Skip to content

Storage Consolidation & Backup Strategy

Generated: 2026-02-01 | Primary: OWC ThunderBlade X12 (21 TB TB5 NVMe)

Current State

Volume Size Used Free Status
ThunderBlade X12 21 TB 7.0 TB 14 TB Pegasus migration ~85%
Promise Pegasus R6 9.1 TB 8.1 TB 1.0 TB 3 rsyncs remaining
Mac HD 1.8 TB ~487 GB 1.0 TB Partially cleaned
13 External HDDs Unknown Unknown — Not yet inventoried

ThunderBlade Current Usage

Folder Size Status
Personal/ 3.5 TB Photos, Family Videos, Dad-Archive (in progress)
_Archive/ 2.0 TB Old backups, legacy items
Media/ 1.0 TB Movies (complete)
Business/ 297 GB DXC, CSC (complete)
Technical/ 213 GB DB/SQL, training, scripts
AI-To-Categorize/ 37 GB Downloads + Mac HD cleanup
_Exports/ 2.8 GB Google Takeout, Notion
Email-Archives/ 1.3 GB PST + extracted
Active/ 8 KB Empty — for current work

7.0 TB of 21 TB used (33%)

Goals

  1. ~~Migrate Pegasus → ThunderBlade~~ 85% complete
  2. Verify all Pegasus transfers
  3. Clean up Google Drive & Dropbox
  4. Import 13 external hard drives onto ThunderBlade
  5. Establish clean taxonomy for unified archive
  6. Set up Corbie AI Personal for tagging and semantic search
  7. Repurpose Pegasus R6 as local backup (Tier 1)
  8. Decommission Pegasus + external drives

Phase Progress

Phase Description Status
Phase 1 ThunderBlade folder structure :white_check_mark: Complete
Phase 2 Pegasus migration :hourglass: 85% — 3 rsyncs finishing
Phase 3 Mac HD cleanup :hourglass: Partial
Phase A Finish Pegasus migration (autonomous) :hourglass: Monitor running
Phase B Verify Pegasus transfers Pending
Phase C Google Drive & Dropbox cleanup Pending
Phase D Inventory 13 external drives Pending
Phase E Unified taxonomy Pending
Phase F Corbie AI Personal integration Pending
Phase G Backup strategy :warning: GAPS FOUND

:warning: Backblaze Gaps (URGENT)

Current Configuration Problems

Source: /Library/Backblaze.bzpkg/bzdata/bzinfo.xml

Issue Impact Fix
ThunderBlade NOT in backup set 7 TB has NO off-site backup Add ThunderBlade volume in Backblaze prefs
Drobo.dmg EXCLUDED by file type Father's irreplaceable archive NOT backed up Remove .dmg from global excludes
Cox upload = 35 Mbps 7 TB initial backup = ~19 days See internet options below
Cox 1.25 TB/month data cap 7 TB = 5.6x cap = overage fees Add unlimited data ($50/mo)

Current config: Account bfrichot@gmail.com (Personal, $99/yr) · Only volume: Mac HD (/) — ThunderBlade NOT included · 10 threads · 100% throttle · Global excludes: .dmg .iso .vhd .vmdk

Internet & Upload Reality

Cox Plans

Cox Plan Download Upload Monthly Data Cap
Fast (300 Mbps) 300 Mbps 10 Mbps $50/mo 1.25 TB
Go Super Fast (1 Gig) ← current 940 Mbps 35 Mbps $100/mo 1.25 TB
2 Gig Extra 2,000 Mbps 100 Mbps $115/mo 1.25 TB
Cox Fiber (if available) 1,000 Mbps 1,000 Mbps ~$100/mo No cap
Unlimited data add-on — — +$50/mo Unlimited

Initial Backup Time (7 TB to Backblaze)

Upload Speed Time to Backup 7 TB Notes
35 Mbps (current Cox 1 Gig) ~19 days Cable upload is the bottleneck
100 Mbps (Cox 2 Gig) ~6.5 days 3x faster, only +$15/mo
1,000 Mbps (Cox Fiber) ~16 hours Symmetrical — check availability

Recommendation: Check Cox Fiber availability. If unavailable, upgrade to 2 Gig Extra ($115/mo) during initial backup for 3x speed. After initial backup completes, daily incrementals are small and 35 Mbps is sufficient.

