e4c9dda0fd
- Marked submodules ai-mcp-pmm-controller, explorer-monorepo, and smom-dbis-138 as dirty to reflect recent changes. - Updated documentation to clarify operator script usage, including dotenv loading and task execution instructions. - Enhanced the README and various index files to provide clearer navigation and task completion guidance. Made-with: Cursor
4.7 KiB
Adding a third and/or fourth R630 before migration — decision guide
Context: You are about to balance load by migrating containers from r630-01 to r630-02 (and optionally ml110). You asked whether it makes sense to add a third and/or fourth R630 to Proxmox before starting that migration.
1. You may already have a third and fourth R630
The repo documents r630-03 (192.168.11.13) and r630-04 (192.168.11.14):
- Status: Powered off; not currently in the Proxmox cluster (only ml110, r630-01, r630-02 are active).
- Hardware (per report): Dell R630, 512 GB RAM each, 2×600 GB boot, 6×250 GB SSD.
- Issues when last used: Not in cluster, SSL/certificate issues, and others — all with documented fixes.
If these servers are still available and you are willing to power them on and fix them:
- Add them to the cluster first (power on → fix SSL/join cluster per reports/R630_03_04_POWER_ON_ISSUES_AND_FIXES.md).
- Then you have four Proxmox nodes (ml110 + r630-01, -02, -03, -04) or three R630s + ml110. Migration can then spread workload to r630-03 and r630-04 as well, instead of only to r630-02 and ml110.
- That gives more headroom and better HA (see below) without buying new hardware.
If r630-03/04 are decommissioned or unavailable: Treat this as “add new R630(s)” below.
2. Does it make sense to add a third/fourth R630 (or bring r630-03/04 online) before migration?
Yes, it can make sense, depending on goals.
| Goal | Add 3rd/4th R630 before migration? | Notes |
|---|---|---|
| Reduce load on r630-01 quickly | Optional | Migration to existing r630-02 (and ml110) already helps. You can migrate first and add nodes later. |
| More headroom long term | Yes | With 3–4 R630s (+ ml110), workload is spread across more nodes; no single node is as hot as r630-01 today. |
| Proxmox HA + Ceph | Yes (3 min, 4 better) | Per PROXMOX_HA_CLUSTER_ROADMAP.md: 3 R630s minimum for HA + Ceph; 4 R630s better for Ceph recovery. You currently have 2 R630s + ml110; adding a 3rd (and 4th) R630 aligns with that. |
| Avoid “just moving the problem” | Yes | If you only move workload to r630-02, r630-02 may become the new bottleneck. Adding nodes gives more capacity so migration actually balances. |
| Cost / complexity | Your call | New hardware = cost and setup. Bringing r630-03/04 back = no new purchase, but time to power on, fix, and join cluster. |
Practical recommendation:
-
If r630-03 and/or r630-04 exist and are usable:
Power them on and add them to the cluster first, then run migration. You get a 4- (or 5-) node cluster and can move workload to r630-03 and r630-04 as well as r630-02. Use reports/R630_03_04_POWER_ON_ISSUES_AND_FIXES.md for the fix sequence. -
If you do not have extra R630s (or they’re gone):
Migration first is still valid: move candidates from r630-01 to r630-02 (and optionally ml110) per PROXMOX_LOAD_BALANCING_RUNBOOK.md. That reduces r630-01 load with no new hardware. If after that you still want more capacity or HA, then add a 3rd (and 4th) R630. -
If you are buying new R630s:
For HA + Ceph, the docs recommend at least 3 R630s (4 is better). So adding a third R630 is the minimum for that path; a fourth improves Ceph and spread. You can add them before or after the current migration; adding before gives more migration targets.
3. Order of operations (suggested)
| Scenario | Order |
|---|---|
| r630-03 / r630-04 exist and you will use them | 1) Power on r630-03 (and -04). 2) Fix and join cluster. 3) Run load-balance migration (including to r630-03 / -04 if desired). |
| No extra R630s yet; migration only | 1) Run migration r630-01 → r630-02 (and optionally ml110). 2) Re-check load. 3) If needed, plan 3rd/4th R630. |
| Buying new 3rd/4th R630 | 1) Install Proxmox and join cluster. 2) Run migration so new nodes take part of the workload. |
4. References
- r630-03/04 issues and fixes: reports/R630_03_04_POWER_ON_ISSUES_AND_FIXES.md
- HA and how many R630s: PROXMOX_HA_CLUSTER_ROADMAP.md — “At least 3 R630s for full HA with Ceph; 4 is better.”
- Load-balance migration: PROXMOX_LOAD_BALANCING_RUNBOOK.md
- 13-node long-term plan: R630_13_NODE_DOD_HA_MASTER_PLAN.md