Files
mcp-nextcloud/docs/troubleshooting.md
T
Chris CoutinhoandClaude Opus 4.7 e78e191818 docs: fix /health → /health/live, refresh stale mode-detection log examples (#766)
Issue #766 reported that the running.md quick-start tells users to
`curl http://localhost:8000/health`, which returns 404 — the server
only registers `/health/live` and `/health/ready` (K8s-style probes).
The same section also listed BasicAuth and OAuth startup log lines
(`BasicAuth mode detected …`, `OAuth mode detected …`) that no longer
exist anywhere in the codebase.

Update running.md and troubleshooting.md to point at the real
endpoints, explain liveness vs readiness, and replace the fictional
log examples with messages the server actually emits today. Also
clarify that the per-session BasicAuth messages only appear after the
first MCP client connects, which is the second symptom the reporter
hit.

Docs-only change; code paths and endpoint surface unchanged.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-12 07:35:49 +02:00

457 lines
11 KiB
Markdown

# Troubleshooting
This guide covers common issues and solutions for the Nextcloud MCP server.
> **Multi-user / Login Flow v2 issues?** See the [Login Flow v2 troubleshooting section](login-flow-v2.md#troubleshooting) for app-password storage, provisioning loops, and OAuth issuer problems.
> **Upgrading from v0.57.x?** See the [Configuration Migration Guide](configuration-migration-v2.md) for help with new variable names.
## Configuration Issues (v0.58.0+)
### Issue: Deprecation warning for VECTOR_SYNC_ENABLED
**Symptom:**
```
WARNING: VECTOR_SYNC_ENABLED is deprecated. Please use ENABLE_SEMANTIC_SEARCH instead.
```
**Cause:** You're using the old variable name from v0.57.x.
**Solution:**
```bash
# In your .env file, replace:
VECTOR_SYNC_ENABLED=true
# With:
ENABLE_SEMANTIC_SEARCH=true
```
See [Configuration Migration Guide](configuration-migration-v2.md) for complete migration instructions.
---
### Issue: Deprecation warning for ENABLE_OFFLINE_ACCESS
**Symptom:**
```
WARNING: ENABLE_OFFLINE_ACCESS is deprecated. Please use ENABLE_BACKGROUND_OPERATIONS instead.
```
**Cause:** You're using the old variable name from v0.57.x.
**Solution:**
**If you have semantic search enabled:**
```bash
# In multi-user modes, you can remove ENABLE_OFFLINE_ACCESS entirely!
# ENABLE_SEMANTIC_SEARCH automatically enables background operations
# Before (v0.57.x):
ENABLE_OFFLINE_ACCESS=true
VECTOR_SYNC_ENABLED=true
# After (v0.58.0+):
ENABLE_SEMANTIC_SEARCH=true # This is all you need!
```
**If you only want background operations (no semantic search):**
```bash
# Replace:
ENABLE_OFFLINE_ACCESS=true
# With:
ENABLE_BACKGROUND_OPERATIONS=true
```
---
### Issue: "Invalid MCP_DEPLOYMENT_MODE"
**Symptom:**
```
ValueError: Invalid MCP_DEPLOYMENT_MODE: 'oauth'. Valid values: single_user_basic, multi_user_basic, login_flow_v2
```
**Cause:** Invalid value for `MCP_DEPLOYMENT_MODE`.
**Solution:**
Use one of the valid mode values:
```bash
MCP_DEPLOYMENT_MODE=single_user_basic # Single-user with username/app password
MCP_DEPLOYMENT_MODE=multi_user_basic # Multi-user BasicAuth pass-through
MCP_DEPLOYMENT_MODE=login_flow_v2 # Multi-user via Login Flow v2 (recommended)
```
Or remove `MCP_DEPLOYMENT_MODE` to use automatic detection.
---
### Issue: Missing TOKEN_ENCRYPTION_KEY when semantic search enabled
**Symptom:**
```
Error: [login_flow_v2] TOKEN_ENCRYPTION_KEY is required when ENABLE_SEMANTIC_SEARCH is enabled
```
**Cause:** In multi-user modes, semantic search automatically enables background operations, which require encrypted token storage.
**Solution:**
Generate an encryption key and add required token storage configuration:
```bash
# Generate encryption key
python -c "from cryptography.fernet import Fernet; print(Fernet.generate_key().decode())"
# Add to .env:
TOKEN_ENCRYPTION_KEY=<generated-key>
TOKEN_STORAGE_DB=/app/data/tokens.db
```
**Why this happens:**
- v0.58.0+ automatically enables background operations when `ENABLE_SEMANTIC_SEARCH=true` in multi-user modes
- Background operations need encrypted refresh token storage
- This simplifies configuration but requires the encryption infrastructure
See [Configuration Guide - Semantic Search](configuration.md#semantic-search-configuration-optional) for details.