Recovery Tiers

Tier 0 — ThunderBlade X12

  • Recovery: 6,600 MB/s (instant)
  • Capacity: 21 TB
  • Cost: $0 (owned)
  • Primary storage. All data lives here.

Tier 1 — Pegasus R6 (Local)

  • Recovery: ~276 MB/s
  • Capacity: 9.1 TB (RAID-5)
  • Cost: $0 + electricity
  • Fast local recovery for irreplaceable data.

Tier 2 — Backblaze Personal

  • Recovery: 940 Mbps (download) or USB drive shipped free
  • Capacity: Unlimited
  • Cost: $99/yr
  • Off-site disaster recovery. USB shipment for full restore.

Tier 3 — Google Drive

  • Recovery: 940 Mbps (download)
  • Capacity: Plan-limited
  • Cost: Current plan
  • Active business docs only.

Backblaze restore options: Web download (zip up to 500 GB per request) · USB drive shipped free (best for full 7 TB restore) · Restore app

Recovery Time Objectives

Data Category Size RTO Backup Tier Recovery Method
Irreplaceable (Dad-Archive, Family Videos, Photos) ~4.4 TB < 4 hours Tier 1 + 2 Pegasus rsync restore
Business-critical (Active, Business, Email) ~300 GB < 1 hour Tier 1 + 2 Pegasus rsync restore
Important (Technical, Media excl. Movies) ~1.2 TB < 24 hours Tier 1 + 2 Pegasus or Backblaze
Replaceable (Movies, _Archive, Old-Software) ~3 TB Days/weeks Tier 2 only Backblaze USB shipment
Active docs < 5 GB Instant Tier 3 Google Drive sync

Pegasus R6 as Tier 1 Backup (9.1 TB)

Content Size Priority Reason
Personal/Dad-Archive/ 308 GB P0 — MUST Father's archive (irreplaceable)
Personal/Family-Videos/ 522 GB P0 — MUST Irreplaceable family moments
Personal/Photos/ 3.5 TB P0 — MUST 25 years of photos
Business/ 297 GB P1 — HIGH Career history, tax/legal
Email-Archives/ 1.3 GB P1 — HIGH Communication history
Active/ varies P1 — HIGH Current work
Technical/ 213 GB P2 — MED Training, scripts, code
_Exports/ 2.8 GB P2 — MED Google Takeout, Notion
Subtotal P0-P2 ~4.8 TB Fits easily in 9 TB
Media/Music/, GoPro/ varies P3 — LOW Can re-rip or redownload

Excluded from Pegasus (Backblaze only): Movies (961G), _Archive (2 TB), AI-To-Categorize

Backup Schedule

Pegasus (Tier 1)

  • Nightly: rsync --delete of P0+P1 folders
  • Weekly: rsync of P2+P3 folders
  • Method: LaunchAgent or cron job

Backblaze (Tier 2)

  1. Add ThunderBlade volume to backup set
  2. Remove .dmg from global excludes
  3. Verify Cox unlimited data is active
  4. Let initial backup run (~6-19 days)
  5. Ongoing: continuous incremental

Cost Summary

Item Monthly Annual What You Get
Backblaze Personal $9 $99 Unlimited backup (Mac + attached drives)
Cox 1 Gig (current) $100 $1,200 940 down / 35 up
Cox Unlimited Data +$50 $600 Remove 1.25 TB cap
Pegasus electricity ~$4 ~$50 On 24/7 for scheduled backups
Google Drive current current Active docs sync
Total (steady state) ~$163 ~$1,949 3-2-1 compliant

Optional: Cox 2 Gig Extra (+$15/mo for 100 Mbps upload) · Cox Fiber (~$100/mo symmetrical 1 Gbps, if available)

