administrator
9c675459fa
.mcp.json: vsechny repo MCP servery na U:\janssen (na Z230 vyresi junction U:\janssen -> U:\PythonProject\Janssen), owa/vbcz-email zustavaji Z230-only. TrilliumMCP: COMPUTERNAME fallback na socket.gethostname(). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
17 lines
1.5 KiB
Markdown
17 lines
1.5 KiB
Markdown
---
|
||
name: project-seaweedfs
|
||
description: "SeaweedFS na Tower1 (192.168.1.50) — kontejner, rozdělení dat, noční backup filer metadat na tower .76"
|
||
metadata:
|
||
node_type: memory
|
||
type: project
|
||
originSessionId: 958cbde4-e6aa-463e-8d0e-cd7c3f7bf9ec
|
||
---
|
||
|
||
SeaweedFS „na hraní" na Tower1 Unraid (192.168.1.50, root, heslo Vlado7309208104++; SSH jen heslem, klíč nejde). Postaveno 2026-06-12.
|
||
|
||
- Kontejner `seaweedfs` (chrislusf/seaweedfs, host network), `weed server -filer -s3`: master :9333, volume :8080, filer :8888, S3 :8333 (anonymní). Limit 1 GB/volume, max 50.
|
||
- Bloby: share `SeaweedFS` → `/mnt/user/SeaweedFS/data`, array disk9–11 pod paritou (useCache=no). Filer leveldb + master raft: `/mnt/cache2tb0225dec/appdata/seaweedfs/` (SSD, bez parity).
|
||
- Noční backup filer metadat: cron 03:30 na Tower1 (`/boot/config/plugins/dynamix/seaweedfs-backup.cron`) spouští `/boot/config/scripts/seaweedfs_meta_backup.sh` — `fs.meta.save` přes weed shell → scp na tower 192.168.1.76 do `/mnt/user/Backup/Critical/SeaweedFS/daily/` (drží 7), neděle navíc `weekly/` (drží 4). Log `/var/log/seaweedfs_backup.log`. Restore: `fs.meta.load`.
|
||
- Tower (.76, root heslo 7309208104) a Tower1 mají vyměněné SSH klíče — z Tower1 na tower jde psát přímo.
|
||
- Pozor: Unraid flash `/boot` je VFAT (case-insensitive) — share cfg `seaweedfs.cfg` vs `SeaweedFS.cfg` je tentýž soubor. Share nastavení za běhu mění `emcmd "query&string"` (jeden argument, nutné `shareNameOrig` + `cmdEditShare=Apply`), live stav v `/var/local/emhttp/shares.ini`.
|