proxmox/docs/04-configuration/MIGRATION_PLAN_R630_01_DATA.md

Migration Plan: r630-01 Data Pool (72% → target <70%)

Last updated: 2026-02-28

Phase 1 completed (2026-02-28): Migrated 8 CTs from data → thin1 on r630-01. data now 65.8% (was 72%); thin1 50.6% (was 43.5%). Migrated VMIDs: 10233, 10120, 10100, 10101, 10235, 10236, 7804, 8640.

Goal: Reduce r630-01 data pool usage (currently ~65.8%, 280G pool) so it stays below 85% and has headroom. Options: move CTs to r630-01 thin1 (same host) or to r630-02 thin5 (cross-host).

---

1. Current layout (r630-01)

Pool	Size	Used%	Free (approx)	Notes
data	280G	65.8% ✓	~100G	After Phase 1 migrations; target <70% met.
thin1	208G	50.6%	~107G	Now holds migrated CTs from Phase 1.

CTs on data (by size, VMID = container):

VMID	Allocated	Role / name (typical)	Migration priority
2500–2505	200G each	RPC (alltra etc.)	Low (critical RPC); move only if needed.
2101	200G	Core RPC	Do not move (deploy/admin).
1000, 1001, 1002	100G each	Validators	Do not move (consensus).
1500, 1501, 1502	100G each	Sentries	Prefer stay; move only if necessary.
10130, 10150, 10151	50–100G	Services	Good candidates (Phase 2).
8640, 8642	50G each	Services	Good candidates.
7804	20G	Gov/portal	Good candidate.
10235, 10236	10–20G	Small services	Good candidates.
10232, 10233	1–3G	Small	Easy wins.
10100, 10101	8G each	Small	Easy wins.
10120	4G	Small	Easy win.

---

2. Migration options

Option A: Same host — data → thin1 (r630-01)

• Pros: No node change; no IP/network change; quick (stop → vzdump → destroy → restore to thin1 → start).
• Cons: thin1 has ~118G free; cannot move all 200G CTs without exceeding thin1.
• Best for: Smaller CTs (total ~50–80G) to free space on data without touching validators/core RPC.

Option B: Cross host — r630-01 → r630-02 thin5

• Pros: thin5 is empty (237G); can move one or two large CTs (e.g. 200G each).
• Cons: Node migration (IP may change unless static); service dependency and DNS/load balancer updates.
• Best for: One or two 200G RPC CTs (e.g. 2503, 2504, 2505) if you need to free a lot of data in one go.

---

3. Recommended phases

Phase 1 — Same host (data → thin1)

Goal: Free ~50–80G on data (get data from 72% to ~55–60%) by moving smaller CTs to thin1.

Candidates (order by size, smaller first for low risk):

1. 10232 (1G), 10233 (3G), 10120 (4G)
2. 10100 (8G), 10101 (8G), 10235 (10G)
3. 10236 (20G), 7804 (20G)
4. Optionally one of: 8640 (50G), 8642 (50G), or 10130 (50G)

Steps per CT (run on r630-01 or via SSH):

# On r630-01 (192.168.11.11). Replace VMID with actual (e.g. 10232).
VMID=10232
ssh root@192.168.11.11 bash -s << 'EOF'
  set -e
  pct stop $VMID
  vzdump $VMID --storage local --compress gzip --mode stop --remove 0
  BACKUP=$(ls -t /var/lib/vz/dump/vzdump-lxc-$VMID-*.tar.gz | head -1)
  pct destroy $VMID --force
  pct restore $VMID $BACKUP --storage thin1
  rm -f $BACKUP
  pct start $VMID
EOF

Automation: Run for each VMID:

bash scripts/maintenance/migrate-ct-r630-01-data-to-thin1.sh <VMID>   # e.g. 10232
bash scripts/maintenance/migrate-ct-r630-01-data-to-thin1.sh 10232 --dry-run  # preview

Suggested order: 10232, 10233, 10120, 10100, 10101, 10235, 10236, 7804, then 8640 or 8642.

---

Phase 2 — If data still >70%: move one large CT to r630-02 thin5

Goal: Free ~200G on r630-01 data by migrating one 200G CT to r630-02 and placing it on thin5.

Candidates: 2503, 2504, or 2505 (non–core RPC; confirm which can be moved without breaking dependencies).

Steps:

1. Confirm the CT can run on r630-02 (network, DNS, NPM proxy, etc.).
2. Migrate (cross-node):
   pct migrate <VMID> r630-02 --storage thin5
   Or: backup on r630-01, copy backup to r630-02, restore on r630-02 with --storage thin5, then destroy on r630-01.
3. Update any static IPs, DNS, or proxy config pointing to the old host.
4. Verify service and re-run audit-proxmox-rpc-storage.sh.

---

4. Scripts to use / create

Action	Script / command
List CTs on data	ssh root@192.168.11.11 'lvs -o lv_name,lv_size,pool_lv --noheadings pve | grep " data "'
Same-host migration (data → thin1)	New script or inline: vzdump → pct destroy → pct restore --storage thin1 → pct start
Cross-host migration	pct migrate <VMID> r630-02 --storage thin5 (or backup/restore to r630-02).
Verify after	bash scripts/audit-proxmox-rpc-storage.sh and pvesm status on both hosts.

---

5. Checklist before migration

• Backup or snapshot critical CTs before Phase 1.
• Run prune on r630-01 data CTs again if not done recently: fix-storage-r630-01-and-thin5.sh.
• Ensure no critical operations (deploy, consensus) during the migration window.
• For Phase 2: document which VMID is moved and update DNS/NPM/load balancer.

---

6. Success criteria

• Phase 1: r630-01 data pool <70% (ideally 55–65%); thin1 still <75%.
• Phase 2 (if run): r630-01 data <65%; r630-02 thin5 used by migrated CT(s); all services verified.

---

7. Reference

• r630-01 data: 280G, 72% used; LVs: 1000, 1001, 1002, 1500–1502, 2101, 2500–2505, 10100, 10101, 10120, 10130, 10150, 10151, 10232, 10233, 10235, 10236, 7804, 8640, 8642.
• r630-01 thin1: 208G, 43.5% used; ~118G free.
• r630-02 thin5: 237G, 0% used (empty).
• Storage docs: STORAGE_RECOMMENDATIONS_BY_FILL_RATE.md, PHYSICAL_DRIVES_AND_CONFIG.md.