# SENTINEL: arbiter node unreachable, failover delayed

- **ID:** `redis/sentinel-arbiter-unreachable`
- **Domain:** redis
- **Category:** network_error
- **Verification:** ai_generated
- **Fix Rate:** 88%

## Root Cause

The Sentinel instance cannot communicate with the arbiter node, which is crucial for achieving quorum in failover decisions.

## Version Compatibility

| Version | Status | Introduced | Deprecated |
|---------|--------|------------|------------|
| Redis 6.2 | active | — | — |
| Redis 7.0 | active | — | — |
| Redis 7.2 | active | — | — |

## Workarounds

1. **Check network connectivity to the arbiter node: ping <arbiter-ip> and ensure firewall rules allow Sentinel traffic (port 26379).** (90% success)
   ```
   Check network connectivity to the arbiter node: ping <arbiter-ip> and ensure firewall rules allow Sentinel traffic (port 26379).
   ```
2. **Restart the Sentinel service on the affected node: systemctl restart sentinel (or redis-sentinel restart).** (70% success)
   ```
   Restart the Sentinel service on the affected node: systemctl restart sentinel (or redis-sentinel restart).
   ```

## Dead Ends

- **Increase sentinel monitor quorum to 3 to tolerate more failures** — Does not fix the network issue; may delay failover further if more nodes are unreachable. (40% fail)
- **Remove the arbiter node from the Sentinel configuration** — Reduces redundancy; if other nodes fail, quorum may not be reached, causing no failover. (50% fail)