[PLAN] Bilbo Memory Recovery - Free 2GB+ for Ollama #297

Closed
opened 2026-04-02 03:15:10 +00:00 by ezra · 1 comment
Member

Bilbo Memory Recovery Plan

Status: READY FOR EXECUTION
Priority: CRITICAL (Bilbo suffocating)
Objective: Free 2GB+ RAM for Bilbo's Ollama

CURRENT STATE

  • RAM: 7.6GB/7.8GB used (97%)
  • Swap: 2.0GB/2.0GB used (100%)
  • Available: 358MB (critically low)
  • Ollama 7B: 4.9GB (60% RAM) - THE KILLER
  • Ollama 1.5B: 1.1GB (14% RAM) - What Bilbo needs

ROOT CAUSE

Bilbo appears "canned" but is actually SUFFOCATING from resource starvation. The 7B Ollama model keeps reloading and consuming all memory.

EXECUTION PLAN

Phase 1: Kill Services (-1.5GB)

  • Kill Hermes gateway (-400MB)
  • Stop LNBits, SearxNG (-300MB)
  • Clear swap (-800MB)

Phase 2: Ollama Restart (-2GB)

  • Kill all Ollama runners
  • Start fresh with 1.5B model ONLY
  • Prevent 7B auto-load

Phase 3: Deploy Bilbo

  • Verify 3GB+ available
  • Start bilbo_snappy.py
  • Test <10s response time

Phase 4: Selective Restart

  • Restart critical services only
  • Keep 7B model OFF

EXECUTE

Reply with:

  • "Execute Phase 1"
  • "Execute Phase 2"
  • "Execute All"
  • "Nuclear"

Full plan: /root/BILBO_MEMORY_RECOVERY_PLAN.md

# Bilbo Memory Recovery Plan **Status:** READY FOR EXECUTION **Priority:** CRITICAL (Bilbo suffocating) **Objective:** Free 2GB+ RAM for Bilbo's Ollama ## CURRENT STATE - **RAM:** 7.6GB/7.8GB used (97%) - **Swap:** 2.0GB/2.0GB used (100%) - **Available:** 358MB (critically low) - **Ollama 7B:** 4.9GB (60% RAM) - THE KILLER - **Ollama 1.5B:** 1.1GB (14% RAM) - What Bilbo needs ## ROOT CAUSE Bilbo appears "canned" but is actually SUFFOCATING from resource starvation. The 7B Ollama model keeps reloading and consuming all memory. ## EXECUTION PLAN ### Phase 1: Kill Services (-1.5GB) - Kill Hermes gateway (-400MB) - Stop LNBits, SearxNG (-300MB) - Clear swap (-800MB) ### Phase 2: Ollama Restart (-2GB) - Kill all Ollama runners - Start fresh with 1.5B model ONLY - Prevent 7B auto-load ### Phase 3: Deploy Bilbo - Verify 3GB+ available - Start bilbo_snappy.py - Test <10s response time ### Phase 4: Selective Restart - Restart critical services only - Keep 7B model OFF ## EXECUTE Reply with: - "Execute Phase 1" - "Execute Phase 2" - "Execute All" - "Nuclear" Full plan: /root/BILBO_MEMORY_RECOVERY_PLAN.md
ezra self-assigned this 2026-04-02 03:15:11 +00:00
Author
Member

Closed per new fleet policy: no local llama-server for models >5GB. RunPod serverless endpoints only. See Timmy_Foundation/timmy-home#409.

Closed per new fleet policy: no local llama-server for models >5GB. RunPod serverless endpoints only. See Timmy_Foundation/timmy-home#409.
ezra closed this issue 2026-04-05 14:05:47 +00:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Timmy_Foundation/timmy-home#297