log 2>/dev/null || echo 'no resume4 log'"

Key Locations

Thing Path on alice
Source code ~/Documents/SSDDataDisk1T/world-generator/
Active config config-50.yaml in above dir
Output root /mnt/DataDisk2T/world-generator-50/
Final world /mnt/DataDisk2T/world-generator-50/earth_1_50/
WP log /mnt/DataDisk2T/world-generator-50/wp_resume4.log
Launch script ~/Documents/SSDDataDisk1T/world-generator/start_wp.sh
Handover doc ~/projects/mc/world-generator/HANDOVER.md on this machine

Tasks

1. Monitor pipeline

Check tile count and disk every few hours. Pipeline is healthy if:

  • Tile count is increasing
  • ps aux | grep wpscript.js shows 8 JVM processes
  • Disk has >50G free

2. If pipeline stops

ssh alice "cd ~/Documents/SSDDataDisk1T/world-generator && nohup bash start_wp.sh > /mnt/DataDisk2T/world-generator-50/wp_legacyN.log 2>&1 &"

Replace N with next number (currently legacy9.log exists, so use legacy10).

3. If disk <50G

Delete osm directory first (441G, safe — QGIS stage done):

ssh alice "rm -rf /mnt/DataDisk2T/world-generator-50/osm"

4. After WP completes (all 64,800 tiles done)

Run:

ssh alice "cd ~/Documents/SSDDataDisk1T/world-generator && PYTHONPATH=src venv/bin/python -c '
from world_generator.config import load_config
from world_generator.tiles import post_process_map
config = load_config(\"config-50.yaml\")
post_process_map(config)
print(\"Post-processing complete!\")
'"

5. Known errors to ignore

  • minutor ... could not connect to display / xcb — headless env, harmless
  • QPainter::begin: Paint device returned engine == 0 — Qt headless warning
  • Stale .world file ... removing and retrying — auto-recovered, normal

6. If you need to edit code

The repo is cloned locally at ~/projects/mc/world-generator/. Edit here, then:

cd ~/projects/mc/world-generator
git add -A && git commit -m "fix: description"
git push
ssh alice "cd ~/Documents/SSDDataDisk1T/world-generator && git pull"

Estimated Timeline

  • ~55,800 tiles remaining ÷ 8 workers ≈ ~7,000 tiles/worker
  • ~30s per tile average → ~58 hours remaining
  • Expected completion: ~2026-05-17 morning

Troubleshooting

WP processes running but tile count not increasing

Check if workers are stuck on a stale .world file. Look at the log:

ssh alice "tail -30 /mnt/DataDisk2T/world-generator-50/wp_resume4.log"

If stuck, kill and restart with N+1.

All WP processes died

Check dmesg for OOM killer. If OOM, reduce wp_parallel_workers in config-50.yaml to 4.

ssh alice "dmesg | tail -20 | grep -i oom"

Memory check

ssh alice "free -h"

8 workers × 6GB = 48GB. System has 78GB, so fine unless other processes consume memory.