SENTINEL redis runtime_error ai_generated true

SENTINEL: failover aborted - no good replica found for master 'mymaster'

ID: redis/sentinel-failover-no-good-replica

Also available as: JSON · Markdown · 中文

80%Fix Rate

84%Confidence

1Evidence

2024-01-12First Seen

Version Compatibility

Version	Status	Introduced	Deprecated	Notes
redis 6.2	active	—	—	—
redis 7.0	active	—	—	—
redis 7.2	active	—	—	—

Root Cause

Sentinel could not promote any replica because all replicas are either stale (data too old), down, or have priority 0.

generic

中文

Sentinel 无法提升任何副本，因为所有副本要么过时（数据太旧）、宕机，要么优先级为 0。

Official Documentation

https://redis.io/docs/latest/operate/oss_and_stack/management/sentinel/

Workarounds

85% success Check replica replication lag and priority: INFO REPLICATION on each replica Ensure at least one replica has lag < sentinel-replica-down-after-milliseconds and priority > 0. If lag is high, wait for sync or manually trigger replication from master.
```
Check replica replication lag and priority:
INFO REPLICATION on each replica
Ensure at least one replica has lag < sentinel-replica-down-after-milliseconds and priority > 0.
If lag is high, wait for sync or manually trigger replication from master.
```
90% success Adjust replica priority to 1 (default) on all replicas: CONFIG SET replica-priority 1 This ensures they are eligible for promotion.
```
Adjust replica priority to 1 (default) on all replicas:
CONFIG SET replica-priority 1
This ensures they are eligible for promotion.
```

中文步骤

Check replica replication lag and priority:
INFO REPLICATION on each replica
Ensure at least one replica has lag < sentinel-replica-down-after-milliseconds and priority > 0.
If lag is high, wait for sync or manually trigger replication from master.

Adjust replica priority to 1 (default) on all replicas:
CONFIG SET replica-priority 1
This ensures they are eligible for promotion.

Dead Ends

Common approaches that don't work:

65% fail
Timeout increase does not fix stale or down replicas; failover will still abort.
75% fail
If the replica is too stale, promoting it can cause data loss and inconsistency.
80% fail
Restarting sentinels does not change replica state; they will still see the same condition.