Unified Taxonomy

ThunderBlade X12/
├── Active/              # Current work, hot data
├── Business/            # All employer/client work by company
│   ├── CSC/
│   ├── DXC/
│   ├── IT-Raven/
│   ├── Freelance/
│   └── <new-employer>/  # Extensible
├── Personal/
│   ├── Family-Videos/
│   ├── Dad-Archive/     # PRECIOUS — Drobo.dmg
│   ├── Photos/
│   ├── Voice-Memos/
│   ├── BSA-Scouts/
│   ├── Documents/
│   └── Fitness/
├── Media/
│   ├── Movies/
│   ├── Music/
│   ├── GoPro/
│   ├── Screenshots/
│   └── Podcasts/
├── Technical/
│   ├── Database-SQL/
│   ├── Oracle/
│   ├── Training/
│   ├── Scripts/
│   ├── ML-Projects/
│   ├── Code-Repos/
│   └── VMs/
├── Email-Archives/
├── AI-To-Categorize/    # Unsorted → Corbie AI processes
├── _Archive/
│   ├── Old-Backups/
│   ├── Old-Software/
│   ├── Legacy/
│   └── External-Drives/ # Raw imports
├── _Exports/
└── _Imports/            # Staging area (temporary)

Taxonomy Rules

  1. Company = subfolder of Business/ — any employer, client, or business entity
  2. Time Machine / disk images → _Archive/Old-Backups/
  3. Photos from any era → Personal/Photos/ (subfolder by source or year)
  4. Videos → Personal/Family-Videos/ or Media/Movies/ depending on content
  5. Music → Media/Music/
  6. Unknown / mixed → AI-To-Categorize/ for Corbie AI to process
  7. Raw drive import that doesn't fit → _Archive/External-Drives/\<DriveName>/
  8. Duplicates → delete (after verification)

13 External Hard Drives

Contents are unknown. This is a discovery phase.

Per-Drive Workflow

  1. Mount and catalog — top-level directory listing, sizes, depth-2 tree
  2. Generate manifest — file list saved to _Imports/<DriveName>-manifest.txt
  3. Identify duplicates — cross-reference against existing ThunderBlade content
  4. Classify content into taxonomy categories
  5. rsync with --ignore-existing — skip exact matches
  6. fdupes second pass — find duplicates by content hash
  7. Log all dedup decisions to _Imports/<DriveName>-dedup-log.txt

Corbie AI Personal Integration

  1. Crawl ThunderBlade folder structure
  2. Extract metadata: file type, date, size, path, EXIF (photos), text content (documents)
  3. Generate embeddings: Azure OpenAI text-embedding-3-small (1536d)
  4. Store vectors in Qdrant — new collection: PERSONAL_ARCHIVE
  5. Tag files with taxonomy labels
  6. Enable semantic search: "find photos from Dad's birthday 2017", "DXC Oracle migration docs"

Implementation

  • agent/tools.py — add personal archive search tool
  • azure_functions/ — add /personal/index and /personal/search endpoints
  • tools/personal_archive_crawler.py — new file crawler

Execution Order

  1. :hourglass: Wait for 3 remaining rsyncs (autonomous monitor)
  2. Verify all Pegasus transfers (Phase B)
  3. Fix Backblaze — add ThunderBlade, remove .dmg exclude
  4. Verify Backblaze is backing up ThunderBlade
  5. Clean up Google Drive & Dropbox (Phase C)
  6. Create _Imports/ staging on ThunderBlade
  7. Inventory each external drive one at a time (Phase D)
  8. Import non-duplicate content into taxonomy (Phase E)
  9. Set up Corbie AI Personal crawler and indexer (Phase F)
  10. Repurpose Pegasus as local backup with scheduled rsync (Phase G)
  11. Decommission external drives after full verification

Immediate Actions

Do These Now

  1. Open Backblaze System Preferences → Add ThunderBlade X12 to backup set
  2. Edit Backblaze exclude list → Remove .dmg to protect Drobo.dmg
  3. Check Cox Fiber availability at your address
  4. Verify Cox unlimited data is already active
  5. Monitor initial backup in Backblaze dashboard

Critical Notes

  • Do NOT delete anything from source drives until ThunderBlade copies are verified
  • Do NOT wipe Pegasus until BOTH ThunderBlade AND Backblaze have confirmed copies
  • Drobo.dmg is father's archive (deceased) — verify checksum after every operation
  • Drobo.dmg is currently EXCLUDED from Backblaze by global file type filter — must fix
  • ThunderBlade has 14 TB free — plenty of room for all 13 drives
  • Pegasus spinning disks max ~65 MB/s via rsync; ThunderBlade NVMe = 6,600 MB/s peak
  • Mac is M4 Max with Thunderbolt 5 — no bottleneck on Mac or ThunderBlade side
  • Cox "1 Gig" = 940 Mbps down / 35 Mbps up (asymmetric cable, NOT fiber)