---
### Issue: Both old and new variable names set
**Symptom:**
```
WARNING: Both ENABLE_SEMANTIC_SEARCH and VECTOR_SYNC_ENABLED are set. Using ENABLE_SEMANTIC_SEARCH.
```
**Cause:** You have both the old and new variable names in your configuration.
**Solution:**
Remove the old variable name:
```bash
# Remove this line:
VECTOR_SYNC_ENABLED=true
# Keep this line:
ENABLE_SEMANTIC_SEARCH=true
```
The server will use the new name and ignore the old one, but it's cleaner to remove the old variable entirely.
---
## Multi-User / Login Flow v2 Issues
For multi-user deployment issues — provisioning loops, app-password storage, OAuth issuer endpoints, scope enforcement — see the [Login Flow v2 troubleshooting section](login-flow-v2.md#troubleshooting).
### Switching deployment modes
```bash
# To Single-User BasicAuth: set NEXTCLOUD_USERNAME and NEXTCLOUD_PASSWORD
# To Multi-User BasicAuth pass-through: ENABLE_MULTI_USER_BASIC_AUTH=true (no creds)
# To Login Flow v2: ENABLE_LOGIN_FLOW=true (no creds)
```
Restart the server after changing modes. The active mode is logged at startup; you can also set `MCP_DEPLOYMENT_MODE` explicitly to fail fast if the env vars don't match.
---
## Configuration Issues
### Issue: Environment variables not loaded
**Cause:** Environment variables from `.env` file are not loaded into the shell.
**Solution:**
**On Linux/macOS:**
```bash
# Load all variables from .env
export $(grep -v '^#' .env | xargs)
# Verify variables are set
env | grep NEXTCLOUD
```
**On Windows (PowerShell):**
```powershell
# Load variables from .env
Get-Content .env | ForEach-Object {
if ($_ -match '^\s*([^#][^=]*)\s*=\s*(.*)$') {
[Environment]::SetEnvironmentVariable($matches[1].Trim(), $matches[2].Trim(), "Process")
}
}
# Verify variables are set
Get-ChildItem Env:NEXTCLOUD*
```
**With Docker:**
```bash
# Docker automatically loads .env when using --env-file
docker run -p 127.0.0.1:8000:8000 --env-file .env --rm \
ghcr.io/cbcoutinho/nextcloud-mcp-server:latest
```
---
### Issue: ".env file not found"
**Cause:** The `.env` file doesn't exist or is in the wrong location.
**Solution:**
```bash
# Create .env from sample
cp env.sample .env
# Edit with your Nextcloud details
nano .env # or vim, code, etc.
# Ensure you're in the correct directory when running commands
pwd # Should be in the project directory containing .env
```
---
### Issue: "Invalid Nextcloud credentials"
**Cause:** BasicAuth credentials are incorrect or the app password has been revoked.
**Solution:**
1. **Verify username:**
```bash
# Username should match your Nextcloud login
echo $NEXTCLOUD_USERNAME
```
2. **Generate a new app password:**
- Log in to Nextcloud
- Go to **Settings** → **Security**
- Under "Devices & sessions", create a new app password
- Update `.env` with the new password
3. **Test credentials manually:**
```bash
curl -u "$NEXTCLOUD_USERNAME:$NEXTCLOUD_PASSWORD" \
"$NEXTCLOUD_HOST/ocs/v2.php/cloud/capabilities" \
-H "OCS-APIRequest: true"
