2.0 KiB

Raw Permalink Blame History

HiveOps Project Memory

Key Facts

CLAUDE.md created at /source/hiveops-src/CLAUDE.md — read this first on every session
Monorepo with 16 sub-projects, all in /source/hiveops-src/
Production server (services): 173.231.252.40 (SSH: hiveops@173.231.252.40)
Production DB server: 173.231.252.45
Production CDN server: 173.231.252.43 (SSH: hiveops@173.231.252.43) — cdn.bcos.cloud + bcos.cloud (DNS confirmed)
Production registry: registry.directlx.dev
Local registry: 192.168.200.200:5000
Production domain: *.bcos.cloud
Git user: directlx <directlx.dev@gmail.com>

Deployment Paths (Production)

Services: ~/hiveops/hiveops-openmetal/instances/services/ on 173.231.252.40
Browser CDN: ~/hiveops/hiveops-openmetal/instances/browser/ on 173.231.252.43
Database: shared/database/ on 173.231.252.45

Browser CDN Release

Use ./scripts/release-browser.sh <version> <win_exe> <linux_appimage> [deb] Files go in downloads/, script updates downloads/browser/latest.json No git on production — scp files directly to 173.231.252.43:

Downloads dir: ~/hiveops/hiveops-openmetal/instances/browser/downloads/
latest.json: ~/hiveops/hiveops-openmetal/instances/browser/downloads/browser/
HTML dir: ~/hiveops/hiveops-openmetal/instances/browser/html/
Files are bind-mounted; nginx serves them immediately, no restart needed
After nginx config change: docker exec hiveops-cdn-nginx nginx -s reload

CDN URL Behaviour (confirmed working)

bcos.cloud → serves html/status.html (HiveOps Production Status — API/incident health)
cdn.bcos.cloud/ → 301 to cdn.bcos.cloud/downloads/
cdn.bcos.cloud/downloads/ → browser installer download page (reads latest.json)
cdn.bcos.cloud/downloads/browser/latest.json → current browser version manifest

Sub-project CLAUDE.md locations

hiveops-agent, hiveops-browser, hiveops-incident, hiveops-mgmt, hiveops-tools/hiveops-generator all have their own CLAUDE.md

2.0 KiB Raw Permalink Blame History