Help: All n8n versions after 1.95.3 fail to start

[Lukasz_Schab]

Help: n8n 1.98.2 Crashes During Startup After Update from 1.95.3

Hi n8n Community!

I’m experiencing issues after updating from n8n 1.95.3 to 1.98.2 and hoping someone might have encountered similar problems or can provide guidance.

My Setup

• Infrastructure: Ubuntu 22.04 LTS on VPS (8 vCPU, 32GB RAM, 400GB NVMe)
• Docker: Using docker-compose with PostgreSQL 15
• Previous version: 1.95.3 (working perfectly)
• Target version: 1.98.2 (latest)
• Configuration: Setup with 25 credentials configured

What I Did

1. Clean update process: Changed image: n8nio/n8n:1.95.3 to image: n8nio/n8n:latest in docker-compose.yml
2. Proper shutdown: docker-compose down → docker-compose up -d
3. Full backup: Database, credentials, and configuration backed up before update

The Problem

After update, n8n crashes during startup with these symptoms:

Logs showing:

2025-06-24T08:48:17.495Z | error | Last session crashed
2025-06-24T08:48:27.501Z | info | Initializing n8n process
2025-06-24T08:48:27.623Z | debug | Lazy-loading nodes and credentials from n8n-nodes-base {"nodes":481,"credentials":384}
2025-06-24T08:48:27.634Z | debug | Lazy-loading nodes and credentials from @n8n/n8n-nodes-langchain {"nodes":93,"credentials":21}

Then it stops/hangs indefinitely

Container status:

• Status: Up X minutes (unhealthy)
• Version check: Returns 1.98.2 when container responds
• Web interface: Returns 502 Bad Gateway
• Health check: Fails consistently

What I’ve Tried

1. Multiple restarts: docker-compose restart n8n
2. Clean shutdown/startup: Full docker-compose down/up cycle
3. Waiting longer: Left it for 10+ minutes to complete initialization
4. Resource check: Plenty of RAM/CPU available

Potential Causes I’m Considering

• Memory issues during loading 574 nodes (481 base + 93 LangChain)
• Database migration problems from 1.95.3 to 1.98.2
• LangChain compatibility issues with existing configuration
• Credential compatibility (though logs suggest they’re loading)

My Current n8n Configuration

# Key environment variables:
N8N_EXECUTION_DATA_MAX_SIZE: 64
EXECUTIONS_TIMEOUT: 7200
N8N_LANGCHAIN_ENABLED: true
N8N_AI_ENABLED: true
N8N_NODES_INCLUDE: "@n8n/n8n-nodes-langchain"

Questions for the Community

1. Has anyone experienced similar crashes when updating to 1.98.2?
2. Are there known compatibility issues with LangChain nodes in 1.98.2?
3. Should I increase memory allocation for the container? (currently using Docker defaults)
4. Is there a recommended upgrade path for production setups with many credentials?
5. Any specific debug steps to identify the exact crash point?

Current Status

I’ve rolled back to 1.95.3 for now (which works perfectly), but I’d love to get the latest features working.

Additional Context

• Production environment with business workflows
• No custom nodes installed (using community standard setup)
• Standard PostgreSQL setup (no custom configurations)
• Regular maintenance and updates applied to the host system

Any Help Appreciated!

Whether it’s troubleshooting tips, similar experiences, or suggestions for a different update approach - all input welcome!

Has anyone successfully updated from 1.95.3 to 1.98.2 with a similar setup? What was your process?

Thanks in advance for any insights!

---

P.S. Happy to provide more logs, configuration details, or run additional diagnostic commands if helpful!

[Lukasz_Schab]

Test Results - n8n Version Compatibility

I’ve conducted systematic testing to reproduce this issue and tested the latest development version.

Environment

• OS: Ubuntu 22.04.5 LTS
• Docker: 27.5.1
• Database: PostgreSQL 15
• Test Method: Isolated containers with fresh database instances

Test Results

Version	Status	Duration	LangChain Nodes	Notes
1.95.3	Working	4s	92	Completes successfully
1.96.0	Hang	300s+ timeout	92	Stops at LangChain loading
1.100.0 (next)	Loop	120s+ timeout	94 → 481	Repetitive loading cycle

Detailed Findings

n8n:1.95.3 (Working)

2025-06-24T17:58:09.307Z | debug | Lazy-loading nodes and credentials from @n8n/n8n-nodes-langchain {"nodes":92,"credentials":21}
✅ Continues to completion, health check responds

n8n:1.96.0 (First broken version)

2025-06-24T17:58:XX.XXXZ | debug | Lazy-loading nodes and credentials from @n8n/n8n-nodes-langchain {"nodes":92,"credentials":21}
❌ Hangs at this point, no further progress

n8n:1.100.0 (Latest development)

2025-06-24T18:09:44.882Z | debug | Lazy-loading nodes and credentials from @n8n/n8n-nodes-langchain {"nodes":94,"credentials":21}

Then continuous repetition:
LangChain nodes: 481
LangChain nodes: 481
LangChain nodes: 481
[Continues indefinitely]

Observations

1. Breaking point confirmed: Issue begins with version 1.96.0
2. Pattern evolution: Problem changed from hang (1.96.0+) to infinite loop (1.100.0)
3. Node count increase: From ~92-94 LangChain nodes to 481 total nodes in loop
4. Persistence: Issue remains unresolved in latest development version

The 481 node count suggests the loading mechanism may be processing all available nodes repeatedly rather than completing the LangChain-specific loading.

All tests used identical configuration with clean PostgreSQL instances and full debug logging enabled.