# Should return XML with capabilities
```
---
## Server Issues
### Issue: "Address already in use" / Port conflict
**Cause:** Another process is using port 8000.
**Solution:**
**Option 1: Use a different port**
```bash
uv run nextcloud-mcp-server --port 8080
```
**Option 2: Find and kill the process using the port**
```bash
# On Linux/macOS
lsof -ti:8000 | xargs kill -9
# On Windows
netstat -ano | findstr :8000
taskkill /PID <pid> /F
```
**Option 3: Stop other MCP server instances**
```bash
# Check for running instances
ps aux | grep nextcloud-mcp-server
# Kill specific process
kill <pid>
```
---
### Issue: Server starts but can't connect
**Cause:** Server is bound to localhost only, or firewall is blocking connections.
**Solution:**
1. **Check server binding:**
```bash
# Bind to all interfaces to allow network access
uv run nextcloud-mcp-server --host 0.0.0.0 --port 8000
```
2. **Test connectivity:**
```bash
# Test from same machine
curl http://localhost:8000/health/live
# Test from network (if using --host 0.0.0.0)
curl http://<server-ip>:8000/health/live
```
3. **Check firewall:**
```bash
# Linux (ufw)
sudo ufw allow 8000/tcp
# Linux (firewalld)
sudo firewall-cmd --add-port=8000/tcp --permanent
sudo firewall-cmd --reload
```
---
### Issue: Server crashes or restarts frequently
**Cause:** Various issues including memory limits or uncaught exceptions.
**Solution:**
1. **Check logs with debug level:**
```bash
uv run nextcloud-mcp-server --log-level debug
```
2. **Monitor resource usage:**
```bash
# Check memory and CPU
top -p $(pgrep -f nextcloud-mcp-server)
```
3. **Use process manager for automatic restart:**
```bash
# With systemd (see Running guide for full config)
sudo systemctl restart nextcloud-mcp
# With Docker Compose (includes restart: unless-stopped)
docker-compose up -d
```
---
## Connection Issues
### Issue: MCP client can't authenticate
**Cause:** Auth flow failing or credentials invalid.
**Solution:**
**For BasicAuth modes:**
1. Verify credentials work:
```bash
curl -u "$NEXTCLOUD_USERNAME:$NEXTCLOUD_PASSWORD" \
"$NEXTCLOUD_HOST/ocs/v2.php/cloud/capabilities" \
-H "OCS-APIRequest: true"
```
**For Login Flow v2 mode:**
1. Verify the server starts the OAuth issuer:
```bash
uv run nextcloud-mcp-server --oauth --log-level debug
# Look for "OAuth initialization complete"
```
2. Verify `NEXTCLOUD_MCP_SERVER_URL` matches the URL clients use to connect:
```bash
echo $NEXTCLOUD_MCP_SERVER_URL
```
3. See [Login Flow v2 troubleshooting](login-flow-v2.md#troubleshooting) for app-password and provisioning issues.
---
### Issue: Tools return errors or don't work
**Cause:** Missing Nextcloud apps, incorrect permissions, or API issues.
**Solution:**
1. **Verify required Nextcloud apps are installed:**
- Notes: Install "Notes" app
- Calendar: Ensure CalDAV is enabled
- Contacts: Ensure CardDAV is enabled
- Deck: Install "Deck" app
2. **Check user permissions:**
- Ensure the authenticated user has access to the resources
- Check sharing permissions for shared resources
3. **Test API directly with Basic Auth:**
```bash
curl -u "$NEXTCLOUD_USERNAME:$NEXTCLOUD_PASSWORD" \
"$NEXTCLOUD_HOST/apps/notes/api/v1/notes"
```
4. **Check server logs for specific errors:**
```bash
uv run nextcloud-mcp-server --log-level debug
```
---
## Getting Help
If you continue to experience issues:
### 1. Enable Debug Logging
```bash
uv run nextcloud-mcp-server --log-level debug
```
Review the logs for specific error messages.
### 2. Test Nextcloud Connectivity
```bash
# Verify Nextcloud is reachable from the MCP server
curl -I "$NEXTCLOUD_HOST/status.php"
# With Basic Auth (Single-User or Multi-User BasicAuth modes)
curl -u "$NEXTCLOUD_USERNAME:$NEXTCLOUD_PASSWORD" \
"$NEXTCLOUD_HOST/ocs/v2.php/cloud/capabilities?format=json" \
-H "OCS-APIRequest: true"
```
For Login Flow v2 mode, see [Login Flow v2 troubleshooting](login-flow-v2.md#troubleshooting).
### 3. Check Versions
```bash
# MCP Server version
uv run nextcloud-mcp-server --version
# Python version
python3 --version
# Nextcloud version (check in admin panel)
```
### 4. Open an Issue
If problems persist, open an issue on the [GitHub repository](https://github.com/cbcoutinho/nextcloud-mcp-server/issues) with:
- **Server logs** (with `--log-level debug`)
- **Nextcloud version**
- **Deployment mode** (single_user_basic / multi_user_basic / login_flow_v2)
- **Error messages**
- **Steps to reproduce**
- **Environment details** (OS, Python version, Docker vs local)
---
## See Also
- [Authentication](authentication.md) - Authentication modes
- [Login Flow v2](login-flow-v2.md) - Multi-user setup, scope reference, troubleshooting FAQ
- [Configuration](configuration.md) - Environment variables
- [Running the Server](running.md) - Server options