Operations¶
This section covers operational tasks for maintaining grite installations.
Overview¶
Grite is designed to be low-maintenance, but occasionally you may need to:
- Run health checks
- Rebuild the database
- Manage snapshots
- Troubleshoot issues
Documentation¶
Health Checks¶
Run diagnostics and auto-repair:
Rebuilding¶
Rebuild the materialized view:
Snapshots¶
Manage snapshots for faster rebuilds:
Troubleshooting¶
Diagnose and fix common problems:
- Database issues
- Sync problems
- Daemon errors
Quick Reference¶
| Task | Command |
|---|---|
| Health check | grite doctor |
| Auto-repair | grite doctor --fix |
| Database stats | grite db stats |
| Rebuild | grite rebuild |
| Fast rebuild | grite rebuild --from-snapshot |
| Create snapshot | grite snapshot |
| Clean snapshots | grite snapshot gc |
| Stop daemon | grite daemon stop |
Maintenance Schedule¶
Regular Tasks¶
| Task | Frequency | Command |
|---|---|---|
| Health check | Weekly | grite doctor |
| Snapshot GC | Monthly | grite snapshot gc |
| Lock cleanup | As needed | grite lock gc |
Triggered Tasks¶
| Condition | Action |
|---|---|
| Doctor warns about rebuild | grite rebuild |
| Database corruption | grite doctor --fix |
| Sync failures | Check Troubleshooting |
Database Management¶
The sled database is a cache that can be safely deleted and rebuilt:
# Check database stats
grite db stats --json
# Verify integrity
grite db check --json
# Full rebuild if needed
grite rebuild
Limits to Consider¶
- Very large WALs slow rebuilds without snapshots
- High push frequency can cause contention on
refs/grite/wal - Each actor's sled database should not be shared between processes
Automation¶
CI/CD Health Check¶
- name: Grite Health Check
run: |
grite doctor --json | jq '.data.checks[] | select(.status != "ok")'
Scheduled Maintenance¶
Next Steps¶
- Health Checks - Start with diagnostics
- Troubleshooting - Fix common issues